Ref Nptel1

INDEX
S. No Topic Page No
Week 1
1 Familiarization with Power Electronic Systems 1
Overview of Basic Power Electronic Circuits from Laymanâ€™s
2 Point of View 19
3 Applications, Definitions, and Nature of Power Electronic Circuits 31
4 Components of a Power Electronic System 59
5 Analysis of Switched Networks 88
Week 2
6 Review of engineering maths for power electronic circuit analysis 109
7 Review of semiconductor physics 130
8 P-N Junction 144
Week 3
9 Power Diodes 171
10 Thyristors 193
Week 4
11 Motivation for rectifier capacitor filter 215
12 Circuit Operation 226
13 Designing the circuit 239
14 Simulation setup for NgSpice and gEDA schematic capture 254
15 Simulating the circuit 281
16 Practicals 283
Week 5
17 Inrush current limiting - intro 286
18 Inrush current limiting - Resistor solution 290
19 Inrush current limiting - Thermistor solution 293
20 Inrush current limiting - Transformer solution 296
21 Inrush current limiting - MOSFET solution 300
22 Inrush current limiting - Relay, contactor 306
23 Three phase rectifier capacitor filter 312
24 Simulation - 3 phase rectifier capacitor filter 314
25 Power factor - motivation 322
26 Power factor - discussion 324
27 Power factor - sinusoidal 340
28 Power factor for rectifier cap filter 344
29 Passive power improvement circuit 350
30 Simulation - power factor improvement 356
Week 6
31 Linear regulators - intro 360
32 Shunt regulator 363
33 Example on shunt regulator 367
34 Non-ideality and solution 370
35 Applications of shunt regulator 373
36 Series regulator 377
37 Efficiency of series 381
38 Negative and dual voltage regulators 383
39 Over current limiting circuits 387
40 Improvements to series regulator 393
41 Regulator performance parameters 399
42 Datasheet of few IC regulators 401
43 Common IC regulator circuits 412
44 Practicals 1 416
Week 7
45 Switched mode DC-DC converter intro 421
46 Volt-sec and Amp-sec balance 428
47 Input-output relationship 431
48 Buck converter - operation and waveforms 435
49 Buck converter - component selection 439
50 Primary configurations 447
51 Boost converter 450
52 Buck-Boost converter 460
53 Simulating the primary converters 471
Week 8
54 Forward converter 480
55 Core reset in forward converter 490
56 Simulating with lossy core reset 502
57 Simulating with lossless core reset 530
58 Flyback converter 535
59 Simulating the flyback converter 543
60 Octave mfile for design 549
Week 9
61 Magnetics design intro 553
62 Magnetics review 559
63 Permeance 570
64 Inductor value and energy storage 574
65 Inductor area product 582
66 Inductor design 588
67 Inductor example 596
68 Transformer design 604
69 Transformer example 619
70 Forward converter design mfile 624
Week 10
71 Pushpull converter 637
72 Flux walking in pushpull 650
73 PWM generation 658
74 Simulation of pushpull converter 669
75 Half bridge converter 680
76 Simulation of halfbridge converter 687
77 Full bridge converter 696
78 Simulation of fullbridge converter 700
79 Area products and mfiles 707
Week 11
80 Intro for drive circuits 720
81 BJT base drive 725
82 BJT base drive example 737
83 Multi-stage base drive 742
84 Base drive with speed-up circuit 750
85 Base drive with isolation 755
86 MOSFET gate drive 761
87 MOSFET drive with isolation 769
88 Over-current protection 774
89 Snubber circuits 779
Week 12
90 Intro for close loop control 787
91 Close looping dc-dc converters 789
92 Simulation of close loop control 800
93 Current control for battery charger application 813
94 Instability in current control and slope compensation 821
95 Slope compensated current control 828
96 Simulation of current control 833
97 Single phase inverter with sinusoidal pwm 841
98 Simulation of sinusoidal PWM 849
Fundamentals of Power Electronics
Prof. Vivek Agarwal
Department of Electrical Engineering
Indian Institute of Technology, Bombay
Lecture - 01
Familiarization with Power Electronic Systems
I am Vivek Agarwal. I am a faculty at IIT Bombay in Mumbai, and welcome to this

course. I am delighted to have you here for this course. We will try to cover several
aspects of power electronics in this course.
Before starting with the formal content of this course, I would like to do a Layman’s
discussion, some questions, some suggestions, some answers. Let us try to start with
some of the very basic things related to power electronics. And what we will do is, we
will just try to first see some of the very common examples around us, which we do not
know by the name of power electronics, but they actually come from power electronics.
This will strengthen our belief that power electronics is an important area. It is a very
important subject.
Now, I am not sure how many of you have very deeply thought about the power
electronic applications around you, but let me ask you a question. Have you ever
wondered what goes inside a power supply that is used in a laboratory? When you were
in your junior classes, I am sure, even in your class in your high school, and even higher
classes senior school, I am sure that you were doing experiments which were involving
power supplies.
1
(Refer Slide Time: 02:02)
So, for example, you must have seen these kinds of equipment in your lab and I am sure
that you have played with some of these knobs. What you see here, which are basically,
this is for the voltage and the current adjustment, which is coming out of this power
supply. And here, there are displays, where you can see how much current and how
much voltage is being coming out, and these are the two terminals from where you tap
the power supply.
So, you adjust your voltage to 5 volts or 10 volts or 12 volts or 15 volts, whatever you
like, by using this knob and through these terminals, you actually are able to feed the
load. So, this is what you did in your lab experiments in your junior classes and even
later on. So, these supplies are actually used at all stages of our academic and research
work.
Now, let me just ask you another question, the desktop computer. Everybody is using a
desktop computer today. Are you aware that, what kind of power supplies they go inside
this desktop computer to make it run? Do you know, what are the various voltages and
the currents, which are required by the various portions, various sections of this power
supply of your desktop computer? Are you aware of this?
2
So, let me just show you one supply, which I just took out from a desktop computer.
Here, so this is the supply, it is a box, as you can see which takes 230 volts, 50 hertz,
AC supply, at one of its ends here, the input socket and it gives out various outputs,
which are coming out, it gives out voltages various ranges 3 volts, 3.3 volts, 12 volts and
so on. So, have you ever wondered, what goes to make up this kind of supply, what is
there inside this supply. Let me explain to you.
Now, I showed you in the computer power supply that, there is a 230 volts input, AC
inputs, so that is a primary supply. And this has the AC supply, this is actually a result of
a war, which was fought in the late 19th century, which is called the war of currents,
where there was a war between the AC systems and DC systems, where people like
Thomas Alva Edison, supporting the DC kind of power systems and people like Nikola
Tesla and Westinghouse supporting the AC systems.
And eventually, it was the AC system, which won and as a result of this, you now see
AC systems everywhere in the world. So, that is the reason. So, we get AC supply, it has
lot of advantages. For example, boosting of the voltage is possible, transmitting over
long distances, it makes the efficiency very high, if you boost the voltage before you
transmit it over a long distance. So, it actually gives you all those advantages. So, that is
why you have a AC system everywhere followed in the world.
3
Now, you have the AC system, but inside the computer you need various DC voltages.
So, what do you do? So, you have a 230 volts AC voltage, which is coming in, which
you plug in to your computer and inside there is a system, which converts just like a box,
which I showed you. It converts this 230 volts AC into the appropriate levels of the
voltage which are required in the various sections inside the computer. So, let us just
draw this and see what exactly is happening.
So, what we are getting from the AC wall sockets is 230 volts rms AC supply. This is
what we are getting. And in India, the frequency of this AC is 50 hertz, which means
that, if I just take one of the time periods between, that is for one cycle, the time period,
T is equal to 1 over 50 which is exactly 20 milliseconds. So, this is the kind of voltage
that we are getting from the AC wall socket.
Now, what we have to observe here is that, this is a bipolar voltage. It has both positive
and negative sides. We are looking at a DC voltage. So, first of all, we want to limit the
direction of this waveform to only one, we want to make it or keep it only positive. We
want to make it unipolar. Secondly, the voltage range which is required in the computer
for the various parts of the computers, and its various circuits is very low. It is about 3
volts, 12 volts, or maybe at the most some 18 volts and things like that. It’s very low.
So, we are getting a voltage which is having an rms value of 230 volts and if it is a
sinusoidal voltage, we can say that the peak is 230√2, that is the peak of the voltage.
4
Now, from such a voltage waveform, we want to derive very low DC voltages. So, there
are two things that we can say here. First of all, we need to reduce the voltage level and
the second thing is we want to make the excursions of the voltage, which is currently
both positive and negative in the AC, we want to make it unidirectional or unipolar.
These are the two things we want to do.
Now, we all know that, to reduce the voltage level, we can use a transformer. In
advanced modern systems, voltage transformation is achieved using high frequency
transformers, which are integrated with the power electronic systems itself. Low
frequency transformers, like the one I will describe now, are seldom used. However, to
keep the explanation simple and till we cover the high frequency converters in this
course, I will explain this computer supply problem with the help of the ordinary 50
hertz independent transformer.
So, let us say, we now apply a transformer and we are looking at something like, 12 volts
DC, that we wish to apply to one of the sections, to one of the portions of these circuits,
which are employed in the inside the computers. So, this is the transformer. Now, if I
apply the incoming 230 volts, 50 hertz to this transformer and let me just decide the turns
ratio of this transformer as something like 25:1, approximately. So, what we will get is
approximately 10 volts rms on the secondary side of this transformer. So, voltage
reduction is possible, but then, as you can see with the transformer, once you have placed
the transformer, you will get a fixed output, given input voltage of 230 volts, you have
fixed the output voltage, once the turns ratio is decided.
Now, how do you make it DC? We have the AC coming in. How do you make it DC?
How do you make it unidirectional? So, for this, let us try to understand and recall. I am
sure, we have all seen the operation of a diode rectifier. The diodes are semiconductor
devices, which came up in the last century and they have kind of revolutionized the way
we process power. So, they are very fundamental and basic units which are used, which
are used in power electronics. So, I am sure that we have all seen diodes.
5
So, let me just form this very customary, familiar looking circuit, we have done it at very
junior classes as well. So, everybody is aware. So, this is nothing but a diode bridge
rectifier circuit, what I have drawn. And what you see are these D1, D2, D3 and D4. These
are the four diodes, which are used in this rectifier circuit. These two points are tapped
outside. Let me just mark them as X and Y and these indicate the output of the diode
bridge rectifier.
And what goes to the input of the rectifier? To the input of the rectifier, we now supply
the voltage, that we got from the transformer, which is about 10 volts rms. So, we have a
AC supply of 10 volts rms. Mind you the frequency is 50 hertz, we did not do anything
to change the frequency. So, it is original 50 hertz, that we are getting with our wall
socket of 230 volts. Now, this voltage is what we are applying, to these various two
points of our diode bridge rectifier. Now, if I activate this circuit, then what would I
expect to see at terminals X and Y? Let me just draw the waveforms, what you will
observe at terminal X Y.
So, now, if I was to draw the AC voltage that we have obtained on the secondary side of
the transformer, then you can see that during the positive half of the cycle, the diodes D1
and D4 would be conducting. And the path of the current in that case, would be as shown
here. And of course, we can always assume that between X and Y, there is this load, this
6
represents that section of the computer, where you want to apply the 12 volts, for which
you are doing all this arrangement
So, let us just call this as load here. During the positive half cycle of the input supply, D1
and D4, get forward biased and they conduct. It’s like opening the gates. The diodes D1
and D4, are now able to conduct and basically somebody is standing at terminals X and
Y is able to see what is there on the input side, through D1 and D4. So, basically whatever
is a portion of the input supply at that time, when D1 and D4 are conducting would
appear across the terminals X and Y. So, this is what you will see.
Now, when the applied AC voltage becomes negative, as here in this case, you will find
in this case that, D2 and D3 will get forward biased, while D1 and D4 will be cut, become
reverse biased. And now, when D2 and D3 are forward biased, the path of the current or
the conduction would be as shown here, with a green color, through D2 anode causes the
cathode through the load back, and now this time, it is not going through D4, it is actually
going to D3. All the way here and then back. This is how it completes the path.
So, what will you see at terminals X and Y during this time though, what would we see?
So now, we will get another lobe, because as far as terminal X and Y is concerned, if you
notice that whether it was D1 and D4, when they were conducting or whether it is D2 or
D3, when they are conducting, they are seeing a current, which is going from X to Y. The
direction of the current has not reversed. So, basically it is a unidirectional current that is
flowing, they will see correspondingly.
Now, the input AC voltage, which is the negative part, but will be seen here as inverted
and so you will find this lobe, which will appear now, the second one and the cycle
repeats. So, D1, D4 conducting here, D2, D3 conducting here, and again, in the next half
cycle of the input AC supply, you have the D1 and D4 conducting, and again D2 D3 and
this cycle goes on. So, now, what would be the peak of this voltage or approximately
what can we say about this voltage? So, this voltage would be having some value, which
we can say is about 10√2, peak value or we can say that the average or the DC value of
the voltage across terminals X Y will be somewhere at midway between 0 and 10√2
volts, approximately.
And if I was to draw this voltage, I will probably draw like this. But the problem is that,
it is not so nice and constant, as it appears in this diagram because of this vast variations.
7
Here the peaks and the valleys, as you can see this, voltage is not going to be like this.
So, the load will actually see a highly pulsating DC waveform. This is not good for the
load. People who are working in this area for a long time, know what would happen to
the various equipment, when you apply this kind of voltage across them. Their life will
be shortened, there will be heating effects for example. So, one has to reduce them. So,
what is the solution for doing this? The simplest solution for this is to basically place a
capacitor.
So, what we do? We just place a capacitor across terminals X and Y and we just denote it
by Cf. It is the term which is used for filtering. So, basically you put this capacitor, which
filters out this very large ripple, which is there in this DC voltage. Now, if the ripple is
very large, your capacitor value Cf tends to be very large. And to reduce that value, some
people actually implement little different forms of the filters. There are various different
forms of the filters and one of the very common is, where you just actually insert one
additional L, here Lf as part of the filter, which tries to reduce the current ripple and
hence the capacitor has to bypass a lower amount of ripple current and its value can be
less. So, if you want to reduce the value of Cf, you can use some inductor a filter inductor
Lf also, as part of the filter in the circuit.
So, what is the result of putting this Lf? If we look at our diagram, where I showed the
output voltage across terminals X and Y, it will be something like this, it is the original
waveform, but now with the placing of the capacitor Cf, it would become like this. If I
still increase the value of Cf, it will become like this. And if I also put one Lf along with
it, then it will become still flatter. So, basically I will be able to get, assuming that we
have put good enough values of Cf and Lf, a nice waveform here, whose value will be
10√2 volts, approximately. That is what, we will get and this is what, we can think of
applying to the various required, and the concerned sections of the computer power
supply.
So, basically, there are various such circuits, that you can imagine are present inside the
power supply of the computer, which I showed you. So, let me just open that supply for
you and try to let us try to see what is there inside.
8
So, I can see some transformers there, and I can see some heat sinking here. So, there are
some heat sinks, which are for dissipating the heat away from the devices, which are
used in the power electronic circuits. I can see the device is mounted there. I am not sure,
whether they are very clear, but actually they are right there. There are capacitors, they
are very small capacitors.
And what you see other than that are some electronic components, some integrated
circuits ICs, some resistors, some diodes very small diode, they are called small signal
diodes. I can see a fan here, which is actually used for fan cooling, for forced cooling of
the devices, which are used in this circuit.
So, these heat sinks will provide limited protection against heat, even on top of that we
would need additional cooling, which is done through this force convection, through the
fan. So, there is a fan there, if you can see. So, I will come to the details of this in, we
will actually look in into much more details of these various components, various things
which are used in the supply, as we go on in this course, but it is just to give you the feel.
So, this is all power electronics. So, as you can see how power electronics is used
extensively to achieve these various voltage levels from a given voltage supply.
Now, one of the problems that we have with the power supply, that we have seen as an
example, just now, we kind of analyzed, we look at its working and so on. One of the
issues is that, for a given input AC voltage, which was about 10 volts rms, we were able
9
to get a fixed value of the output voltage across terminals X and Y. It is frequently the
case that, we have to vary the voltage across a load for controlling it. For example, a
motor or a light bulb etcetera. The uncontrolled AC to DC system, we discussed so far,
cannot achieve this. We not only need different voltage values to apply to the loads, but
also need to regulate them or hold them constant irrespective of the changing operating
conditions.
So, this is the voltage regulation problem, which is a very fundamental problem,
addressed in most of the power electronic systems in some form or the other. So, what is
the solution? And what exactly happened? Why is it that we are not able to vary the
output voltage? Well, the reason is that you are using diodes, and the moment there is a
forward voltage occurring across them, they conduct and they cannot be controlled, you
cannot control once the forward voltage is appearing across the diode. It will conduct.
There is nothing you can do, to prevent that.
So, the question is, if I can control the diode, so I can somehow have a mechanism by
which, I can have the forward voltage held up, or opposed by the diodes. In the face of a
forward voltage also, there is a way that this device does not conduct, till I ask or till my
circuit requirement asks it to do. Then I can probably achieve a constant voltage at
terminals X and Y, which appears across it.
Now, going further ahead, let us assume that we are using this imaginary device. And let
us just say that, these are the properties of this device. So, I will just redraw the device,
which I introduced last time.
10
And it is a three-terminal device, where the terminals are marked like this. Now, let us
make certain assumptions about this device. Let us say that the conduction of current
through this device can only take place from 1 to 2. Now, this is quite understandable, if
you compare it with a diode. The conduction of current in a diode can always take place
from anode towards cathode. So, based on this assumption, this device can conduct only
from terminal 1 to 2. It cannot conduct from 2 to 1. It can conduct only when you apply a
control pulse to terminal 3.
Now, based on these assumptions and probably another assumption, we can make is that,
when there is a conduction taking place across this device from 1 towards 2, let us also
assume that there are no net I2R losses which are taking place. Now, why do I talk about
this? This device after all is made up of some material, and we will see eventually that,
this is nothing but a semiconductor device, made up of some material. So, depending on
the technology used, and the material used and its operation, and its mechanism, it will
actually offer some resistance to the current that is going from 1 to 2. And there will be
I2R, joules heating effect, that loss will take place, and actually it is this loss, which
appears as heat across the junction of the semiconductor devices, which actually makes
the thermal management very important.
11
So, in the computer supplied, power supply which I showed you, I did try to tell you that,
there was a heat sink onto which these devices were mounted, and I also mentioned
about forced cooling. These are all components of a thermal management system, which
are trying to take care of all the unwanted undesirable losses, like the one shown here,
the I2R losses.
So, let us say for, at least for now, that this is not there. So, we actually have an ideal
device, where R is 0. So, it does not offer any resistance to the flowing current. So, such
is this device. So, let us say, we are talking about an ideal device, all these devices which
I mentioned, they have, if they are ideal, they will fit this description, which I just gave
for this device.
Now, let us go back to our example of the rectifier, but now replace with this special
device. So, let me just redraw this modified rectifier with some voltage VAC and I am just
showing another terminal. So, instead of diodes, I am putting this third. This is a special
device. Later on, we will see that, some of these semiconductor devices, which I
mentioned to you, could be actually replacing this device. As before, there is a load
which is connected to X and Y, which are again marked here and my AC supply goes
here and goes to this midpoint. So, for all the devices I have marked a control terminal.
This is the terminal three of the special device.
Now, let us see how the operation of this circuit would take place and what we will
achieve or obtain at the terminals X and Y in this case. (Refer Slide Time: 30:12)
12
This is just the background waveform. This is not what is going to appear. This is just for
reference. Now, let me just use a red color and let us say that the pulse to the third
terminal, the control terminal is not applied up to this point. The individual pair, the
corresponding pair of these special devices, are not supplied up to this point which is
marked as cross.
So, what will we actually see at the output? Assume that this load is very nice and a
purely resistive load, for our understanding. Now it is very easy to see that by adjusting
this delay that we give to the control pulse, before we apply it to the control terminal of
the device, let me just call this as td delay time. So, we can say that as td is increased, this
red waveforms or these red pulses become narrower, which means that the net DC
voltage will come down, will keep coming down, as you increase the value of td.
On the other hand, if you decrease the value of td, you will find that this value goes up.
Therefore, we can conclude that, by suitably adjusting this delay time or by suitably
adjusting the control pulses applied to the controllable devices, I can then control the
output voltage or in other words, I can control the voltage or I can regulate the voltage
across terminals X and Y. So, as the load varies, I can still vary this voltage to and I can
control this voltage to remain constant at the desired value. So, if I want to keep it at 12
volts. I can keep it at 12 volts irrespective of the load that appears across the terminals X
and Y, this is what I mean. And this I can do by adjusting td.
So, having understood this part, let me just now explain to you the very basic voltage
regulation problem. So, basically now what we have is I can just explain this in terms of
the block diagram.
13
So, I have a block which represents a rectifier system, where I have the capability to
adjust td. So, this td can be varied. So, I have a rectifier here. This block represents a
rectifier, and these are the two terminals X and Y. And I can all the time, irrespective of
the load, I can sense the voltage across X and Y, I can sense the voltage. I can compare
it, this is basically a comparator block, this is a representation of a comparator. And if I
want to always keep the voltage across the terminals X and Y equal to 12 volts, I
compare it with a 12 volt signal, which I call as a reference, and then based on the output
of this error amplifier, I do something. I do some calibration. I do some manipulation of
signals and I control td.
So, basically what I do is that, by adjusting the value of td, in response to the difference I
obtain between what I get at the output, and what I actually want at the output, which is
nothing but 12 volts. Based on that difference, based on that error, I can adjust the value
of td in such a way that I can get the desired voltage of 12 volts, I can get the desired 12
volts out for now, that is the irrespective of this load. This is what is called the voltage
regulation.
Now, what do you see here is that, there is a closed loop, which is involved in the
system. You can see that there is a signal, which is sensed and fed back into the system,
which then governs the control action, which takes place and corrects the output voltage
14
as per that. So, basically the voltage regulation problem is a control problem and there
are various aspects of it, that we will study as we move ahead with the course.
So, let us just now assume that, for a minute, we have a device which is controllable,
which can block a forward voltage. So, it is a modified, let us say that it is a super diode,
super form of diode which actually can block the forward voltage
And let us just say that, this device, let us call it or show it just by a block. So, instead of
a diode, now you have this block, which is representing the device. You have the
terminals, let us call them as 1 and 2, the two terminals and let us have a third terminal.
Here the diode did not have a third terminal, the diode has just two terminals, as we
know it is having an anode and a cathode. So, it has got only two terminals.
But now in this device, let us say, I have a third terminal also. And the purpose of this
third terminal is that, even though you have applied a forward voltage across 1 and 2, as
in a diode, you have still not supplied the activating pulse through the control terminal 3.
So, what will happen in this case? In this case, what will happen is that the input supply
which is going like this, now at this point looking at our previous diode rectifier circuit,
our D1 and D4, would get forward biased and they are ready to conduct. But now instead
of these those diodes, those four diodes D1, D2, D3 and D4, we are using this special
diode device switching device, which is shown here, with an additional terminal third
terminal, which is called the control terminal.
15
So, what will happen, if I have not applied, if I have used instead of the diodes, I have
used this device, but I have not applied any control signal to 3. So, my circuit would
keep waiting for a control signal to appear across 3. Now, let us say, we apply it at this
point. Let me just show you, with a colored pen. So, let us say, we apply the control
signal, at point 3 to diodes D1 and D4. So, to this special device, which is in place of
diodes D1 and D4, we apply it at this point. Immediately as soon as we apply this control,
now this device works as a diode, and you will find that at the output terminals X and Y
of the previous circuit, you will get a lobe of this type.
Now, similarly, let us just replace diodes D2 and D3 also, with this special device and we
again at this point will see that this point of the input AC voltage, the device would
become for all these two devices, which are placed in place of D2 and D3. They will be
forward biased and they will be ready to conduct. But, let us say, they are not given a
control pulse till another point which is up to here, which is symmetrical to the first flow
let us say. So, as soon as you give the control pulse to the terminals 3 of the two devices
placed in place of D2 and D3, you will find that a lobe will appear at the output terminals
X Y, which will look like this.
So, what you see here is your input supply, and what you see with the red on top is the
voltage X Y, and the cycle continues. So, again in the third lobe, suppose you have not
given you have delayed the pulse, control pulse through terminal 3 to some point, you
will get. So, this train would continue. So, this leads us to the understanding that, if we
can control the pulses which are applied to terminals 3 of all these special devices, which
are now constituting or comprising our rectifier, we can actually play with the output
voltage value. So, these pulses will become closer or they will separate out.
Their portions will be more and more portions of these output voltage waveforms, we
will be cut out, and you will be able to play with the output voltage.
Therefore, the conclusion is that, if we can use a controllable device rather than a diode,
then I can vary my output voltage, by controlling the pulses applied to the control
terminals of the various devices. A switching device like a diode, is called an
uncontrolled device, because there is no way to control this device, once a forward or a
reverse voltage is applied across the diode. But if you look at this special device, I talked
about there is a provision through terminals 3, that we can actually control this device.
16
So, it is not just a matter of applying a forward voltage across it. It is also a matter of
applying the control pulse to the device. So, when we apply the control pulse would
determine what exactly appears to the output.
So, this leads to the requirement and then development and inventions of the various
types of the power devices, which can then be used as controllable devices. So, the
special device that I described to you, as a fictitious device, actually subsequently as
research progressed on semiconductor devices, they all develop into devices, such as the
silicon-controlled rectifier.
The bipolar junction transistors, the MOSFETs, which is the metal oxide semiconductor
field effect transistor, and a combination of some of these devices, like the BJT and the
MOSFET, it led to what is called an insulated gate bipolar transistor and so many other
devices which came up.
So, as the course will proceed further, we will see the development of these devices. We
will study the properties and behavior of some of these devices. But one thing is clear at
this point, with the help of the example I gave you, is that these controllable devices, by
virtue of their controlled terminals, give us the flexibility obtaining any desired power
levels or any desired voltages or currents, that we need for our application.
17
So, in the next lecture, we would see some more examples, we will try to understand
power electronics in very basic terms. We will continue to do that and then we will start
with the formal content of the course.
Thank you very much.
18
Prof. Vivek Agarwal
Lecture – 02
Overview of Basic Power Electronic Circuits from Layman’s Point of View
Welcome back. So far, we have seen the importance of Power Electronics by looking at
some examples which are very close to us. We saw an example of a laboratory power
supply, and then we saw a computer power supply which actually I took out from a
desktop computer, and I showed you that this power supply is actually meant to cater
various voltage levels and various power levels in the various sections of the computer.
So, there are computers, a very complex machinery and there are lot of circuits inside
which need different levels of voltage and power.
So, we saw that typical power electronics system is required, a complex power
electronics system would be required. But I actually gave you an idea that how such
reduction in the voltage levels, and how a change from AC to DC form of electricity can
be achieved by using a very simple power electronic circuit which we studied as a
rectifier. We took the example of a diode-based rectifier first, and it gave us, after all the
filtering and all the other explanation, it explained to us that how we can get a nice and
constant DC voltage at the output terminals which we described as X Y of such a
rectifier system.
But there is an issue that we observe that there was no way to change this voltage. There
is no way to control this voltage across terminals X and Y if diodes are used in the
rectifier. So, the only thing if you want to change the voltage across terminals X and Y is
to change the input AC voltage itself that is all you can do. But then I introduced
fictitious imaginary device, a three-terminal device. I said that it has terminals 1 and 2 in
the same way as a diode. But it also has an edit third terminal. Unless and until we
activate the third terminal, my device does not conduct. This fictitious device does not
conduct even though, it may be forward biased. So, basically what I am saying is that it
has the ability to block the forward voltage. It is not like a diode which will start
conducting the moment it is forward biased.
19
Now, let us move ahead from the example of the rectifier, and let us now say that instead
of a AC voltage which is available to us, we are generating power supply we have a
generating power with batteries. Now, we all know that the batteries will give you a DC
supply, so it will give your DC voltage. When you apply a load to it, then we extract the
DC current through it. So, it is basically a DC supply.
Now, suppose you have a DC supply which is either through batteries or it could be
through DC generators or it could be through some other ways like for example, fuel
cells. Fuel cells are currently the front running candidates for being used as renewable
energy sources. And this could be actually a solar photovoltaic source. They all give out
DC voltage. In fact, even a wind energy source they are all used with rectifiers. So, they
also give a DC source.
So, let us say we are in a place we have where we have the primary source of power as
DC. So, we have the DC which is being produced. But we want to use it in some
processes, in some applications which need another level of DC voltage. What do we do?
So, one very simple way of looking at it is by way of this example. So, let us say if we
have something like 48 volts which is being produced by the battery or the DC generator
and I just want to get 24 volts, then what can I do? So, let me just draw this particular
diagram.
20
So, this is the Vdc which I am denoting, my primary supply which is of the DC form in
this case and let us say it is not a regulated power supply, it is a raw power supply. So, it
maybe it is just whatever the source is giving out, it is just what is it reflected here and
there is no regulation closed loop regulation associated with it.
Now, let me just put a simple switch here, and let me just denote at the output a load. Let
us say again for simplicity that this is a pure R load resistive load. Now, the switch that
we see is turned on and turned off continuously. So, basically the switch has these kinds
of pulses which are applied to its control terminal. By the way this switch is again a
controllable switch a three-terminal device, one the kind of one that we assumed. This is
the same three-terminal device.
And what we see here are the control pulses applied to the terminal 3 of this device.
Now, when that happens assuming that there are no losses, there are no other issues, they
are device is very ideal and so on and everything else is ideal in the circuit. What we can
see at the output is that when the switch is on then the Vdc will appear at the output
across the load and when the switch is open, that this part is broken. So, this load cannot
see the voltage at all, so you actually get a 0. So, what you get at the output, it is
basically these kinds of pulses, these kinds of output voltage pulses which are as you can
see in the shape in the appearance, they are same as the control pulses applied to the
switch or the switching device the controllable device.
Now, you can see that by changing the time for which these pulses are applied to the
controllable device the switch, so I have just put these arrows just to say that this right-
side boundary it can be shifted, it can be shifted actually to this side or to this side. So, as
that happens, as you shift on to this side, you extend the boundary you stretch the
boundary of the control pulses what you will see is that here these pulses will also stretch
out. While if you move to the left side you will find that they will shrinking.
So, what is the result of this? Now, if I was to look at the overall result of doing this what
I observe is that, if I was to denote the average of this as the DC voltage content of the
output I can see that by stretching the control pulses right or left I am able to move my
average value of Vo up or down, Vo is the voltage across the load; Vload or Vo we can just
denote it the way we like. So, that is what is going to happen. So, this is a very simple
example.
21
Now, if I want to get for a 48 volts DC voltage, I want to get a 24 volts output I can see
that the control pulses should actually be half of this duration. If they are exactly half of
this, we will get about 24 volts here. But the problem is that with this circuit how do we
get a voltage which is more than 48 volts across the load. Is it possible to get a voltage
which is output voltage Vo which will be more than the applied 48 volts? The answer is
with this circuit no, but power electronics also gives you a way to do that. So, what we
use is basically in this circuit we modify this configuration, and as we will see in the
course that we apply certain energy storing elements like inductors.
And we use the mechanism of storing the energy into the inductor and then releasing it,
storing the energy into the inductor and then releasing it, more like a spring action
compress release, compress release, and we will see that this section leads to the
boosting of the input voltage. So, you actually are able to get a voltage which is much
higher than what you have the input. But just to understand the very basic example of
power electronics for this kind of a situation where you have a DC input or the DC
power as the primary source, how do you actually get a variable output voltage is this
example that I have given.
And, mind you that regulation problem which I explained with the help of a rectifier is
equally valid here because I can say sense this output voltage as before, I can compare it
with a reference and based on the error I get from what I want Vo to V and what I am
actually getting from that difference I can always modify I can always manipulate and
control the pulse to the terminal 3 of the switch shown here. So, this is an example where
you can see that from a given from a DC source how you get variable levels of DC, how
you get a different level of DC voltage and how you can control this voltage.
Now, talking about the DC source; now, there are several applications where it is
desirable to have a DC source, and then manipulate it, and then get the kind of power
that you need to supply your load. Now, one example of this is in AC drives, AC motor
drives. Now, we all are aware of the electric machines. We have the two types the DC
machines and the AC machines. In the DC machines also, we have DC generators and
we have DC motors. And this motor works when you apply a DC supply. And to change
the speed of a DC motor, so suppose your application is such that you wish to change the
speed of your motor, you change the voltage supply, you change the DC voltage supply
apply to the armature of the motor, DC motor.
22
Now, when you go to the AC motor and you want to actually control this as the motor,
so you apply AC and AC as is an alternating current waveform AC waveform, you apply
to the AC motors. Now, one way of controlling let us say the speed of these AC motors
is just by changing the voltage, but the problem with these AC motors is that if you just
change the voltage levels but you do not change the frequency of the voltage supply that
you are applying to the motor. Then you actually are running the risk of saturating the
core which makes of them of the machine.
Therefore, you must have heard, at least some of you must have heard about what is
called a V by f control where the voltage applied to the AC motor windings to stator and
its ratio with the frequency. So, the voltage rms voltage applied to the three-phase stator
of an AC motor to the frequency of this voltage supply that is always kept a constant.
Now, this basically means that we need a mechanism that given a DC supply, primary
DC source we want to create a AC supply whose rms voltage we can vary and whose
frequency also we can vary in such a way that the rms voltage to the frequency ratio
remains constant.
Now, I will just show you with the example very again we will take the layout the
Layman’s approach we will not talk about very specific terms very technical, involve
technical terms in this just for understanding will come to these things slowly in due
course. So, I will just give you the example how this is done.
23
So, let me just start with a DC source, and let us say that I have that t terminal device
with all the ideal assumption. And I show its control terminals with this as you can see
there is a line with a small speck segments with a bubble like this. And I just apply it to
this bridge which I have form of these of these special devices the controllable devices.
These are different from diodes mind you as I told you. It has got a third terminal which
can control which can control precisely when these devices will start conducting and
when they will stop conducting the current.
The load is connected here. So, let us see here this is the output and I just mark this as A
and B, and let us say that this is going here and let us say that this is going to one of the
phases of my AC motor, let us say. So, let me just call this as the first as phase let us say
A of my three-phase AC motor. Now, for there are three-phases in the AC motor; there
are single phase motors as well but let us talk about three-phase AC motors right now.
So, basically what we need is we need to create another two phases here. But for now, let
us assume that we are just talking about one of the phases of the three-phase AC motor
and what will happen.
Now, let us say we just have; so let me just mark these devices as 1 2 3 4 and let us call
the control pulses to these various devices which are applied to the various terminals as
g1 and g4, and g2 and g3. So, there are control pulses to these two pairs that I want to
draw and show you. So, let us say these are the control pulses which are given to the
control terminals of devices 1 and 4, and let us say these are the pulses which are given
to the devices 2 and 3. Now, when this happens, when you apply these kinds of control
pulses to the various devices what would you observe? At or across terminals A and B
which is supplying power to the one of the phases of the stator windings of the AC
motor. What kind of waveform we will get? So, we will see that if we apply this
waveform, let me just draw another.
So, I will basically get a waveform of this type, which is shown with a black color and it
is; so assuming that we are taking all precautions to keep symmetry between the various
waveforms. So, this is what is going to be the output. And we can just check this here, so
if this voltage is Vdc that we have applied, then when the devices 1 and 4 will be applied
to control pulse and they will be conducting we will find that Vdc will appear across
terminals A and B directly, and that is why this level Vdc will appear here. And when
devices 2 and 3 will be applied the control pulses and then they will conduct 1 and 2, and
24
1 and 4 are not conducting at that time as you can see. Then at that time a minus Vdc will
appear across terminals A and B. This can be very easily verified by assuming the ideal
assumptions, we have made for the special device.
We have assumed that there are no drops, there are no losses in these devices. So, you
will get this negative sign. So, like this this will repeat and you will get this square wave
it is not sinusoidal but it is an AC nevertheless, positive, negative. Now, if I can play
with the frequency if I can bring these pulses control pulses shown with green and a red,
if I can bring them closer or if I can take them away I can play with the frequency. So, by
controlling their width of these pulses I can play with the rms output voltage and by
bringing them together or away I can play with the frequency. So, in effect this kind of a
system it actually gives me starting from a DC voltage it is able to give me a AC output
voltage whose rms voltage I can control and whose frequency I can control. Such a
power electronic system or circuit is called a DC to AC converter or simply an inverter
and has several applications.
Now, let me just give you the last example which will kind of complete this family of; it
will exhaust all the permutations and combinations of the power conversion. So, this one
is actually if we have a primary source of electricity as an AC source. So, AC source is
the primary form of electricity and we are trying to get another level of AC voltage at the
output, then we call it as an AC voltage controller. And I will just explain to you how
this works.
So, just like we have seen the examples of earlier the AC to DC rectifier, then DC to DC
converter and then DC to AC converter. Now, let me just draw an AC to AC converter.
25
So, first of all our controllable device, the special device which I introduced calling the
two terminals the regular terminals as 1 and 2 before between which the currents would
flow and a third terminal which is marked as 3, and it is this terminal which is called the
control terminal and it is to this terminal we apply the control signal and that is when the
device this device starts to begins to operate, begins to conduct.
We assume that this device is capable of withstanding, a large negative voltage between
terminals 1 and 2 we also assumed that the resistance offered by this device as the
current flows from 1 to 2, terminal 1 to 2 is 0. So, we are just considered an ideal device,
and also we assume that the current cannot flow from terminal 2 to 1, it can just flow
from 1 to 2. So, it is a unidirectional device.
Now, let us say that we have this kind of device with us, and we connect it in the
following manner. So, let us say we have this primary form the primary form of power
which is the AC in this case. And let us say we are taking it from the 230 volts, 50 hertz
the mains supply that we get very commonly from the wall socket. And what I do is that
let us say in one of the lines let us in the top lines, I connect this special device I use two
of these and the way I connect them is anti-parallel. So, what I have here is this device
this is terminal 1, this is terminal 2 and this as terminal 3 and the other branch that we
have made here. Let us say this device is again input but now it is the other way round.
So, here 2 is basically connected here, terminal 1 is here, so this devices as you can see
26
has been reversed in the direction in which it can conduct and it has again the terminal 3.
Let me call it as 3 dash the control terminal.
And these are connected here terminals 2 and 1 or the two of these controllable devices
they are connected, and let me just put some load that I want to drive which has a need
for a variable AC supply to be applied across it. So, it is that load. For simplicity for now
just assume that this is just a simple R load, simple R load. Now, such an example where
you would need a variable AC supply to be applied to a load could be for example, a
room heater or it could be a dimmer where the lights for example, in the on the theater
and the stage, people try to dim it or maybe make it bright. So, they apply a variable AC
to it, these are, so there are many examples of this.
So basically because we will see later because, this supply is going to be full of
harmonics. So, basically, we try to operate only crude type of AC loads with it not very
sophisticated because otherwise they would have to take all the harmonics and their life
would reduce. So, such is a situation what I have shown here in this diagram. So, this is
my diagram. Now, let us see what happens, how this would work and what we will get
across the load. Are we really going to get a variable AC supply across the load that we
have actually started that was over basically our objective?
Now, when the AC supply on the input side is positive like this, we will find that the top
device, will have a forward bias across terminals 1 and 2. When the input AC supply is
27
negative at that time the bottom device, that is going to be forward biased, which means
that during the positive half of the AC cycle that we are applying as input. The top
device is ready to conduct and when the voltage applied the AC supply voltage at the
input is negative during the negative half, the bottom device is ready to conduct is ready
to conduct. But as I told you I must supply a control signal to the corresponding control
terminals 3 and 3 dash. Now, let us look at this with the help of the waveforms.
. So, what I have drawn is the input AC supply this is the input AC voltage which is
having 50 hertz or 50 hertz which is applied. Now, as I told you just now, when the AC
input is positive during the positive half the top devices ready to conduct. Now, let us
assume that at is it is at this point, where I give the control pulse to the terminal 3. Now,
what will happen? As far as the load is concerned what will it observe. It will observe the
corresponding part of the input AC supply voltage. So, what will it see? Let me just draw
it. This is typically what it will see.
Now, let us move on to the next half which is the negative half of the input supply
voltage AC supply voltage. Now, by symmetry let us say approximately here is where I
give the control pulse to terminal 3 dash. So, if you look at the previous our diagram
here. So, I gave during the positive half I applied the pulse to 3 and now, I am saying I
will apply the pulse to 3 dash when the input supply voltage is negative.
So, when I do that now, this device will conduct the lower the bottom device will
conduct; and what will the load see? What voltage will appear across the load? It will be
something like this. Now, I again in the next half cycle which is the positives half cycle,
I again give the control pulse to terminal 3 at this point and I again get this kind of
voltage across the load. So, if you see carefully this waveform the green waveform and if
I could maybe make it a little dark, and also a bit and also continuous. Then what we
observe is that in this kind of circuit which we are calling as an AC voltage controller
AC to AC, but AC of one level to the AC of another level.
So, what we mean by this is that if you see this green waveform, the green waveform is
as you can see both positive and negative, so it is AC. But if you do a root mean square if
you try to find out the root mean square value of the AC voltage which is in green you
will find that that if I just say because it is appearing at the output across the load if I just
denoted by Vo; so Vo rms is going to be much less than or less than the input supply rms
28
voltage. Now, it is very easy to see that as we move this point at which I am triggering or
I am controlling the corresponding devices into conduction if I just move these if I shift
these points left or right, I am going to vary I will be able to vary the rms voltage that
appears across the load. So, this is how I am going to control the voltage that will come
across the load.
So, what would be the, typically what should be the control point or what should be the
trigger point to the control pulse or to the control terminal if I want to get the maximum
rms output voltage across the load. The answer is that I should make this almost here 0.
So, it there should be no delay from this point up to the point I trigger of the control
terminal, likewise here, so it should actually move all the way up to this point this point
you should move all the way here. And as I need less and less rms voltage at the output
across the load I can keep on going on the right side. And you can see that if actually we
were to trigger at this point, particularly for an ideal resistive case you will find that we
will actually end up with a 0 voltage across the load. Now, this is a principle which is
very commonly used in today’s fan regulators.
So, I will just show you one fan regulator. Today a modern fan regulator which is used in
almost all the households; so, you can see this there is a there is a knob here which you
can move and it has all these gradations 0 1 2 3 4 and so on and 5 which is the
29
maximum. So, what it does is, now inside this; now what I want to do is I am just trying
to apply a different rms AC voltage to the fan that is what I am doing with this regulator.
Now, in the earlier ones, in the earlier regulators that we used the big ones, we were just
using a resistive potential divider system and as I probably mentioned in one of my
earlier lectures you actually end up really dissipating lot of power when we use that kind
of an arrangement to control the speed of the fan but now these modern fan speed
regulators they have nothing but as you can guess now, from the various things which I
am saying that it is going to I have something like this arrangement inside. So, if you see
here there is this arrangement of this very special three-terminal device which is
connected in a specific way, and later on I will tell you that this is called this device is
called a Triac.
And we will see what its structure is, but the principle is going to be somewhat what I
have already told you, this is what. So, basically our fan regulator modern fan regulator
today it actually uses a circuit like this one which is shown here, here which is shown
here and just by varying this knob you are basically controlling the that delay at which
you are tricking the control terminals, the various control terminals. So, you can see that
if you are at 0 your delay is maximum you have taken it all the way to the right side at
180 degree point, and when you want to increase the speed you try to bring that control
point the trigger point towards the left more left. Then you go to 2, more left then you
further increase the output AC rms voltage, you further increase and this is the maximum
you can get. So, actually here you are applying the full blast of the input voltage across
the fan.
Having reviewed the AC voltage controllers and all the other basic types of power
electronic converters, from the next time we will now get a bit formal and look at some
other details of power electronic systems.
I thank you for your attention.
30
Prof. Vivek Agarwal
Lecture – 03
Applications, Definitions and Nature of Power Electronic Circuits
Hello and welcome to Fundamentals of Power Electronics. So, in the earlier lectures, I
introduced you to some of the basic day to day life examples, pertaining to power
electronics. The idea was to familiarize you about the usefulness and importance of
power electronics. Now, I want to talk to you a little bit about some advanced
applications, just to motivate you a little bit more.
So, what you see on your screen right now is the mars exploration rover, which is on the
mars surface. You can see 6 wheels there actually. Are you aware how these wheels are
driven and how they are controlled? There is a very extensive use of power electronics in
the driving of these wheels, in the control of these wheels each of which is having a
dedicated motor which is controlled power electronically. Apart from the wheels, there
are several other tasks which are performed by power electronics. For example, rock
drilling for experimentation, it controls cameras, their orientation and so on. So, they are
all actually carried out by power electronics.
31
Now, let us look at this Mars orbiter, our own India’s own Mars orbiter, the Mangalyaan
which has been orbiting the Mars. Now, what you see at the front are the solar panels
which are used to power this orbiter.
As is the case the electricity produced by the renewable energy sources such as the solar
photovoltaics that is seldom compatible with the loads with the electrical loads which are
supposed to be fed from it, and same is the case here on this orbiter. So, the electricity
produced by the PV panels is actually conditioned appropriately, to actually get it in the
right form DC or AC and the right level of the voltage before it is applied to the various
nodes inside this orbiter. All this is done with the help of power electronics.
32
Now, the next image shows you robotic arm that is used for the repair of space shuttle.
So, there are several drives, electric drives which are controlled through power
electronics which are working in this arm.
After some of these space applications let me show you some more familiar applications
of power electronics such as the home appliances. So, what you see is a bunch of kitchen
appliances like the mixies, and the grinders, and you see certain other appliances such as
the refrigerator, and the air conditioner. All of these they use power electronics. To
33
increase for example the efficiency, to reduce the energy consumption, the motor that is
used for the compressors in the refrigerators, and the air conditioners is nowadays
controlled with the help of power electronics.
Same is the case with the motors which are used in a mixie for example, so that is done
through power electronics control and we try to conserve the energy. And we also try to
improve the power quality the power drawn from the socket by these machines, so that
they do not spill distorted waveforms into the power system.
Another very famous example is of the electric vehicle, Now, what you see here is an
electric vehicle which is parked and it is actually being charged. Inside this electric
vehicle we do not have the conventional internal combustion engines which are run by
gasoline, by the petroleum products. Rather there is basically one or two or more electric
motors which are driven by the battery, and the battery is being charged through
charging stations. Such an arrangement obviously, does not cause any pollution in the
environment environmental pollution which is a matter of great worry these days and in
future we can expect many of these vehicles to be actually electric vehicles which are
using electric motors.
34
The picture shows electric, locomotive it shows and you can see the pento graphs on the
roof of the locomotive which are trying to trap the electric power from the overhead
lines. This power is tapped through the pento graphs and inside the locomotive this
power is appropriately conditioned and applied to the motors which are then used to
drive the wheels of the locomotive. There are several other applications in inside this
locomotive. Inside these electric trains which are all actually managed through power
electronics for their high efficiency and compactness.
Moving to another arena, now let us look at some of the examples of the renewable
energy systems. Now, everyone is concerned about the deterioration in our environment
quality, the ozone layer is depleting and people are actually exposed to harmful
radiations. So, there are health hazards. And many organizations who are actually
concerned about this have already raised and around. So, there is a constant search for
alternative energy options which will actually reduce or eliminate the use of for example,
thermal power plants which run on coal and emit a lot of toxic gases into the
environment.
35
What you see in this picture is a series of windmills, the windmills they are connected,
the shaft is connected to the electrical generators which generate AC power. And
because this AC power does not have a fixed frequency it varies with the speed of the
wind, it is converted into a DC form of power by using rectifiers and then using another
power electronic stage a power converter which converts DC to AC. This is converted
into AC and then actually it is mixed with the power grid to transformers.
Here is another example of a renewable energy system in which a solar PV panel is

mounted on a rooftop. Since the electricity produced by solar panel is a DC electricity, it
must be converted into an AC form using an appropriate DC to AC converter
configuration, so that it can be either used for these AC loads or it can be fed into the
grid.
36
Now, you will notice that apart from the fuse box which is used for protection there is
also a meter. The specialty of this meter is that it can actually measure power in both the
directions. So, basically you can measure the power which is coming from the grid into
the household or you can measure the power which is going into the grid from this
household through this solar PV system. Hence you are able to do what is called net
metering and are able to calculate the tariffs appropriately.
37
Having seen several examples, now let us try to see how we can define power
electronics, we would like to actually assign a formal definition to power electronics. So,
in the very simplest form power electronics is actually concerned with the processing or
the conversion of electric power, through electronic means. When we say power
processing or power conversion, we mean converting the electric power from one form
to another, that is AC to DC or DC to AC or changing the voltage, the current or the
frequency or a combination of these. So, all this actually constitutes what is called power
processing or power conversion, and this is what power electronics does for us, using the
electronics and using the control.
Now, the definition of power electronics has taken various forms over the years. There
are various authors, who have actually defined power electronics one way or the other.
Now, these definitions might appear different in terms of words, but they all actually
amount to the same basic definition.
Professor William Newell says that power electronics deals with the use of electronics
for the control and conversion of large amounts of electrical power. So, this is actually
one of the earliest definitions, given by William Newell, who is considered to be a very
towering personality in power electronics. He seems to emphasize the use of electronics
for the control and conversion of power.
38
Then Peter Wood has described in his book, power electronics to be a branch of
electrical and electronic engineering that is concerned with the analysis, design,
manufacturing, and application of switching power converters. So, not only that he has
given it the definition a different form, various brought in analysis design and
manufacturing, he is also focusing on the application and on the switching power
converters, which are actually the heart of any power electronic system.
Professor Ned Mohan says that, in broad terms the task of electronics is to process and
control the flow of electric energy by supplying voltages and currents in a form that is
optimally suited for user loads. So, he is actually stressing on the importance of the right
configuration and the right kind of control, which can then provide the exactly the right
form of current and voltage, which a load needs.
Likewise, there are many other definitions, one of the definitions that I will like to quote
is from Professor Wilson, which says power electronics is a technology associated with
the efficient conversion control and conditioning of electric power by static means, from
its available input form into the desired electrical output form. So, this is what is the
definition given by Professor Wilson, in one of his papers.
Now, here in this definition, he seems to emphasize something that is different from
other definitions, he seems to emphasize when efficient conversion. So, basically we are
trying to see; that means, less energy loss and also he is talking about static means. So,
39
basically he seems to be ruling out anything that will be used for energy conversion
which will have, let us say, some wear and tear associated with it, because of its moving
nature.
Like for example, rotary converters are one such possibility to do power conversion but
then they are less efficient and they also have moving parts and are, therefore, subject to
wear and tear, which reduces the overall life and reliability. So, Professor Wilsons
obviously, emphasis seems to be on a lot of efficiency, reliability, and longevity, and
also availability of the power electronic system.
Now, one thing that is very important about power electronics that I would like to share
with you is that it is truly a multi-disciplinary area. One of the very important things,
which Peter Wood says in his book is that the reason people use power electronics, the
reason people need power electronics is because people do not use electricity. What it
means is that, we use electricity not directly as electricity. We actually convert this into
mechanical energy or heat or light. So, therefore, there is a need for not only power
conversion but also to control the rate of power conversion.
So, you should be able to control for example the intensity of the light, you should be
able to control the heat. For example, you should be able to control the speed of the
motor. So, basically this is where the power electronics lies and the power converters
seem to do just this, they are able to do power conversion, they are able to change the
form either AC to DC or DC to AC or they can keep the form same, they change the
levels of the voltage for example, or they change the frequency plus they also almost
always are able to control the rate of energy conversion.
Now, looking at the power electronics as an interdisciplinary field we can actually divide
electrical engineering into three major disciplines - the power engineering, the control
engineering and the electronics engineering.
40
And actually, we will see that the power electronics borrows concepts from the power
engineering field, the control engineering field, as well as the electronics engineering
field and that is how the entity of power electronics comes into being. So, you can see
that in this picture, power electronics is shown as an overlap of the three main sub
disciplines of electrical engineering, the power, the control and the electronics
engineering.
The electronics engineering deals with the devices, power devices which are used as
switches and also with circuits which are used for processing information. The power
engineering which consists of rotating equipment and static equipment related to the
generation, transmission and distribution as well as utilization of electric power and
power electronics, actually plays a role big role in that as well. Similarly control, so
power electronics uses several concepts of control, there are many, in fact, the majority
of power electronic systems are closed loop systems and the control engineering
concepts are required to control these systems, to understand the stability issues and to be
able to see their response characteristics.
41
When we talk about electricity, we talk about three phase or poly phase AC or one-phase
AC or DC. However, it is a poly phase AC which is the most general form and other
forms can be derived from it. While one-phase AC is clearly a subset of poly phase AC,
the DC form can be thought of as a special case of one-phase AC with zero frequency.
Keeping this in mind it is possible to represent any type of power converter by a
switching matrix, a matrix of switches.
Now, let us say that we have m input lines and n output lines of a power converter
system. Now, on the screen what you see is a special case, where m is equal to 4 and n is
equal to 4 is considered. So, you have 4 input lines and 4 output lines. So, the way
actually this whole thing is configured is shown. I will just draw this diagram and show
that.
42
So, these are the m input lines and with the red pen let me just draw the output lines. So,
these are n output lines and m input lines. Right now, they are just actually plays in this
crisscross fashion. Now, if I connect at each of the points where the input and the output
lines meet, one switch.
So, then I would actually have a switch, which will be here. So, we will have all these
switches and so on. So, basically if there are m input lines and n output lines, I can say
that there would be m*n switches, which will be required to complete this matrix. Then
whenever any switch is closed or turned on, we can say that that particular input line is
connected to the output line. So, for example, if I choose this particular switch, if I
connect or if I turn on this switch, I can say that this line if I can just if I just mark this as
Vsm that is connected to this line which I can just mark as I1 its connected to there.
Now, since the devices the switching devices are usually unilateral, they are only
conducting in one direction, and if we want the power to actually be able to travel both
ways, in both direction, we actually must also connect one more switch along with each
of these switches in anti parallel which means that it will be having the other polarity.
So, basically the number of switches will now, become 2 m*n.
Now, in many converter systems there is a possibility that one of the lines like for
example, this line in the input and this line at the output, they might be connected as a
common ground reference. And in that case, one of the switches which will be
43
connecting this line and this line will actually not be needed, and this actually
corresponds to the degeneration of this line.
So, we will actually have at the max 2*m*n number of switches which will be required.
In the switching matrix, which actually represents a power converter, it actually uses a
combination of the input lines and it produces an output at the output lines.
So, basically it depends on which of the switches are activated, what kind of control
signals are applied and what type of switches are used. Now, one of the requirements of
high efficiency is that you cannot have the basic control element the switch in this case
to be actually a continuous variable, you would actually want this to be working as a
switch, as an on off device and it cannot be work as a continuous variable, because it will
then result in loss of energy and will actually result in poor efficiency. Therefore, we
must always operate these switches or these switches are always operated in what is
called switched mode or on off manner.
Now, what is the effect of this? What is the effect of having this basic control element
the switch operating in the on off, in the discontinuous manner, not as a continuous
variable? What is the result of this? The result is that, what we get at the output, it is
synthesized to appear at the output, it may be close to what is desired but it is not
exactly what is desired. The reason is that these switchings, the on off switchings or the
44
switched mode operation of these switches as we will see later, it actually causes certain
gaps to appear.
Likewise, these gaps may also cause distorted currents to be drawn from the input side,
which are which is not desirable for the supply. So, the result is or the consequence of
this is that we use filters. So, you see in the in this picture that you see on your screen,
you see an output filter and an input filter.
Now, what is a filter? A filter is nothing but a combination of the energy storing
elements, such as L and C, which are appropriately actually connected and they form a
filter. These energy storing elements, they store energy and whenever it is available and
then they use it this energy to fill this gap these values which might appear due to
switching of the switching matrix. So, that your output actually approaches the desired
form.
Similarly, on the input side this filter, the filter components the use store energy to
supply those components, those values, in the waveforms they try to fill those values
which would have otherwise been drawn from the input side, which is not desirable.
Hence filters are very important. But at the same time, filters contribute to the weight
then the cost and the volume of the system because they are made up of magnetic
components such as L.
45
Now, one of the things on which the filter size would depend is the rating of the system
of course, the other way is the frequency of the system. So, if you are operating these
switches at high frequency, the gaps that will appear in the output or the input
waveforms, input current waveforms and the output voltage waveforms. They would be
very small, these gaps will be small and they would need much less energy and hence
much smaller L and C components, to fill those gaps.
Therefore, operating a power electronic system above much above a nominal frequency
or operating it at a higher frequency is highly desirable because it helps us to reduce the
size, the weight and the volume of the filters. But it must be remembered that when we
increase the switching frequency or when we increase the operating frequency, we
actually are increases our losses which takes place when switching devices is switched,
we will see these details a little later in the course.
Now, an important aspect of the switches, though we have not done the switches in
details here. But I must tell you at this point something that, I also mentioned in the first
one or two lectures, that there are basically two types of devices, the switching devices,
and today most of these switching devices, they are all semiconductor devices, they are
all made up of semiconductors.
Uncontrolled and controlled, these are the two types of switches, when we talk of about
diodes, these are uncontrolled switches. And we call them type one devices, when we
46
talk about a thyristor and or when we talk about a bipolar junction transistor or a metal
oxide semiconductor field effect transistor or we talk about an insulated gate bipolar
transistor, we are talking about the control switch. But within the control switch there are
two categories, one of them is the thyristors, which we call as type two devices and the
other is the other switches such as the BJTs, the IGBTs, and the MOSFETs which we
call as type 3 devices.
Now, the difference is that with a thyristor has a forward voltage blocking capability, so
till a gating signal is applied it would not start conducting but the problem is that it
becomes like a car without a break. Once it starts when it actually once it has a current
which is gone beyond the latching value, then the thyristor does not switch off, till this
current is brought to 0 and for a stipulated time, a reverse voltage is applied across the
device. It cannot be turned off through its control terminal.
Now, this is not the case with other devices like MOSFETs, BJTs, and the IGBTs which
can be simply turned on as well as off, by applying appropriate control signals to their
control terminals. A diode on the other hand is a completely uncontrolled device, it has
no forward blocking capability; the moment a forward voltage comes across the diode it
conducts. So, we have all these types of switches, which are used in power electronic
circuits.
47
Now, very briefly now we will come to this point again, we will just briefly see how a
typical power electronic circuit looks like. So, here is a typical power electronics system.
So, what you see at the heart or at the center of the system is a power converter and as
we have been every, as we have seen before it is able to change the form of electricity.
So, you can see that it mentions DC to AC, AC to DC, DC to DC and AC to AC, these
are the various forms which the power converter is able to convert. Now, at the input
side and the output side, are the filters about which we discussed some time back. These
are the storing element, elements L and C, which are connected in some combination, LC
filters, as LC filters which actually are able to fill up the valleys and prevent the gaps,
and reach and actually result in the desirable wave forms.
Apart from this there is a controller which is shown at the bottom, which actually takes a
feedback signal from the output site, from the load side. This is a typical feedback
system I briefly talked to you about the output voltage regulation in one of the earlier
lectures. And this is just exactly a feedback that is taking place as part of some sort of a
regulation. It also, the controller also accepts external control inputs or into signals such
as for example, set points. So, if, we want to have the fan running at a certain speed, so
we can actually just move the knob and we can just externally adjust the set point and
then the output actually tries to follow that.
You also see on the diagram one input signal which is drawn from the input of the source
side of the system this could be for example, the AC waveform which is feeding this
entire power electronics system and you are using these voltage information voltage
templates, just because you need to decide the switching instants for the devices which
are present inside the power converter, based on this information that you get in terms of
the input voltage AC voltage.
While discussing the AC voltage controller, in one of the initial lectures, I mentioned
about the delay and advancement in the control pulse application to the device to control
the output voltage. And, there are other applications like for example, the input signals
are also used, when somebody is trying to do what is called a feed forward control. So,
we will see some of these aspects later in the course. But this actually diagram should
give you an idea that how a typical power electronics system looks like.
48
Now, one way of categorizing power electronics is by looking at this diagram which
shows the three space three-dimensional space of voltage, current and frequency. Now,
what it means is that, basically it means that, if I consider a basic device which is, let us
say existing today with maximum current and voltage capability. And then I say that, I
just use this device for my power converter, then the resulting power electronic system
actually will come in what is called medium power domain. The unit, the device which
are available they are actually compatible for their basic rating, the basic highest rating
which is available.
Now, it is not uncommon that we want to actually go to sometimes higher voltages and
higher current ratings. So, in which case these devices can be connected in series and
parallel, they connected in series to get a higher voltage and there they result in as I will
show you. So, this is, the power level that is achievable in this domain which is the
medium power domain the basic device with the highest rating of current and voltage.
But now, if I just add in series, too such basic devices, I actually go into what is called
the high voltage domain, which is actually shown on it on this side.
And it is also not uncommon to parallel the devices, to increase their current ratings in
which case I go to the high current domain. And sometimes I want to do both series and
parallel combination of the devices, to go to both high current and high voltage domain,
which actually leads me to what is called the high-power domain, which is several mega
49
watts as you can see from here. Now, this is one way in which we can actually categorize
power electronics it forms in the medium power domain, high power domain or we can
say that high current domain, high voltage domain, high current domain and medium
power domain and high-power domain, these are the four.
Similarly, the frequency also plays a role what you see typically is that when you are
actually going for very high frequency operation usually the power levels are less from
here this diagram; what you see here this one. So, the power levels are actually less. So,
there are all sorts of devices which are coming up almost every with different speeds,
with improved speeds, with improved current and voltage ratings. So, therefore, there is
nothing sacrosanct about the values used in this graph. The idea is to show you that how
qualitative categorization or the various power electronic systems or power electronic
circuits can be done using the voltage, current and frequency information or capability of
the devices.
Now, a special comment about, and AC to DC converter has been, it is among the first
circuit it is called actually a rectifier, it is among the first power electronic circuits, which
were developed in early 19th and early 20th century. And they actually are more
interesting because they have several implications, and they have several applications.
Now, what you see of this diagram this is a very important aspect associated with power
converters.
50
You know that is the capability to operate in how many quadrants. Now, these quadrants
are defined by the voltage and the current that appear at the terminals of interest. Now, in
an AC to DC rectifier we obviously, have one port with AC the other port with DC. So,
if I look at the DC side, and I look at the AC, if I look at the voltage and the current then
if both the voltage and the current are positive, it basically means that the power is going
from the AC side to the DC side. So, this is actually a rectification which is taking place.
It is also called a positive rectification.
But if you go to quadrant 3, you will find that the voltage as well as the current, they are
negative, both are negative. So, the power is still because the product of current and
voltage is still positive, so the power is still flowing from AC to DC, and the rectification
is taking place, but here we call it negative rectification.
Now, what you see in the second and the fourth quadrants is what is called a
synchronous inversion. Now, what is happening here is that you will find that the power
that you see the DC terminals is actually negative, because either the voltage is negative
or the current is now negative, but both of them are not negative or positive. So, you
have a negative sign coming with power, which means that the power is now flowing
from the DC side to the AC side. Now, this is something which was actually called as
inverter by people like prince, who was the first to invent such rectifier circuits. But
actually this inversion is not to be confused with the dedicated DC to AC conversion,
which takes place and for which we need either force commutation of the specific
switches such as thyristors or we need to use certain even advanced switches like the
BJTs and the IGBTs or the MOSFETs. During the synchronous inversion, the thyristors
that is type two switches undergo natural commutation.
Now, having talked about the various quadrant of operation, having talked a little bit
about the devices and their capability. Now, let us try to see what are the various possible
types of power converters which are realizable, and how they are associated with the
switch properties the switches which are used in making the power converters or the
power matrix or the switching matrix. Now, the first quadrant rectifiers, which are
having both voltage and the current is positive, they actually can be realized with just
diodes, and we just call them as type 1 switches. The diodes is a type one switch or the
uncontrolled switches.
51
But if you go to the second and the fourth quadrant AC to DC conversion operation,
there you can use you will need type 2 switches such as the thyristors, as we will see.
Several frequency changers and AC voltage regulators, converters that you can do with
type 2 switches as well as with part of these can be realized with force commutated
inverters.
Similarly, the DC switching and the regulators, the DC to DC converters that is, they are
realized with the type three switches the IGBTs, the MOSFET from the BJTs or with the
type two switches, the thyristors which are actually subjected to forced commutation. So,
this is how we can actually define the various types of power converters and associate
them with switch properties.
52
Now, what you see on the screen is another representation for power converter. Now,
clearly this is very different from the switching matrix representation of the power
converter, it is a different approach. Now, what you see here is that a power electronic
converter, of static nature it can be represented something that lies between the source
and the loads and each of these paths, which connects the source to the loads, is actually
having a variable impedance, an impedance that varies and we need to now see, we will
see now, what are the property of these impedances, what is the property of this variable
impedance network.
53
So, the property of these of these impedances or the switching of impedance should be
actually what nature it should have, that is depicted by these pictures. So, what you see
here a resistive switch, which changes the impedance from very high to very low value,
and that is shown in as a first image on the left, first figure on the left side.
Similarly, there is a possibility of a switch which is an inductive switch which actually

switches an inductive impedance to a very high value, from a very low value. And here
this is also called a saturable reactor because actually what we do is when we want the iL
to be extremely when we want the inductance will be extremely low we saturate the
inductor and when we want this to be very high the inductance to be very high, we bring
it out of saturation.
So, this works this is as inductive switch and similarly a capacitive switch. So, the
various power converters, that we have the four basic forms they can make use of one of
these three types of impedance switchings. It is important to know that high efficiency
demands that the switching of the impedance should be as fast as possible.
So, for example, if you are talking about a DC to DC converter or if you are talking
about a DC to AC converter, this negative resistance active switch, which actually
switches between a very high value of resistance and a very low value of resistance, is
shown. You can see that we have drawn the same switch characteristics on the right side.
So, this is basically, so, you can see here the characteristics of the switch that we have
54
also seen in the previous slide and this is a switch, this is the switch. Just we talked about
several switches, which will be placing at the intersection of the input and output lines in
a switching matrix. This is that switch.
Now, it can be shown that if this resistive switch which is realized through a transistor, a
bipolar junction transistor is actually switching between two operating points V1 I1 and
V2 I2; it is possible to show that the pulsating power which can be processed by the
switch, by this switch is given by this expression
Pac = d1( 1-d1 ) (V2-V1) (I2-I1)
(V2-V1) and (I2-I1) we are switching between the two extremes of the voltage rating and
the current rating of the device, then obviously, this is the complete V I rating of the
device. So, if you have d1 and (1-d1) is 0.5 you actually have this maximum power that
can be processed by the switch which is given by this expression.
Now, this is a little involved proof, which can be shown, but if you see graphically on the
right side on this picture, this particular rectangle which is formed by this, this area
which is one by fourth of this this particular rectangle, that constitutes the maximum
power that can be processed by the switch.
Now, here is an example of an inductive switching. So, what you see here on the right
side is a load which is supplied by a Vac or source and on the left side we have coupled a
55
DC source along with a control resistor. So, actually you apply a DC current into this
coupled inductors. So, you can see here, these coupled inductors. You are actually trying
to inject through this a DC current and in the process what we do is we saturate the core.
In the positive half cycle, at a certain AC voltage one of the cores saturates and in the
negative half, the other core saturates. Since the two windings are in parallel, when the
core saturates, a low inductive impedance is offered. And you actually have the power
that is flowing into the into the load from Vac and when you do not have saturation, then
this actually offers a very large inductive impedance and actually nothing can flow into
the load at that time. So, by using a very small DC control here, DC level control here,
we are able to actually control large chunks of AC power. Now, this is highly suitable for
AC to AC power conversion.
This is another way in which a magnetic amplifier can be coupled. Here you can see that
the load side circuit is not connected in parallel, it is actually connected in series. And
this was widely used this concept of amplification was widely used in magnetic
amplifiers, which became very popular in the 1920s, 30s and had played a major role in
the World Wars, World War II, many of the applications were actually run with the help
of magnetic amplifiers. However, the magnetic amplifiers importance, it waned when
people actually brought in the vacuum tubes and the vapor tubes into the market in in
around the same time in 1930s and 40s.
56
So, just to quickly summarize that you can actually have AC to AC conversion, when
you can have a variable resistance or capacitance or inductance. We saw the example of
a magnetic amplifier, where it was an inductive impedance that was switched from the
maximum value to the minimum value, and vice versa, to control power.
Likewise, you want to have a rectification like AC to DC, we will have to use these non-
linear resistors. So, if you look at this graph, it shows you the property of such a device.
So, the diodes, for example, in which case will be an uncontrolled rectification and the
thyristor there see as will have these kinds of characteristics, and they actually lead to
controlled rectification. Because we are having an AC source, the use of thyristor is fine
because it actually provides opportunity for natural commutation of these devices.
When we are talking about DC to AC conversion, in fact, even in DC to DC conversion

the use of a negative resistance device, the kind of one which we saw just now, some
time back, which switches between a very high resistance to a very low resistance, just
like a BJT example that we gave, they can be used for DC to AC applications. And in
DC to DC applications also, we have to understand that there is an intermediate stage,
which actually pierce a DC voltage, into a pulsating power first. And the same principles
apply for the switch, as we have discussed before.
So, with this I end this particular session. And in the next one, we try to go back to the
diagram of the power electronics system, and we will try to see what the various
57
components of this power electronics system are and how they work. We will look at
some of the basic aspects.
Thank you very much for your attention.
58
Prof. Vivek Agarwal
Lecture – 04
Components of a Typical Power Electronic System
Welcome back to Fundamentals of Power Electronics. Any power electronic system

consists of two stages, a power stage and the control stage. The power stage, which is a
very common name given to the power circuit or the power electronic system, it has the
switching devices including diodes, it has got resistors; also whether we want them or
not, we do need resistors for various applications.
We have capacitors and now these days even super capacitors have become important
and batteries are also playing role, a big role in several power electronics systems these
days. The other components are the inductors and the transformers. The control stage, it
actually is mostly electronics, it consists of controllers, sensors, device drivers, signal
conditioning circuits and most of them are actually in the integrated circuit IC form and
it also includes these days, surface mount technology devices. So, this is all you can
expect in a power electronic system to be present, a modern power electronic system.
59
Here, you see the various packages in which the power devices are available. So, what
you see towards the back side are two IGBT modules. So, these are fully controllable
devices, meaning that they can be controlled; through their control terminals whether it is
turning on or turning off, they can be completely controlled through this controlling
terminal.
And , so these are like available either as module, sometimes they are actually having a
module which is consisting of 2, such devices IGBT’s they also have modules, which are
available with full 6 devices. So, that they can form some very standard power electronic
circuits; such as the DC to AC inverters. Now what you see at the forefront are some of
the other packs like the TO 220 packs, which actually are showing various types of
devices. We have the transistors here, bipolar junction transistors; we have the metal
oxide semiconductor field effect transistors; we have on the very right side a stud type
diode.
So, we have various forms in which these devices are available and which are actually
used in all circuit layout depending on the requirement and the application.
60
This is another picture of showing the power devices. On the left side is a thyristor;
thyristors still remains the device for the highest power levels required in the power
industry. So, though the thyristors have a very distinct disadvantage, in this in the sense
that they cannot be turned off; after they have been turned on, it is not possible to turn
them off, without using forced commutation or unless there is a natural switching which
is unless there is a natural commutation which is taking place.
Type two switches, thyristors or the SCRs, once triggered into conduction cannot be
turned off through the gate control; they can only be turned off. When the current
through them goes to 0 and a reverse bias voltage appears across the device for a
minimum duration of time. This is often achieved by using commutation circuits; by
commutation we mean, that we are trying to divert the current through the thyristor into
an alternative path, so, that the thyristor can be switched off, this is called the process of
commutation.
Still these devices they remain important and that is why it is important to familiarize
you with such devices. On the right side for comparison, we just put a TO 220 package, a
plastic package of a very low power, it is a 13 ampere 600 volts thyristor device, while
the one on the let just stud type the left side one, is actually a 800 volt 100 ampere
thyristor device.
61
So, this is a comparison between the two and they have been shown in the backdrop for
comparison with a lead pencil. So, you can just estimate their size; heat sink is another
important component of the power stage. Heat sink is a high heat capacity body on
which the power device is placed. This allows the power device to stabilize the
temperature, the heat sink is usually a metallic material which provides a high heat
capacity.
So, that the device temperature does not go very high in a short time and it gets the
necessary time to stabilize with respect to temperature before experiencing very high
temperatures.
The slide shows the power devices mounted on the heat sink. You can see very clearly
the fins which are actually provided to facilitate the easy passage of heat into the
ambient.
62
The second slide shows, a component which is mounted on the heat sink along with a fan
which is used for forced cooling. In several applications, a fan is required for forced
cooling, apart from a heat sink.
Now, resistors they are used for various purposes ; the resistors are used for example, as
snubbers, they are used for voltage balancing, they are used for current measurements,
we all know that we use current shunts and then the voltage that appear across these very
low resistance current shunts is then used to present a current. They are also used for
damping; if there is a unnecessary energy surplus energy which is actually circulating in
a system; then you use these resistors to dump it; less it actually kills some other useful
component of the circuit. So, these are some of the not all, but some of the very
important applications of resistors in power electronic systems; power electronic circuits.
63
Now the basic relationship of resistance remains the same irrespective of what is the type
of the resistor and how it is made. So, the resistance is given by this formula, where if I
consider, this as the as a solid bar and I say that this is the area of cross section through
which a current is flowing i and it has got a length l and ρ is actually the property of the
material which actually makes up this bar this rod. So, the resistance is given by the
product of resistivity into the length by the area of cross section through which the
current is passing.
Now, depending on the application it can, various types of resistors, they are they are
available the ones for power applications they are typically can be divided into three
64
varieties. One of them is a solid carbon composition resistors, the other is the film
resistor and the third one has wire wound resistors.
Solid carbon composition resistors are made by crushing the carbon and ceramic and
mixing them in the right proportion to achieve the desired resistance value. However,
these are not very accurate values and could vary up to 5 to 10 percent. The film type and
the wire wound resistors have a cylindrical or tubular core made up of pure ceramic.
For the film type resistors, the carbon or the metal film is placed around the ceramic
body and precisely cut in the form of helical grooves; which gives very accurate
resistance values up to 0.1 percent tolerance. The wire wound type resistors use the
nickel chromium or the nichrome wire. One of the specialty of the nichrome wire is that
its resistance does not change much with the changing temperature. The wire wound
resistors can give an accuracy of around 5 percent.
Another issue with wire wound resistors is that they also have a large parasitic
inductance. Therefore, they cannot be used in applications where the parasitic inductance
might be a concern in the working or the measurements. One of the properties that we
are looking at resistors is they know that their resistance do not change significantly with
temperature.
They are they are they are having a good noise characteristic they have a good voltage
rating. So, basically when we apply the material when we use the material in a resistor,
its dielectric strength should be enough that it does not break down, when we apply the
voltage that you need to apply to a given circuit. So, basically these are some of the
tradeoffs when you use these different resistors.
65
Now, I will just show you some pictures. So, these are the solid carbon composition
resistors about which I mentioned, which are actually made with the mixture of ceramic
and carbon and the ratio of the carbon and the ceramic decides the resistance values. So,
these are some of the solid carbon composition resistors and you can see the colored
bands. And these colored bands are as per a code which can be decoded and they actually
represent the value of the resistance represented by each of these resistor. So, there are
golden and silver bands which actually tell whether these are 5 percent or 10 percent
tolerant resistors.
66
Here is a film resistor, and you can see that there is a carbon coating which has been
done around the core; a ceramic core, on the 2 sides you can see the nickel caps; the
nickel metal caps. And you can see the grooving’s made in a helical structure, and these
are very accurately made with the help of computers. So, you can actually very precisely
control the value of resistance. We can also see the leads how they have been fused with
the caps of the two sides and how this resistor has been realized.
These are some examples of the wire wound resistors, their various types what you see
on the lower left corner is actually a variable resistor. So, it is basically used as a
rheostat. And what you see on the right side right bottom corner is actually a wire wound
resistor which is actually mounted on a heat sink; because a lot of heat is actually
dissipated when the resistor operates because of i2r. And there are some resistance
particularly high power ones and wire wound resistors are used for higher power rating
as well. So, that is why these heat sinks are also there. So, these heat sink option wire
wound resistors are also available.
67
The other important element about which we are going to talk about is a capacitor; and a
capacitor it is nothing, but when you have two conductors; which are lying close to each
other, one of them is having a set of plus charge and the other is having. So, these are let
us say two conductors as I have just drawn here, let us say there are some positive
charges here and there are some negative charges here and you actually separate them
through an insulator, such as a dielectric.
If you put that there then this actually amounts to a capacitor. The voltage of this is V1
and the voltage of V of the second conductor is V2 then the potential difference between
the two is V1-V2. Now we can actually very easily show that the charge Q, where, Q
actually is a positive Q charge on the left conductor and a negative Q charge on the right
conductor, Q by V is actually a constant.
We can show that, this Q by V is a constant and actually this is what represents
capacitance. So, the capacitance is actually defined by Q by V and we can see that for
very low values of voltage, a given capacitance can hold more charge little mathematical
manipulations. We can easily show, that the capacitance is given by
C = εo A /d
68
So, if you just see the slide on your screen. It actually shows a very special form not the
very general form that I showed you in the beginning with 2 conductors. Now these are 2
parallel plates which are carrying a charge plus Q and minus Q and there is a electric
field; which goes from positive side to the negative side E and there is a potential
between the two which is given by V. Then, if the other dimensions d, that is the distance
between the 2 plates and A the area the surface area of these plates is known, then the
capacitance is given by
69
C = εo A /d
Now, if there is air between these plates. This is the capacitance that you get, but if you
actually insert a dielectric; which is actually having several dipoles which actually get
aligned when you place them in between the field that is created between the 2 plates,
they actually get aligned. And they actually, result in the net reduction of this electric
field which actually has a effect of causing an increase in the capacitance. So, basically
your capacitor now becomes C
C = K εo A /d
So, if you see this expression, K epsilon naught A by d ; it is K times C which was the
original capacitance with air used as dielectric. So, the basically we want to use a
dielectric, which will have a dielectric strength, which can be are a higher voltage
without breaking down or also it will give a higher capacitance. It will be able to give me
a higher length dielectric constant which will be then scaling up my capacitance value
drastically. So, actually I will be able to reduce the size for a small area or within a small
area I can then pack a bigger capacitance that is the main idea.
Now there are 3 types of major types of capacitors which are used in power electronic
systems depending on the specific applications. One of the things is that the dielectrics
when you insert between the 2 plates, it actually incur losses. Therefore, they are
specified as with RMS current rating, because the i2r losses are there and also dielectric
losses are there.
There are two important points when we talk about capacitor one, when we talk about
the dielectric strength, the e field strength at which it will break down. So, you would
actually apply how much potential before the electric field between the plates, it breaks
down. So, you actually need to use a material dielectric material that will not happen.
The other is the dielectric constant which actually ends up multiplying the capacitance
net capacitance value of a given 2 plates.
Now one very important factor apart from the dielectric strength and the dielectric
constant is the dissipation factor d. Now this d is actually if you represent the capacitors
70
admittance as G plus G times B then, it is a ratio of G to B the real part to the imaginary
part that actually determines what losses will take place. So, actually because G is a real
part which will actually be responsible for the losses. So, you would actually want the
admittance corresponding to a given capacitance; it actually is not having any real part.
So, if I was to represent this on a diagram; it would look something like this. B shown on
the y axis and G shown on the x axis and this is the resultant admittance. If we denote the
angle between the resulting admittance and the y axis that is B by delta, then
G/B = tanδ
which is also called the loss tangent and is a measure of the losses in a capacitor.
Now ideally we would like the delta to go to 0, so that the real part which is this, this
axis is 0. So, there is no projection of this you admittance onto the real part which
actually means that it is a completely imaginary admittance.
71
Now, there are 3 types of capacitors which are film type, electrolytic type and the multi
layer ceramic type capacitors. The film capacitors which are the first in this category,
these are used usually for very large current low capacitance values like for example,
resonant power circuits or in snubbers; where you need very small values of the
capacitance, but the amount of currents that you handle is very large. Actually it is an
unpolarized capacitor, it does not have a positive or a negative terminal.
So, you can connect it either way. The way it is made is that you actually take too long
foils which are made up of some conducting material; which actually 2 foils and between
which you actually sandwich some dielectric material and then you roll this sandwich
this entire thing into a cylinder.
So, I will show you some pictures which are there with me which will actually depict
film capacitor. Now there are three types of dielectric material which are in use; one of
them is the polystyrene, the other one is polycarbonate which has a dielectric constant of
something like three and polypropylene is a very useful, it has really turned out to be a
very superior quality dielectric, it has a dielectric constant of like 2.3. The dielectric loss
is very less in both polycarbonate and polypropylene type of capacitors; in case of film
capacitors which can be used for high frequency applications as well.
72
it is observed that, a 50 micro farad polypropylene film capacitor would actually show a
resonance peak, add something like 100 kilohertz. So, this is a peak that it will show and
if you take for example, 10 nanofarad film capacitor it will show a peak at something
like 10 megahertz. So, it will just show a peak at that it will resonate at that.
Now these things of course, are important to understand because these when you talk
about capacitors when you talk about inductors or when we talk about resistors, they are
not pure resistors or capacitors or inductors. They have the other elements also for
example, a resistor would also have parasitic capacitance and inductance associated with
it for example, a wire wound resistance is also having an inductive component a major
inductive component therefore, we have to see what applications we are doing and then
only we should decide whether we can use. So, if there is an application where the
presence of inductor will create problems, I cannot use a wound wire resistance.
Similarly, here when we see the capacitance case, we need to be sure that we are away
from these resonant points. So, there are parasitic within the capacitor which are going to
cause a very high gain at these points when the frequencies will match; and result in this
kind of a resonance. So, this can create problems in the circuit and deteriorate the
performance. So, that is why it is important to see the parasitic equivalent resistance of
all these elements.
73
Now, the second form of capacitor after the film capacitor is the electrolytic capacitor,
now an electrolytic capacitor can give you a much higher capacitance per unit volume
compared to the film capacitors. So, in that is the reason why they have a high series
resistance. They are also suffer from the inability to withstand high reverse voltages and
that is why they are actually unipolar capacitors, you cannot actually choose the
terminals for plus and minus arbitrarily. There is a fixed positive terminal and there is a
fixed negative terminal. So, these are basically what are called polarized capacitors and
they are used basically for DC systems for DC circuits because of this.
So, when the electrolytic capacitors they are formed with either aluminum they use
aluminum or they use tantalum. Tantrum electrolytic capacitors are able to give more
capacitance per unit volume, but they are also more expensive because of the material
itself; tantalum is more costly than aluminum. And the first step, there are various ways
in which these aluminum and tantalum capacitors are made.
The first step is if we take the aluminum electrolytic capacitors the first thing that is done
is to chemically etch the aluminum foil. The idea is to increase the porosity to increase
the area of this and actually it is increased almost 100 to 125 times that area by this
etching process. Then a thin layer of aluminum oxide Al2O3 which is having a dielectric
constant of something like K is equal to 8.4.
74
Which means, that now the capacitance that what we are going to make will be 8.4 times,
what it would be if there was air which was used here instead of Al2O3. So, that is used a
form on this etched aluminum foils body. Now this oxide then is actually covered with a
liquid electrolyte, on all over this and it constitutes a plate of the capacitor. And next to
this is actually a second foil of aluminum is actually placed which actually is used for the
electrical contact, with the electrolyte and the other side of this aluminum foil is used as
a negative terminal of the capacitor.
So, basically what you have is the aluminum foil which is etched, then next to it is the
aluminum oxide Al2 O3 with a dielectric constant of 8.4 which is working as a dielectric.
Then, you have actually a plate form with the help of a liquid electrolyte and then you
have a second file of the aluminum coming in, which is in touch with this electrolyte.
Now, it is this electrolyte which is present here, which is the cause of low life of these
aluminum electrolytic capacitors. Over the years over the time this electrolyte liquid
electrolyte it actually escapes out of the unit and the capacitors loses its properties and its
no longer a capacitor. The tantalum oxide it actually has 3 times the dielectric constant
that an aluminum oxide layer can provide so; obviously, these are giving a significant
difference when it comes to the size for example, size of the capacitor compactness and
so on, but then as I said they are all they are very expensive. The third type is the multi
layer ceramic capacitors which are made by stacking many layers of thin ceramic sheets
and actually on these sheets a metal screen is printed.
So, actually you print what is called a metallic screen on top of this and there are several
of these layers which are stacked and sintered to form a solid capacitor or monolithic
solid capacitor. So, this is a ceramic capacitor that is how it is going to look like and
usually, it will be marked with an NPO or COG symbol. So, whenever you see NPO or
COG mark (Refer Time: 27:12) that this is a ceramic capacitor. And usually they are
actually available in 2 classes, class 1 and class 2. Which actually depends on their
capacity to withstand temperature, what is the range of temperature in which they can
work and also how much is the variation in their value.
So, how much is their value at 25 degree centigrade and how much it varies at a given
temperature. So, usually this is given in the form of a data and this is also alpha
numerically, it is actually printed on the capacitor; which can be decoded and we can
75
find out what is the minimum and the maximum temperature; and what kind of deviation
in the value of the capacitance that one can actually expect to happen. The thing is that
these ceramic capacitors typically you have very low resistance values. And so they are
highly desirable, but at the same time they are very expensive. So, compared to
electrolytic capacitors for example, they are very expensive.
So, here are some pictures just to show you how they look like. So, what you see on the
screen at the film capacitors, the one on the left side is a polyester film capacitor a
polystyrene film capacitors. So, basically polystyrene has been used in the film
capacitors as dielectric. And on the right side is a polypropylene capacitor, so, you have
actually the polypropylene plastic or the polymer which has been used between the
metallic plates; or the foils when this particular capacitor has been made and we already
discussed their properties and usefulness and relative advantages.
76
These are electrolytic capacitors, these are the most popular and common ones on the left
side at the top you see a typical aluminum electrolytic cell capacitor, on the right side are
the tantalum electrolytic capacitors and at the bottom, you see a full assortment of
various types of electrolytic capacitors, that consists of aluminum as well as tantalum.
And of course, depending on whether the electrolyte is wet or solid, you also have the
categorization of the electrolytic capacitors as wet electrolyte or with solid electrolyte or
the dry electrolyte. So, these are dry capacitors as they are called.
77
And here there is an assortment, which shows a ceramic capacitor on the right side is
where you can see the stacking of the various ceramic layers with the metal printing on
them. So, they are all stacked as you can see and the terminals are drawn electrodes are
drawn so that you can connect the capacitor to the external circuit and the bottom, you
will see a variety of ceramic capacitors of all types. A capacitor is never pure in nature, it
has a parasitic component of resistance and inductance. A widely used equivalent circuit
considering the effect of these parasitic elements of a capacitor is now drawn.
C is the main capacitance or the pure capacitor, Re represents, the insulation resistance
or the dielectric resistance which is used between the plates of the capacitor, Rs is the
parasitic resistance, which is on account of the lead resistance, Ls is the lead inductance
which induces a parasitic inductance in the capacitor. While analyzing the capacitors or
capacitor based circuits, it is important to consider the effect of these parasitic
components as they affect the performance of the system significantly.
78
Now, super capacitors is a very special case energy storage devices, they are the super
capacitors are actually also capacitors, but they have very different properties. One of
them is that they are different from the conventional capacitors, that we just discussed
like electrolytic capacitor and electrostatic capacitors. In super capacitors the charge is
stored in an electrochemical interface that is formed between the electrode and the
electrolyte. Super capacitors are also called double layer capacitors; what you see at the
bottom there is a set of various types of super capacitors which are shown there; these
are of various farad rating, usually the super capacitors are available with very high
capacity.
So, it is not like micro farad or something they are actually farads, but the problem is
that their voltage ratings are less; they cannot go for no more than 2.5 to 2.6 volts
because, the distance between the 2 ends or the 2 plates, which in this case happens to
be the electrode and the electrolyte is very small, it is only some molecules which are
separating them and this causes a very high electric field to be present, if you actually go
for a higher voltage that will actually break and the capacitance property will be lost.
That is why these are available only for very low voltage values, but their capacity is
very high.
79
The other important components of a power electronic system are the inductors and the
transformers. The inductors work as filters or snubbers or energy storage elements;
which are fundamental to the working of certain power converters. The transformers are
used for voltage or current transformation and galvanic isolation.
So, we are going to actually consider these things in details, but let me just tell you what
are the various geometries in which these inductors and transformers are available. So,
they are actually, an inductor is nothing, but a wire which has been actually wound into a
coil that is actually related with the flux linkage in the current. So, basically if I say
lambda is a flux linkage then I can say
λ = Li
where L is the proportionality constant between the flux linkage and i.
80
To understand the principles of inductor, let us consider a torus made up of plastic, there
is a coil which is wrapped around it all over the periphery as shown. And there is a
source connected to this coil which is causing a current i to flow in the coil. There are N
number of turns, the width of the cross section of this torus is d, the distance from the
center of this torus to the midpoint of the cross section of the torus is a, the dotted red
line represents the mean flux part inside the torus. This is the basic structure of an
inductor.
Now, if we apply the amperes circuital law; which says that if there is a surface and there
is a current passing which is going through the surface let us say i1 i2 i3; then if I take the
line integral of H the magnetic field intensity across this closed part which encompasses
these currents. Then I can say that H dl line integral over this path would be equal to the
sum of all these currents which are enclosed by it, using this amperes circuital law and
applying it to this toroid this inductor that we have just defined. We can write
ɠ H.dl = ∫J.dA
where, dA represents an elemental surface area.
81
Assuming that, d is much smaller than A. We can write
H (2πa) = Ni
H = N / (2π a) = Ni/l
.L is the mean magnetic flux path length. Since, a is much larger than d, it is to assume
that the flux density B is uniform across the cross section of this plastic torus.
The magnetic flux density B is related to the magnetic intensity H by the relation
B = μo H
Where μo is the permeability of the free space. And we can therefore write,
φ = Bπd2 /4
Where we used the fact that
B = φ /A
Now, we can say that this flux is established due the Magneto Motive Force defined by F
F = Ni = ∫J.dA = ɠ H.dl
F = H (2πa)
82
F = φ / (π d2/4) . (2 πa/ μo)
Φ = F/R
Thus,
R = 2 πa/ μo(π d2/4) A/wb
83
Now, looking at the torus and this inductor arrangement with N turns through which i is
flowing causing a flux phi. We can say that, each turn is associated with an emf turn
given by
Eturn = N dΦ/dt = d λ /dt
The value of inductance L can be obtained as
L = N2/R
Now, instead of the plastic torus suppose this torus is made up of a magnetic material,
then it is found that the value of the inductance that would be achieved would be much
higher. The use of core is therefore, a very highly desirable things when we try to design
and fabricate inductors and transformers. The magnetic core is made up of magnetic
material which is usually a ferromagnetic material or its alloy. Each of these magnetic
material is actually characterized by its unique BH curve. Since we know that B is the
flux density flux per unit area and H is nothing but, N into i the number of turns into the
current we actually can get a BH curve by plotting i versus φ.
I will just draw a typical BH curve, considering only one curve in a typical BH curve
would appear something like this; let me also mark the N into I, on the x axis which is
nothing but H and also phi over A, on the y axis which is nothing but B and that is why
its called a BH curve. So, what we see on the curve is that after a certain threshold value
84
of the current, assuming that the number of turns and the dimensions of the core are
fixed. We can say that the flux φ saturates, no matter how much we increase the current
after that the flux does not increase it said it remains at its saturation value, which we can
denote as φsat.
It is important that whenever we are dealing with a magnetic system we do not operate
our system at currents more than a threshold value. So, if I call this to be a threshold
value ith my current should not exceed ith to avoid saturation of the core. Saturation of the
core leads to the shorting of the electromagnetic system and it actually will draw a large
amount of current, which might cause other safety issues with the equipment or with the
circuit. A transformer is just like a couple inductor.
Now, we just saw an inductor a coil which was wound around a torus. Now, if there are
more than one windings more than one coil then this could be a coupled inductor or this
could be used as a transformer; we will look at these details a little later. It is important
to consider the equivalent circuit of an inductor taking into account the parasitic
components that might be present in the inductor. Therefore, an inductor is never a pure
inductor it also has a parasitic R and a parasitic C in it.
The equivalent circuit of an inductor maybe drawn like this; so this L is the main
inductor, but along with it there is a parasitic inductor Rp and a parasitic capacitor which
is because of the inter turn capacitance Cp. And whenever we are considering the
85
inductor, we must consider the effects of these parasitics. So, that we know that the
operation of the circuit where this inductor is used will not be affected by these parasitic
components.
There are various types of course, the various types of geometries which you can see on
the left side top is laminated low frequency core that is used for low frequency
applications. Typically this is a metal, which is made of silicon steel or which is also
called transformer steel for our regular low frequency transformers and inductors, this is
useful. What you see in the middle on the right side in the top row is a C type core.
Now, another type another C type core is taken and merged with it and the winding is
faced in between them, so that the flux created by these coils it actually is limited by this
core which is a C type core. What you see on the rightmost side in the top layer on the
top row is a toroidal core, now here there is a problem with the winding because, they are
not open, it is a closed system.
So, you need really a patience winding though it can be done through machines; it can be
done in an automated manner, but it is then turns out to be expensive and cumbersome.
What you see on the lower left corner is a pot type core. So, basically you can see a pole
in the edge of centre or on which the winding is placed and then this entire thing is
closed from the top.
86
So, there is a exactly analogous element part which is exactly going and sitting
symmetrically on top of this and will provide a closed magnetic path to the to the flux;
that will be caused by the current that will flow through the coil. On the right side of the
center of the lower row, you see an E type core and it works almost a same way. As the
pot core the explanation is always the same only thing is the shape is different and this is
actually an open core.
Now, what I show on the right side, bottom corner is actually an image where I have
shown a bobbin. Now this bobbin is a structure which is realized either of hard paper
board or plastic on which you can actually bind the your coil your windings, you can put
on that and then you can insert the magnetic core like for example, it is shown an E core
is being inserted.
So, one has been inserted already from the front right side and there can be one which is
inserted from the left side as you can see and these are symmetrically inserted inside and
due to the dimensions they exactly match with each other; they touch each other inside
the bobbin and a closed path is formed for the flux which is created by the coil which is
been wrong on this bobbin. And on the leftmost side there is a; there is a metallic
clamper, which is used to actually hold the two E’s of the core together, so that they do
not move.
So, that is how you realize. So, this could be if it is a single coil, this could be just an
inductor or this could be actually places 2 coils where it could be working as a coupled
inductor or it could be working as a transformer. Likewise more and more windings can
be associated with any of these magnetic systems and you will therefore get the
corresponding magnetic component and it actually varies as per the application. So, for
various applications, you will have various different types of these magnetic components
which would be created.
So, with this I conclude this session and now in the next session, we will start looking at
some actual power electronic circuits and we will try to see, what exactly is required to
analyze? What exactly is required to design them? What exactly is required to
understand them? So, we look at those aspects in the next lecture. I thank you very much
for your patience and attention.
87
Prof. Vivek Agarwal
Indian Institutes of Technology, Bombay
Lecture – 05
Analysis of Switched Networks
Welcome back. Now, we have learned so far an interesting fact that power electronic
circuits can be represented by a switching matrix. We have also seen another
representation of power electronic circuits in terms of the switching impedance which
give additional insight into the nature of switches and the desirable characteristics of
their operation for the four basic power conversion operations. Indeed several power
converter configurations and topologies are available, for each of the basic power
conversion configurations and they can all be derived from the general switching matrix.
However, besides comprehending the operating principles of the power converters, their
analysis is equally important not only for enhanced understanding, but also for the design
and practical realization in an optimal manner.
So, obviously, the question arises as to which analysis method and tools we should use,
which ones would be the most suitable? To answer this question, it is important to have
some idea about the nature of the power electronic circuits, at least the general
representative circuits. The switching matrix is certainly a good and compact representation
of power electronic circuits as we have seen, but it does not provide the feel, the physical
feel of the actual power electronic circuits.
Now, we begin this session by looking at some of the important representative power
electronic circuit configurations which are frequently encountered. Their nature and
working will give us clues about the appropriate analysis methods that can be used to
analyze the power electronic circuits. The presence of one or more switches is what
distinguishes a power electronic circuit from non-switched electrical networks. The ON-
OFF operation as of the switch brings the power electronic circuit in the category of the so
called variable structure systems. So, this is what usually the control engineers prefer to call
such systems which change their form or configuration while operating.
Now, what exactly does a switching operation in a circuit it achieve, what does it do? So, it
may connect an energy source to the network or to the circuit, it may disconnect an energy
88
source, it may take out of the circuit or it may alter or modify the circuit configuration in
any desired manner required by the operation. The actual switches in power electronic
circuits are the diodes, the BJTs, SCRs, IGBTs and the MOSFETs what which I briefly
mentioned in my earlier lectures, but for now it is enough not to really worry about their
specifics.
So, let us not worry about the specifics on the switch. Let us assume the switch to be
general ON-OFF device with all ideal specifications and characteristics for its use in the
given circuit. Once we develop a basic understanding we will replace the ON-OFF switches
by the actual switching devices and that will be really convenient, and it will help and
enhance over understanding. Now, let us just review quickly some of the commonly
encountered situations in power electronic circuits.
Let us say that there is a DC source Vs that is supplying power to an RC network and there
is a current that flows whenever the switch is turned on or closed, let us say it is done at t is
equal to 0. Now, if we analyze this circuit by using the very fundamental and basic
concepts, we can say that a long time after switch has been closed; the capacitor voltage Vc
will be equal to Vs. Now, due to the kvl, the Kirchhoff’s voltage law, the voltage across R
during the steady state will be 0. But what we observe is that the steady state solution that
you observe here is not really predicting the value of the voltages, the voltage across the
resistance, for all times after the switch has been closed.
89
This is because when the switch is closed at t is equal to 0, the voltage across the capacitor
cannot rise suddenly. And the Kirchhoff’s voltage law demands that the entire source
voltage that is applied; therefore must appear across the resistance which means that the
initial value of resistance at some point 0 plus immediately after the switch is closed should
be equal to Vs. But this is not predicted by the steady state solution that we achieved
through our basic understanding. There are other circuits also which are important, but we
will find that it is difficult to, there is a problem in predicting their complete response if
they are switched; it is not possible to predict the values correctly for all times after the
switch has been closed.
A similar thing can be seen in this example where a DC source Vs feeds an RL network.
And as we will see later once again if we try to use the our basic understanding about the
steady state operation, we are not able to predict the operation of the circuit at all times, all
instants after closing the switch. This is observed in several other networks also. We can see
that the same situation will happen whether it is AC source or DC source, whether the
network is a simple first order system or it is a second order system, or it is even more
complicated, we will not be able to, we are not able to predict the complete response as it is
called by just using the steady state solution.
So, what do we observe, what is our conclusion from all these samples. So, we can say the
first observation is, it is true that the network may be non-linear in power electronic circuits
90
when we take it as a whole along with the switch. However, between the switchings the
network is linear. So, you can see that this is just a combination of R, L and C; which are all
linear, linear elements. It therefore, follows all the rules of linear systems.
When we say a linear system, a given system is linear when it follows or it obeys what is
called the superposition principle and the homogeneity principle. Now, fortunately this is
the case with power electronic circuits, and this is what we will see now. Now, one of the a
important things about the linear systems is that if you are looking at the response of a
linear system, if the forcing function f(t) is sinusoidal, then we find that the response is also
sinusoidal. The only difference that might come is that maybe the magnitude of the
response and is different from the input, and maybe it has a phase with respect to the input.
Also we must note that the linear time invariant systems, they produce the same frequency
as the source or the forcing function. So, basically the response has the same frequency as a
source of the forcing function.
What else we observe from the general examples we considered of power electronics. So,
the general expressions for voltage and current, we observe that they could be written as in
terms of linear differential equations with constant coefficients. For example,
Vs = Ri + L di/dt.
91
So, here we can see that this is R and L, these are constant coefficients, and they do not vary
with time. We also observe that the circuit steady state analysis using the basic concepts, it
does not lead to a correct solution for all values of t, after the switch has been closed. It is
expected that some correction factor is required to be added to the steady state solution to
get the complete response, which will be valid for all values of t after the switch has been
closed.
So, let us say I have a source Vs, and it is connected to a network of R and L. And there is a
current that flows in this system. I am not really right now saying that it is a switched
system. To make it a switched system we will see what operations we need to do. So, this
could be a DC source or it could be an AC source anything is possible. Now, if I want to
know how the nature of current in this circuit what I would do, and how I would actually
analyze. So, by using the Kirchhoff’s voltage law, I can say
Vs = Ri + L di/dt.
Now, as you can see this is a differential equation. And we can also see that it is a linear
differential equation, but it is a non-homogeneous differential equation because this is not 0.
And remember that the coefficient of the various coefficients of this differential equations
of various terms, you see R here L here they are all we are assuming that they are time
invariant, they do not vary over time. So, basically this qualifies to be a non-homogeneous
linear differential equation with constant coefficients.
92
Now, if we look at a switched circuit - a typical power electronic circuit, then I can say that
it is going to have a switch here between this point. And let me just denote this switch by
sw. Now, a typical problem that would occur because of the continuous ON and OFF of
these switches of the switching matrix or the power electronic converter would be the ON
and OFF of this kind of a switch. It will continuously switch ON and OFF. Now, we want
to see that once the switch is turned on when the switch is closed and my current starts
flowing how my current going to look like, so that needs basically an analysis and solution
of this differential equation. We need to solve this differential equation to do that.
If I just put one C here, you can see that I will have an additional term that will come here
that will be having a derivative term actually, because we know we are trying to write
voltage. So, voltage across what will be the voltage across the capacitor, it will be 1 over C
i dt, where C is the value of this capacitor. So, this term will come here. Apart from any
initial conditions of an initial charge that might be present on this capacitor or any initial
current that might be flowing through this inductor. So, the question is how do we solve
such differential equations? And the very good point is that we have very nice and
established tools to handle these kinds of systems. Let us just review how they are
particularly solved.
So, use this mean of what is called operational calculus. Now, let us try to go a step ahead
and understand or review what operational calculus is, and how it helps us in analyzing or
93
solving system equations that I just described to you. So, one of the things is that there is
this concept of an operator which is usually defined by p, and it actually means d/dt so, this
is also called a differential operator. We define this.
And if I define this to be equal to d/dt then in my equations in my differential equation

wherever I have the presence of a derivative, I can replace that by p. So, what will be the
result of that my previous equation which is
Vs = Ri + L di/dt.
Now, clearly if this was d2/dt square, then this would become p2. If it was d3/dt cube, then
this will become p3 and it goes on.
Now, what is the advantage of in this case, coming back to the same equation again, what is
the advantage of doing this is that now because you are able to write
Vs = Ri + Lpi.
Actually you can deal with it as an algebraic equation. So, the introduction of the
differential operator p has helped us in converting a differential equation, a linear
differential equation into an algebraic equation which can now be solved. So, basically if I
am trying to look for the solution for i, then i would say it is simply as R+pL, simply it
would be like this. So, you can clearly see that the introduction of p this operator has really
94
helped us in simplifying the differential equation - its solution and as well as its
representation. So, it can be represented as an algebraic equation.
Now, not only that you can use p to denote d/dt, but you can also denote the integration or
the integral functions using this p. So, if you want to do an operation of this type, all you
need to do is that you just do 1 over p, multiplication just by 1 by p. And this will actually
give you the integration of a function.
Now, let us say we have a case where a source e is applied to a resistor, then I can say that
the
i=e/R by Ohm’s law
But let us say that instead of a resistor, this was an inductor, then we have to see that
e = L di/dt,
And this value of the inductor is L, earlier the resistance I used to values R let us say. Now,
we are considering the second case with only inductor being supplied by this excitation e.
So, as you can see
e = L di/dt from Kirchhoff’s voltage law.
Now, using our differential operator p as before I can write e = Lpi. And now I can say
95
i = e/pL
Now, what does this equation show, this again you can see this is ohm’s law just like you
have ohm’s law here. This is also ohm’s law, but if it is ohm’s law and this is applied
voltage and this is current, and obviously this is nothing but resistance. But we know that
this resistance is not having the same nature as the resistance R that we have employed
before this is a very different type of resistance.
And we will see the details probably later, but it is usual to call this one not resistance, but
impedance. And it is all it is often called because it actually is using this linear of this
differential operator p, it is actually called operational impedance, operational impedance.
Just way as we saw the example of an inductor, we can see that for a capacitor.
So, if that from excitation of e is applied to a capacitor, then we can say
C de/dt = i
And then using the operator again we can say that Cpe = i. And from here actually we can
manipulate this equation a little bit, and we can say
i = e/1/pC
96
Again this is ohm’s law. Voltage is there; current is there. So, this must be resistance or as
we learned in the case of inductor this must be an impedance. And that impedance value is
given by 1/pC.
Now, we will see in due course that whenever the excitation becomes sinusoidal, whether it
is this case or the previous case of inductor, or it is any other case or indeed even more
complex with any complexity switching circuits, where sinusoidal excitations are involved
there this we will see that this p will actually be replaced by omega. Omega is nothing but
the angular frequency of the sinusoidal excitation applied to the network. So, you will see
that this will actually get replaced by that.
So, if a sinusoidal function is there, so this is a pretty general representation, where we have
not mentioned what is the form of e actually strictly speaking I should be writing maybe e
of t unless it is a DC value, in which case of course, there will be a 0 coming in here at this
point. So, once again 1/pC is the impedance, and it is also called an operational impedance
because it involves p. Now, to distinguish it from the conductive impedance, the case
previous case where inductor was involved, this is called an inductive operational
impedance. And the one in this case where there is a capacitance, this is called a capacitive
operational impedance. Let me show you one example.
So, let us say we have a circuit we have a network which is supplied from a source e and let
us say it has a switch now. So, talking about the switched circuits in power electronics, let
97
me just bring in this switch here. And let us say there is an inductor L and a resistance R.
Now, the moment this switch is closed, the applied source function e which is in this case a
battery e gets applied to this network which consists of R in L, and it results in certain
current and; obviously, it will result in some voltage across the inductor and the resistance.
Now, what we want to do is we want to determine let us say the voltage Vo across the
resistance R after the switch has been closed. So, let us say the switch is SW. We want to
determine in this output voltage after the switch has been closed. Let us say we are
interested in knowing how the output voltage Vo across R varies as a function of e.
Using the potential divider I can straight away write Vo by E is equal to R over R plus p L,
where pL represents the inductive impedance of L. Now, in this expression, we can actually
say that E is nothing but the forcing function and we may denote it by f(t). Vo may be
considered as the response and we may denote it as y(t). Then we can say that y(t) is equal
to some function of pf(t). Now, it is this function which actually comes here which is
denoted by G(p), because we can see that this expression is a function of p. G(p) is called
the network function or the transfer function.
When we talk about the response y of t, it actually would consist of two components; one of
them is the forced response component, and the other we will see would be the natural
response component. The first response component is actually obtained by using the non-
homogeneous linear differential equation; it is called the force response because it is it
98
remains as long as the forcing function remains. However, it becomes prominent in the
steady state. In the initial transient state, after the disturbance has come like the closing of
the switch, it actually is dominated by the natural response.
The natural response which is the second component dies down with time. And the solution
or the response y(t), it actually becomes equal to the steady state response a long time after
the switch has been closed in the steady state. Now,
G(p) = N(p) / D(p)
these are two polynomials in p. So, if we can see this general network which will obviously,
get more and more complex, if we see more complex networks which will not just be a
simple first order system where p you can see is only having a power of 1, it may have
many more. So, it will be a general polynomial in N. So,
y(t) = ( N(p) / D(p) ) * f(t).
Now, the determination of the natural response is done by considering the homogeneous
equation corresponding to the original non-homogeneous will be a linear differential
equation representing the system. So, basically what we do is we make f of t is 0. And when
we do that we actually end up with
D(p)*y(t) = 0.
And since we are talking about a non-trivial response or solution, we can say that this
actually means D(p) = 0. Now, when we solve this polynomial equation, this equation is
also called the characteristic equation of the system. And it actually governs the behavior of
the natural response of the system as we will see in examples that will follow now.
Now, depending on the order of the system which is actually determined by the number of
energy storage elements like how many L s and C s are there, the order of this characteristic
equation is determined by that. So, if you have one c and one L, the order is 2. If you have
only one L and there is no C, then the order is 1. And if you have two Ls and one c, then
you have an order of 3. Depending on the nature of the roots that this equation gives
basically which actually has come down to D of p equal to 0. The solution that it gives let
us say it gives for a second order system 2 roots, p1 and p2. So, it is the nature of these
roots p1 and p2 that determines the response the nature of the response, the nature of the
99
natural response of the system. We will see these things in details with the help of some
examples in due course.
Now getting back to our original problem that we wanted to see how Vo varies with e or we
wanted to determine Vo by E we can write the following expression again. What is Vo over
E using the potential dividers and using the differential operator p we can straight away
write that
Vo /E = R / (R+pL)
which I can just manipulate a little bit and write it in a different form, this way. Now, this
gives me
(L/R)dVo/dt + Vo = E.
Now, in this steady state in a long time has passed, long time has passed after the switch sw
has been closed, then we can we know that our derivative term will go to 0. So, we can say
that this is 0. So, I can say that my Vo is equal to e long time after switch has been closed.
Now, this is what is called the forced solution. So, basically by putting the derivative term
equal to 0 for a DC excitation source, what we mean is that in the steady state, we do not
expect any time variation in the response. And that is why if we take that term if you make
that term 0, it actually means that we are making p = 0.
100
Now, similar shortcut methods are applicable when other types of excitations or sourcing
functions are used. For the DC excitation, if we put p is equal to 0 in the network function
as you can see in the slide the first force response method that we have put there, now we
can straight away get the forced response to a DC excitation. Similarly, for an exponential
excitation, we have written that how a forced response can be obtained by putting p is equal
to the coefficient of the exponential power in the network function. Similarly, we can obtain
the force response to this sinusoidal excitation also. So, these shortcut methods are very
handy in quickly getting the force response.
So, Vo after a long time the switch has been closed is equal to E, and we are saying there is
called the forced solution. And let me just say that VoF for the forced part is equal to E, let
me just write this. Now, the question is that when the switch was closed, when switch was
closed, immediately after that immediately after that let us say at some time p is equal to 0
plus what would be the current in the system, the current i in the system i will be 0, because
there is an inductor in the circuit and we all know that the current through the inductor
cannot change suddenly.
So, initially if there was no current in the inductor, it will remain like that even after closing
the switch at a very small time at every infinity similar time after the switch has been closed
at t = 0. Now, if that happens if i is equal to 0 immediately after the switch has been closed,
then what will be Vo. Now, we know that Vo = R*i. And if i is 0, that means, Vo is 0 also.
101
So, the voltage across the resistance which we are saying is output voltage is 0, because the
current through the inductor cannot change suddenly which means that the solution which
we just obtained we just obtain this solution by which we said is long time of the switch
has been closed is no longer valid.
As per this solution what we have seen before VoF should be E. But as per the solution that
we have seen now by doing this simple explanation, because there is an inductor the current
is 0 by a simple analysis and understanding, we found that Vo is equal to 0. So, basically
what we are saying is that this solution is not a complete solution is not a complete solution.
There must be something else in this solution which will make it valid for all times starting
from t is equal to 0 plus all the way up to a long time, which is actually called the forced
state or it is called steady state. Now, everything has steadied after a long time, everything
is in a stable steady form now and we call it is steady state.
So, clearly the solution we got what we called as a forced solution is not working out when
we try to analyze the circuit for a very short time after the switch has been closed. So, what
has gone wrong, what is the problem? So, problem is that we have not considered the
response of the elements of the circuit themselves, we have not considered the solution that
we have got the first solution that we got, we did not consider the property of inductor L.
And for getting a complete response or to get a complete solution which will be valid for all
times after the switch has been closed, we must actually do something to obtain or include
the response of the inductor and the resistor also in the solution.
102
Now, how this can be done? Now, this is very simple, complementary function represents
the natural response of the system. So, to get this complementary solution complementary
the second part of the solution as I mentioned some time back, you just do
(L/R)dVo/dt + Vo = E.
And this is a corresponding homogeneous equation, but the original non-homogeneous

equation that we derived at the very beginning. Now, how can we solve this equation? As
you can see that it gives us some ideas as you can see that if I take this Vo onto this side
some and also I take L/R term onto the right side, it will be like saying that there is some
constant term and some variable which you are getting after doing the time derivative of
that quantity.
So, you do a time derivative of Vo, you get some constant and Vo itself. Now, if we all
know that there is only one function, the exponential function which has the capability to
satisfy this; that is if you do the derivative of an exponential function, it will actually give
you the exponential function itself. So, therefore, it is very customary to find Vo to assume a
form for this solution to be Vo. And let us mention t here. The earlier one I mentioned VoF
for forced, now I am saying Vo transient. So, let us mention this t to indicate that this is a
transient solution or the natural solution the due to the presence of the elements L and R in
the circuit that solution, I am trying to find out.
103
And let us assume by the explanation that I just gave you that this has the form k est. So,
this is I am assuming that it is some exponential function with some constant. Now, if we
substitute this assumed value into this equation that we have the homogenous equation, I
can say
( Ls/R + 1 ) k est = 0
And from here I can say that this term must be 0 which means that s = -R/L.
Now, in a moment we will see the significance of this particular value i. So, this actually is
very important relation, it is a very important analysis that we are doing.
So, basically if I get s = -R/L then I can say that my assumed natural or the transient
solution. And considering our earlier discussion and our conclusion, we can say that the
total solution, therefore the
Vototal = VoF + Vot
And this particular term is a transient term which is due to the natural response of the
system it goes to 0 with time.
So, as a time a long time elapses after the switch has been closed, this term will go to 0. So,
what you by putting the values of VoF and Vot that we have obtained we know that
104
VoF = E, and
Vot = k e –Rt/L
So, we got this as a solution, the total solution. Now, the question is k is still unknown in
this expression. So, how we can find this value of k, how we will find this particular
expression for k? We note that we use the initial condition. We note that at t is equal to 0
plus i is equal to 0. And therefore, Vo = 0 also because Vo is nothing but iR. And if i = 0, so
Vo = 0. Using this initial condition; using this condition because this equation the total
solution must satisfy and under all values of time after switch has been closed.
So, considering this particular initial condition, I can say
Vototal = 0 = E + k e -0
So, at t = 0, we are using this condition that the output voltage is 0. And I have just
substituted t = 0 in the rest of the expression on the right side. This yields
k = -E
And hence I can write down now, I can write down the total solution to be equal to
Vototal = E (1- e –Rt/L ) = E(1-1) = 0
105
And likewise when the t tends to infinity, you can say that the Vototal = E the forced solution.
So, long time of the switch has been closed only the forced solution which is just because of
the applied excitation function will remain the transient term has actually decayed to 0.
So, now let me just generalize this. So, how do we get the total solution? We first will get
the forced solution or the steady state solution. And then we will get the natural solution or
the transient solution. And then we will add this transient solution to the forced solution,
and there will be a constant. Now, then how many constants will be there in the previous
example there was only one k. But if you have more number of elements which can store
energy like you have more number of Ls and Cs, then you will find that the number of
constants which we need to determine, they will also increase. And in fact, the number of
constants that we will have to determine is directly equal to the number of storage elements
used in a network which we are analyzing.
Now let us graphically see how these two will response the force response and the natural
response looked like, and what the addition of the two leads to. So, what is the total
response looks like.
106
So, as you can see if I was to just plot these on a graph so let us say this is my time axis and
this is the axis on which I am plotting the output voltage of interest in the circuit that
appears across R. Then let us say if this is the point E in terms of magnitude, then I can say
that this is the forced response. And if I say that this is the minus E point the graph, then I
can say that this is the natural response. And when I combine these two that is so I can
actually mark this as Vot the transient response, and if I combine these two, if I add these
two, I can say that I get nice response like this shown by green. And it actually becomes
same as E, because you can see that after this point here my transient response has become
0. So, after some time after an initial phase, the switch has been closed here at this point.
So, this is in the time, this is the instant at which the switch was closed.
So, after some time the natural response has died down, and what remains is only the forced
response. Now, this we will find this kind of an action, this kind of process, this kind of a
phenomenon, happening all the times in power electronic circuits. Power electronic circuits
making use of several switches, they do go into these kinds of transitions. And depending
on what is the network that we are considering there, we will get different response.
With the concepts learned so far, can you try to solve the above example, can you try to find
out the complete response of the network shown above? Just to give you a hint as this is a
second order system we expect that it will involve two constants whose values will have to
107
be determined by using the boundary conditions. So, I can actually provide you the solution
and I would like you to verify that indeed this solution satisfies.
So, you will find that
k1 = E/3, and k2 = -4E/3
The complete response of the voltage
Vototal = E ( 1+ e-4t / 3 – 4 e-t/3)
So, you can verify this. And you can even try to plot this response. So, you can have the
individual response, the natural response and the forced response, you can plot them and see
what you get.
I thank you very much for your attention. In the next session, we will see some other
aspects, some other tools which will help us to analyze the power electronic circuits.
108
Fundamental of Power Electronics
Prof. Vivek Agarwal
Lecture – 06
Review of Engineering Maths for Power Electronic Circuit Analysis
Hello and welcome back. So far we have seen some examples in which we considered
circuits which were fed by DC sources and we represented them by linear differential
equations with constant coefficients and we solved them to get the desired response of
current or voltage.
Now, let us see an example in which an ac source is feeding the network.
So, the AC source or the forcing function is supplying through a switch network that
consists of R equal to 1 ohm an inductor which is 2 milli henry and a capacitor which is
0.5 milli farad.
Now, when the switch is closed at t is equal to 0, a current results in this network. We are
interested in determining the complete response of this current that is for all time after
the switch sw has been closed.
109
Now, the first step is to apply the Kirchhoff’s voltage law which will straightaway lead
to the following equation
L di/dt + iR + 1/C ∫idt = Vm sinωt
Now, if we differentiate this equation on both sides, we can see that it actually turns out
to be
L d2i/dt2 + R di/dt + i/C = Vm ω cosωt
We are assuming for this circuit that the current in the circuit at time t is equal to 0
i(0) = 0 vc(0) = 0
So, these are our initial conditions.
By rearranging this term, we can write this in a slightly more organized form as given
below.
d2i/dt2 + R/L di/dt + i/LC = Vmω/L cosωt
In order to obtain the natural response, we will have to work with the homogenous
equation corresponding to this non homogeneous equation.
So, we will get
110
d2i/dt2 + R/L di/dt + i/LC = 0
So, this is the corresponding homogeneous equation. Now, many times people use what
is called the standard form of the second order system the standard form of the
characteristic equation of the second order system which is obtained by using zeta,
introducing zeta as a damping factor and define this as
ξ = R/2 √C/L
Similarly, the natural frequency of oscillations is defined by
ωn = 1/√LC
So, I can just write here the natural frequency. So, if we use these parameters that we
have specially introduced, we can actually get the roots m 1 and 2. So, there are two
roots
m1,2 = -ξωn ± ωn√1-ξ2
We can also write the roots by using the original expressions in the terms without
making any substitution.
So, now here if you see that this there is a square root quantity. So, now depending on
whether you get inside the root whether you get a positive, negative or a 0 quantity, you
actually have three different behavior in response of the system. So, if zeta is equal to 1,
we call this as a critically damped response. If zeta is greater than 1, we call it over
damped response. So, you will not see any oscillatory response in what in any of these
two conditions and if you have the zeta less than 1 which means that now this square root
quantity would actually become the inside quantity will become negative and hence, you
will get imaginary roots.
Now, this as because of our experience with Euler’s formula, we know that this will lead
to some sin and cosine terms and hence, we can say that this is going to be an oscillatory
response also called as under damped, not damped which means it is going to oscillate.
So, under damped response.
111
So coming back to our original equation, so, either we can use the roots in this form that
we obtain or we can still continue with the same form
m1,2 = (-R/L ± √(R/L)2–4/LC)/ 2
By putting these values, various values of R and C as is given in the problem in the
example we then find m1 and m2 and then, determine the response.
Now, if we want to get the complete response of the system, complete response of the
current let me just mark the current i. So, we want the complete response of i of the
switch has been closed at t is equal to 0. Now, there are as we know now we know very
well that there are two components that will actually constitute the complete response.
One is the forced response which is present long time after the switch has been closed
and a natural response which will be present only at the time of switching and the
switching takes place and this response remains for some time, but then afterwards it dies
down.
So, basically it is a transient response which leaves for a very short time also called the
natural response as it depends on the circuit parameters. Various element that are used L
R and C, it does not depend on the applied function the forcing function.
112
So, the natural response if you want to determine we know that p1 or we can just call the
general expression m which actually denotes the roots of the quadratic expression that
we will get by substituting the right hand side the forcing function equal to 0 in the
previous equation. So, basically Vmω/ Lcosωt term vanishes to 0. So, this becomes the
homogeneous equation and then, we are trying to find the roots of this equation. So,
there are two roots which are given by this very standard expression and then, we can
then write the two roots because we have been using p1 and p2 to define the roots.
We can say that p1 p2, they turn out to be
p1,p2 = -250 ± j 968.25
So, these are the two roots we get now depending on what we get in the square root term
here. So, these could be for example, it could be 0 that we are getting in which case we
will have the real roots. In one case it might turn out that the left hand term R by L
square is greater than 4 by LC in which case we will be getting real roots.
Now, one of the things that we can immediately try to correlate here is that because they
are imaginary roots and from the Euler’s formula, we know that something that is having
an imaginary component we have seen that how it is related to sin and cos. So, it should
actually give you an idea already that this is going to be an oscillator response. So, it will
be an oscillatory natural response that you are going to get.
113
So, the general form of the natural solution for imaginary roots expressing or actually
depicting or showing oscillatory behavior can be given by in or it , t for transient and n
for natural. So,
it(t) = e-250t (A1 cos(968.25t) + A2 sin(968.25t))
Now this is the form of the natural response and please remember that the natural
response is going to be the same whether we are having a sinusoidal source function or
whether we are going we are having a DC forcing function or source function, the
natural response is independent of the applied source.
So, this natural response is same and we will see you will find this also covered in the
previous example, where we consider a dc source based systems. Now, let us find out the
forced response also for this given network, so that we can then add the natural response
and the forced response solve for the constants A1 and A2 using the boundary conditions
and then, substitute back in the expression and get the complete response. So, what is the
forced response in this case? So,
if(t) = 1.714 sin(ωt+80.11)
This is the forced response and how we get this? We just got this simply by using the
fact that the current, the forced current long time after the switch has been closed would
be given by the applied voltage vac divided by the net impedance z and we have just
written the expression for complete expression for the voltage ac which is given ac
voltage which is given as z.
So, we see that there is a current which will have a phase with respect to the applied
voltage and when we solve this, this is a straightforward process because all the
parameters are known to us. Now, what is the next step? Next step is to write the
complete response.
114
And now we know it very well that Itotal the term I have been using to denote the
complete response is nothing, but it transient response plus if(t) the forced response and
we just add these up using the expression that we already obtained for
itotal(t) = it(t) + if(t)
= e-250t (A1 cos(968.25t) + A2 sin(968.25t)) + 1.714 sin(ωt+80.11)
One important thing that you might have noted that because we are considering linear
networks and we have said that these are the networks that power electronic circuits have
to deal with the switchings that is between the two switchings that take place, we can see
that when we apply a sinusoidal forcing function of a response also is sinusoidal and its
frequency is same as the applied forcing functions frequency. Only thing is that you will
find that its magnitude and the phase would be different from the applied forcing
function.
So, these things I would advise that you verify and be confident about this point. Now,
we know the boundary conditions already. We are given that
i(0) = 0 vc(0) = 0
115
So, we will be using some of these boundary conditions now to determine the unknowns
A1 and A2 in the complete response equation. So, if i is equal to 0 or i at 0 is equal to 0,
then if we solve the equation number 1 above, we will get
A1 = -1.68
So, you just substitute it t is equal to 0 and the left side i total is equal to 0 at that time.
So, we just got the values of A1, we have all the values of the unknowns are there in the
equation 1. Now, we also know that if we differentiate i and we see what the derivative
of the current is at time t is equal to 0, then we find that that is also equal to 0. You can
show this by knowing that the capacitor voltage at the instant 0 is 0. So, this you must
verify is very simple and straightforward that you can verify.
So, when you use this condition that is you differentiate equation 1 in both the sides, put
the left hand side equal to 0, then you will find that
A2 = -0.53
We can then get the complete response
itotal(t) = e-250t (-1.68 cos(....) -0.53 sin(….)) + 1.714 sin(ωt+80.11)
This is what would be the final response and you can always try to plot and see also.
So, you can because you can see sin and cosine terms, you can see that this is going to
have an oscillatory response as far as the natural response is concerned and then,
afterwards it will steady, it will steady down to an ac response which is given by v by z.
So, this is one example that I wanted to do. Now, having seen the classical method, now
let us look at some other tools that we can use and make life simple for us. Now, about
the additional tools I would like to mention the Fourier methods.
Now, one of the things is that an electrical engineer is fascinated that if there is a non
sinusoidal, but periodic function and if it can be decomposed into sinusoidal components
and it is applied to the linear networks as we see in power electronic circuits, then he can
get the response which is which is going to be same form that is going to be sinusoidal as
well with the same frequency. And then, you can get the response to all the decomposed
116
sinusoidal terms and you can get the response to all of them and then, by superposition
because the system is linear, they can all be added to get the final response.
So, this actually helps, it actually provides a very nice tool to the electrical engineer and
in fact, in various other branches to be able to study the response of a linear system to a
periodic, but non sinusoidal function. Now, non sinusoidal function there are several
examples which are there in electrical engineering. Formally we can actually use what is
called Fourier series which is which actually falls in the category of Fourier methods
where if your function, the applied function is a periodic function with let us say p of t,
then the Fourier series representation of f(t) can be decomposed into sin and cosine terms
which is of the form as you can see on slides.
So, you can see that f of t can be shown equal to some term a0/2 which actually
represents the DC or the average value or content in the function f(t) plus it is you can
see a summation that goes from n is equal to 1 to infinity of cos and sin terms which are
added and they both have an and bn as their coefficients where these coefficients an and
bn they vary an and bn, they vary with the value of n.
So, you get various values of you can get for an by putting various values of n and
similarly for various values of n for you can get for bn, the expressions the general
expressions for obtaining the an and bn coefficients is also given on this slide T by 2 is
actually the half time period of the periodic function which we are trying to decompose
117
into sin and cosine terms. Now, before we move on with one quick example on Fourier
series.
I would like to mention that there are some ways in which we can actually simplify the
Fourier series analysis. These rules are they make use of the symmetry of the given
periodic waveform. So, there is something called an even symmetry or if the function is
even, you can then use this even symmetry property. Now, if a function is even, a given
function f of t is having an even is an even function.
Then, basically what we mean is that I just write here and show you. So, even function,
so a given function is even if
f(t) = f (-t)
and the example for even function is a cos function.
So, we know that cos(theta) is equal is equal to cos(-theta). So, this is the even function.
Similarly if a function is odd for an odd function
f(t) = -f(-t)
So, if that is the condition which a function is actually satisfying, then it is an odd
function. Similarly we also have situations which are referred to as half wave symmetry
and the quarter wave symmetry.
118
In the half wave symmetry we have
f (t) = - f (t-T/2)
and you will find that the wave forms which have a half wave symmetry, they do not
contain even harmonics and an example of this is the inverter output. Similarly, we have
the quarter wave symmetry which exhibits both even symmetry and half wave symmetry
or odd symmetry and half wave symmetry.
So, we can make a similar simplifications as a matter of rule when you have this kind of
a situation. So, using these various conditions it is possible to simplify the Fourier series
analysis. So, as I mentioned as far as power electronics is concerned, Fourier series will
help us to decompose the voltage and current waveforms which may be non sinusoidal
into the corresponding sin and cosine terms summation of the corresponding sin and
cosine terms and as you will see later with several examples, they are very useful for the
harmonic analysis of power electronics systems.
Now, here you see a very typical example which I would like to leave as an exercise to
the participant. You can see that this is a half wave rectifier. When the input AC supply
is positive, the diode d will conduct. When it is negative, the diode does not conduct
because it gets reverse biased during that time. So, what you will see the voltage across
the resistance r will be this half wave rectified. So, you have these alternate loops which
appear from the input AC supply.
Now, somebody would want to know that what is the decomposition of this output, this
half wave rectified output? How do you decompose it? How does it look when you are
when you want sin and cosine terms to represent this? So, obviously, this is a periodic
function, but this is not a sinusoidal function. So, if you can actually do this analysis
further, then you will be able to get the term that is shown right at the bottom of the slide.
119
So, this will actually need some simple calculations and manipulations which I expect
you to do to learn this and also I encourage you to do many other examples. So, there are
several other sources which are available which talk about Fourier series expansion
which you must try.
These are what you see on the screen are the Fourier series expansion. So, you have
actually in this case the function is actually broken into a Fourier series, but now instead
of an and bn you have only one coefficient which is Cn and you can see the Cn expression
120
is written at the bottom on the right side in terms of the time period of the periodic
function. So, this is a much more compact form and we have made use of the fact that
sinωt = (ejωt – e-jωt )/ 2j
So, these things actually follow from the Euler’s formula and everybody knows there.
So, there is an expression for cosωt also which you can see which is given on the left
side. Now, there are many functions which are obviously not periodic.
Fourier transform is a tool which has been derived from the Fourier series expression
which considers the time period of such functions to be infinite and with this assumption,
the Fourier transform method it can be used to transform periodic or non periodic time
domain signals into frequency domain signals.
So, it is definitely a more powerful thing than the just Fourier series expansion. Fourier
transform can be used to determine the harmonic spectrum of the voltage and current
wave forms in a power electronic circuit. This is what it does and you can see that there
are the corresponding formulas which are given. So, f(t) is a given function which we
have represented as you can see the bottom and you can also see the Fourier transform of
that which is a capital F(ω) given by
F(ω) = ∫f(t) e-jωt dt
121
The right side plot you see a peak which is there at the fundamental frequency
corresponding to the frequency of the given sinusoidal function. So, you see a peak there
and rest all the frequency components are 0.
Likewise, if you see the bottom example it is just a single pulse which has been given
which is now non periodic and if you actually do a Fourier transform using the formulas
given before, you actually end up with what you see on the right side. that kind of a
harmonic spectrum that is what you will get by using Fourier transform. The problem
however is Fourier transform is good, but it does not exist for those signals which
increase with time. So, for example a ramp signal. So, you will find that the Fourier
integral in this case the Fourier integral is the integral which was involved in the Fourier
transform expression. The two expressions which I showed you a short while ago that
Fourier integral is not finite. Now, it is a solution of this because ramp signal and
similarly such many signals are very important in power electronic domain.
So, obviously if we cannot analyze signals such as a ramp signal, then Fourier transform
has a very big drawback. It actually has a lacuna which must be overcome, we must be
able to overcome this drawback. Now, Laplace transform is actually a modification of
the Fourier transform method which now ensures that even the signals which are rising
122
with time which are increasing with time such as a ramp signal, they also would
converge. So, Laplace transform it uses the finite Fourier integral.
So, basically the limit from minus infinity to infinity is changed to 0 to infinity. So, there
are no restrictions when we use the Laplace transform and therefore, Laplace transform
is the one which is used extensively.
So, you can see that if there is a function f(t) and if you are doing the Laplace transform
of this function, then you can say that it is denoted by capital F of s a
F(s) = ∫f(t)e-st dt
And here you will find that this function always converge and after you have used this
transformation, you have actually solved your problem, the time domain problem by
converting everything into the frequency domain using the Laplace transform and you
have come to a point where you have got the Laplace transform of the solution that you
are looking for you can actually do a Laplace inverse. So, on the right side you see how
you actually get the function f(t) in time domain back from a given Laplace transform F
of s.
So, we will look at some example of this, so that this becomes clear.
123
Let us again consider our earlier example of a switched RL network, where we have a
battery source, a DC source which is applied through a switch sw to a network that
consists of R and L assume and this switch is let us say closed at time t is equal to 0 and
that is the time when this applied this dc source is applied to terminals x and y that is
right across the network. Now, once the switch is closed by using the Kirchhoff’s
Voltage Law KVL, I can straight away write that the applied source function
E = Ri + L di/dt (1)
So, this is how the governing equation will look like once a switch is closed and a
current begins to flow here. Now, because there is a switch in the system and we have
seen in one of the earlier lectures that when you are closing the switch, when you are
closing the switch it is like applying the applied source or the source function e in a
stepped manner. So, if I just denote this kind of a function, so this is the time axis and
this is the let us say the magnitude and if I denote this by u of e, then this is called the
unit step function.
So, there is this pulse, it rises to a magnitude of 1.0 at this point and we said that this
point is exactly 0 and then, this pulse remains 1 for all times with greater than 0, for all
times, then 0 that is in this direction that is e less than 0. This function u of p is 0 we
know this. So, a complete representation of this equation that we wrote let us say this is
124
equation number 1. A complete representation or more complete representation of this
governing equation should have this kind of an expression
E u(t) = Ri + L di/dt. (2)
Now, let us see how we solve this. Let us see how we solve this and what difference it
will make to bring the Laplace transforms into this solution. So, now let us begin, let us
introduce the Laplace transforms and let us see how it makes our life easy. We have to
solve this linear differential equation which is given by two. So, to begin with let us take
the Laplace transform of both sides of equation 2.
So, on the left side we will get E and the Laplace transform of u of t we already know
how u of t looks like.
E[Lu(t)] = R L[i(t)] + L L[di(t)/dt] (3)
Now, this old style capital L is used here to denote the Laplace transform.
So, coming back to our equation what we were solving using the fact that the Laplace
transform of u of t can be shown to be equal to 1 by s. The Laplace transform of i(t) let
us say it is I(s).
So, this is what we use this notation is what we use and this is the Laplace transform of
the unit step function. Then, using this back in equation 3, I can write down
125
E/s = R I(s) + L (sI(s) – i(0-)) (4)
Now, what we have done is here we had this derivative term. This derivative term of the
current the response function in this case. So, we need to write the Laplace
transformation, we need to write the last for Laplace transform of the derivative of i(t).
So, this just this is nothing, but the Laplace transform of this derivative term and you can
see that it involves the initial condition. So, when we will use the Laplace transform, you
will find that it to really simplify the three step solution. The classical solution into a
much simpler one, it into almost one or two step process more straightforward it, but of
course you will also have you will have to still work with the initial conditions and as
you will see later also with the roots of the characteristic equations.
So, these are several things that we will have to compute apart from the other things that
we are using the Laplace transform for.
Now, the question is before I solve this further, I need to know what I put for I of 0
minus in this case. So, this is 0 minus which is shown just to kind of denote just to kind
of highlight that we are talking about a time just before we close the switch. Now, there
was an inductor in the circuit before the switch was closed. We know that there was no
current in the circuit and after the switch has been closed also because of the presence of
the inductor, there would not be any current.
126
So, basically
i(0-) = i(0+) = 0
So, using this in the previous expression which I can show you, so this is the last
expression. We got equation 4, I can write. So, using this using our initial conditions we
can say that
E(s)/s = I(s)R + LsI(s) (5)
So, basically you can say that from time domain in a way we have just gone to the
frequency domain. So, this is number 5, equation number 5 and we can now just kind of
rearrange this just like an algebraic equation and we can just write we can get.
L(s) = E/L / s (s+R/L) (6)
This is what we get. So, very nice we got an expression in s domain, we got an
expression for the current or I should say the Laplace transform of current we got an
expression for that.
Now, I do not understand this language. I do not know what this means I am used to
working in the time domain, I am not used to working in s domain. So, I must get this
thing back or inverted into a time domain expression and then, I will be able to
understand what it means.
So, it is already, so solution is right there in front of us, but it is in s domain. Now, how
do we do this? As you can see that there is a product of these expressions which both
have s and the way to do this is by using what is called the part partial fraction
expansion. So, we use, so let me just call this is expression 6.
127
So, now solving it using the partial fraction method that we have briefly seen before we
can say that the trick is to say that
I(s) = (E/L) / s(s+R/L)
I just write this as a sum of two expressions or two partial expressions k0/s plus
k1/(s+R/L).
Look at this expression.
So, we have actually written this in this manner. So, we say that this actually is
representable in this form. The only thing is I do not know what k0 is and what k1 is. It is
very simple. As we have seen before that we can straight away find the expression for k0
which will be nothing, but this expression. So, because we are finding out k0, we will
take the term on this side which has this particular factor over there.
Thus by putting s = 0, we get k0 = E/R. and by putting s = -R/L, we get k1 = -E/R
Substituting k0 and k1, we get
I(s) = E/R [1/s – 1/(s+R/L)]
128
So, coming back to our objective that we want to find out from here what is I of t. So, all
I need to do is now I have now these two expressions and these are very nicely
identifiable with a set of such similar expressions that are available in standard tables.
usually what people do, they will just find out the Laplace transforms and the inverse
Laplace transforms of some of the very common functions and they will actually put
them in a table. So, here I will show you such a table.
So, all you can do is now you can identify what is the inverse Laplace transform of this
term and this term.
So, basically from the inverse Laplace transform, we get
i(t) = E/R (1 – e-Rt/L)
Now, this is a time domain expression, this equation is a time domain equation and in
fact, we can see that we have got how the response function i(t) the current in the system
would behave. We note that the classical solution is a three step process while the
Laplace transform is just a one step process to get the complete response. I hope that the
tools we have learnt would be useful for analyzing the power electronic circuits that are
to follow in this course.
I thank you very much for your attention.
129
Prof. Vivek Agarwal
Lecture – 07
Review of Semiconductor Physics
Welcome back. So, today let us try to see certain facts about semiconductors and how
they can be used to produce some useful devices. Particularly we have to see how that
would be useful for Power Electronics. Now when we talk about solids, a very good
definition which really helps us relate the microscopic properties of a given material to
the macroscopic or external properties and characteristics of a material, the definition of
resistance is very helpful. Now, let us just see the definition of the resistance for a
minute.
R = ρl/A
where I have just assumed that it is a rod let us say made up of some metal or maybe
some other material semiconductor or maybe an insulator I do not know; I just taken a
general example. So, this area what you see is A and this length is l. Now, if I want to
determine what is the resistance offered by this? by this rod or this bar of rectangular
cross section to a current that is trying to flow.
130
So, let us say there is a current that is trying to flow through this bar and come out like
this, what is the resistance that would be offered to this current this particular expression
for the resistance gives us exactly that. It actually tells us what is the resistance offered
by this rod to a current that is trying to flow through it? As you can see area of cross
section l, the length of this bar; these are the physical properties and these are all
externally measurable observable properties.
So, looking at this definition we have constant of proportionality which is the rho which
is also called as a resistivity. Now this resistivity is a function of the internal properties
of this bar or the material which comprises this bar, which makes up this bar. Now
talking about the resistivity and the resistance, the solids, the solid material can be
divided into 3 main types.
And that is roughly based on the resistivity value.
131
So, you have basically one particular category of solids which have extremely low
resistivity. So, the rho is 10-5 Ohms meter, this is the typical resistivity that we have for
one category of solids. There is another one whose resistivity is extremely high it is in
excess of 105 Ohms meter. And lying between these two ranges is another class of solids
which has typically resistivity ranging from 10-5 to 105 and mind you the units are Ohms
meter.
So, basically what we see here is that looking at the basic definition of the resistance we
can see that any current or any charge carriers that will try to flow through the bar that
we have considered in our example. It will see a very low resistance when it is passing
through this particular category of solids it will see a very high resistance when the
charge carriers are trying to flow through this type of solids. And here it is going to be in
the in the intermediate range in the intermediate category it will be the resistance will be
actually modulated.
So, basically it will depend on some external conditions that will apply that there will be
the charge carriers would experience a resistance that will vary between a certain range.
So, maybe R1 to R2 between this range, now it is customary to actually call these solids
which actually offer a very low resistance to the flow of charge carriers as metals and it
is very common for this the other extreme category of very high resistivity, very high
resistance as insulators and in between these two, this intermediate category its common
132
to call this as semi conductors, some people also call them as semi metals; so, semi
conductors.
Now, examples of metals of course, there are many copper, silver, gold and so on,
insulators again mica, quartz and so on and semiconductors as you will see there are
many, again there are many examples. The most common being the silicon and the
germanium which are the pure semiconductors in the pure form, but then there are also
certain compounds like for example, we have the gallium arsenide or we have the
cadmium telluride or we have the cadmium sulphide and so on. These are the inorganic
compounds of which are having semiconductor properties and we also have organic
semiconductors like several organic polymers which are there so organic semiconductors
as well.
Now, to understand the properties in more details and then try to see what is the
relevance and usefulness of the semiconductors, let us go back to a very fundamental
concept.
The concept of Neil Bohr’s atom.
133
So, we all know from what we have studied in our earlier classes that as per the Neil
Bohr atom, the atoms nucleus lies in the center and is having all the positive charges of
protons inside. And around this nucleus are revolving several electrons, just like the earth
revolves around the sun, these electrons are revolving around this nucleus and if I just
mark these electrons as e1 e2 e3, then I can see that each of these electrons, they occupy a
unique orbit.
They occupy a unique position, a unique orbit around the nucleus which uniquely
determines the energy of these electrons. So, unique energy and we all know from our
past knowledge that we can uniquely define each of these electrons which are revolving
around a nucleus by using the concept of quantum numbers. So, we all know that there is
principle quantum number, there is azimuthal quantum number and so on.
So, each of these electrons can be uniquely represented by a set of these four quantum
numbers. So, each of them has a fixed energy, now when an electron let us say e1 gets
some energy from outside. So, let us say there is a packet of energy which has come and
hit electron e1.
So, let us say a photon, which is the fundamental particle postulated by Newton as what
the light consists of the dual nature of light people know that light has dual nature, it has
both the particle nature as well as the wave nature, as proposed by Huygens and the
other scientists.
134
So, let us say that the light is falling on this atom and let us say there is a photon which
has got an energy which is more than a certain threshold energy associated with this
electron.
So, if that energy is transferred to e1, this e1 will move on to a higher energy orbit. So, it
will absorb this energy. So, let us just say this is delta E which has been imparted to e1
and e1 will just move on to from position x to x’, a new orbit that corresponds to the a
new different energy which is a higher energy than what e1 was associated with at the
previous location.
Likewise, if let us say e3 loses energy, it will fall. So, it actually this e3 it actually loses
energy. So, let us say delta e it will fall down from its original location y to y’ to a lower
energy, now both these energy transitions what we have seen from lower energy level to
higher and from higher energy level to lower, the transition from lower to higher energy
or from higher energy to lower energy.
So, this is what is the concept that we get from the from the Neil Bohr’s atom the model
that he postulated. Now, the point is that the way we look at the Neil Bohr’s atom
model,it is like an isolated atom we are looking at, we are not talking about the
interaction of these electrons with some of the other electrons that might be present in
the vicinity.
135
And this is what really happens when we talk about solids and maybe liquids and gases
because then we are talking about several atoms which are coming closer to each other
of course, in gases this closing in or the distance between the atoms is very large in
liquids it is really relatively less as compared to the gases. But still very much higher
than what it is there for solids, but for solids we have many of these atoms they come
very close to each other and they form a crystal structure.
Now, we know that how these various atoms are arranged as you can see in space
occupying very fixed locations and a pattern which is repeating. Now, these are the
atoms and now if we consider each of these atoms having the same model as what we
discussed sometime back about the Neil Bohr’s atom, then we find that it is not possible
for us to really, see what is happening because there is some electrons which are
revolving around all these atoms and we do not know how they will interact and its
extremely complicated.
So, the best way to look at this situation that arises when the solids are formed, that is
when the atoms come closer is by way of the concept of band energy. Now when we say
band energy what we mean is that lets say we just talked about solids what we mean is
that we will just try to define, we just try to kind of ascertain the charge carriers and their
energies and put them in various categories.
136
So, let us say how we can do that. So, in the band energy concept if I have the energy
which is plotted here in the y direction then I can if I just maybe divide this in certain
zones. So, let us say I just define this in these 3 zones and I just say that this is a zone
which is having further divisions.
Now, this is energy your y axis is energy. So, basically you are talking about various
energy levels here, likewise in this band which I am not yet mentioned what this band is
though I am sure many of you can guess. Again , let us say there are several sub energy
levels and there is this band here which is in the middle of these two bands, which is let
us say does not have any charge carriers it does not have any electrons or holes.
So, this band what we are calling which is sandwiched between the top and the bottom
band is what is called the forbidden band which basically means that none of the charge
carriers electrons and holes are going to occupy this band. Now all the electrons if you
look at basic diagram here where we have shown. So, so basically there are these atoms
which are very lying very close to each other and we know that is say only electrons here
are sticking very close to the their parent atom and sticking very close to it.
So, let us just say that those electrons which are very close to their parent atom and not
free to move, let us say they are lying in this in this band; let me just call this as A band
this is B and lets call this as C. So, C is the one which is forbidden you have seen A is,
let us call this as valence band.
So, these are basically this band will contain all the electrons which are kind of stuck
with their parent atoms they are bound to them and they are not free to move likewise B,
the band which is shown at the top here with the higher energy level is called a
conduction band is called the conduction band. And this consists of charge carriers
which have been actually transported from close to the atom, they somehow found some
energy maybe thermal energy which then helped them to kind of jump over on to this
band.
So, all these green lines they are showing how the charge carriers when they are in
valence band, but when they are imported energy they somehow find energy they are
actually transported, they are actually all going into the conduction band and there in the
conduction band they exist as free charge carriers and they are free to move.
137
So, basically this is the basic concept of semiconductors and as we go along, we will
understand this even better. Now inside the semiconductor these charge carriers, they are
in constant motion back and forth. So, they will find some thermal energy, some of these
carriers will actually be able to jump over this forbidden band and will now become a
free charge carrier when it goes to the conduction band and when they collide they lose
energy they actually then fall back. So, I could probably say that they are coming back
here back to the valence band where they combine with poles and get neutralized. So,
this process is all the time on.
At temperatures greater than the absolute 0 temperature it is possible that some of the
electrons they find enough energy to break away from their covalent bonds and become
free. Now electrons from vicinity may then come and occupy these missing electron
positions and similarly we may have more electrons coming in and occupying this latest
missing electron positions and the process goes on. So, it would seem that as electrons
are moving towards one direction the missing electron position which we have also
called as hole is moving to in the opposite direction.
Now, thermal equilibrium the densities of electrons and holes are same, let us say this
density is ni, then this is governed by a very important relation.
Which is
138
ni2 = c e(-qEg/kT)
where Eg is the band gap energy and is equal to 1.1 electron volt for silicon, q is the
magnitude of the charge of the electron k is nothing, but the Boltzmann constant and T is
the temperature in degree Kelvin.
So, what is the situation though in the conduction band with this. Now, it is observed that
in metals even at room temperatures, there is a huge amount of charge carriers which are
there in the conduction band which we have marked as B and that is why the metals
copper, silver, gold and so on. They show an enormous conductivity even at room
temperature and once you have applied an external electric field these charge carriers
they will just flow as per the field and will not be, you will not be able to control them,
there is no control.
Now, if you look at the insulators which is the third category; let me consider you will
find that the conduction band has very few charge carriers and in fact, they will hardly be
any charge carriers and hence when you try to apply an external field there are no free
charge carriers to move around and that is why you will see an insulator conducting very
less current; it will hardly have any current. Now semi conductors on the other hand
which lie between the first in the third categories of metals and semiconductors they
actually have a very unique property, they actually present as we will see now under
certain conditions very unique opportunities to allow controlled flow of charge carriers,
139
we can actually control these carriers. So, let us just see now how this is achieved and
how useful semiconductors are formed.
Now, as we expect all our electronic circuits for which eventually we are going to use
many of these devices made up of semiconductors, we expect them to actually be
operating at room temperatures. Now the semiconductors at room temperature, I am
talking about in their pure form, silicon germanium many of those. They actually have
few charge carriers available or present in the conduction band at room temperature. In
fact, at absolute 0 minus 273 degree centigrade, in principle, there is the 0 charge carrier
in the conduction band of a semiconductor material. Now what do we do? Now that is at
this point and here that we introduce what are called impurities in these semiconductors.
So, what let us understand this phenomenon of adding impurities in the semiconductors.
So, I have let us say a solid, a semiconductor solid block and we know from the band
theory that there is there are 3 layers, 3 levels the valence band d and the forbidden
band where no charge carriers can be present. They are not allowed to occupy any of the
energy levels here and the conduction level, we know this. Now, what we do with some
technique we try to bombard, the semiconductor solid, with an impurity there are many
trivalent or pentavalent impurities. Now what I mean by this is trivalent like for example,
boron or aluminum.
140
So, if you look at their outermost shell, if we talk about the now you just talk about the
Neil Bohr’s atom as an isolated atom and if you look at these elements in which are
many more actually in this group, then what will happen these guys, they have three
electrons which are there in the outermost shell. And if you are trying to basically
bombard this material with what are called pentavalent impurities, pentavalent means
something like phosphorous and arsenic and so on these guys they have 5 electrons in
their outermost shell.
So, let us this is the situation, now what happens when we bombard this trivalent
impurities here on the sample? What happens, let us see that now this particular sample,
the original semiconductor which is not ideal created at all you have not mixed anything
with it is a pure semiconductor we also call pure semiconductors without any
adulteration, without any mixing as an intrinsic semiconductor. So, we have this piece of
intrinsic semiconductor and we know that let us say we are talking about a silicon
semiconductor here. So, we are going to have this is the situation silicon silicon silicon
silicon silicon and silicon silicon silicon and we can see that the outermost shell of these
silicon atoms they have 4 of these electrons and they form covalent bonds, they form
covalent bonds with their neighbors.
So, let us just see the case of this and this atom, these two atoms. So, you can see that
there are 1, 2, 3, 4 electrons around this silicon and there are one 1, 2, 3, 4; 4 electrons
across around this silicon. So, what these two atoms do they are their neighboring atoms
of this crystal structure they form what is called a covalent bond. There are other types of
bonds also as we know ionic bond for example, but that is not what we are discussing
here, they are not relevant to our semiconductors.
So, likewise this silicon atom it can form a bond here, with a neighboring which is at the
bottom below this can form a covalent bond there, likewise it can form a covalent bond
here and likewise it can form a covalent bond here. So, it is basically these two electrons,
these two electrons, these two electrons and these two electrons they are shared these are
constantly interacting, they are just back going back and forth and they are just. So, this
is a theory which has been given for the covalent band which are formed. So, this is how
the intrinsic semiconductor or silicon semiconductor structure would look like.
141
Now, on that you bombard your trivalent the boron or aluminum, what would happen? If
you have actually many of these atoms bombarding here at this place is that these atoms
they will go and maybe they will dislodge one of these or many of these at different
places silicon atoms. So, what basically you will have is a B coming in here, a B coming
in here. Now what will happen that when the B comes here because it is one electron
shot it has got three electrons in its outer orbits, it is actually having an affinity to have
another electron.
So, basically what happens that it actually will pull out one electron from some atom in
its vicinity, now that when it moves and it completes the boron’s valency requirement,
but it does it at the expense of basically extracting one electron. So, this electron let us
say move here and when it moves here it basically leaves out what is called a hole.
So, as you can see these electrons gone there leaving a miss missing electron position of
the hole. So, many of these atoms of the boron trivalent impurity when they are actually
when they actually replace the atoms which are originally sitting in the intrinsic
semiconductor, they actually generate many such missing electron positions of the holes
in this entire; in this entire structure. So, such atoms like the boron these are called
acceptor atoms or acceptors or this is called an acceptor type impurity because it accepts
an electron and generates a hole.
Now, when many of these traveling atoms are added to this structure this structure
becomes rich in holes and we because the hole is missing electron which we associate
with a positive charge. Therefore, this entire structure becomes rich in positive charge
carriers and is called a p type semiconductor.
It is an extrinsic semiconductor now, but a p type semiconductor an extrinsic one, the

one which has been adulterated, which has been mixed with impurities or in terms of
semiconductors we call them as doping. So, it has been doped, an intrinsic silicon
semiconductor has been doped with trivalent impurity to generate p type semiconductor
material. Now likewise in the same way it is easy for us to see that if a pentavalent or a
atoms with 5 electrons in their outer most shells when they are going to be mixed with
the intrinsic semiconductor, then at that time we are going to have the atoms which will
go and sit and replace the original silicons, they are going to have one additional
electrons.
142
So, these are called donors they actually are able to very easily donate their electrons and
these electrons then become free to move, just like we have the holes becoming free to
move in the p type semiconductor. Here we have electrons which are free to move, just
because these pentavalent atoms have acted as donors and have donated several electrons
in this material.
So, this kind of semiconductor this is called an n type semiconductor, even after doping
the thermal ionization process which causes the breaking of bonds is not disturbed. This
is because the doping concentration is much less compared to the intrinsic semiconductor
concentration therefore, even after doping the product of the thermal equilibrium
densities or the electrons and holes which we now define as n0 and p0. So, their product
n0 p0 = ni2
Now this is a very important relation and it is also called the law of mass action. So, we
will see now in the next lecture what more we can do with these semiconductors and
how we can form some useful devices with them.
143
Prof. Vivek Agarwal
Lecture – 08
PN Junction
Hello and welcome back. Having understood the formation of these you know now
modified semiconductors which are also called extrinsic semiconductors the p-type and
the n-type, let us now see what happens when we try to bring you know these p-type and
n-type semiconductors close to each other. So, basically what we are saying is that let us
take as an example of silicon crystal.
You actually bombard let us say one part, you bombard it with several trivalent atoms
like the borons, you know aluminium and so on, those kinds of impurities. And as we
know we will be ending up with several holes. These are the holes which are created, but
of course we also have with each of them one negative ion.
So, the overall material remains neutral. So, if it is a hole which is formed a
corresponding negative charge has been introduced also. So, this is let us say boron. And
here let us say on this side we bombard with phosphorus atoms, let us say phosphorus
atoms. And this gives rise to as we know these are the donor atoms so there are many
electrons which are now available, in this particular region. And of course, they also
144
have their corresponding ions which are plus, they are positively charged ions
everywhere you have this situation. The overall material is neutral no net charge.
Talking about the P-N junction let us say we have a crystal on which on one side we are
growing an n-type material and on the right side or the other side we are growing the p-
type material. So, let me just draw this energy level diagram for the n-type first.
So, let us say these are the various energy levels, let us say that this is nothing but my
conduction energy band level and this is my band gap this entire what you see here, this
is the valence band level and so on. And let us say now let us make use of a very
important parameter which is used in such systems which is called the fermi energy
divided by EF. Now, what is fermi energy? Fermi energy is nothing but a chemical
energy you know which is actually remaining constant for the material throughout you
know in the equilibrium condition.
EF is approximately equal to the average of the conduction band energy level and the
valence band energy level. So, we can say that it is equal to Ec plus Ev by 2, where Ec is
the conduction band energy level and Ev is the valence band energy level. So, which
means basically that even when the two systems they merge with each other they end
from the p side they are going to have a constant fermi energy. And to do this we will see
that how the energy levels shift. So, I am going to draw these diagrams further and
actually I am going to explain that to you.
145
So, let me just draw the fermi energy level. This is the original level without any doping.
So, let us me let me call this as EF for the n side, EF of n. Now, when we do doping what
you will see is that we actually are doping with the donor atoms whose energy is very
close to the conduction band. So, we are here. So, basically the net result of this is that
you know my fermi energy shifts up. So, this EFn actually shifts up to a new value which
is shown by this green colour. Now, let us see what will happen correspondingly for a P-
N junction.
So, now let me just do the same thing for a for the p side materials. So, they are the other
side of the crystal semiconductor crystal where I am trying to now grow p-type material.
So, let me just draw again the energy diagram.
So, let us say this is the as before this is the conduction level and let me just say p here
just to you know mention that this is the p-type and here also let me just say n just to
show that this is the n-type in the previous diagram. And also let me just say Ev of n
here. Now, on this side if I further draw this I can say that this is my valence band Ev p-
type and like this.
So, what will be the fermi energy for just the p-type? It will be somewhere here, this is
the fermi energy. Let us call this as for the p-type EF. Now, when we do the doping we
actually bombard the acceptor atoms and as you know that the holes are associated with
the valence band, so you are actually ending up you know or introducing these acceptor
146
atoms or the holes on this side, which means that my fermi energy for this side p-type
actually will shift down and you know it will it will come somewhere here. Let us say
this is here. So, this is the new fermi energy.
Now, if you bring the two sides the p-type and the n-type together, I just keep them
together and I just go by my definition, fermi energy should have the same value for the
combined n p sides under the equilibrium condition. Then, what we see is that there is a
basically shifting of the other levels. You will see that the conduction band level, the
valence band level on the p side they actually shift, they actually come on to this side and
I am just trying to show you this. So, what you see here is that you know this is how
things will be. So, basically I can just say that I can combine my green lines on the two
sides which means that the fermi energy levels are same, while as far as the other levels
are concerned, they will all shift you know upwards as you can see I will have to just
make this kind of a structure over there. This is how this P-N junction would look like.
So, we can actually see that there is a gradient, there is a potential gradient which has
been created and this is what actually we denote as the junction potential phi c. So, we
can see that there is a transition region between the p and the n-type material. As we will
see later and analyze in due course that this is the depletion region or the space charge
region of the P-N junction.
Now, let us just see here very close to this junction. Now, we know this is the p-type
semiconductor we have created this is an n-type semiconductor we have created and let
us just discuss this junction point where these two material they are just opposed against
each other. So, let me just draw this again here. So, let me just draw this and now let me
just say that this is the junction region. So, this region which was here, has been now
drawn here again.
Now, if I also see this region and analyze and see what is going to happen? What is
going to happen is that now that there are the two types of charge carrier movements
which will take place when this situation happens. Now, one of them you know is what
is called the diffusion we can see them in the left side that many holes, their
concentration is much higher than their concentration on the right side. Likewise, the
concentration of electrons is much higher on the right side compared to the left side. So,
there is a concentration gradient of holes. So, basically what happens that the holes try to
147
flow because of this concentration gradient from left to right. So, we have basically an
influx of holes from left side to right side. Likewise, if you see the right side which is
rich in electrons and as a much higher concentration of electrons compared to the left
side, we will see that there would be a flow of electrons.
The second phenomenon which will cause the flow of charge carriers is the drift
phenomenon, right. The I have just written the second type of you know phenomenon
which will cause these charge carriers to flow as the drift phenomenon which is under
the action of an externally applied field or a created or a generated electric field. We will
come to that. Now, what will happen? Diffusion is taking place, the holes are trying to
flow into the junction area from left to right, and the electrons are trying to flow into the
junction area from right side to the left side.
So, basically what happens if I just draw this again if I just draw the junction again that
the holes moving in from left to right, they would neutralize you know the some of the
ions which are present along with their electron pair. So, basically you know the
situation will be that you have these ions and each of them you know these are positive
ions, each of them is balanced by an electron. So, basically when your holes flow from
left to right, they neutralize these electrons, we combine with them. Likewise, if you see
on the left side of this junction you have negative ions, this is what we had particularly
this is what is there here which are having holes, as pairs to neutralize them.
148
Now, when the electrons flow from right side to the left side they end up neutralizing
these holes. So, basically what happens is that you end up with a region, so p-type, n-
type and you actually end up with a small region between this p-type and n-type
material, which is made up of these ions, positive ions on this side and negative ions on
this side, and there are no charge carriers in this region. This region is what is called as
the depletion region, no charge carrier, or space charge region.
Now, it is very easy for us to see that you know this kind of a structure is nothing but a
capacitor. So, this is nothing but a capacitor. And we all know that a capacitor this is the
let us say the positive plate and this is a negative plate it will have a field, and depending
on this distance it would have a potential between these two plates. So, this is what has
happened because of this diffusion current which has flown in terms of holes and
electrons. Formation of a depletion layer which is devoid of charge carriers which is
having an electric field as a result of this electric potential.
Now, the question is how long will this diffusion current continue to flow. Now, the
answer to this is that the diffusion current flows as long as this electric field which is
created in the depletion layer or the potential that is created in this layer does not balance
out the flow of the diffusion current. So, basically the tendency of this electric field that
we have created is to basically cause a current that will flow now in the opposite
direction to the diffusion current. So, if I was to say that my diffusion current direction,
net direction is this, assuming the direction of the holes to be the positive current and the
actual current. Now, what will happen is that drift current will try to flow the other way
around. And therefore, this process of diffusion would continue till the drift current is
equal to the diffusion current and balances it out, and that is where the whole process
stops.
Now, that there are two types of currents which are flowing, one of them is the diffusion
current, the other is a drift current. And you have seen that at the equilibrium point when
the two currents become equal that we call as a thermal equilibrium condition. And if
there is no external voltage applied, then we find that this equilibrium is maintained,
otherwise it is disturbed and a new equilibrium point is created.
149
Now, if the applied electric field is on let us say towards the right, this is the E field then
we know that the holes, they are going to experience or they are going to cause a hole
current in the same direction, the same direction is the field shown with a red colour. The
electrons on the other hand will actually constitute a current which will be in the opposite
direction to the applied field.
The net current due to the applied electric field is created because of the net result of the
electron and hole current on account of the drift. And the expressions can be given for
the electron drift current we can say that it will be you know i electron drift we can say
that this is nothing but two times mu n into n into E, where q is the charge the electron,
mu n is the mobility of the electron, n is the concentration which is the per unit volume
concentration the number of you know the electrons per unit volume and E is applied
field. Likewise, we have the hole current on account of the drift which is given by q
times mu p into p and E, where q is a charge of the whole, mu p is a mobility of the hole,
p is the concentration per unit volume of the holes and E is the applied electric field. So,
it is the net sum of these which constitutes the total drift current.
Now, similarly if there is a gradient of the concentration from one side to the other. So,
for example, if there are more holes on the left side compared to the right or there are
more electrons on the right side compared to the left side, they will actually flow from
the high concentration area to the low concentration area. And I can actually show this
150
with the help of these two, you know these two diagrams where I can say that the y axis
is actually the concentration of holes and here also the concentration in the next plot y
axis shows the concentration of electrons, and since we are doing the analysis usually in
the x direction let us say this is the x direction. So, we can see that there is a there is a
diffusion gradient that we see as we go from left to the right side both the cases.
And in the case of the holes, we will find that the current that will be caused because of
the diffusion of holes on account of diffusion will be given by you know minus of q into
D p which is the nothing but the diffusion constant of holes into dp by dx. And for the
electrons it will be given by q into D n which, where S n is nothing but the diffusion
constant for electrons into dn by dx.
Now, there is a negative sign that will come along with the expression for holes that is
because you know we have the concentration gradient for the holes which is in this
direction, and which is negative. So, we have a negative direct gradient of the diffusion
as we go towards the right, but the conventional current is also in the direction of holes.
So, wherever the holes are flowing that is the direction of the current, so there is a minus
sign which actually remains there. But in case of electrons because the direction of the
current is considered to be opposite to the direction of the electrons and the gradient of
the diffusion gradient of the electrons is considered to be again negative in the in this x
direction, that is why you get a positive q n into D n into dn by dx that expression. So,
basically these two currents because of the holes and electrons diffusing from one side to
another are added up and we get the net diffusion current across the P-N junction.
151
Now, this analysis part of the diffusion current drift current, it actually gets simplified if
we use a very important relation which is called the Einstein’s relation which is given by
Dp, the diffusion constant of holes by mu p the mobility of holes equal to the diffusion
constant for electrons to mu n and this expression is actually given by kT by q. So, this
ratio and this ratio these two ratios are equal to kT by q, where kT by q is often denoted
by VT by some authors. So, this may be a very important relationship that we may be
using in the future.
So, if you take a P-N junction which is formed by you know juxtaposing p and n-type of
semiconductor material you actually end up with this kind of depletion layer, which will
have a field and a contact potential. So, having understood the you know the formation
of the P-N junction and the mechanism involved let us now see how a P-N junction
works when an external voltage is applied across it.
152
The P-N junctions are used as diodes which do not have a control terminal. Now, if you
remember in my earlier lectures, we have grouped diodes under the category of type 1
switches, now they turn on and turn off depending upon the magnitude and the polarity
of the voltage that gets applied across them you know by the circuit in which they are
connected. You can see that you know there is a junction and on the two sides of that you
can see the p-type and the n-type semiconductor. And you can also see in red colour the
two metallic contacts which have been provided you know for connecting the wires.
Now, p-type material or the p-type side, it is actually called anode and it is denoted by A.
While the n-type side it is actually called the cathode and denoted by K. Now, depending
upon the polarity of the voltage you know a P-N junction may be subjected to either a
forward bias where the voltage VAK is greater than 0, you know or it could be subjected
to a reverse bias when a negative voltage that is VAK less than 0 is applied.
Now, what you see right now is the picture of a reverse biased P-N junction which I have
drawn sometime back. Just below the P-N junction you can also see the distribution of
the minority charge carriers. In reverse bias, the p side is connected to the negative
terminal of the supply or the battery while the n side is connected to the positive terminal
of the battery. For these conditions, the holes on the p side they move towards the left
attracted towards the negative terminal of the battery.
153
Similarly, electrons on the n side, they move towards the positive terminal of the battery
thus both holes and the electrons they move away from the junction. As the holes in the
electrons move away from the junction the depletion layer width increases. Now, the
process will stop when all the holes will flock the negative terminal and all the electrons
will flock the positive terminal. And the applied potential will be completely balanced by
the potential developed by the depletion layer. Now, there is no current flow obviously,
because there are no carriers and the only way these carriers could come is probably you
know the p-type carriers, the minority which are in minority in n-type material that they
actually cross the junction over the p side and then travel to the negative terminal, but
you know there are too few holes on the n side.
Likewise, there are too few electrons in the p side, so this is not possible. However,
thermally generated carriers holes in n-type and electrons in p-type do keep coming up
and they move over the junction constituting a small current which is called the reverse
saturation current it is denoted by inaught or is. Now, this current does not depend on the
value of the applied potential however, it shows a very strong dependence on the
temperature.
Now, talking about the reverse biasing of a diode it is important to also highlight and
mention the phenomenon of voltage breakdown or the device breakdown when the
voltage in the reverse direction exceeds a certain voltage. Now, this will be reviewed
again when you will look at the I-V characteristics of the diode. Now, when we talk
about the breakdown, there are two types of breakdown mechanisms which are found in
diodes. One of them is a Zener breakdown the device which happens if the breakdown
voltage is in the range of typically 5 to 6 volts or less. The avalanche breakdown is the
second type of breakdown mechanism which is absorbed in P-N junctions. This happens
if the breakdown voltage is typically 6 to 7 volts or higher.
154
Now, what you see is a P-N junction which is connected in the forward bias manner. So,
you can see that the p-type side or the anode is connected to the positive terminal of the
supply and the n-type which is the cathode is connected to the negative terminal of the
battery.
So, as you can see in the forward bias condition the applied potential is in a position with
the contact potential of the junction. This results in the lowering of the height of the
potential barrier across the junction which causes the holes from the p side to diffuse
across a junction or to the n side. And likewise, the electrons can diffuse from the n side
to the p side across the junction.
Now, under the equilibrium condition I have plotted the concentration of the minority
carriers holes on the n side and the minority carrier electrons on the p side let us just
consider the n-type side. Now, because of this concentration profile that you see there is
a diffusion gradient and the holes they tend to flow from left to the right, as they flow
from left to the right they are actually neutralized by electrons. To maintain this
equilibrium profile which you can see here in the green colour electrons are replenished
by the source negative terminal and there is no problem with it because there is no dearth
of electrons now. So, there is no dearth of the supply of the charge carriers to
compensate this.
155
Now, similar explanation holds for the electrons on the p side. On the p side we will see
that the positive terminal of the supply will actually be replenishing the holes which will
be used for recombining the electrons which are trying to diffuse towards the left side.
So, you can see that this could actually result in a continuous process where the electrons
and the holes can be continuously supplied. And it is very much unlike the reverse bias
condition that we seen before. Hence, it is a forward biased configuration which is used
for the conduction of the diode.
It may be noted that the current during the forward bias condition is induced by the
minority charge carriers. They follow the mechanism of diffusion. Since both electrons
and holes are involved in the process of current flow that is why this device is called a
bipolar device, such devices are also called minority carrier devices because the current
is induced by the minority charge carriers.
Now, let us derive the analytical expression for the diode current, you know which is
actually called the diode current equation which will be applicable during both the
forward bias and the reverse bias conditions that we discussed just now. Let us look at
this diagram which I have drawn just a few minutes ago to save some time. So, what you
can see is here the excess minority carrier or distribution you know in the end in the p-
type.
156
So, you can see the see the distribution of the p-type you know carriers here on the n side
and you can see the distribution of the n-type minority carriers in the p-type region here.
And here in this is the depletion layer, and we have just marked the two edges of the
depletion layer as minus xp and xn as before as we have done in the previous analysis
and here x is equal to 0 it actually is the metallurgical junction.
Now, there is a very famous law which is used in the semiconductor physics, very
extensively which is the law of the junction and what it does is that it actually gives an
expression for the concentration of the minority carriers. Like for example, if you see
here at x is equal to x n which is the edge of the depletion layer, so it actually the law of
the junction it gives the concentration of the minority charge carriers at the edge you
know of the depletion layer.
And so, you can get see that the holes in n the subscript, n actually denotes that it is now
at about n side. So, p minority carriers p in n at a location x of n is given by this
expression which I have numbered as equation number 1 here as you can see. So, this is
actually equal to p and 0, where p and 0 is nothing but the thermal equilibrium value of
the minority charge carrier concentration at this point at this edge of the depletion layer.
So, this is basically you know the point we are talking about, and this entire thing this
height is nothing but the thermal equilibrium concentration of holes in the n-type region.
So, I can write a similar equation you know for n-type charge carriers in p, at x of p I can
write down a similar expressions there. So, now, we know the concentration of the
minority carriers at the two edges. So, I can we have it at x n and we also have it at
minus x of p.
Now, as we go towards the right side into the you know the n-type or the p-type material
you know there is this relation which is given by 2, again a very standard relation which
we are given here for the holes that shows the how the concentration of the holes in the
n-type it exponentially decays. It starts from a p n 0 which is the steady state offset due
to the thermal equilibrium value. And as per this relation now what is this relation here?
This is nothing but this particular height. So, this into, so this is just an into sign here into
this exponential of x minus x n divided by L p. Now, x minus L x n is actually the
variable which is taking us on to this side of x. So, we are just called it as expression 2.
And it is not difficult for you to see that a similar expression can be found for n, minority
carrier n in the p-type as you can see on the left side and we can find a similar relation
157
almost same expression with proper substitution of the variables we will get the
corresponding expression for n p.
Now, in this expression 2, what you see is L p is nothing but the diffusion length of
holes. The diffusion length of electrons is denoted by L n. And what it actually shows is
that how long into the n-type can the holes penetrate before getting recombined with
electron or getting neutralized by electrons. So, this actually is a very important
parameter smaller is the value of L p, faster would be the decay of this exponential curve
on the right side and the same thing with electrons in the p side.
Now, if you look at the this distribution, minority carrier distribution again in this
forward bias junction and what happens is that as holes move due to diffusion towards
the right side they combine with electrons. And to maintain the equilibrium condition the
external source, which in this case is negative here and positive here which is connected,
so the negative source here must supply the majority carriers electrons, depending on
how many electrons are used up in neutralizing the holes the minority carriers which are
trying to diffuse onto the right side. And similarly, on the right side we can see that as
electrons are neutralized by the holes the holes have to be supplied by this, so that
equilibrium profile is maintained.
This diffusion length L p, it is actually related to the lifetime the minority carrier life
time. So, what is the lifetime of let us say holes in the n-type if we denote it by tau p,
then we can say that the diffusion length of the holes in p-type is actually given by D p
into tau p, where D p is nothing but the diffusion coefficient of the holes. That we have
seen when we were doing the diffusion, diffusion current, and tau p I just told you is the
lifetime of the holes in the p in the n-type material.
158
So, these their standard values are actually available which from many sources which
can be referred to. So, basically now these minority carriers these holes they actually
give rise to diffusion. Current density J of p equal to q times D p by L p, p n 0, e raise to
power V over V T, where V T is nothing but kT by q that we have seen before minus 1 e
raise to power minus of x minus x n by L p. And in arriving at this equation at this
equation you know we have of course, made use of you know the equation that we just
wrote the equation number 2. And the other expression we have used is the expression
for the diffusion current which is nothing but just to kind of quickly recall for your ready
reference minus q times D p into dp by dx.
So, when we actually do a differentiation of expression tool and we actually use this in
the diffusion current equation along with you know the and when we use this, we
actually get this expression. So, let us call this as expression 4, and let me just call this as
expression 3.
Now, clearly the number of electrons supplied by the external source should correspond
to p n of x n which is the maximum concentration of holes on the n side. Hence, using x
is equal to x n in equation 4, the current density due to the holes is given by J p equal to q
times d p by L p, p n 0, e raise to power V by V T minus 1. Likewise, a similar analysis
can be applied on to the p side material and we can actually write J n you know equal to
q times D n by L n, D n is nothing but the electron diffusion constant and L n is nothing
159
but the diffusion length of the electrons, and E raised to power sorry there would be a n p
0 term into E raised to power V over V T minus 1. So, let us call these two equations as 5
and 6.
Now, we know that the net current that will flow across a P-N junction, would be a sum
of these. And moreover, if A is the cross sectional area through which this current is
going to flow from the density, the current density J we can actually get the absolute
value of I. So, if we can write I is nothing but A times q D p into d into p n 0 divided by
L p plus q times D n p 0 divided by L n, e raised to power V over V T minus 1 and let us
call this as 7.
So, this is the total current on account of the electrons and the holes which are
constituting the current as minority carriers, the explanation that we have just seen. Now,
a further understanding into this expression can be obtained by noting that p n o which is
the thermal equilibrium value of holes in the n-type region can be given by n i square the
intrinsic divided by N d and we can say n p o is equal to n i square divided by N a. So,
these are like when we have seen the doping, at that time we have obtained these
relationships and we have in fact, given some standard expressions and relations which
are very useful you know for our further analysis, works.
And when we do this substitution, then we actually get I you know the current due to
electrons and holes the total current that is A q into n i square, in the bracket it is D p the
160
diffusion constant for holes divided by L p into N d plus the second term within this
bracket which is D n over L n, the diffusion length of the electrons into N a the acceptor
atom concentration into e raised to power V over V T minus 1. This is the expression
that you know we will get. There is a small change that you can do the bracket actually
should come like this not here it would come like this and e should be inside. So, it is e
raised to power V over V T minus 1 here, it is e raise to power V over V T minus 1 here
and they are in the bracket just for easy readability.
So, the final current we can if we recognize that you know this dome here denotes the
saturation current, the current that will flow for example, when the P-N junction is
reverse biased let us say if this is I s, then I straight away get I is equal to I can just write
I s and I can say e raise to power V over V T minus 1. So, if I just number these previous
equations as let us say 8 and this one as 9, then I get this as a 10; question number 10
which is nothing but the famous diode equation which is used extensively in the analysis
of P-N junction.
The ideal characteristics of the diode based on this diode current equation that we
derived you know it can be drawn as follows.
What we will see is that let me just say this is the x axis and this is the y sorry this is the
y axis, let us say I is plotted here and this is the x axis or which the voltage across the P-
N junction is plotted.
161
Then what will we observe in the case of the forward bias is that there would be an
extremely small drop. So, there will be a small cut in potential here which is usually
defined by V gamma and its typical value as I said is physically it is a contact potential,
and the applied voltage that appears across the diode in a forward bias case must be
greater than this voltage which is V gamma or also called as its threshold or the cutting
voltage. And after that so there is no current till that time, because there are no charge
carriers which are able to jump. Till they have at least got energy to overcome this
potential contact potential, and after that it actually rises the current rises like this.
Now, in the reverse biased case where we are actually having a negative voltage which is
applied across a P-N junction, as you can check from the polarities of the battery which
are connected there you will find that there is no possibility of any current flowing, so it
remains 0, current remains 0. After a certain threshold voltage of the negative side there
is a sudden current that builds up in the negative direction, this actually follows the same
phenomenon you know that we discussed some time back about the breakdown of the P-
N junction when a large reverse voltage is applied.
After a certain threshold point, let us call it as thresh threshold for now a very large
negative voltage, so you are increasing the negative voltage across the P-N junction
slowly and a time comes you know when large current flows in the negative direction.
So, because of this threshold voltage it is more common to call as a breakdown voltage,
because what has happened is that the diodes, the diode has basically broken down the
mechanism has broken down now there is a reverse current which has taken over and
which is actually a very high value. So, it a huge current would flow unless it is limited
and would actually destroy the P-N junction.
Now, having seen the working of a P-N junction. Let us now analyze the P-N junction.
Let us try to see how the electric field and the potential they vary you know across the
junction length.
162
Now, what you see on the slide is you know a plot of the charge density on the two sides
of the metallurgical junction defined by x is equal to 0. It is assumed that the P-N
junction has a step doping profile. Further it should also be noted that thermal
equilibrium condition has been assumed. We have also plotted the variation of the
electric field and the potential. Now, what we are going to do is we are going to actually
get analytical expressions for the electric field and the potential, which we will later on
use in the further analysis and understanding of what we can do with the devices where
we use the P-N junction.
163
We begin by the famous Poisson’s equation. As you can see del square V is equal to
minus rho by epsilon is the Poisson’s equation, where rho is the charge density. So, what
we have done is we have just written some of these equations beforehand just to save
time. So, you can see that the Poisson’s equation may be written you know as del dot of
del V equal to minus rho of epsilon. Now, knowing that E is nothing but the negative
gradient of V, we can actually further simplify this and get del dot E equal to rho by
epsilon.
Now, we very well know as you can see in equation 5 how we define del. Del is defined
you know as given by this expression in terms of the partial derivatives of the unit
vectors i, j and k.
Now, if we use 5 equation 5 in equation 4, and we just put these partial derivatives
expression there dot E as you can see here, so this dot E you get a rho by epsilon. Now, if
we take this 3 dimensional analysis then it will get a bit complicated and difficult to
understand. So, to make a life simple let us just consider the analysis only in the x
direction, in which case the expression 6 you know instead of using these partial
derivatives in the x, y and z direction we can just take now in the x direction and this
partial derivative I can write as a full derivative dE by dx is equal to rho by epsilon
which is our equation 7. And we can treat this as a starting point of our analysis.
164
Now, let us consider the interval minus x p to 0. You know you can refer to the picture
on the slide and you can see that this is the interval towards the left of the metallurgical
junction into the p region. So, now, if we use the data for this left side interval in
equation 7. We can say that dE by dx is nothing but minus q N a by epsilon where minus
q times N a is equal to rho which is nothing but the charge density. The negative sign
comes because of the fact that the charge is negative for the interval minus x p to 0.
Now, this can be further simplified as per equations 9 and 10; you know and we can
actually get a very nice expression for the electric field in the left interval to the left of
the metallurgical junction.
Now, we can similarly find an expression for the right side of the metallurgical junction
which is which we can find and which will be valid for this interval and this is the
expression.
Now, E x which is given by 10 and 11 is what you see plotted actually on the slide. So,
you can see that in the left interval the electric field starts from 0 and goes down all the
way to the maximum value E max and then in the right interval it again goes down all
the way to 0. So, the maximum value of the electric field occurs at x is equal to 0 which
is nothing but the metallurgical junction. Now, apart from this electric field variation we
would also like to see how the potential phi would vary across the junction length.
165
Now, we know that phi of x you know it is given by the minus integral of E x with
respect to x if you are considering only x direction analysis. So, we can actually split this
integral on the right side into two parts one for the interval minus x p to 0 and the other
for the interval 0 to x n, and then for these E x values within this and this these two
components left interval we use expression 10, the expression that we have obtained for
E of x and we use 11 for the right side expression of E of x. So, when we actually
substitute these expressions for E x using 10 and 11, we actually get this equation 12
which on simplification. We can see gives us this expression.
Now, one interesting thing that you can try is you know is that you may use the what are
called the equilibrium conditions, where we say that the current due to holes across the
junction because of diffusion and drift will add up to 0 under equilibrium condition.
Similarly, we can say that the electron current due to diffusion and drift will also add up
to 0. So, you can see these two expressions and once you actually fill up the values and
you equate them to 0, you can actually show a standard expression that we get for the
contact potential phi c. You may use the Einstein equation obtained earlier in this lecture.
Now, during the forward bias the applied voltage across the diode or the P-N junction
would actually oppose the junction potential phi c. So, it is not really very critical for us
to know what happens to the depletion layer width, or the drift layer width. But during
reverse bias where the applied potential adds to the contextual contact potential that time
166
the depletion layer expands and it is important for us to evaluate what is the value of W
analytically and that is what we are going to do now next.
Note that if you see this picture which I have drawn here, I have shown that x p you
know the depletion layer into the p region and x n the limit of the depletion layer into the
n region they are not same, and they are not required to be same also it will actually
depend on the doping levels of the p-type and the n-type. So, this is something that is
true actually generalize the analysis that we are presenting.
167
Now however, in spite of this a symmetry, the charge equality requires that two times the
accepted atoms concentration into x p. Please refer to the picture on the slide. Must be
equal to two times the donor atom concentration into x n, and let us call this as 14. And
15, so you can see N a by N d is equal to x n by x p.
Now, as we have said before that for the reverse bias case the applied potential minus V,
it actually adds up to the contact potential. So, basically you know this whole thing it
becomes the net potential across the device and this we can equate the expression we
obtained in 13.
Now, if we use equation 15 and 16 we can then obtain x p and x n, and then if we add x p
and x n if you again refer to the image, we can get the expression for W, which is the
width of that depletion layer. And we can show that very simply that it will actually
given by this expression 17, now this left side term this term the one which have
encircled with red, this let us denote it by W z 0 and this is actually the width when there
is no external voltage bias applied. Now, we would like to see what is the value of the
maximum electric field that will occur. So, that we can get very easily by using an
expression 10, you might use even expression 11 if you want to try both will give the
same answer put x is equal to 0 and actually you end up with this relation. So, this is the
maximum E feel.
168
By using 15 and 17 we can obtain the x value of x p, and substitute in 18 to get this final
value of E max.
Now, when the reverse bias voltage becomes so large it goes to all the way to the
breakdown voltage, then at that time my electric field maximum value also goes to the
breakdown electric field value, and then I can actually get you know this final expression
for the breakdown electric field. And you can see that there is a negative sign here which
was there if you see equation 19, but that has got actually accommodated because there
is a negative sign associated with V BD also because there is a negative voltage which
we are applying.
Now, let us call this bring down electric field expression equation to be equation 20.
Now, using expression thirteen for phi c and the equation I have numbered as star in 20
we get this final value of the breakdown voltage across the P-N junction. So, you can see
that we can actually say that this approximation will have to be applied to simplify the
analysis and we can actually get this V BD approximately equal to epsilon into N a plus
N d into E BD square divided by 2 q N a into N d. So, this is the final expression that we
get and later on we will use them to further enhance our understanding about different
aspects of the P-N junctions.
So, next time we will see certain more aspects of these. I hope that I will be able to cover
some very interesting things in the coming lectures.
169
170
Prof. Vivek Agarwal
Lecture – 09
Power Diodes
Hello and welcome back. Power devices are the power switches which we have
considered or we have talked about when we first introduced the switching matrix to
represent a power electronic circuit. We will start with power diode. The reason is that
power diode is probably the most basic device out of all the power devices and if you
understand the power diode, then it becomes easy to understand other more complicated
power devices.
Now, one of the things that differentiate between what are called the power devices and
the low power signal level devices such as diodes and transistors is the power level. So,
here in power devices we are talking about very high power levels. Talking about power
devices in general in the power electronic circuits, we assume all the ideal properties of
the power device.
For example, we assume large big down voltage; we assume fast turn on and turn off
times; we talk about the low drop voltage drops and resistances across the device when it
is conducting and we also talk about a capability to dissipate a large power by these
devices. You will see that in due course when we do some analysis that these properties,
they are actually having some sort of a trade off between each other.
So, basically there is no single device which will have all the properties and therefore, a
person who is actually designing a power electronic systems and he is trying to use
power devices he will have to use what is the best available device with all the best
possible features. He will not get all the features and at the same time he may have
therefore, more than one choice. So, it may not be a very unique device that he can use,
but he can probably choose from given set of devices. Therefore, it is important that a
design engineer he understands what exactly is the working principle or the power
devices. What would be the meaning of a certain trade off?
171
Suppose and design engineer decides to compromise on the losses, in favour of getting
high breakdown voltages then what is the price he will have to pay for that and how he
will actually compensate in his design. So, all these things actually demand or in other
words I should say that a good and optimum design, it demands that we know the
working of various power devices and power diode it actually serves as a very good
precursor to such a study.
So, now let us just try to see what the structure of a power diode is and let us try to see
how it is different from a low power signal level diode or we will just call them a signal
diodes now. So, what you see right now is the structure of a power diode. The symbol of
the power diode is also given on the right side. So, what you can see that just like the
signal diodes you again see that there are two layers, one of them is p, the other is n. So,
there are two p and n type materials which have been brought together, but now you also
see a third layer which is sandwiched between the regular p and n layers and this is
actually a lightly doped n layer that you see.
Now, the doping levels are actually you can see in this slide these are the typical highly
doped and lightly doped concentrations which are given. So, we begin by an n-type
substrate which is heavily doped with something like 1019 atoms donor atoms per
centimetre cube on which is grown an epitaxial layer of a lightly doped n material. So,
basically you have much reduced concentration in this region of the donor atoms. It is
172
only 1014 atoms per centimetre cube. Over this is grown a highly doped p layer. So, again
if you see the concentration it is 1019 atoms acceptor atoms this time per unit per
centimetre cube.
And, this p layer is connected to the anode which is serving as anode of the device
through a metallic contact and at the bottom you can see that, there is a metallic contact
which is providing cathode connection to this device. The symbol of this diode is same
as the low level low power level diode, there is no difference, but I would like to mention
two or three important differences between the signal diodes and the power diode that
you can see on your slide right now.
One of them is in their i v characteristics. So, one of the things that you will see is that if
I draw the i v characteristic of the two diodes and let us say I use a green pen to draw the
characteristic, i v characteristic of a signal diode. So, you can see that it actually is going
to rise exponentially it drawn and this is what is called a cut in potential or contact
potential and the device actually starts conducting only after that.
What you will find in the power diode; however, when you actually try to measure its
characteristics i versus v? You will find that first of all the value of this cutting potential
is going to be 1 or 1.5 volts. So, this is going to be more; let us say I just kind of magnify
and show you. So, the characteristic would be something like this. So, what you see in
this in this drawing is that the red colour characteristic or the power diode is more linear.
173
Now, this linearity is coming because you have an n minus layer in this device which
actually is obviously, having much less charge carriers and therefore, contributes to more
voltage drops, ohmic drops which actually brings in this resistance type characteristic.
So, that is why you see instead of an exponential in the green one, you see this linear
curve or linear plot in case of the power diode.
While looking at the i v characteristics of the diodes, it is important to note that even
though the cut in potential or the threshold potential or the contact potential of the device
may be given by a standard constant value as we discussed in lecture 8. The actual
voltage drop across the device may be higher something like 1.5 volts or even 2 volts due
to the ohmic property of the device. Hence it is important that we do not confuse the cut
in potential or the contact potential values with the actual voltage drop across the device.
The other difference between the two is of course, the major difference as you already
what have noticed that you have only p and n layer in a signal diode while here three
layers you have an lightly load doped n layer, which is sandwiched between highly
doped n and highly doped p layers. So, that is of course, another difference. The third
difference point is that the doping levels in power diodes, they are higher both for p and
n type they are higher than what you observe in case of signal diodes. So, these are some
of the differences between the two types of diodes.
Now, why there is a need for n minus layer? Let us just study this part a little bit. Now,
we have seen the forward bias and the reverse bias operation of the diode earlier and in
principle it remains the same. The physics of the power diode operation remains the
same, both in forward bias and in the reverse bias. So, there is no difference between a
power diode working principle. However, let us see what happens because of the
presence of this slightly doped n layer in between the p and the n layer, the main p and
the n layers.
So, in the last lecture that is lecture 8, we had obtained an expression for the breakdown
voltage V BD which was given by if I remember correctly by equation 21. So, you may
refer to your lecture 8 for this equation. Now, assuming let us say that we consider a
special case now where we say that the p side, of the p-n junction is more heavily doped
compared to the n side. Now, which actually means that Na is much larger than N d.
174
We can show that
VBD = K1/Nd …… (*)
let us call this as an equation star. Now, similarly if you use equations 10, 11 and 13
from lecture 8, and if you use the star equation that we just obtained we can also show
very easily that the width of the drift region actually is about you can show
approximately
Wdrift = K2 VBD ….. (**)
where K2 is basically a constant which will come by substituting the numerical values so,
the various variables which are there in the equation.
Now, let us call this as equation double star. Now, what are the conclusions that we can
draw from these two equations star and double star. So, one of them is a very important
observation that shows that if you need a large voltage breakdown capability; so, large
VBD if we need it means that the intensity of the donor atoms must be as small as
possible as per the equation star, which means that we are saying that we need our n
layer to be lightly doped.
And, the second observation we can say is that a large VBD will also require a large width
of the depletion layer. So, which means that my W drift should be large if we want to
175
have a large V BD which basically this means that when well I will apply a large voltage
I across my p-n junction, then my drift layer should be able to absorb the entire the
depletion layer.
So, this is the major thing about the power diode basically the idea of using this n minus
layer is to provide a high reverse voltage capability to this device and this you will see
would be very apparent when we use applications such as rectifiers where you actually
have large reverse voltages coming across these devices. So, you will see this happening
even in the case of SCRs, the silicon controlled rectifiers and you will see that they also
need a large negative voltage capability because of this.
The other devices that we will study in due course such as the MOSFETs, the IGBTs and
the power BJTs they lack this particular feature they lack this capability of being able to
withstand a large reverse voltage across them and we will find that because of this reason
their application in processes like rectification is going to be limited.
Now, the question is that if power diode is able to withstand a large negative voltage
reverse voltage across it, then and we are achieving it by actually inserting a lightly
doped semiconductor layer n minus layer in it then obviously, you are having a region
which is not having sufficient charge carriers.
176
So, if I draw the situation let me say that this is the diode and let us just divide it into
three regions p plus n plus and in between n minus and now, you have let me just use this
blue pen now and now let us say we have the charge injection which is taking place. So,
let us say we are trying to study the diode now during the forward conduction, when it is
conducting as the forward bias case that is a time when the current flows and that is the
time and you expect the drops would take place what happens.
If we just go by what we have seen so far without applying any other explanation we will
find that it turns out to be a device with a huge ohmic drop a very big ohmic drop and
probably the device would be impractical for use, but a very good thing that happens in
bipolar devices such as a power diode or the power bipolar junction transistor and even
the insulated gate bipolar transistor that we will see later is what is called conductivity
modulation.
So, when you have your diode which is conduct which is actually connected in the
forward bias; so, you can see the polarity of the external voltage source that is applied,
now what will happen that holes will be injected from p region to n region during the
forward bias.
Now, because n region is not having enough electrons definitely not as many as the
corresponding holes in the p region because that it is highly dome if the injection of
holes from p region to n region or n minus region is large your electrons in the n minus
region will not be able to neutralize the holes, and therefore, there would be an actual
accumulation, at least temporarily of plus net plus space charge in this region in the n
minus region.
Now, the moment there is a net space charge which is positive in this n minus region you
have electrons which actually start jumping and going into this n minus region they all
cross over. So, what happens basically is the net result of all this is that you are this
particular region which was actually devoid of charge carriers because of which we were
actually afraid or an anticipating that there would be a big ohmic drop is actually has
become rich in the charge carriers.
Now, this phenomenon of injecting the holes from the p type to the n minus layer and
injecting electrons from the n layer to the n minus layer we call this a phenomenon of
double injection and it actually leads to increase in the conductivity, which is also called
177
conductivity modulation. We will find that this is going to be the case with all the bipolar
devices. In fact, all the bipolar devices will have this phenomenon and therefore, they are
usually having much less on losses, on-time losses are much less in bipolar devices.
Let us try to see that, what exactly properties of diodes are, when it is used in power
electronic circuits because that the power electronic circuits require this power diode to
actually go on and off. So, this is going to be working as a switch an uncontrolled switch
in a power electronic circuit. So, it is important that we see the switching characteristics
of a power diode when it is actually used in a power electronic circuit.
So, what I have done is that on this sheet I have just drawn the typical switching
waveforms for the voltage and the current across a power diode. So, what we are doing is
we are looking at both the turn on and turn off switchings of the diode. So, what you see
on this side on the left side if I just draw this on the left side is a turn on and on the right
side is the turn off switching of the diode.
During the turn on let us say that the reverse voltage across the diode just when we
started to turn on the device just before that it was let us say -VR and the current was
obviously 0. So, the current starts rising from 0, and this is what this line shows here it
goes up with a slope which is defined as diF/dt and would be actually determined by the
external circuit elements which are there like the inductance. This could be the legitimate
inductance or it could be these free inductance that might be present in the circuit.
178
Now, at the same time during this up slope of the current you have the excess carriers
which are all getting accumulated in the lightly doped n minus layer. Now, what is
happening to the voltage across the diode at that time the voltage is also rising. It is
rising starting from -VR and it is going up.
Now, ordinarily we would have assumed that it would just go and get settled at some
value here which is the forward voltage drop, but there are a few issues one of them is
that during this time which I have marked as t1 during which the current through the
diode is building up the excess charge carriers have still not built up completely. So, still
there are there is a high resistance across the n minus layer or the drift layer of the
device. So, that is one thing.
Secondly, there are stray inductances on a corner of the silicon wafer as well as the leads.
So, this voltage what you see it actually tends to o shoot and it actually you can see that I
have marked this as VP it actually reaches a value which is VP. Now, what happens at the
interval after t1? So, it is marked this as t2, now what happens during the t2 is that my
current I the diode current has settled.
It has settled to a constant value here. So, let us say this is the constant value IF. The
excess carriers have now all built up. So, the losses or the voltage drops across the n
minus layer have also reduced drastically, and they have actually because of the
conductivity modulation that we have seen before you have very low drops there and that
is the reason you find that the voltage across the diode it falls quickly from it is value of
V P and it comes to it is on state value which is it is regular normal, this it actually comes
to the V ON value.
Now, this is where it continues as long as the diode is conducting the forward current IF
which is shown here. Now, let us consider the turnoff switching of this diode, but before
that I just want to say that this t1 plus t2 it actually represents the forward switching on
of the diode or we can just call it as the forward recovery time as it is called by many
authors. So, this is a forward recovery time which a diode needs to turn on.
Now, coming to the switching off characteristics of the diode let us say that at the
beginning of this interval t3 which I have just marked here let us say the circuit
conditions are now such that the diode is now about to start turning off. So, you can see
that the current falls starts falling and let us say that the current now as the current is
179
falling let us denote this current by iR and let us say that this load therefore, will be
diR/dt. Again, it will be determined by the elements of the external circuit in which this
diode is placed.
Now, till the end of the t3, the diode current continues to fall and it actually comes to 0,
but unfortunately the excess carriers which have actually got built up during the turn on,
they are still there and they are still not removed. So, what happens that the current
continues to flow? So, you can see that the current actually reverses here after t3 the
current has reversed; it is now negative.
And, you can see that this is all because of this stored charge that we have seen here and
at a certain point, we find that this current reaches it is maximum which we are denoting
as the reverse recovery current Irr and beyond that all the excess carriers are getting
stripped off because of the reverse current now there are hardly any excess carriers left.
So, the diode, it obviously, cannot meet the demands of some external inductances which
might be there in the circuit and so, it actually quickly falls to 0 marking the end of this
recovery time. So, I have just marked these intervals when the current became negative,
and after losing all the excess charge carriers it has quickly dropped to 0 I have just
marked them by interval t4 and t5.
Now, typically this t4 plus t5 this interval is actually referred to as a reverse recovery
time and we denote this by trr and this triangle of course, I have to say that this is just an
approximation if this will actually have a very complex waveform. So, this we have just
shown as a triangle because it is enough for our understanding and analysis.
So, this triangle it actually this area it represents the extra charge that has got
accumulated during the on condition of the diode which has now been swept out and the
p-n junction of the diode has regained, it is capability to block the reverse voltage. So,
basically now we are back into the condition where the junction can now block any
reverse voltage applied to it and hence we can actually reverse bias diode.
Now, what I have written here is that how much is this charge we can approximately
right by just taking the area of this triangle and we can just say Qrr is nothing, but half
into Irr which is nothing as we have seen is the peak the peak reverse current Irr into trr
which is nothing, but the base of this triangle. So, half into the peak current Irr into trr
180
that will give me the area which is nothing, but the excess charge carrier that has been
removed during the turn off process before the diode is ready to take the next condition.
When the current is just falling during the interval t3, there is hardly any impact on the
voltage. So, the diode it continues to be in its normal on state because the current is still
flowing, and it continues like that the junction of the diode is not yet reverse bias. In fact,
the junctions must remain forward biased. So, that there is a possibility for the exit
carriers to be swept across swept out.
Now, at the end of interval t4 when the charge carriers have been swept off the diode
junction it becomes reverse bias and when it gets reverse biased you can see that the
voltage suddenly falls down, and again because of the stray inductances shows some
over shoot we just call this peak value as VRR and after that it just goes back and it just
stays put at the reverse voltage that has been applied to the diode to reverse bias it that
completes this process of the switching off or turning off of the diode.
So, this interval t4 plus t5 this is called the reverse recovery time which is actually
associated with the turning off of the diode now the forward recovery time when the
diode is turning on is not really a very critical thing because we can see that this actually
can take place very fast and it does not really have any delays as such associated with it.
However, it is the reverse recovery time at when we diode is getting turned off that we
are more interested because that is more critical because unless the store charges are
actually removed or swept out, we cannot say that the diode has actually switched off
and it has a direct bearing on the speed at which the diode can actually work. So, what
should be the frequency what is a frequency at which my circuit is operating, where I
want to use a particular diode which; obviously, we determined by this reverse recovery
time that a diode actually exhibits that we have just seen.
181
So,
t4 + t5 = trr
Now, I can say that peak reverse recovery if you look at the diagram that I have shown
before that the switching waveforms of the current and the voltage that I have shown
before. So, if you look at the current waveform at the time of turn off, you can actually
see that the Irr valued can actually be given by this term t4 that interval time duration
into the slope of the current at which the current was actually falling down. So,
Irr = t4 diR/dt ……. (A)
Now, we know that
t4 = trr - t5 = trr / (s+1) …… (B)
where s is nothing, but the ratio of the time t5 to t4.
So, this is that ratio. Now, often we denote this t5 by t4 the people who work with diodes
they call this as a snappiness factor snappiness factor and usually t5 is much less than t4
and we say that the such a diode is a snappy diode, which can really turn off fast that is
what it is, but if you want to have this t5 to be comparable to t4, then we got what are
called non-snappy diodes and these are also called soft recovery diodes.
182
Now, one of the important things here is that we would like the current during t5 not to
fall very suddenly in certain cases because this might cause large voltage spikes to the
other devices in the circuit. So, if we I if somebody really analyzes the circuit as we will
see later on in this course we will see that a very small time t5 if the interval t5 is very
small we will find that the current would be so snappy it will be so fast that the di by dt
will be very high and this will actually cause large voltages to appear across some other
components and the devices in the circuit, which our circuit may not allow and that is
why we might have to go for what are called soft recovery diodes.
Now, using this expression that we just got if we just call this as B this relation is B, and
we also know that Qrr the charge that got accumulated which has to be actually removed
during t4 plus t5 interval
Qrr = Irr* trr/2 ….. (C)
and let me just call this as equation C. Now, if I use the value of Irr from a from equation
A and if I substitute this into C, I can say
Qrr = diR/dt =trr2 / 2(s+1) ……. (D)
D can lead to this more comprehensive in expression for the diverse recovery time. So,
trr = √2 Qrr(s+1)/diR/dt …… (E)
183
Now, if I use E in A use basically this expression that we have just obtained E in
expression A then we can also get straight away Irr, the peak river reverse recovery
current; we can get this equal to
Irr = √2 Qrr diR/dt/ (s+1)…… (F)
Now, during the forward bias when the when we were looking at the on time
characteristics of the diode, we find that when the current I F was is flowing in the diode,
obviously, it has caused a lot of excess charges charge carriers to get accumulated in the
device mainly they the drift region, the drift layer.
Now, let us say that this charge which has got accumulated is QF let us say this is the
charge. So, this is the charge which has got accumulated during ON operation of the
diode. So, this QF we can show is also equal to nothing, but the forward current into tau,
where tau is nothing, but the average lifetime of the carrier.
So, basically when a carrier is there in a certain region how long it can survive before it
recombines either with a hole or with an electron. So, this charge carrier if it is a hole
how long it is before it recombines with an electron or if it is an electron how long it is
before it actually recombines with a hole. So, this average time is actually considered to
be their lifetime how much they can live in a particular region of a semiconductor. This
is how it is defined is the most basic term.
Now, so, basically what we are trying to say is if we know the lifetime tau and the
forward current we know how much charge has been accumulated in the device. Now,
some charge when you are switching all the device some of this charge QF. So, this QF
actually has two parts; one of them is let us say QF1, which is actually cleared off by
recombination.
But, there is another component which is as we have seen the current actually reverses
and that is what is clearing it. So, that it is basically getting cleared by the reverse
current. So, that let us call as the component number 2. So, clearly both QF1 and QF2 are
smaller than QF individually.
184
Therefore, from what I just discussed or I just explained I can say therefore, that the Qrr
which is swept out or swept away where the reverse current is less than Q F, that is what
we just concluded. Our QF2 part which I just explained is nothing, but Qrr in this case.
So, that is less than this and let us call this as an expression H.
Now, once we know this I can say trr is less than just using my expression E that we
have obtained earlier I can say
trr = √2τIF/diR/dt …… (*)
and similarly, we can say
Irr = √2τIFdiR/dt …… (**)
So, these are very nice and compact relations that we have got let us call them as star and
double star; these two expressions. Now, this is an inequality, this is an expression which
is an inequality expression both of them, but in the worst case I do not think that there is
any confusion about this that we can just replace these signs less than signs or inequality
sign by an equal to sign.
There is another type of diode not a silicon p n junction diode it is a different type of a
diode which is used extensively in power electronic applications and this is actually
called a Schottky diode.
185
Now, a Schottky diode is actually represented by this symbol. This is a Schottky diode
symbol and what it actually has is an interface of metal with semiconductor. Now,
usually both n or p type semiconductors can be used, but in practice it is the n type
semiconductor usually it is the n type semiconductor which is preferred for this, and the
metal which is usually used for this application for the Schottky application it actually
could be molybdenum or it could be tungsten, but they have their own advantages and
disadvantages. Schottky diode is actually a majority carrier device. It is not a minority
carrier device like the bipolar power diode that we have considered so far a p n junction
that we have considered so far it is actually a majority carrier device.
Now, when we talk about a majority carrier device that basically means that its
mechanism is such that it does not really need to pile on or kind of accumulate the
minority carriers on to the other side of the junction. Actually the flow of the current in
this case is constituted by electrons which are actually generated through thermionic
emission.
So, both metals and the semiconductor they actually have this thermionic effect because
of which the electrons are generated. One of the very big advantage of the Schottky
diodes is that they actually incur very low forward voltage drop as a diode is conducting.
The other advantage of course, is that because there is no accumulation of minority

carriers there is no reverse recovery time problem here. So, these are really fast they
186
switch really fast, but the Schottky diode has a reverse leakage current that is much
higher than the silicon p n junction diode I want to actually show you some typical data
sheets.
Now, for the power diodes typically we would be interested in looking at some of these
parameters. One of them is the junction operation temperature. So, the specification what
is the maximum junction time operation junction temperature or junction operating
temperature. So, what is the maximum temperature allowed as per the datasheet. This is
a very important parameter for the design; then what is the reverse recovery time trr; then
what is the repetitive peak reverse voltage.
So, for example, in rectifiers if you will see you have these reverse voltages which are
going to come across the device when it is reverse biased. So, what are these peak
voltages which are repetitively coming and appearing across the device? This is again a
very important parameter and it is also important to see what one time maximum reverse
voltage that will appear across the device, then what is the peak forward search current in
the forward direction, what is the maximum search current that is allowed? What is the
average current with the device will allow the diode will allow to flow? What is the rms
current that it will allow? What is the peak current that it will allow? So, all these are
very important parameters that will somebody will typically look at when looking at the
data sheets of the power diodes.
187
You can see there these parameters the you can see the forward average current the
forward RMS current and you can see that for this example which actually happens to be
a 150 ampere stud type diode. You can see all these values are given you can also see a
very interesting thing here which is saying i squared t at50 or 60 hertz. So, what is that
unit which is basically the kilo ampere square of seconds? What does it show? It
basically means that we are trying to put a protective device we are trying to put some
protection for the diode.
Now, i square t is given to you. So, some value is given to you let us say do we just call
it value and it is also given to you that how much time a certain current is allowed. So,
you can actually get from here i square to be equal to value divided by e and then you
can just get the value of i by square root of this term. So, i is equal to square root of value
over t. Now, this will give you basically this is the limit that is given by the datasheet.
So, if your current is going above this you must have a fuse in place that your device
actually gets cut off from the supply so that it is same otherwise you might end up
destroying the device.
188
We can also see on this slide which is the second page of the datasheet of this particular
diode the maximum forward voltage drop which actually is about 1.47 and you can see
that all the other values are given at which this particular forward voltage drop will hold.
So, it also gives a temperature of 25 degree centigrade it says that what is the peak
current which is 600 mps and so on under those operating conditions it says that the
voltage drop will be 1.47.
189
And here there is a datasheet that is of a Schottky diode basically they are calling it an
ultra fast recovery time. So, basically you have the trr to be extremely low in this diode.
Schottky diodes as I told you are actually famous for being very fast with very low
reverse recovery time. And, you can see the maximum peak repetitive voltage, and
maximum working peak reverse voltage which is about 200 volts for this particular
device, then you can see a forward current average which is IF average is given, then
there is an rms value of the current is given which is allowed through this device.
And, on this sheet you can see what the reverse recovery time is. So, it clearly says that
this particular Schottky diode is having a trr reverse recovery time of 17 nanoseconds
when a forward current of 100 amperes is maintained and there is diF/dt, the rate at
which the current is varying is also given as some minus 200 ampere for microseconds
look at the really fast slope there, and the reverse voltage maximum is 133 volts which is
allowed and even the temperature is given as 25 degree centigrade.
So, under these conditions it is given that what would be the reverse recovery time of this
device and if you see the third row, it also gives the maximum reverse recovery current,
the Irr value that we have involved in our analysis.
Now, based on the ongoing discussion that we are having about the speed at which the
diodes can switch and about their maximum voltage capacities and so on, it is customary
to actually classify or categorize the diodes into three types.
190
These are actually the general purpose diodes which are typically used for rectifier
applications 50 hertz, 60 hertz rectifier applications. So, here the speed of switching is
not really important. So, these diodes can be slow. The other type of the diode is what is
called the fast recovery diodes, now, these are the diodes which we want to switch fast
and therefore, we expect their trr to be very low or extremely low.
These are the ones which are typically used for inverters and our DC to DC converters or
chopper applications and the third one are the Schottky diodes. They can switch really
fast as the switching times are typically nanoseconds and they are used in switch mode
power supplies.
The problem is that because we are not able to go to high power the Schottky diodes they
have certain limitations with the voltage, the maximum voltage that they can go to they
are not really available for use in other applications like inverters and choppers. So,
schottky diodes the third type, I can just say that it is mostly the switch mode power
supplies of the appropriate power levels where it is used.
Now, I will just write down some typical numbers here. So, I will I can say that the trr if
I just say this parameter and compare these three. So, for trr I will say that the general
purpose diodes would have something like 25 to something like 50 microseconds. The
fast recovery diodes would have typically 5 microsecond of trr reverse recovery time,
191
while the Schottky diodes would have in nanoseconds. So, we just saw when Schottky
diode which had something like 70 nanoseconds is the reverse recovery time.
The other parameter the important parameter that we know is a forward current which
can be carried by the diode and for the general purpose rectifier the forward current
rating typically is up to a few kilo amperes. It is really huge, thousands of amperes can
carry for fast recovery diodes typically it is less than 1 kA 1000 amperes typically less
than that there is also pretty large. But, they are of course, not having the same capacity
as the general purpose diodes, and the Schottky diodes will typically have less than 300
amperes as their rating allowed rating, alright.
About the rated voltage that these diodes these three types of diodes can take the rated
voltage the general purpose diodes typically 50 to 50 kV, 50 volts to fifty kilo volts. So,
huge voltage they can take the fast recovery diodes can take something like 3 kV up to 3
kilo volt and the Schottky diodes as I said before they are capable of only up to about
100 volts. Now, there might be diodes which are now coming Schottky diodes where
people are actually are able to sustain higher voltages.
With these words I end this session and in the next one, we will actually start with a new
device, new power device and we will see what other benefits we can read from using
those devices that we follow.
192
Prof. Vivek Agarwal
Lecture – 10
Thyristors
Hello and welcome back. Now, after having seen the working of a power diode, its
switching characteristics, its construction, so on, now let us move on to a bit more
complex power devices which are more involved and actually they can do lot more than
the diode which as is an uncontrollable switch. Now, let us begin with this type 2
switch, the Thyristor.
Now, thyristor actually is a generic name for usually the 4 layered devices which came
up in1950s. And the salient feature of these 4 layered devices is that they have an
internal regenerative process which actually results in their turning on. Now, we will see
that under the thyristor family, there are several devices which were developed; one of
them is silicon controlled rectifier which we are going to see in details. The limitation of
the silicon controlled rectifier though it was possible to turn it on by using the gate
terminal, it was not possible to turn it off by using the same the same terminal, you
cannot have any control you cannot exercise any control over the turn off of the device.
So, basically, I termed it as a car which we were driving which has actually whose
brakes are failed. Now, the question is how to stop this car. So, this kind of a challenge
was always presented by the SCR.
Then a variant of this was actually developed which was called the gate turn-off thyristor
GTO which actually could be both turn on like a SCR, but could also be turn off through
its gate terminal. So, we are going to look at these mechanisms how this is done. So, this
particular session is mainly devoted to the working the construction of the SCR and its
close variant the GTO, the gate turn-off thyristor.
193
So, what do you see on the slide here, on the screen is actually the structure of a silicon
controlled rectifier that, we just draw this here very quickly. So, that I can use this to
explain a few things to you later on.
So, this is a structure and it has 4 layers as we talked sometime back, I just mentioned
that. It has a p1 layer, we call it p1 and p2 because there are 4 layers and two of them are
p and two of them are n-type. So, p1 plus, plus means very highly doped, n1 with a
minus sign which means it is a lightly doped layer and we have the other p which is a
194
moderately doped layer and this actually is going to be the gate of the device. And we
have this n2 plus layer. So, the gate I can just probably show here as a gate. And this n2
plus, so this second n layer which is very highly doped is actually through a metal
contact becomes the cathode, which we just usually because of the week connection we
actually prefer to noted by K. And similarly, this lower layer p1 plus highly doped p1
layer we just show that this is p anode denoted by A.
But for the reasons of working of this device, the basic principle that control the working
of these kinds of devices, it is usually the case that the gate and the cathode structure
there highly inter digitated. So, you actually have the gate and the cathodes which are
interwoven with each other.
So, if I was to just show you the top this portion here, I can show you something like this
p2 and of course, there are other layers here p2, and into this p2 you will see that there
are several n2 plus layers regions. So, you can see that they are all the time, they are they
are actually distributed all over this p region and that is how what we what we mean by
saying that they are highly interwoven. You can actually I could actually mark the
cathode and other place things also here.
So, let me just say a metallic contact here, a metallic contact here, an n2 plus and another
one here and so on the many of these, they will all we connected at one place and this is
called the Cathode. Likewise, you have the gate structure here, the gate structure here
and they are also, so let me just use the green colour here sorry I should have used a
green colour here its these are connected, and these are giving you the this has to the gate
terminal. So, you can see that the cathode and the gate they are extremely interwoven
with the each other and this actually plays a very prominent role, and we will see that it
is a very critical role in the operation of the device.
Now, looking from the top this structure of the anode and the gate being interwoven with
each other or inter digitated with respect to each other can be actually represented. We
will just have a very simple appearance here, like this.
195
So, I will just use a red colour may be to denote the cathode part, this all cathode and I
just use a blue colour to denote the gate. So, you can see that this portion is gate and here
there is cathode. So, there are some silicon controlled rectifier devices which will have
this kind of a appearance, a top view of the device. So, you can see how the gate and the
cathode are distributed with respect to each other.
Now, there is another one which is a little bit more complex and that is specifically used
for the cases, where the silicon controlled rectifiers are meant to use for high frequency
operations. What you see just now this one here is actually for low frequency operations,
really when you are operating at a very low frequency. These high frequency ones are a
little bit more involved, so let me just draw them a little bit more carefully. So, you can
see that this is how; so, you can see this is how the gate would lo in this case and the
remaining thing is your cathode. This is a very important aspect for SCRs and the GTOs.
Now, let us try to understand how the SCR works. So, let me just draw this picture.
196
So, this is my SCR I am just trying to draw it now horizontally. So, on the left side let me
say this is a metallic contact to the anode and on the right side let say we have a metallic
contact with K and you have these 4 layers that we know very well there are 4 layers
different types of semi-conductors. And we have already seen that this is p1 very highly
doped with a plus sign n1 with a minus sign means it is a lightly doped n layer, p2 here
without a plus or a minus means it is a moderately doped or layer of p-type
semiconductor and here it is n2 plus, so highly doped n-type layer which is there. So, this
is how this los in horizontally SCR.
Now, let us see what the various modes in which the SCR operates are. It will actually
helps us construct the I-V characteristics of the silicon controlled rectifier in due course.
Now, there are two ways or there are two conditions under which the SCR can be
actually off. So, it is not just that when the device is reversed biased that your silicon
controlled rectifier would be off. It can actually be off even when a forward voltage is
applied to the SCR and this is called the forward voltage blocking capability or what
differentiates the fully controllable devices or semi controlled devices with a device like
a diode which is uncontrollable switch or device. So, there are two cases let us see what
are those conditions when SCR is off.
Now, this can happen as I said both when there is a plus VAK applied across the device
or there is a minus VAK applied which means it is a reversed bias across the device and
197
plus VAK this is a forward bias across the device. So, let us just put A here. Now, when
VAK is negative with G open, the G is open I sorry I did not mark that G here I think I
should mark that here. So, with G open we come to this case, so what happens under this
particular condition? We will find that J1 and J3 are reverse biased, and J2 that is
forward biased where I can just mark here also these junctions J1 and J2, I can say this is
J1, this is J2 and this is J3.
Now, the point is that this J3, if you see this junction J3 it is actually surrounded by two
heavily doped layers. You can see that both p2 and n2 are heavily doped, p2 is of course,
moderately doped, but it is still very heavily doped relatively very high doping still. So,
J3 is obviously, should be concept that we have learnt while we were studying the power
diode, we know that the J3 junction cannot withstand a large breakdown voltage. So, it
actually we will find that it actually breaks down at very low voltages, very low negative
voltages that will come across this junction it will not be able to pair them. But J1 if you
see that is surrounded by one highly doped layer p1 plus, but the other one is n1 minus
which is a lightly doped layer. This one here when it is a reverse biased this junction J1,
the depletion layer can be accommodated in this slightly doped n1 minus layer and that is
why the capability of J1 to stand a large negative voltages very high.
So, J1 is where therefore, will be able to absorb this negative voltage, there will be a
applied across the device. Usually the in practice we will see that n1 minus this layer
here that is kept to be bit long. So, that it can absorb the depletion layer, because as we
said it is going to be basically, withstanding all the negative voltages because J3 is not
capable of doing that.
Now, let us see what happens in the B case that is when VAK is positive, and, again let
us say that G is open and assume VAK to be less than V BO the break over voltage in
the forward direction in these two conditions. Now, we will find that J1 and J3 are
forward biased while J2 is reverse biased, but this is not a problem that J2 is a reverse
biased because J2 also has this n1 minus layer on to its left. So, just like J1 has n1 minus
layer on to its right J2 layer also the J2 junction also has a n1 minus layer to its left. So,
basically, what we are trying to say is that both J1 and J2 can withstand a large reverse
voltage and they will have a large voltage break down capability. By the way if the break
down occurs in these devices it is mostly through avalanche break down.
198
Now, we want to understand little bit more about this device. And it is observed that in
this context it really helps if we can actually split this 4 layer device into two transistors
which are connected in a very specific manner that is what we are going to see now. And
it is very easy with the help of this what is called two transistor model to understand the
working of the the several aspects of the silicon controlled rectifier.
So, let me just first prepare the ground for creating this two transistor model. So, please
see that this is just the one of the half the anode side half that I have drawn, and I have
just kind of sliced through this device and the other side, will be also will be just
analogous to this we cathode side that is k and I can just say that this is my n1 minus
layer and this is p2 and this is n2 plus these are the two layers and here is the gate. But I
will also have to obviously show the connection between these two. So, we know that
this p2 is actually connected with this and this n minus n1 minus is connected through
this connection.
Now, what we see here on the left side if I just lo at this structure and also lo at this right
side structure these are nothing but two transistors one of them is a p n p transistors the
other is an n p n transistor. So, if I give them their symbols that are used to represent
them and I redraw this I can probably make it much simpler to look at. So, let me just
draw that. So, this let me just draw the p n p type transistor first. So, let us say this is the
anode of the main device of the SCR and this is the first transistor which is the p n p
199
type, let us call this as T1 and in the bracket I will just say p n p type and it actually is
having a current ic1 the collector current there will be associated with it and of base
current ib1 there will be associated within. Of course, we also have the emitted current
which is associated with it.
We can similarly draw the transistor corresponding to the second half of the SCR shown
above. So, this is that n p n transistor and it is not very difficult for you to see that this
base of the second transistor which I am calling as T2. Now, this is n p n-type first of all
transistor and this one the base of this is connected to the collector of the first transistor.
And we can also say that this is getting a base current ib2 and let us say I should
probably use another colour may be a blue colour here just to say that this is the i G
which is probably coming from some circuit which is designed to provide the gate
current for control of the device SCR. So, this here is going to k.
Now, let us quickly understand a few things. Let us supply a forward voltage across the
device that is the anode is more positive compared to the cathode. Assume that the
applied voltage is less than the forward break over voltage V BO which we will discuss
when we come to the I-V characteristics of the device. When iG is 0 both p1 and p2
transistors are off; when iG is increased slightly, let us see what happens. We find that
this actually means ib2 increases which means ic2 results, but that is connected to the
base of p1 to this means because ic2 is equal to ib1, ic1 begins to flow and the last
important point here is that as this ic1 grows.
It actually pumps in extra base current into T2 which in turn increases its collective
current and which in turn increases the base current of the first device and so on. So, you
can see that this is a cyclic process this is a chain reaction which is starts and it continues
till both the transistors they actually get into this saturation mode and completely that
turned on.
The turn on of the SCR can also be explained using the Aber Balls model according to
the Aber Balls model I A. So, let me just write this equation.
200
The current flowing through the device
IA = ( α2 IG + ICO1 + ICO2 )/ (1 - α1 + α2) …….. (A)
where ICO1 and ICO2 represent the leakage currents of the device and α1 and α 2 are the
base transport factors associated with the two transistors.
Now, in the absence of the gate current and assuming that VAK we applied forward
voltages less than the break over voltage VBO, α1 and α 2 are small and therefore, IA is
equal to ICO1 plus ICO2; just two leakage currents add up and it is a very small current.
When the forward voltage applied across the device VAK tends towards V BO the
reverse bias layer J2 it expands into the n1 and the p2 layers. As this depletion layer due
to the reverse bias it expands into n1 n p2 layers the base areas of the two transistors they
reduce. This results in a higher gain value and the value of the αs also go up.
Now, as α1 and α2 values go up α1 plus α2 it tends to 1, and if I just number this

equation as A and I use this fact that α1 plus α2 is equal to 1 in equation A, we can see
that I A the current through the device it raises which means that the device turns on.
Now, as the that the two transistors the p n p and the n p n transistors they go into
saturation and the anode current of the device it builds up, there is a time when this
current re is actually exceeds a certain value which is called the latching current. And I
can just say if I just denote the device current by IA, I say the moment my IA is greater
201
than or equal to IL, where IL is nothing but the latching current. Now, the device
becomes uncontrollable through G.
So, now even if we remove the G source or even we apply anything to the G source to
somehow stop the device from conduction it will not help that control is lost.
Interestingly, to understand the concept of the latching current we need to go back to the
original 4 layer structure of the SCR.
202
Now, I will just quickly draw this 4 layer structure. So, these are the 4 layers and if we
have the anode on this side and the cathode on this side I can just mark these various
layers as n2 p2, n1 p1.
Now, as the device turns on the p2 layer it gets flooded with electrons coming from G
the gate and together with the n1 layer these two they actually constitute what is called
an intrinsic low doping density n layer of the device. Thus, the SCR it actually reduces to
a power diode structure which we have studied before. Now, we know that a diode is
uncontrollable. Now, that is why the gate of the SCR it loses control over the device ones
it has crossed a certain value of current called the latching current. In fact, after the SCR
has turned on, the two transistor model is not really valid.
Now, one very important thing is about the switching off, as we said the gate control is
lost ones the device current has exceeded the latching current. Now, let us say that by
some mechanism this current is brought down. Now, while this current is being brought
down the device actually is not turned off is not considered to be off, as long as it has not
gone beyond another very important parameter current parameter which is call the
holding current. So, while we are trying to switch off the SCR my I A must go below this
current which is called the holding current and the holding current is less than the
latching current, and for the ideal case it is very common to take I H to be equal to 0.
So, you actually want your device get in to fall to 0 and now you must maintain reverse
bias voltage across the device for a certain stipulated time t q, which is called the circuit
commutated turn off time t q for the device to completely switch off. Otherwise, if you
again apply the forward voltage to the device may again get turned on it may remain in
the turn, it may not never switch off it may turn on. Again, actually the turn off
characteristics of an SCR are similar to a power diode that we have seen before. Hence,
during the turn off the excess charge carriers must be removed in the case of an SCR also
this takes time. There is a trade-off between the value of the tq and the forward voltage
drop of the device.
If t q is more typically more than100 microseconds as the case with the rectifier grade
SCRs then the forward voltage drop is relatively low, typically 1.5 volts on an average.
However, if the value of t q is less typically 20 to 30 microseconds as is found in the case
203
of inverter grid SCRs, then the forward voltage drop goes up and is found to be typically
2.5 volts or even higher in such cases. So, that is an important thing.
Now, the other very important thing about is SCRs is that they have because there are the
surface areas are large usually use for high power applications, and though the gate and
the cathode structures are, we have actually make them highly inter digitated we have
tried to make them distributed. So, that the gate has excess to all the regions of the
cathode as is required. Because if you are trying to put a control pulse through gate it
must be able to reach all the regions of the cathode to be able to then only it will be able
to exercise control over the device.
So, in spite of doing this in SCR we will find that this does not really happen and
whenever the gate trigger is given the current through the device actually it starts from a
very small slice, in a region which is very close to the gate, and then slowly it spreads to
the other areas. Now, since the external circuit in which or the circuit in which the SCR
is connected that has nothing to do with what is that region or what is that slice a portion
of the device which is actually conducting. It has a large current let us say to conduct, it
wants a large current to flow.
Now, that current tends to come out from this very small slice very close to the gate
region. And this can cause that very high current to pass through that slice very thin slice
or break of the device which may actually causes lo likes heating, hotspots and may ruin
the device. Therefore, there is a limit which is called the di/dt limit which must be
followed and it is actually part of this specification (Refer Time: 29:52) of any SCR that
you should not exceed a di/dt beyond a certain value. Now, typically the di/dt is
controlled by using an inductive snubber in series with the SCR. That is how you limit
the di by d t, so that you do not have hotspots and you do not ruin the device. So, it is a
very parameter.
Now, just like we saw there is a limit on the di by dt if you recall one of the things is that
when the devices are applied a forward voltage it was actually the J2 which was reverse
biased. So, J2 junction which was a reverse biased.
204
Now, if we just see the device here, in this way and we just say that there is a J3 here,
J2 here and J1 here my anode is on this side, my cathode is on this side, this is p1 n1, p2
n2 that I have now J2 is reverse bias when I have applied a positive politely externally
through an external source through this.
Now, it is basically, as a reverse biased junction; J2 is a reverse biased junction it

actually acts as a capacitor. And we know that we can write this relation q is equal to C
J2 let us say the junction two capacitance when it is a reverse biased C J2 into V. Now,
dQ by dt from here will give me C J2 dV/dt. Now, what happens that if this dV/dt it
actually is more than a certain value you will actually have this dQ/dt which is nothing
but the current and which I can actually show with the green colour here that this is the
current which is actually passing from n1 to p2, this is the current that is because of this
junction J2 capacitance. So, this I this current and I can actually call this as ic, that is it.
So, what we will see that this current that flows its value will depend on how large is dV
by dt from this expression and if this value is very large my i c will be large. And if you
see this diagram carefully and lo at the previous constructions carefully you can see that
this i c is working just like giving a gate current to the device which will actually it
triggers this device into conduction. So, basically, you will have to limit dV/dt otherwise
the device may get into conduction, arbitrarily at any time. Not true a proper control at
205
the time that you desire through a proper gate signal. So, that is why there is a limit on
the dV by dt rating as well.
Now, one very nice way of actually handling this is actually obtained by is actually
realised through shorted emitter connections which I will just show here. I will just show
the relevant part. So, this is my cathode and my gate is here, let us say and n1 layer is
here so this is p2 basically. And now what is happened is that my emitter is shorted at
many places. So, what I will actually see is something like this is the emitter or the
cathode.
So, what will happen is that when this i c will come as I have shown in the previous
diagram, it will actually find a path directly to the cathode, it will directly find a path
because if I just draw here this is all connected, this is all connected here, this cathode
through this plate it is its they are all conduct. So, basically, you will have this current i c
which comes and goes just its bypasses the layer which is n2 it bypasses the n2 layer and
it actually straight away goes to the cathode.
And in this process, it reduces the effect and you can actually have more dV/dt. So, if
you want to increase the rating dV/dt you can actually have a structure of your emitter
which is made like this and therefore, this current that is being generated through the
reverse biased junction J2 this will actually get distributed all over the place. It will
bypass the n2 layer and in the process it will allow you to have more current which
means more capacitance which means that it will allow you to get a or to work with a
higher dV/dt. So, we will just quickly see the characteristics.
206
So, let us say we are plotting VAK on to the right side, minus VAK on this side and we
have this the device current I A which is discrete here. Let us say that this is 0 the origin.
Now, the various device characteristics are going to lo like I verses V for the SCR. I will
just I am going to use another colour here. So, let us just use a red colour. Let us see we
apply a forward voltage.
When we apply the forward voltage let say it goes like this, the leakage current increases
little bit and it comes to a certain value V BO when iG is 0, I should say and now I am
not applied any gate current signal. And the moment it crosses V BO the saturation
current to actually initiate the region that you process of the device that I just describe to
you with the help of the two transistor model. So, actually you come here and then from
here you have this very unstable mode which I am just showing with a, let me just use
black colour like this and then the device actually just takes off, like this.
Now, instead of giving no gate signal or gate current if we had given some gate current.
What we will find? That this point, where the conductions release starts will actually
come a bit earlier. So, you will actually have for slightly more i G you will have the
characteristic like this, for still more i G you will have the characteristic like this and so
on. So, this is the forward side or this is a forward direction characteristics of the SCR.
Now, what you see on the left side if you apply a negative voltage to the device then
what you find that it will have let me just use another colour now, a green colour that it
207
will have a characteristic which will be something like this. Of course, this current here
magnitude will not be so much, I have just magnified it for the understanding part and
then beyond a certain value which is the maximum reverse voltage or some people also
call it as reverse working maximum voltage negative sign negative voltage, it will
actually go there and then you actually the device goes avalanche breakdown. So, this is
how the characteristic of the of an SCR would lo like.
Now, the characteristic of a GTO is also similar as far as a forward side is concern. Now,
only thing is that in GTO we happen to do a compromise with the left side
characteristics. So, the capability of a GTO to withstand large negative voltage, that is
much less compared to an SCR. For an SCR in general the rating of the VBO, this rating
and this rating here both they would be symmetrical they will be same. With GTO this
side rating would be less.
Now, to conclude about SCR here are some interesting facts. Fact number 1, the SCRs
are often used is (Refer Time: 39:07) pairs to reduce the gate current requirements. And
once such pair you can see that I have already drawn here where you can see that there is
a main SCR, and there is an auxiliary SCR and the auxiliary SCR is actually used to
draw a the main SCR. It suddenly reduces the gate requirement required at this terminal
G for the overall device, but it actually has a disadvantage of high forward voltage drops.
208
Fact number 2, there is variant of SCR which is called the a symmetrical SCR where the
reverse voltage capability of the device is not same as the forward break over voltage
capacity of the device or the forward break over voltage rating of the device because
usually that is the case with the SCRs. Now, the reverse voltage blocking capacity is
actually compromised through special design in the structure. So, as to achieve a lower
turn of time t q and to achieve a higher dV/dt rating for the device.
Point number 3, there are special variants of SCR which are actually called the light
activity SCRs, which actually get triggered by flashing light on the gate region. Now,
this is the very important and useful device for high voltage applications where several
of these are actually stacked in series and providing triggering signals becomes a
challenge. Now, after having done this let us now see the most important and popular
variant of SCR that is the GTO.
Now, coming to the GTOs which I said that they take away the inconvenience of the
SCRs. They as I said the SCRs they cannot turn off, in spite of the fact that their basic
construction do have some sort of interviewing between the gate and the cathode region,
but it is not so significant.
The GTOs were designed with even more prominent inter digitization of the gate and the
cathode region. And what you see on your screen right now is the example of the gate
turn off thyristor GTO, which is nothing but a variant of thyristor.
209
So, we will see that its structure is almost the same as an SCR with some minor
modifications done here and there, to be able to be switched off. So, what you see here,
for example in this slide you can see that this is a very highly inter digitated structure. I
will just mark and show you um. So, if you see here this is a very highly inter digitated
structure what you see on the screen here. So, this is your cathode, and you can see that
the p layer, which is the gate of the device, that is you can see that is highly inter
digitated. In fact, the cathodes are actually made like islands. They jut out from the
surface, in between the gate area and they are all connected through a common metallic
plate and this is what you see on your screen right now.
So, how we are able to switch off a GTO, but not an SCR. So, one of the important
things that really is contributing to this is the fact that we have a very highly inter
digitated structure. So, this is something that is very important for us to understand.
Now, to understand this mechanism of switching off the GTO, let me just draw this
structure and let us just try to see how this can be done. So, let me just draw one more
time this two transistor model, remember the structure of the GTO is same as that of the
SCR. So, it can also be represented by a two transistor model just the same way as we
have seen in case of the SCR.
So, here there is a, so, that is your T1 as before and this is a diode and there is another
diode here. We just representing the various junctions or the transistor by diodes this is
210
T2, and this is connected here this is nothing but K and here this thing let us say it is
connected to an external supply, VAK.
Now, we are trying to turn it off. So, we assuming that we are applying to the gate
terminal a negative voltage, we are applying a negative source. So, let me just denote
that using a red colour this is the G of the device and I just connect this voltage here
voltage source here whether negative here and a positive here. Now, the moment this
source is applied as you can see in this diagram that this particular junction it becomes
reverse biased. It gets reverse biased and when that happens this transistor T2, it actually
goes to from saturation, it goes to cut off more.
So, of course, our assumption is that in the beginning before just before we apply this
negative pulse to the gate the device is conducted, and when it is conducting we have
seen before I told you that both T1 and T2 are in saturation. So, when a negative voltages
are applied across this junction represented by diode then you find that this particular
transistor T2 it goes into cut off. Now, when that happens, then this current that is
flowing here in this branch which is the collector current as far as the T2 is concerned,
but the base current as far as T1 is concerned we find that reduces.
And when that reduces T1 also goes to cut off. And it again it is see that it actually
causes a cyclic process, it is a chain reaction. So, this will cause a reduction in the
current that will be flowing through T1, further reduction in the current that will go into
the gate terminal T2 and then again further reduction in the current of T2 in the
collective region and so on. So, this chain reaction continues till the 5 l G current I A, if I
just mark this as I A it goes to 0. So, this is how GTO turns off.
Now, a lot of help for this actually comes from the very peculiar structure that it has.
Now, one of the things that is done as you will see in the slide again, if you see the anode
region actually is inter digitated between p plus and n plus layers. The idea of doing this
is to basically, increase the speed of the GTO, this give works as follows.
211
As n1 is extended into p1 layer in the form of fingers as you can see in this diagram. So,
you can see this diagram which I have drawn this red colour is actually the extension of
the n1 layer into the p1 layer. So, basically, this effectively reduces the 4 layer device
into a 3 layer n p n bipolar transistor therefore, the (Refer Time: 47:32) deep process is
not required it is just like switching off normal (Refer Time: 47:36); therefore, the
switching off time or the turn off time of the GTO comes down.
However, there is also a disadvantage as I have tried to show with the help of this two
transistor equivalent circuit that I have drawn below here, so this one. Now, these fingers
the extension of n1 into the p1 layer it actually amounts to shorting the emitter and the
base of the top transistor T1 in this two transistor model. Now, it has the following
repercussions. It actually causes a higher voltage drop when the device is conducting.
There is also disadvantage with respect to the reverse bias voltage capability of the
device. Since, the device behaves like a diode without a likely doped n layer its reverse
voltage blocking capability gets compromised.
Another important point when, basically understand here is that a GTO would usually
have a higher forward voltage drop when it is conducting compared to an SCR. And it
will have the same sort of limits on di/dt and dV/dt as we have seen in the case of SCR,
and the one important point however is that I like to mention that there is a maximum
212
limit on the anode current which the GTO can actually turn off. So, which we can
actually turn off through the gate control of the GTO. Now, this is an important thing.
Now, typically when there is a current I A which is flowing. So, let me just, I have not
shown you the symbol yet of the GTO. So, I will just draw this.
This is the symbol of the SCR. If I put a gate terminal and if I just use this. So, these are
both gates. So, this is SCR and this is GTO. These are the symbols. Now, let us say it is
conducting the GTO is conducting a current I A. Now, what is that current that we need
to pump through the gate terminal to turn the GTO off? It actually is determined by this
what is called the turn off current gain, which will denote by beta off of the GTO. So,
that turn off current gain or the beta of GTO it is defined as the I A that is flowing what I
am showing here, divided by ig; the current that we need to pump through the gate
control to turn the GTO off.
So, you can see from here that if the typical values of beta off are something like 3 to 4
then you can see that we will need a very high current up to 25 percent all the IA will be
required to actually switch off the device. So, there is always a limit on this what current
you can pump through the gate. So, that is why there is a very critical value that you can
actually turn off your GTO. So, I A if it is beyond a certain value you will not be able to
turn off through the gate control as is usually taught.
213
So, these are some of the things I wanted to tell you. Of course, one of the very important
thing is that for the GTOs we have to if we are trying to turn on and turn off, turning on
is same as what we do in SCR. So, typically the waveforms would appear something like
this. So, this is for turn on of the GTO and let us say this is nothing, but ta and tb. And
when you are trying turn off then you need to give pulse, a high current pulse which will
be determined by our beta off; what I just explained to you. And this is how it is going to
lo like. So, this will be t b and t c. This will be the duration when you are trying to turn
off.
So, basically, you will have to give a continuous current waveform to the GTO unlike the
SCR to turn it on. While as far as the turning off is concerned you can give a high current
pulse which will be determined as per the wave if you will lo like this, and will be
determined by the limitations imposed by the beta off value, which is nothing but the
turn off current gain of the GTO.
So, I think we will just conclude this session here. With this I thank you very much for
your attention and we will continue to work on the power devices in the coming lectures.
Thank you very much.
214
Prof. L. Umanand
Department of Electronics Systems Engineering
Indian Institute of Science, Bengaluru
Lecture - 11
Motivation for rectifier capacitor filter
In this video capsule, we shall look at rectifiers and capacitor filter. We will be focusing
on the full bridge rectifier and only the capacitor filter for the filtering circuit. If you
understand that there are many circuits that convert AC to DC; how the voltage rectifier
capacitor filter combination is one of the most popular AC to DC converter that you will
find and you will find it almost in every product. Even though the full bridge rectifier
combination has lot of disadvantages like very low power factor, it draws peaky currents
things like that. It has the advantages like very low component count, low cost and very
high volumetric efficiency; it is very compact.
Because of these this has become very popular and we will study this particular circuit.
Later on I will describe to you some of the problems that you will encounter with the
circuit and the corrections needed and the protections needed to handle those issues and
probably later on and we talked about DC converter. I will also talk about the power
factor improvement techniques and methods. So, now, we will talk about the rectifier
circuit, its operation and the waveforms.
215
So, first let us start with the source. The source is generally the wall outlet 32 volts 50
hertz grid. It is sinusoidal in voltage wave shape and this needs to be interfaced to a load
which expects unidirectional voltage wave shape. The key point here is that through the
load, the current should always flow in the same direction; whatever may be the polarity
of the source. Let us now interface the diodes step by step. First when the source is
having a polarity; positive polarity, this is positive with respect to the other terminal
diodes need to be connected. In this fashion such that the current flow completes as
shown here. Notice the current flow direction here through the load which is in this
direction.
Now, let us say that the polarity of the source is reversed. Now it is in the blue signal
level. These 2 diodes will become reverse biased and turn off. I will explain a little later.
Let us first see the current flow direction. This is positive with respect to this; so, you
will see that. Here, this is positive during the negative half and the flow of current will be
in this direction and again entering here at the same point terminal through the load
resistor and then, comes back to the source.
Note again, here importantly that the current direction has been retained. Thereby, you
will see that the voltage across the load is uni direction whether the source voltage is
positive or negative. Now, let me just tell you how this diode gets reverse biased here.
So, you saw that when the blue colored diodes are conducting, this is positive with
respect to this terminal.
So, when this blue diode is conducting, this terminal, this point is at a positive potential
with respect to this. So, this diode see the reverse biased voltage, reverse voltage;
therefore, it is reverse biased and is off likewise this point is at a positive potential
compared to this point. Again, this diode sees the reverse voltage and it is reverse biased
and in the off situation.
By a similar argument, it can be seen that the other two diodes that is this diode and this
diode will be turned off, when the voltage polarity is positive, when this is positive with
respect to this. So, these other two diodes will be conducting and will make these two
diodes go off. So, the total schematic of the rectifier is like this as shown. Now, to this if
we how to connect a filter because this is a varying voltage we need to connect a filter.
216
It is a simple matter of connecting a capacitor like this as shown. This is a capacitor C.

Now this capacitor will act as a filter and trying to pass current through load which is
more or less at DC. Now, at this point may be interesting to see some real components of
the rectifier and the capacitors.
I would like to show you some common rectifier diodes. Now, this is a very common
rectifier diode used in most of the products 1 and 400 11; 1 and 400 7 kind of series kind
of diodes. It can handle just about 1 ampere so and pretty popular and common. They
217
come strips like this and then, you will be mounting 4 of them to form a full bridge
rectifier.
Like this you also have still smaller diodes these are 1 and 4 and 4 8 type diodes much
smaller handles; only around 500 milliamps or 800 milliamps in that range for much
smaller circuits. You have some fr series high amps; higher amps one. Also you see here
this is the one on one which I am right now showing is around 3 amps range FR series
diode; rectifier diodes can be used for rectification and then this is a 6 amp diode, what I
am right now showing.
So, this can also be used for rectification. An interesting thing here would like to show
you this you see this is a full bridge rectifier completely. It is having 4 diodes within it.
So, 2 of the leads are connected to the AC and 2 of the leads are connected to the
unidirectional port that is plus minus and to that unidirectional port pins you connect the
load resistor r u and the filter capacitor.
So, this come this comprises 4 diodes within is the full bridge rectifier component
available to you. Another interesting piece is this you see here, it is having 5 pins. This is
actually a 3 phase rectifier. You connect the a phase, b phase, c phase to these 3
terminals and then, you can take the plus and the minus output from this and connect the
capacitor or the load at this point. This is a 3 phase rectifier and you will see that the
back side, there is an aluminum facing which can be used for connecting to the heat sink
for thermal regulation.
So, the heat sink can be of this type like or any other type depends upon the design and
one will mount it in this fashion to get a better thermal flow. So, like this rectifiers I have
few capacitors also to show to you.
218
So, these are the aluminum electrolytic capacitors which will be used. In the smaller
ones, you will use these kind of aluminum electrolytic capacitors; the one with the line
mark is the negative polarity usually and a slightly bigger one with this is 4700
microfarads. This is also aluminum electrolytic. Most of the time it is aluminum
electrolytic that will be used as still bigger one normally used as d-ceiling capacitors in
high power circuits, in inverters and things like that one, where there are terminal lux.
You connect the lux there and this is from Aerovox this is 450 volt DC high voltage,
1000 microfarads capacitor.
This also aluminum electrolytic; these are some of the components actually. You see
these are the 2 sets of components that will be using in the rectifier circuits, back to the
circuiting. So, we have the full bridge rectifier and the capacitor connect as a filter here
across the load.
219
Now, let us name the parts we will call the input sources V in, diode D 1 and diode D 4
connect during the positive half cycle of the input. Diode D 2 and diode D 3 conduct
during the negative half cycle of V in. There is a capacitor C and a load R naught are
now we have put a resistive load. The point that is of interest is here and that is called V
c.
V c is a voltage across the capacitor is of importance to you because that is the voltage
that the load will also be seen and this actually would be the filtered voltage. Another
point of interest of; points of interest or the 3 currents that flow at this node. This is the
current coming out of the rectifier, this is the current that flows into the capacitor and this
is the current that flows into the load.
So, here we expect that at this point; we expect the current to be a D C and around here,
we expect the current to be spiky peaky current and here in the through the capacitor, we
expect the current to have gone down by the load current value something like that.
Anyway, we will discuss this current wave shape in more detail and the slides to come.
But this is generally the idea that I want to convey that this is a current which is having
an average value plus an AC component.
The average value part goes into the load resistor and the AC component part, the one
which does not AC average value 0 average value flows through the capacitor. Must note
that in the steady state, the capacitor can handle only a pure AC current that they should
220
not be an average value; otherwise there will be charge buildup. So, under steady state
the 0 average part of the current or the only the AC component flows through the
capacitor. So, that is the action on the filter.
So, now let us see the waveforms in detail before that let us name the parts. So, the
current that is coming out of the rectifier, you will call it as i. The current that goes
through the capacitor as i c, the current that flows into the load R naught has i naught.
Let us have this template. This is the sine wave shape of each rectified half. So, this is a
positive half of the input source.
This is the negative of the input source rectified and placed in this fashion, the
unidirectional fashion. Let us start from this point to speak point as a reference and then
we will come back to it and extend it and let us say it is in a steady state which means
that there is some charge accumulated in the capacitance. So, if I say it is we are starting
at this point, the chart, the voltage at the capacitance will also be of this value V in value.
So, at this point, the capacitor is discharging through the load with some time constant.
So, this is the discharge part of the capacitance; whereas, the input wave shape is
following this path. So, if you look at the diodes in the full bridge the load voltage here is
higher and the voltage on the input side is lower and all the diodes are reverse biased and
off. So, during this part and up to this part you do not see any diode conducting because
are there reverse biased and the capacitor portion is isolated from the input proportion.
221
So, capacitor is discharging on its own. Till it reaches at this point, when the input wave
shape has taken at and it is rising up the moment the input rises beyond this, the diodes
conduct.
So, in this case the blue colored diodes conduct and the capacitor follows the input. The
moment the diodes conduct the capacitor node point is directly connected to the source
and it will just follow the source up to this point. And then again, after this point the
period repeats and you will see the capacitor discharging in this fashion and the moment
capacitor starts discharging like this and the input also starts to flow down along the
sinusoidal path.
You will see there is a differential difference and potential the diodes are reverse biased
and they turn off. Likewise so and this would be the steady state waveform wave shape
for the output at node V c. So, this is also the ripple that the output would see across the
load and this ripple is controllable through design; we will see that later.
Now, if you see only during this portion, we see conduction in the diode; only during this
portion, we see conduction of the diode. So, let us mark that off and during that only we
will see current. Here, the current i is flowing. So, what would be the type of the current?
So, the moment the diodes turn on at this point, the diodes turn on at this point, a voltage
source is seen connected to a capacitance. So, there is two direct connection of two
potential devices. So, there will be a huge current rise here and it is limited only by the
track inductance diode impedance and any (Refer Time: 17:52) or the capacitance and
the inductances here; these kind of non idealities only will limit the current.
So, the current wave shape will be something like this. It will have a very sharp rise and
it will have a time constant based on only the series impedance is coming in the path.
Now this current will repeat every half cycle as shown here. What happens during the
first cycle when the capacitances start charged, we will look at that a bit later. Let us first
clarify the other components of the current. Now we see that this is the wave shape of the
voltage with a ripple and this is the current that is flowing output of the rectifier just
before the filter.
222
Now, during this time D 1 and D 4 are conducting; other two diodes are conducting
during this period D 2 and D 3. So, basically this would form the diode currents and if
you calculate, if you draw the waveform for i in, the input current. This is the wave
shape for during the positive half cycle and during the negative half cycle, this would be
in reverse like this. Now, these are the wave shape for the input current. Now, let us look
at the load current and the capacitor currents.
223
Now, the load current will follow the voltage at that node V c node. So, that V c value
divided by R will be i naught and it will have some ripple as seen here. We can
approximate the averages right through this as shown and for practical purposes, we can
take this i naught as the current flowing through the output ripple free. But be aware that
there will be a small ripple.
Now, if you remove this average component, from this i current, you will get the current
that flows through the capacitance. So, let us draw the capacitance waveform during this
period of conduction. Then, you will see that the capacitance would have gone down by
an amount of i naught value as shown here. In the steady state, you should note that this
area is equal to this area.
I would now like to revisit these waveforms to discuss about the current during starting.
So, let me remove this part of the waveform. Now, you see at the time of starting the
capacitor voltage is 0. So, the probably it will be at this level. Now if we apply a voltage
and if by chance that voltage is starting at this point, then you will see that the capacitor
build up the charge gradually from here on up to this point. So, as the capacitor is
building up the charge there is current flow out of the rectifier and therefore, the rectifier
current will be for this whole half cycle.
So, you see the conduction period will start increasing and before the complete half
cycle. Likewise, the current wave shape also will get modified something like this. But
224
you should understand that the capacitance had 0 voltage and therefore, 0 charge; unlike
in the case of the steady state condition. At this point the charge on the capacitor was
finite and there was some energy. It was only making up the difference energy. Now, this
whole energy has to be made up. So, you will probably land up with current waveform at
the first start up value which would be much higher than the following and the city state
currents.
So, this may be the shape of the wave shape you will see from startup for the currents.
But a problem could occur what if at the time of turn on; the turn on was not
synchronized with the 0 of the input voltage, but it occurs somewhere here; which means
the turn on occurs somewhere at this point and which means that the capacitor voltage is
0 and the grid voltage is pretty high value and you are connecting to potential devices
together, there could be a huge surge current.
So, what happens to the current wave shape? The current wave shape may look
something like this, but the aptitude may just go out of the screen, it may be a very large
current and then, you start having the steady state. But the effect would be that by the
time the first cycle is over the current will be so large, the devices the diodes may blow
off and once the diode is blown off, the rectifier circuit will not see the second half cycle
at all.
So, it never even goes beyond the first half cycle. So, it is very important that we do limit
the first current search. Of course, I will discuss about the protection circuits and see how
we limit the turn on current surges. But keep this in mind that all is not rosy and all is not
well and good with the circuit that we have just discussed, the turn on current are just can
be pretty high and it could be big drain even on the wall socket outlet. They have to be
rated for such high currents. Therefore, we would like to put a series impedance
somewhere here and limit the current surge.
225
Prof. L. Umanand
Lecture - 12
Circuit Operation
So, till now we have studied the Rectifier circuit, capacitor filter circuit. We have seen its
waveforms; we have learnt how to simulate the rectifier capacitor filter circuit. We now
have to do one important activity which is designing the components, you have only two
types of components; one is the diode, rectifier diode and the other one is the capacitor.
So, these two components, you will have to rate such that they will be able to handle the
electrical stresses and the thermal stresses. In this course, we will discuss about rating for
the electrical stresses like the current and the voltage stresses; how the thermal stress
calculation was will be another course in itself.
So, therefore, it is out of the scope as far as this course is concerned to calculate and rate
the components for thermal stresses. There is another point which is also there while you
are designing circuits that is to design for life. This is also out of the scope designing for
a given MTTB MTT for MTPF mean time to say how work keep that in mind that these
are aspects that you will have to consider when you are designing a practical circuit. In
this course now, we will look at how to design for the thermal; how to design for
handling the electrical stresses.
226
Here, now back again to the whiteboard where we have the waveforms, we have seen
these waveforms earlier. This is the output voltage waveform with the ripple Vc; what
we call Vc is the output voltage waveform across the capacitance. Io is the load current.
We have shown the average value of the current, there will be a small ripple which will
have the same similar shape as that of Vc/R. This green waveform is start of the current
output of the rectifier as we saw and we need also to define few more parameters which
we will do now in order to do the calculations for the values of the capacitance and the
diode ratings.
Now, we have this is half the wave shape of the full sinusoid. So, this must be having an
angle of pi. Remember that this is whole 2 pi and this much is pi which means T/2;
period by 2. Then, another parameter that we will define is the conduction period, the
period for which the diode conducts and we will name it as alpha. Of course, from this
point to this point is also pi, I can repeat every cycle. Another variable 25 is the peak to
peak ripple. So, this is the peak to peak swing of the output voltage. Therefore, we can
call that one as the peak to peak ripple.
This definition Im is the peak current that is flowing out of the rectifier as shown here.
You see normally when we are rating or designing the components, there is not much
that we would get by trying to calculate exactly the nature of this curve. Because we
need to anyway give some safety factors that we will normally be over rating the
227
devices. If you take consider this rectangular green rectangular box, if the current wave
shape were something like that. If you design your components for this flat topped green
rectangular pulse, then definitely it would handle the inner shape shaped pulse as it
comes out of the capacitor filter rectifier.
So, therefore, normally in design practice what we do, we use engineering judgment here
and say that for rating the devices, we will design the devices to handle this flat top
complete flat top rectangular pulse much easier to design for that and if we design for
that, it will definitely handle this shape of current. So, that is what we would be trying to
do for. This is the current ratio that we will assume for design; only for design purposes
not for any analysis. So, now, we shall define one more variable called Vo. Vo is
actually the average value of the output voltage Vc which you see at the capacitance
node. So, if this is the peak value and this is the main value of the ripple, this plus this
divided by two the average value will be Vo.
So, this would be the average value of current, the average value of the voltage is given
here. Two more variables which is the peak value that the capacitor will take Vm1 and
the minimum value it will discharge to Vm2 in the steady state. So, using these variables
definition, we shall now calculate what should be the value of the capacitance and also
what should be the value of the currents that should flow through the diodes and such.
Now, for this to write down the equation, it is now better for me to use the writing board.
So, I will go to the writing board and start writing the equations so that you will be able
to follow me.
So, I have with me the same waveform figure with all the parameters named here with
me and we shall use this for writing the equation; design equations. Now, first of let us
calculate the value of the capacitance. Let us note down; write down, what it that is given
to you for design purposes the specifications is. So, the specs that are given to you are
the following. Specifications; one of the things that are given to you would be what is the
value of Vo; Ripple spec, this will be given to you. It will be given to you as peak to
peak ripple variation, delta Vr as that as we have marked..
228
Then, what else that would be given is you will have the power Po output power we also
have the input spec. Now, the input voltage as two parts; one is see would have you
would probably you have something like 230 Volt rms which implies 230√2 peak. This
would be your V m, but it is not just that your input voltage will swing from place to
place from minus from 180 degree 180 Volts to 270 volts. So, plus minus 20 percent,
plus minus 30 percent; these kinds of swing you will find real voltage means great. So,
you should also specify the input voltage tolerance. So, this is generally given like 10
percent or 15% of your nominal value. So, it will vary from minus 10 percent to 30 Volts
rms to plus 10% to 30 Volts rms which means 230 Volt plus or minus 23 Volts things.
So, this tolerance has be given. So, let us say we call it has percent tol as the variable and
then of course, you need to have frequency. The frequency supply frequency is known it
is always 50 Hertz at least in our country (Refer Time: 11:13) bother much about that.
So, these are the specs that are given to you; these are specs that are there with you,
using these specs, we have to now arrive at the value of the capacitance C that is our first
job. Let us start at this point. Let us see what is happening at this point or shall we come
down here corresponding point. So, at this point, the capacitor has an energy half C V m
1 square and from here to here, the diodes are off. Likewise here from here to here, the
diodes are off capacitance is doing only the job of discharging into the load.
229
So, from here to here it has discharge and to the load, nowhere else and it has reached
energy lower energy level of half C V m 2 square. So, what has happened to all this
energy lost from here to here, it has gone to the load. So, that is our starting point that we
will use.
So,
C Vm12 - C Vm22 = Po
So, during the time Vo into Io that much amount of power is being put for a period of
time T by 2. So, if you if you look at this equation pi minus alpha by pi into T by 2 is the
weighted period into Po the power put. So, this is actually the energy Watts into time;
Watt seconds. So, that is the energy the amount of energy that you see here half V half C
V m 1 square minus half C V m square 2 square that is put into the load. So, let us
simplify this half C V m 1 square minus V m 2 square which is pi minus alpha by pi Po
pi 2 f; f is the supply frequency.
Now, this can be split into C V m 1 plus V m 2 by 2 into V m 1 minus V m 2. So, this is
actually a square minus b square form a plus b into a minus b. Now, this is nothing but
let me move the screen up; this is nothing but Vo the average value. This is nothing but
delta V r. So, therefore, we have C equals pi minus alpha by pi Po by 2 f Vo delta V r
coming in the denominator, but Po itself is Vo into Io. So, Vo and Vo cancel and
230
therefore, you can write the value of C as pi minus alpha Io by 2 times supply frequency
into delta V r.
So, this is the value of the capacitor that you will have to put in order to get this
particular ripple and for this particular load current. However, you should note that there
are few other variations that can come into the picture; you should calculate for the max
value of Io. So, if you calculate the max value of Io, the value of C would be sufficient to
handle that. You should also calculate for the ripple that is minimum. So, the for the
minimum ripple, I will get higher value of C. So, once you have taken care of the worst
case condition for minimum ripple as for value of Io, the value of C will hold good..
Another very important criteria that you will have to take into account while you are
putting the value of C in a real circuit is when you buy an aluminum electrolytic
capacitor, it has a very large tolerance minus 40 percent to even 100 percent those kind
of tolerances.
So, normally what you calculate and then, what you buy and then actually measure can
have a very large and significant variation. So, normally what is done in practice is once
you calculate let us say you get some 15 microfarad or 10 microfarads, you will take 3
times that value. So, 30 microfarads then put that value. So, that even if even the
capacitor that you have bought is minus 40 percent down, it will be able to handle this
kind of variations. So, this is how you calculate the value.
Let us now calculate the rms current that is flowing through the capacitance because that
is one parameter which we that is one parameter which we need to calculate because that
will indirectly affect the heating of the capacitor.
231
So, to do that let me go back again to this conduction period that is when the current is
flowing. This alpha time is the period which during which the diodes are conducting; do
we know this value of alpha, we know this value of peak V m 1. This is the cos wave
from here. So, V m 1 cos of alpha will be V m 2. So, we know V m 1, we know the
ripple, we know V m 2 and therefore, we should be able to know alpha.
So, how we can start this, like this V; I will use the blue ink V m 1 equals sorry V m 2
equals V m 1 cos alpha. Alpha equals cos inverse V m 2 by V m 1. V m 1 is known
which is V rms value into root 2. V m 2 is known from the spec which is V m 1 minus
delta V r. All these are coming from the input spec then once you know the value of
alpha, you can calculate the current that is flowing through the capacitance. The current
that is flowing through the capacitance now as I said we are going to make this
rectangular approximation and you have to make use of the condition that the average
value of the current through the capacitance is always 0 in the steady state.
So, this is the area that is that is the charge; charge up of the capacitor when it is
conducting when diodes are conducting. When the diodes are not conducting capacitors,
capacitor is discharging and the area is this. So, we know that this is Io and therefore,
this would be Io into pi minus alpha by pi into T by 2 at in that area and this area of
course is I m minus Io alpha by pi into T by 2. So, this T by 2 is added. So, that this is
the charge and this is the charge.
232
So, if you remove out the T by 2 just to find out the rms value of the current. So, I c rms
I c rms is root mean square. So, for the positive area, I m minus Io whole square with the
duty cycle of alpha by pi plus Io square pi minus alpha by pi this whole thing is. So, this
would be rms r m s value of the current that flows through the through the capacitance
that this rms current square into the esr value of the capacitor will give you the heating
affect within the capacitor. So, this may be useful especially when you want to do
thermal management and thermal design.
So, once you know this you will be able to specify the capacitance.
So, the capacitance can be specified as follows. We know the value. The C value is given
by the equation that we just now derived Io by 2 f into delta V r. So, this Io; Io you
should take Io max; maximum possible Io that flows through that and minimum value of
delta V r min. Under these condition what is the value of C that you would get? This
variation why I am mentioning this variation is that the input voltage varies from a
minimum value to a maximum value. So, find out the delta V r min whatever the worst
case condition and then plug these in.
Now, the capacitor voltage rating; so, the max voltage that the capacitor will ever see
will be V rms into root 2. This is V m. Now, V m itself could be swinging to upper end
because of the tolerance I say 1 plus percent tolerance value whatever 10 percent 20
percent by 100. So, this would be the maximum value that the capacitor will see. Of
233
course, you also have the I c rms rating as we just now saw I m minus Io whole square
alpha by pi plus Io square pi minus alpha by pi. So, this would be the rms value.
So, with this and then you can say the make type electrolyte type would be aluminum
electrolyte. We do normally use tantalum. So, this would complete the electrical design
of the capacitance.
Now for the diode, the diode if you see when the diodes are off, they would see a
maximum voltage the peak inverse voltage; peak inverse voltage would be V rms root 2.
So, if it is 230, it would be 230 into root 2, 325 Volts to that you give the extra the
tolerance swing for that particular place; this would be the max. So, when you choose a
diode, you have to choose a diode which is definitely much having a peak inverse
voltage rating much greater than what you would calculate using this formula. There are
two other values that you need to calculate for the diode one is the average I d average
and the I d maximum.
So, I d maximum is same as the current the my peak current that is flowing through
rectifier . So, the I d maximum will be same as this. This is the I d maximum. I m
average value through the diode will be you see that through every diode, it flows
through alternate cycles. So, through d 1 and d 4, it will be through this d 1 and d 4
again, it will flow in the next yellow half cycle. d 2 and d 3, it flows here and d 3, d 2 and
d 3 again flow in the next blue half cycle. So, therefore, the average for this is I m duty
234
cycle will be alpha pIo pi by alpha by 2 pi because this repeats every 2 pi. So, for the
diode we can easily find the average in the following manner.
I m alpha by 2 pi this would be the average current that flows through the diode and
when you chose the diode, it should have a rich current rating greater than this and I m
itself is the peak current rating which is flowing or very small period, around alpha for
this current kind of a pulse approximation. So, by this now you have the diode rating and
you also have the capacitor electrically designed and these values can be your guiding
values for you to choose these devices. So, what you can do is that we can put these
equations into a script file so that when you want to do an iterative design you keep
changing the specification then you check keep checking things and how you may not
get a particular component and then, you may want to change things. So, it is good to put
them in a script file like something like octave or matlab.
So, we will be using octave script file and the run the script file repeatedly to see what
are the designed values so that so that we do not have to manually calculate these values
them and again which will lead to boredom. So, I will now show you how to put them
into a script file and the thats not automate the design courses. I shall show you know
how to make a sample script for designing the designing the values of the components.
So, this is our folder dialogue will go into DCDC folder, I have already created a sample
here.
235
So, this is the m file, I will explain to you. I shall keep it on to one side, at the same time
I would also like to open octave. I will open this octave. I shall keep this on to one side;
maximize the space. So, I have the workspace here. So, it is like this. Let me go into the
same folder here ok. The way you, I have already written it so that we do not waste too
much time. You have to classify your script into 3 parts. This syntax is very similar to
exactly similar to what you would do in Matlab; it is also having an extension dot m pi.
You have the specification of the circuit written first then you proceed with the
calculations all these are calculations and then, followed by display.
This is what you would like to display at the end of the day after you finish the
calculation. So, keep it into these three parts. So, for the specification, what is it? This is
where you will change for doing iterations; all these calculations are based on these
variables. We have specified the I have used appropriate variables here because I cannot
use subscripts and superscripts. So, V I rms is the input voltage, the tolerance 20 percent
tolerance. V r is the peak ripple voltage of the output means supply frequency and the
power Vo and the calculations you first try to calculate what is the minimum value and
the maximum minimum value of the peaks have the maximum value of the peaks. So,
that you can get the worst case condition; go through these equations just like we
discussed, but I have now used the tolerance values also.
236
Similarly, you will get two values for alpha; alpha min alpha max. Choose the max value
of alpha for calculating the value of C, then here to calculate the voltage that the C will
see the output current max, output current the capacitor rms value and then, the diode
selection peak inverse voltage value, I d average and the rms value. Then finally, the
display have used the f print of statement; one I have used. So, that it is put on to the
display standard output. You could put it to a file also and this is just like your C syntax
specifications, output power capacitive selection values and diode. So, these are the
parameters that you would like to. So, this I will also upload it into Google drive. So, that
you can have a look at it and then try to make your design script files along these files.
So, we go into the octave word space, what you would do is just run the script file. So,
we know that the name is red underscore filter dot m which as to that that underscore
filters. Do not give the dot m h extension just run it. So, you will see that this gets
executed. So, probably may be good to clear the screen before you to save that.
So, let me do that execution once again. So, the specifications this is the output power all
those things and capacitor selection values 100 and 114.95 microfarads so on and so
forth. The diode selection have the (Refer Time: 35:22). I have multiplied here by 1e6 so
that I can express it in microfarads rather than having a very long floating point number
here.
237
So, this way you can keep doing the iterations any number of times change the power
value make it 1000 Watts and then rerun the script. So, you will see things changing and
then, you can keep experimenting with it and then go back to the simulation plug-in the
values and see what the values that you would get are. So, this could give you a lot of
insight into the rectifier circuit; rectifier filter circuit in such. So, this is how you would
do about analyzing and designing. The two important components in the rectifier filter
circuit is the diode and the capacitor.
238
Prof. L. Umanand
Lecture - 13
Designing the circuit
In this video capsule, we shall look into the installation process for installing the open
source software packages under our system such that we will be able to make effective
simulation for this course. The plan is to install gEDA software tool set. This is the
electronic design automation tool set and then we shall install NG spice a very popular
package for circuit simulation and then after this we shall install octave. Octave is a
MATLAB like environment which is very helpful in scripting and automating many of
your design tasks.
So, these three set of software packages we need to install I will show; I will show you
by taking a walk through and then configuring the system to our needs. I have a Fedora
23 Linux desktop. So, I will show you by show by taking a walk through the process in
this particular operating system; my students have checked it out on Ubuntu also.
So, I guess that it should be fine even in other Linux distributions. On Windows I have
not tried nor have my students tried. So, I am not too aware on how it will perform on
Windows. So, I suggest that if anybody of you are trying it on Windows, that is EDA,
ngspice and octave you can try and then let us know how it performs or you can use
alternative packages whatever is available on the Windows platform.
With this we shall start now the process of installing these software packages. We shall
begin the installation process by first downloading the resource from the Google drive.
This is the Fedora desktop and let us open the folders and here I have created a DCDC
converter folder and within that folder I have downloaded this resource01 dot zip folder
file.
This is located in Google drive and the link for this is given on the course website. This
has two folders and one file readme dot txt file there is a bin folder and the symbols
folder of course, I will explain to you later what are the functions and how we will use
239
that one. So, for now extract unzip it and within that you see this folder. Readme dot txt
gives you a step by step instruction on how to go about these process of installation.
So, open that read me dot text file and keep it to one side. And all you have to do is
follow through the steps the commands are given here you just may have to copy and
paste it if it is a Fedora system or equivalently typing the command line commands if it
is any other Linux system.
Again go to the desktop and open a terminal let me adjust the size so, that I am able to
see that. Now let us copy this and paste it here, it will ask for the password and then
execute it. So, in my case I have already installed all these packages, it will search and
look for the packages in the repository and install them. So, you have the gschem check
the gschem docs, netlist symbols utils g shell the geda g schematic this is what we will
be using at great length in this particular course.
So, dependencies resort nothing to do everything has been installed. In your case if you
have not installed geda you will see that it will give a list of packages and ask you
whether to install or not you say yes and it will get installed. Next we need to configure
geda you can In fact, open g schematics and see what happens what pops out. So, let me
type gschem enter and you will see h u i something like this open up.
240
This is a status dialog which opens which just gives you an indication of the paths which
from where it takes this is the g schematic r c file gafr c file where it is located it is
located and user share gEDA; however, these are in the this need root permission to
modify these files.
However if you want to customize you need to have copy of these files in your user dot g
e d f folder which I will tell you about shortly. Close this this is how the g e d a user
interface will look like. This is having a dark background some people may like this
background fine, but I would prefer to have a light background and I would like to have
the default gschem opening with a light b g e.
241
(Refer Slide Time: 07:33).
Now here is the components and this is where you have the libraries, where a list of
components coming here and then you can choose these components to place onto the
sheet. Close this for now we know that g e d a is installed working and running and it is
opening up. At this point I would like to configure it I will show you how to configure it
to make it into a light background 1, you may want to change it or you may not want to
change it you may want to retain the dark background it is up to you, but let me show
you how to go about doing that.
So, you see here the readme dot txt file gnetlist just go through this command syntax,
copy and then paste. So, this is just a g g e d a command what it would do is that, it will
create dot gEDA folder in the users root directory and into that directory is where we
would like to have the r c files. So, if I go into home and to control h are alternate the
hide it will show the hidden files with the dot prefix. So, you see that this has got created
dot gEDA and into this you have logs and a geometry file. Here we have to put the
gschemrc the g a f r c files anyway.
We will come back to that later and you will see what to do next. Next we would like to
make a g a f directory and this is where we would like to place the symbols of g e d a
symbol the custom symbols that we make that has to be placed within this. So, we will
give it a place. So, just copy this and paste it here onto the terminal and press enter. So,
242
what it would have done now is create this g a f directory g a f. It is right now empty, but
we want to fill it up with symbols custom symbols.
Next we create within g a f another subdirectory called symbols that is where we shall
place our symbols. Copy paste and enter this if two then redirected to errors not log is
just an indication if there are any errors to see you can go and look at this log file see
what happened. There are no errors nothing will be entered there.
So, now after having run this particular line within g a f you would have this and inside
that let us place the symbols. So, what we shall do? We go into this DCDC converter
folder and within that we have the resource folder and in that we have a symbols folder
in the symbols folder, there are three subfolders one is called a block a bond and a
comps.
A block has many symbols associated with add clock gain integration multiplication pid
pwm bridge, pwm single phase sample and hold s r latch summations so, on and so,
forth.
So, these are custom blocks more something more logic oriented not much of power flow
and interesting to have them. I will come to this later we may not be using this, but
anyway you I would like to share that with you.
243
The a comps are the component block, the default are the generic components capacitor
Default diode, power diode, source Flow source current source Gyrator inductance,
power switch resistance SCR, signal switch transformer a transformer bridge,
transformer for bridge another flyback transformer, forward converter transformer PUSH
pull transformer in the normal transformer. So, these are symbols which will be useful.
This is particularly symbols used for bond graph simulation. So, people who are familiar
with bond graphs and have done bond graph modeling and simulation they may use it.
244
But right now we shall not use it, but; however, I will share this thing this folder also
with you, you copy all these copy and go into the g a f go into symbols and paste them
and now we are in business we have all the symbols in the proper directory.
Now, we have to assign the proper path. So, now, that we have done that, we will assign
the paths and create the respective files. Now there are two commands here one is a
command component library search, where it will search through this path dollar within
braces HOME is a environment variable, which will take the root users r and into gaf and
into symbols now this should be the path of all the component library search and they
should get reflected in your g g schematic when you open it. And this is pushed into a
file, which is residing it dot gEDAand into a gafrc file if gaf r c is not there it will get
created.
So, this is your own custom gaf r c file it will look such far there and then see this and
then set the path environment path. So, we can have this copy and paste it here and run
that one. So, once you run that if you go down into the hidden dot gEDA file I have
unhidden that. So, into that you will see that there is a gafrc file that is created. So, into
the gafrc file see whatever we have typed.
Common library search that would have come in here; so, when gEDA opens up it will
look into this file see this path declaration and then appropriately remember the path.
Next we would like to make the background light. So, that is this load build path gEDA
245
gschemrc into that gschemrc we will put this gschem color map lightbg. So, instead of
that bg we want a light bg and then put that into dot gEDA gschemrc file. So, we will
copy that and paste it here. So, that would have got entered in there. So, you see that
gschemrc has been created and within that this has been copied.
So, it will look in there and automatically put in the proper background color. Now I
have to control h and hide all the dot prefix files. So, that it does not clutter up the
folders.
246
Now, again now you can run gschem.
Now, you see the background is light, see that it takes gafrc from the local users dot
gEDA directory gschemrc also from the local. So, it is working properly you can open it
to full screen you can use the wheel button of the mouse to zoom in and zoom out.
And then go into the components, now you see that our custom libraries ablock aborn
acomponents are reflecting here. So, if I open it they are all coming in here the symbols
247
and that is very nice and then the component now close this, and back again to the
terminal going back to the readme dot txt file.
The next part of the simulation is installing ngspice. So, installing ngspice is pretty
straight forward you will just copy this, dnf install ngspice and let me clear this terminal
screen control l and that is cleared test. So, do install ngspice enter the password and of
course, in my case here it is already installed and in case you yours is not installed, it will
say that this particular package has to be installed whether to install or not click yes and
then you will be through.
And then after that you need to install tclspice also though actually you may not used
tclspice, there are some libraries in tc lspice that are useful for ngspice and kindly to
install tclspice too and here again it is already installed here and then you will get such a
message. So, now ngspice is also in place and if you want to test that you just type in
ngspice dash dash.
248
So, that will give you a menu of options and how to use go through that and I would
recommend that you download the ngspice manual and read through it, if you are not if
you do not know how to use spice. It is very very similar to p spice of the windows it is
actually a word out of the berkley spice and quite universally used by many people..
You can type just plain ngspice and it will go into the ngspice environment. It is ngspice
version 26 plus and a developed by Berkeley CAD Group and then you have the ngspice
environment prompt. So, I will quit from this here knowing that ngspice is working.
249
And now we have one more package to install and that is octave. Octave is a very very
powerful tool you should have that it is a open source equivalent of MATLAB
sometimes I find it much more powerful than MATLAB, but anyway that is a good
package to have on your system.
Paste it here, I have used a dnf install I have used dnf everywhere doing this process
because this is Fedora 23; Fedora 22 onwards we have to use dnf for the package
installer as the package installer before Fedora 22 it used to be yum. So, people having
Fedora version earlier than 22 kindly use yum y u m. And the case of Ubuntu people can
use the apt get package installer.
So, use this press enter and the rest the package installer will do the job and here again
you see everything is already installed and dependencies are resolved nothing else to do.
Now, to test whether an octave is running properly or not just type octave; you will now
see a GUI popping up very nice GUI.
250
It looks much like the MATLAB GUI this is where the workspace window is and this is
where you will probably do all the work. So, octave is working this is version 4 plus.
Most of the earlier versions octave they may not have a GUI integrated, but it will work
on the terminal just like the ngspice environment. So, if you want to work in such an
environment which I also prefer you type octave dash dash no gui.
So, then you will get it in the terminal this is very fast and you will be able to do
equivalently, good programming with this terminal environment to. So, let me quit that
251
and with this we have installed three packages one is the EDA Electronic Design
Automation the next is the spice that is the ngspice and tcl spice have been installed and
then octave and we have installed it on a fedora system. I hope that you will not have a
problem and other distributions to kindly follow it and then look into help on the net to if
you run into problems and there is of course, a forum to discuss. There is one last job
which we need to do which is the step 4 and that is something which you need not do,
but something that I would like to share and which may be of help
So, this is not compulsory. So, what I request you to do is, if you do not have a bin folder
in your user home create one create a bin folder. So, you will see that a bin folder is
created. Then go into the DCDC there is this bin folder there are six files here; ngplot,
ngsim qplot, qsim, spice read file and this update net dot m. So, you just copy all of them
and go back there into the bin and paste fine; ngsim and qsim should be made
executable, they both are octave scripts octave scripts meaning that it starts with a hash
exclamation user bin octave. So, it calls it runs octave directly from the terminal. So,
these are useful scripts I will tell you more about this at the time and they are needed, but
its better you keep it ready.
So, to make these two executable in the readme file, I have put these two commands
change mod u plus x qsim. So, they will be made executable. So, what you have to do?
252
Go into the terminal cd bin. So, you see that it has the six files chmod u plus x ngsim. So,
that will be made executable chmod u plus x q sin that also will be made executable.
So, that is it your software installation portion is over, next we can start on building our
simulation schematics and simulate the circuits. This portion this installation portion is
one time. So, it is worth putting the effort to do this work neatly and properly now and
then forget about it next only you can think on circuits and simulation.
253
Prof. L.Umanand
Lecture - 14
Simulation setup for NgSpice and gEDA schematic capture
In this video capsule we shall take a walk through the simulation process. We shall draw
a schematic and we will take the full bridge rectifier capacitor filter circuit as an
example, generate the netlist and use n g spice to simulate it see the waveform, plot the
waveforms and try to get some insight of the circuit. One thing that you have got to
understand is that simulation is for mature people. The values of the simulation or the
results of the simulation is as good as the models that we put in.
So, remember that many a times, we will put kind of idealized model for the diodes or
transistors or various components, and expect the simulation to do waveforms which you
would get on the oscilloscope of a real system. However, there will be considerable
difference because there are many non-idealities and many of the uncertainties which we
will not be taking care in the simulation.
Keep in mind that simulation used to give you good insight of the concept of operation
of the system. It can be used for design the rating of the various components, but the real
test is when you put them into the hardware physical hardware and realize the actual
waveforms and results on the real hardware. So, use the simulation simulation and
simulation results with a pinch of salt and use it for understanding and design purposes.
So, with this kind of a limitation, you try to go ahead with the simulation.
So, let us now begin the simulation work using gEDA for generating the schematics and
n g spice for simulating. Let us begin the simulation process by opening the gEDA
schematic. Last time we had used the terminal to open, but we could also go into the
show applications menu there is the gEDA schematic startup, click icon, click on that
one you will have gEDA open up. We have a light background gEDA as we have set it
up last time. Use the mouse wheel to zoom in and out dynamically. So, first let us save it
as a file that we can name it as a rectifier dot schematic. So, what we can do is go into
file, we will do a save as I have this DCDC here. I have cleaned up DCDC, there is
nothing else we can name it as rectifier dot s c h.
254
So, it will be located in that place.
Next we go to the components. And try to draw the components from the libraries. So,
first let us go and draw the circuit components which is the diodes, the sources and the
capacitors load. The acomps folder library which we had installed, yes in the last session
contains the generic components; you can use it for most of the simulation because there
are generic models. Let us pick the capacitance we need that. We need a diode, a power
diode, we need four of them.
255
So, let us install 4, I am using e r to rotate the devices.
I will rotate this device too. Then we will need load resistance. We will put one
resistance here; I would also need another resistance to put act as track resistance,
otherwise you will land up in convergence problems. So, I will place that here we will
see how we will connect it into the circuit. Then we need to take draw the source. So, the
source you can take from the spice simulation elements library.
256
The source is a sine source; we will pick a sine source. We also need to pick one more
item here, there is spice include directory. We need to include a file which contains the
library of models custom model that we would be creating. So, we will give a pointer to
the file that that would be our library, our custom library.
Next we need to take in from the power rails library. The generic this is a generic
marking symbol which can pick of the potentials of any points or circuits.
257
We will keep on there, we need a ground symbol. This is important without the ground
single spice will not work. So, I think we have most of them in place. So, let us position
these elements. So, let me put the source here. Let me have a series impedance. This
would be a very small impedance. Let us position the diodes something in the form of
(Refer Time: 07:04) full bridge nature just position them, and then you can adjust it later
on like this, there were more diode.
Then we can have the capacitance position like this. I will rotate this load resistor and
position it like this here. This we will leave it for now. Now, let us draw the wires the
wires can be drawn using this ad nets. So, click on that now you are in the ad net mode,
you will see that it searches and links to this net.
258
I will start putting the wires click right click. Now, now will go here, click, click, right
click, click, right click, click , click, click, click, right click and completing the wiring by
right clicking click, click, right click, click, right click and so on you make the
connection. You make the connection for the diodes too. The diode is a sub circuit model
that is why we have an x mark, x at the beginning all sub circuits will have x at the
beginning.
Then the source is connected to the center portion of the bridge. And we have the circuit
in place change over the cursor. Now, the ground we could probably see ground is a
virtual concept. It is just a reference or the zero node with respect to which the voltages
are measured. If we keep the ground here, then when you measure this node, you will get
directly to output voltage. We give the ground here, then you will have to get the output
voltage will have to take voltage of this minus the voltage of this nod, that is something
which you already know circuit analysis.
Now, let us place these labels here, placing these labels are important because when you
run the simulation in spice the netlist is generated according to the labels that you would
provide. If you do not provide the labels, then it will assign some numbers. And then it
will be difficult for you to keep track of the circuit node points. If you consciously assign
labels, then you will be able to get more insight, analyze and know the netlist is much
better.
259
So, let me call this as node a and I will copy this paste it, I will rotate it, and I will call
this node b. By this what I have done, I have given a conscious label to this node. Now,
when I want to refer to the input source I will say V a minus V b which will give me the
value of the result of that waveform. Then we could give a node number here and let us
say we call this as, and we could give a node for this point. They call this as e; we will
call this as c. We, we can give one more node. Here I would like to place a track
imedance here to introduce a non ideality later on. So for now I will just not give this
label, I will give this I will remove this label and denote this label as o, so that we know
it is an output label.
So, we have the labels; we have the ground; we have the components on place; we have
wired it; we have to give the component names. So, you double click on those.
260
You will call this as R s series resistance. And it should have a value. And let us not give
100 Ohms; it is too much of a series resistance a 0.1 ohm, the reasonable value to take
care of the tracking impedance.
We will we will come to the source later we will give names for this diodes x D 1.
261
We will call that D 1; will call this D 2.
D 4.
262
D 3.
And the diodes are labeled. Then after that let us label this capacitance, we would call it
a filter capacitance C f.
263
We will call this as load resistance I will call it as r r naught. So, we will give some
value, we would keep this kept around 25 Ohms.
So, I that would see some amount of current they will put 1000 micro farad capacitance.
264
Like what we saw yesterday the capacitance value. And now the source we need to give,
some value let us first give it the name V in.
So, we have the name V in. And we need to provide the parameter for the sign attribute.
265
So, here we will have to give the sine wave parameters. When in doubt you can always
refer to the manual, I had recommended in the last session that it is good practice to
download the n g spice manual and keep it by your side. Just to look at the parameters of
the sign, now let me go to the manual.
This is the n g spice manual you see the general syntax sign, V offset, V amplitude,
frequency, delay and the damping factor. So, you can give a damped sine wave also. This
is an example of the sign you have the V in between the positive node. And the other
266
node sign 0, offset 1 amplitude 100 MEG frequency. What we need to give is 0 offset 3
V m √2 or 325, 230 √2 or 325 volts amplitude 50 hertz 00, so that is what we will do.
So, we shall put it as sign spices case insensitive 0, 325 volts is what we have V 1 50
hertz. What we want no delay and no damping, so that is it that is source is set.
Now, to the intro directory at a directive you make correct s. value 1 and the file. So,
right now I will type in a file e d t 01 dot s u v, e d t is electronic design technology 01
dot s u v, you can have your own name whichever is convenient and close. So, it will
267
reflect here, and this is the file that will get included when you generate the netlist, so
good time to save, save the file. Let us close the circuit.
The schematic, and now you see that in the DCDC folder. We have rectifier dot s e h. So,
we have now the schematics ready, but we have to add two more files, one file we will
save as into the DCDC.
I will name it as rectifier dot c ir. Always keep the same, the same name as the two
schematic file name, and only the extension you change it to cir, save that.
268
And here first line is always a comment. So, start from the second line, we walk we
would like to do, sorry, we would like to have a branch and analysis perform step time of
50 micro second up to 100 millisecond using the initial conditions. Also include the
rectifier dot net rectifier for dot net is not yet generated, but we will shortly generate that
I will let you know how to do that.
So, these two control statements for simulation include it in a separate file never include
it into the net file. It is better if you do it in this kind of an organized way, so that later on
for modification complex simulations it would be very, very easy, save that and close
this file. So, now you have rectifier dot cir also here. One more file we will include you
remember that we had a e t d 01 dot sub file which is supposed to contain the models. So,
let me save this in the same directory.
269
And this time e d p 0 1 dot s u v and save that. So, this we have linked in as an include
directive. And what should this contain. This is a library of my custom components. So,
you could have DESE science Bangalore can have your own thing. What is it that we
need to provide a model, we used a diode there and it was a power diode let us have a
model for that. So, the power diode is consisting of two parts, a model for the diode T
40, diode and the model sub circuit model for the power diode.
So, I will just write this. I think you know how to use spice how to make sub circuits. So,
I will not right now teach about, how to go about making sub circuits, because it is
outside the scope of this course. I assume that you know spice modeling model I will
give a name def now d this is a default diode model of a default diode. Now, let me
model the power diode. Now, it is a macro model. It is it is useful to how this macro
model sub circuit model, because there are a lot of non idealities which will not take care
in the default ideal diode which you can take care in the subcircuit model, and it will the
convergence numerical conversions will be much better.
So, dot sub circuit I just put 13 caps just recognize that however, spice is case
insensitive. So, let me write down the model first and diode. Let me take a normal ideal
diode say a default power diode can give some descriptive names I put a option between
one naught two and three and some high value 10,000 ohms. And C shunt one not three
one naught one 0.01 microfarad dot model, default power diode D, which has RS series
270
0.01 junction capacitance of 100 p f dot end s end the sub circuit. So, this is the sub
circuit save that.
So, I think you recognize this you have the diode between one naught one and one
naught two and you have resistant shunt resistance, then again in series with a shunt
capacitance. So, R shunt and R c shunt are in series and across the diode dx. So, this can
be closed. And we have required files. This is referred already in the schematic. This
rectifier dot c a r gives you what is the type of analysis that you want to perform.
And we need now to generate the netlist. So, generating the netlist pretty simple, open a
terminal and we will go into that folder through the terminal. You see that all these files
of there. So, let us generate netlist from the schematic. The command is like this g
netlist, slash g spice s d b dash o rectifier dot net. We want the output to rectifier dot net
from rectifier dot s c h schematic. So, when you run this, it will generate the netlist. So, if
you look there is a rectifier dot net generated. Now, these three files are the ones that will
go into n g spice or simulation.
271
So, if you look at the not let you see the generated netlist, the models taken from the e d t
sub because of the intro d dt dot s u b, you see the v in between a and b we have applied
the labels and it has taken the label names for the nodes r s, r naught the powers power
diodes and the capacitance. Now, we just have to simulate. So, let us go through the
simulation process. You are already in the DCDC folder; you have to call n g spice. So,
you say n g spice rectifier dot c i r you call a rectifier dot c i r because within c i r, you
already have said include rectifier dot net. So, it will appropriately take the netlist from
here.
272
So, now you go into the n g spice environment. So, the n g spice environment has the
circuit taken into its memory space, workspace, run type the command run.
It will run the circuit, and these are the nodes plot vectors that you will have. What we
have to do next is let us plot v naught or v output. You can always open the schematic
for reference keep it by your side here, and let us say we want to plot this plot p naught
plot v v naught. You will see the plot come up you can double click on that.
273
And then you see the capacitor charging from the starter and then the rebels as we had
discussed in theory close that you want to probably see the source current here. So, let us
say plot v naught and also I would like to plot the current of which branch v in branch v
in branch here and then you will see that you will see the current.
Now, the current through the source is always considered in the opposite direction. So,
therefore, you see that it is shown minus, but actually the current is in flowing in this
direction. So, whenever you take a source branch, you can give minus of that. By for
example, you can say plot v naught minus I v in sorry plot v naught 0 minus v in. So, this
was the proper direction of that kind. You see that initial starter current is higher than the
normal steady state current as we had discussed in the last session. So, after this quit and
you are out. So this completes the one simulation process. There are a few points and tips
that I would like to mention at this point which I will do so now.
Now in the case of the simulation after generating the netlist use the n g spice. One issue
with the n g spice plot, n g spice rectifier dot cir, and the (Refer Time: 28:47) see that
plot, take some example v naught though the plot is of a good quality and you can
expand it compressive.
274
And you see that you can also zoom into a particular part, you will get an expanded you
also of that, all these are possible. The only way you can take this for documentation is
by making a screen capture short of that so which can be pretty high memory when you
are making a document. And when you make the document, you will get this black
background. And hard copy is only for directly taking it as a printout. Therefore, for the
purposes of documentation, this n g spice plot is not very comfortable I would say. So,
you can do some documentation with that that is one issue.
275
And the other issue that we saw is at the time when we were generating this netlist. So,
when we generated the netlist from the schematic, we used a command like this g netlist
dash g e spice dash s e b so on so on. But every time typing is every time you make a
change in the schematic, you have to type this out and then generate the netlist dotnet
file. What you can do is put this into a script file, and then keep calling it in a generic
way for any such simulation.
So, combining these two problems and making a script such that these two problems are
solved using octave, we have generated the script file and that is what you had initially
copied and put into the bin directory here. So, in the bin directory, in the last session, we
had copied six files and put it into this folder from resources. You see that there is qsim
and then there is a n g n g sem. Now, this qsin is an octave script which takes care of the
issue of this netlist command generating netlist command.
And also to show you the waveforms using octave or MATLAB type of commands, and
also to put an output in the form of a PDF file.
276
Qsim what it will do is it will take the netlist do the spice simulation, and then put the
results of the simulation in to the raw file. And the raw file is passed as a parameter to
the spice as spice read file (Refer Time: 32:16) dot m file. This, this is a open source tool
function available from n g spice which which can be downloaded, but I have included it
in the resource you can use it.
So, it will read this raw file, and then the octaves g n u plot is used for plotting g nu plot
is very powerful and you can have different types of outputs PDF outputs or s v g output
from that. So, that is what we do. And this is a small script file written by us here. It is
there are there is lot of scope for improvement in this it is just a sample type of script file
which goes to show what you can do and there is lots that you can. N g sim is actually as
also similar to qsim only that it also generates the rectifier dot c i r file thus many times
we forget, forget to forget the syntax what you want to write or plan you know is the step
size ending time n u i c, but if it is an a c analysis or if it is bias point analysis so many
other analysis which are there you may forget the syntax, so that way n g sim helps you
do that one.
So, I will leave you to go through that and use it if you feel that it is comfortable. I will
just show one small example with qsim we are in the DCDC folder. I will run qsim slash
h for health, so that you know what to do. So, you have to give it in this form slash h will
give you this particular screen, qsim file name without the extension. So, qsim rectifier
277
without the dot s c h just say qsim rectifier. It will appropriately take the appropriate files
generate the net file, and then the n g spice or raw output file which will be used by
octaves g n u plot, it will be read by the spice, read find m dot m, and then a cube plot
dot m will be used to plot the outputs.
So, these are some of the syntax which you do not need to now use these two syntax
because the g n u plot itself will give you an output for whatever format you want. So, let
me just run qsim rectifier. Now, you do not need to have generated the net file. So, what
we will do I will just delete this netfile to show that that was not needed immediately
after you do those schematics, you have, this you should have this. And of course, edit
zero one dot sum qsim rectifier will generate the netlist, will generate a raw file output
which contains all the results of the simulation.
Now I if the default time, so I can give it as v naught, now here are to become a
separation I can say current through v in things like that I can also give the v v a minus v
b. So, all this let save like to see.
278
So, you will see it something like this g n u brought output, and you see here you have
you can save it into the file.
So, now let us say save as test dot s v g save, and then I will come out of qsim. So, here
you will see the test dot s v g which can which can be taken into any of the graphics
package like vector graphics package like inkscape.
279
Or anything similar packages and you can do further editing and then put in to the
documentation. So, this is how we go about doing a complete simulation starting from
schematics and n g spice and the plot outs.
280
Prof. L. Umanand
Lecture - 15
Simulating the circuit
Today we will discuss a very important circuit which is the Diode Rectifier Capacitor
Filter Circuit. Every power electronic equipment be it a low-power power electronic
equipment or a high-power power electronic equipment ranging from few watts to
megawatts you will find diode rectifier capacitor filter circuit in it. So, it becomes very
important for us to understand the pros and cons of the diode rectifier capacitor filter
circuit, how it operates, what are the issues involved and how to go about designing it.
So, that is what we will focus on this week.
We will also use octave which is a clone of MATLAB for writing on the design
equation, so that you can do iterative design. And for simulation we will use ngspice or
at least I will show the simulation on ngspice, it is open source spice platform. I will be
working on Linux; fedora Linux, but it does not matter which operating system you are
using it may be windows or any other operate distribution of the Linux. You can
appropriately use that simulation package that you are comfortable with.
LTspice is another nice simulation engine very fast and very good, you are encouraged to
try that out. Pspice is also another simulation engine. If you are having MATLAB and
simulink you can use simulink, sim power system to do your simulation exercises and
learn more about the circuits. So, that is what we will be doing this week.
281
A typical power electronic system will look something like this. You have the AC source
here it gives you an AC voltage. So, voltage source 230 volts and the power flows into
rectifier which is normally a diode bridge rectifier get a rectifier output passes through a
filter and the filter filters out the output voltage to a filter DC. The filter can be an LC
filter or just a C filter and most cases it is just only a C filter.
The unregulated DC is passed through the core power electronic system which is a DC to
DC converter or DC to AC inverter and the power flows through it to the output load
which may be in the DC form or an AC form.
In this course we are going to discuss this entire power electronic system the various
aspects of it and the various circuit that make up this. This week we will focus on the
rectifier filter alone, that the capacitor rectifier filter circuit is what we will deal with
now.
282
Prof. L. Umanand
Lecture -16
Practicals
A lab on Rectifier Capacitor Filter will have this following typical tasks. This is the code
circuit, this is the diode with rectifier capacitor filter. You code design for a typical
specification something like this 100 Watts and there is a list of tasks here.
283
The task number 1 and 2 design the circuit and simulate. Now, this is important you do it
before coming into the lab. So, 1 or 2 days before coming to the lab you will have to
spend time on this design the circuit, get the values of the C and select the diodes and
then use those values simulate the circuit in ngspice platform or Pspice platform or
LTspice platform or MATLAB environment.
After you have understood the circuit by introducing non-idealities, you can introduce
non-idealities here, you can introduce non-idealities here, track inductance and
resistance; track inductance and resistance and see the effects of all these non-idealities
on the operation of the circuit and the waveforms at various points. After you have
become conversant with the circuit you can then begin the other portions of the tasks.
Rig up the circuit and implement in the lab, if you are doing a hardware lab, plot the
output voltage waveshape and study it is ripple character and see what the effect of the
non-idealities on the output voltage waveshape is. Plot the input current with respect to
the input waveshape and study it. You can also study the currents that are flowing here.
If you are not able to measure the current probe you can put a very small resistance
fraction of ohm and measure the voltage across that resistance which you will put here
which will give you an estimate of the current. You should understand the circuit after
having include a non-idealities also.
284
The last one is very important measurement of the power factor because this is a non-
linear circuit and the currents are not sinusoidal, even the voltage maybe sinusoidal. The
current drawn by the power supply is non-sinusoidal peaky and this is known to be a
very poor power factor circuit. So, that is con, that is the disadvantage of this circuit even
though it is used most popularly everywhere and that is basically because of its low cost,
low component count and best volumetric efficiency.
The power factor is bad, how do you calculate the power factor, it is a non-linear circuit
and the wave shape is non-sinusoidal. Of course, you could use the harmonics, then from
the total harmonic distortion you can calculate the power factor, but you will need a
spectrum analyzer for that. So, here let us discuss sometime next week on how to
measure the power factor from the oscilloscope. So, this we will discuss on this in more
detail in the next week discussion.
And typically, a submission to your instructor should be of this fashion. You need to
submit a report file which is the way you have done your lab circuit implementation, and
the tabulations, the specifications and the results oscillograms should be put in there
along with your conclusions.
This is the design file which I said that you can use octave or MATLAB and putting the
equation, so that you can do iterative design for any given specification. And give these
two to your instructor as part of your lab work. However, lab work or physical lab work
is not in the scope of this particular online MOOC course, but this is just a
recommendation of how you should go about doing it in your own respective labs.
285
`Fundamentals of Power Electronics
Prof. L. Umanand
Lecture - 17
Inrush current limiting - intro
Let us now discuss a very important topic and that is Inrush Current Limiting in Rectifier
Diode Capacitor Filter Circuit. This inrush current limiting is a very serious problem in
the rectifier diode capacitor filter circuit. So, what it basically does is that in the circuit as
we see here you have the source connected through a switch and through a diode bridge
rectifier like this when you have a capacitor here and this is the V naught across the load
and the load may be resistive or any other power supply itself.
And the problem here is that the source the current through from the source passes
through the switch during on condition which is a very low resistance, passes through the
diodes in the forward bias condition which is again very low forward resistance and
passes through the capacitance and back through this diode and the source. Likewise, in
the negative direction also it takes the alternative path.
The problem here is that you have very low negligible impedance in the path the current
flow path for the current flow from the source to the capacitor, to charge it up. So, at the
time of startup when you want to switching on the switch, you can expect a very large
286
inrush current to be drawn from the source and this large inrush current can blow up
these diodes or even the capacitor. And we have designed all these diodes and capacitor
to operate only in the steady state condition.
So, this implies that even though the diodes have diodes and the capacitors have been
designed for steady state operation the moment you switch on the circuit, by switching
on the switch the components will blow and the circuit will not operate. So, every time
you switch on the circuit there will be components blowing. So, how do we overcome
this problem of inrush current limiting? Now, that is the discussion now that we will see.
You can now consider the source Vi connected through the switch flowing through the
capacitance and connect them this fashion. So, the source Vi having a voltage Vi and if
you switch on the switch close the switch you expect a current to flow through this path
and this at the time of startup is called the inrush current and this can be very high.
Why will it be very high? Now, let us say that you have Vc as the variable indicating the
voltage across the capacitance here. So, inrush current is difference in this potentials, so
that is (Vi – Vc) divided by whatever is the path impedance. Now, the path impedance is
comprised of the impedance of the source which will be very low the impedance of this
closed switch which again will be very low, esr of the capacitance again low and track
impedance wire impedance and so on. So, therefore, the impedance in this path is very
low and that Z is tending to 0.
So, let us imagine a situation where when you are starting. So, at the time of starting V c
is 0, so this is 0. And the inrush current is given by Vi by Z and Z is very low. And as
chance may have it at the time of switching this ON Vi worst case possible can be V
peak, the max possible value in the sine wave. So, that value divided by Z can be a very
very large value, can be 1000s of amps and this is 1000s of amps these diodes are not
designed for and they will blow up. So, every time you switch this on, this diodes will
blow up or the capacitor will blow up and this circuit will not function at all.
We have designed these diodes and capacitor to function in this steady state, but we are
not able to take the circuit into the steady state because of this inrush current problem.
Why should the circuit operate during the steady state? So, at the steady state the
capacitor is fully charged in this; this is fully charged. So, this is close to Vipeak. So, Vi
peak minus V c which is close to ipeak could be a delta V divided by a small value
287
divided by this Z will give you significantly lesser inrush current. And that value is what
we have design these diodes for, the capacitor for to handle these currents.
So, in the absence of this charge here you can expect a very large inrush. And how do we
make the transition from the switch ON transition dynamics to the steady state dynamics
where these components can handle the current, now that is where we need to put some
circuits here which will make the transition, startup transition and then hand the bottom
over to the steady state portion and the steady state portion the diodes and the
capacitance will behave as we have discussed in the last week.
So, let us now see what circuits we need to interpose here to make the circuit operate
even during startup and transit nicely and smoothly into the steady state operating mode.
Now, let us see what happens to the wave shape when the switch is closed. So, let us
draw the x and y axes and let us say that this is the rectified input sine wave, it has an
amplitude of V m. So, it is basically V m sin omega t and we are taking a benign case,
where at the time when this switch is being closed that is t equal to 0, the input wave is at
0 and it rises in a sin fashion. So, you are synchronizing the closing of the switch with 0.
Under such conditions what happens? So, this is the t axis and you will see that there is a
large inrush current and that is corresponding to this i here, I will mark it here.
288
So, large inrush current here and what happens to the voltage wave form, output voltage
wave form? This voltage wave form will rise here along with Vi up to this point and then
the Vi is falling the diodes are reverse biased, then this will start going down in this
fashion depending upon this load. And then again when Vi crosses and forward biases
the diode here you will see that it charges the voltage follows the input wave shape, so
and this repeats. So, here actually you have reach steady state. So, here there will be a
current flow as we have seen in the previous modules.
So, this will correspond to the charge state of the capacitance. So, you will see that this
peak amplitude is correspondingly smaller compared to the starting peak, ok. Now, this
is a benign case where the switch turn on is synchronized with the 0 crossing of the input
which will not happen in most cases, it will not happen in a particular situation. When
you turn on it may so happen that the input is at its peak at the worst case condition and
what happens.
Now, consider the worst case condition where at the time of switch ON of the switch,
switch ON of the switch, it so happens that the input voltage is at its peak value the V m
is at this point suppose sine wave. So, this could happen this the worst case condition if
this happens what will be the peak inrush current.
The steady state portion there is no change there will be similar to what it was here.
During this point it is Vm and it is Vm by impedance Z there V c is 0 that is a charge on
the capacitor is 0 therefore, no voltage. So, you will have very large overshoot I just want
to show the relative magnitude will be something like that will be much higher than what
would be for a synchronized switching where Vi starts from 0. So, this could be a very
large value. It could go to 1000s of amps.
So, this is the current here and the capacitor voltage quickly builds up to this value and
then the diode cuts off and then the steady state starts. So, you see that it is during this
time the entire problem occurs and during this time itself the diodes will blow and you
will not be able to take it to the steady state. You can in fact, verify this by simulation. In
the previous modules, I have shown you the simulation of rectifier circuits in ngspice,
you could probably get some insight by doing some simulation on that. Probably I will
show you a quick example.
289
Prof. L. Umanand
Lecture - 18
Inrush current limiting – Resister solution
This simpler solution for inrush current limiting problem is by putting a resistor. As we
saw in the simulation putting a resistor will definitely limit the inrush current. So, let us
put the resistor around this place here let us shift it out and then include a resistance like
this. So, this will be the surge limiting resistance R S.
So, how to find the value of R S? So, R S will be V i minus V c, V c you consider it as 0
worst case condition and at startup V i max. So, let us say V i max divided by R S will be
the inrush current limit. So, therefore, V i max divided by in limit, limiting value. So,
this value can be obtained from see while doing the steady state analysis and steady state
design you would have chosen the diode from the data sheets to the diode, you will get
the non repetitive peak current rating of the diode and it should be less than that peak
current rating of the diode; non-repetitive peak current rating.
So, once you have selected a diode let us say it has a non-repetitive peak current of let us
say 50 amps then you can choose I in limit at around 45 amps or even 50 amps so that
you are within the safe limits during the steady state operation.
290
Now, with respect to Vi what to choose see the input voltage is to 230 Volt rms that is
the nominal value, but it has a tolerance plus minus 20% in most places. But this plus
minus 20 percent tolerance is a standard number, but it can go to much higher values of
degrees of tolerances and at different places. So, you can accordingly take the tolerance
limit according to places. So, let us say we have a tolerance limit 230 Volts from the
nominal it can go plus or minus for as the worst case for this application is plus 20%.
So, if at the time of switching it was Vimax which means 230√2 + 20%. So, you will see
that the range is from 260 volts to 390 volts corresponding to the plus 20 percent max.
So, use this value 390 volts here divided by I in limit what you want to be the limit, so
that you are in the safe region for the diodes. So, this is how you have to choose this
surge limiting resistor. So, if I put mark the voltage measurement like this and is you are
keeping the common probe here in the positive the probe here you get VRS because
there is a current I in flowing through that and the I in flows in this fashion as we saw in
the simulation, right, in steady state.
The job of R S is only during the starting inrush current after that it is not needed, but it
is there in the circuit, so because of that you will see that there is a voltage drop which is
coming across at due to this nature of current that is flowing through it. So, you will have
a voltage drop in across the resistance in this fashion having similar wave shape as the
current and that is VRS which is i into RS. Now, this drop is coming across this so
which means that the voltage across the load V naught will be lesser by that much
amount.
So, therefore, you will see that the voltage is lower; the average voltage at the output is
lower compared to the one without having put RS. So, one of the main advantage of this
solution is simplicity. So, that is the main probe, simplicity, you just need one
component nothing more and it is robust and it will solve a problem and it will make the
circuit operable during startup and steady state.
However, there are disadvantages, one of the main disadvantage is that it is lossy
because there is going to be a drop current flowing through it there will be a loss and
therefore, the efficiency of the entire circuit will suffer will be low. And the second
problem is the available voltage at the output is lower, so lower V naught as compared to
291
the case where you do not have RS. So, let us look at few other solutions where these
disadvantages are removed.
292
Prof. L. Umanand
Lecture – 19
Inrush current limiting – Thermistor solution
Another solution slightly better solution than the resistor one is the use of a Thermistor.
Thermistor is actually a variable resistance which wherein the resistance varies with
temperature. And it is a negative temperature coefficient thermistor, but you need to use.
I will explain what it means.
Symbolically, the thermistor can be drawn like this in a electric circuit this passion. So,
this is R ts, so this would be a thermistor, this indicates that the resistance varies with
temperature. So, NTC means Negative Temperature Coefficient, what it means is that the
value of this resistance will decrease as the temperature increases. This is a core concept
as the temperature increases the value of the resistance decreases. What does this mean?
So, let us say at the time of start we have not yet closed this, the cold resistance R ts has
some value.
Now, on closing this Vi by Rts it will limit the current. So, inrush current is limited and
then quickly moves into the steady state operation and there is a current flows through
this and they drop across this. So, as this dissipates this starts becoming hot; Rts will start
293
becoming hot and as the temperature rises the resistance will start reducing and it will
reduce to a point where the power dissipation becomes very low. So, one of the problems
of the resistor solution where we said that it is lossy in the steady state and therefore,
efficiency is low this thermistor avoids the problem by using this automatically varying
resistance as it becomes hot, the resistance value goes down and becomes less lossy. So,
this is one of the advantage by using the thermistor.
Let us see how this thermistor looks like. So, this is picture of a typical thermistor NTC
negative temperature coefficient device. It has 1 ohm as its cold value. So, at low
temperature or room temperature it will be 1 ohm and as it starts heating up dissipating
this will become hotter and this this resistance will start going down from 1 ohm
onwards.
So, if you plot on the x-axis temperature in degree centigrade and on the y-axis R ts the
resistance of the thermistor and let us say it is 1 ohm at 25 degree centigrade it will start
reducing as the temperature increases temperature of the device increases because of
dissipation. So, it will come to a low value of series resistance and in the steady state the
loss will not be as much as it was in the case of a fixed resistance. So, you get
thermistors of a various cold resistance values and you also get various current handling
capability of thermistors. So, look into the data sheet and you can choose appropriate
thermistors.
One problem that you need to be aware of while using thermistors is that let us say you
switch it on, this thermistor has a high cold value resistance it will limit the inrush
current and then passes on to the steady state operation works nicely. Now, let us say the
power goes off, once the power goes off the capacitor discharges the capacitor goes to 0
voltage condition. If immediately in a very short time gap the power comes back again
capacitance of discharge condition the thermistor has not cooled enough to go back to its
cold resistance value and therefore, the current limiting at that instant will not be as
effective.
So, you have to be careful there because when you switch it off you have to give
minimum time gap for the thermistor to recover its cold resistance value, only then the
next time you switch it on it will perform the inrush current limiting action. So, this is the
only problem when there is a power going off and then coming back in a short span of
294
time, there could be issues in the thermistor not limiting the current because of not
regaining its cold resistance value. Otherwise, this is a very good solution.
295
Prof. L. Umanand
Lecture – 20
Inrush current limiting – Transformer solution
Another inrush current limiting solution is by using a Transformer. So, let us interpose a
transformer in between. So, let me include a transformer like this. So, you have the
primary, you have the secondary with a dot polarities like this as shown and let us say
the primary to secondary turns ratio is 1 is to n. So, if you have a Voltage Vprimary
Vsecondary is equal to n times Vprimary. So, that is what it would mean. So, this is a
transformer and this is how it is connected and this has very many advantages.
First of all it is not lossy, it is not a lossy element and therefore, efficiency is not
affected. Secondly, there is galvanic isolation what it means galvanic isolation is that
there is a physical isolation. The conductors here on the output side and the conductors
on the input side are physically isolated which means that if there are any surges here
due to any part the circuit on this side or there are any surges on this part of the side they
are not communicated to the other side. So, therefore, any surges appearing on the grid
or the mains will not go and destroy the components on the load side. So, that is one
major advantage of galvanic isolation.
296
Another advantage is that the primary and the secondary coils can have different turns
number of turns, as a consequence you can have a turns ratio which will result in scaling.
So, let us say for example, you have 230 Volts here and I want here a 15 Volts to give to
a 15 Volt power supply. You can adjust the turns ratio such that it is a step down
transformer or if this is a lower Voltage AC and you would like to have a higher Voltage
here it can be a step up transformer. So, n is degree of freedom that it gives to the
designer to choose a proper scaling factor to make the Voltages between input and the
output compatible, where as in the case of the one without the transformer the output
Voltage is link tightly to the input Voltage amplitudes and fluctuations
Another advantage is the inrush current limiting and that is why we propose the solution.
So, inrush current limiting is a natural spin off by putting the transformer. I will tell you
how this comes about if we draw the equivalent circuit you will see that there are leakage
inductance was which will prevent the first cycle current inrush happening.
One main, a one and only main disadvantage that I can say at this point in time is that the
low frequency transformer is bulky and secondly, it is costly. Now, these two these two
points bulky and costly are the ones which I can consider as major disadvantages more
from the expense point of view from the point of view of theoretical operation putting a
transformer is really good, one of the best because it is bulky and costly, we normally
limit this kind of a circuit putting a transformer to low power circuits, low power
applications where you do not need to spend too much on the transformer in terms of real
estate and price.
Now, let us see how this transformer will solve the inrush current limiting problem.
297
So, what we shall do? We shall replace this transformer with its equivalent circuit. I will
take n is equal to 1 for now it really does not matter even if n is other than 1. So, we will
replace it by a transformer T network equivalent where this is the mutual inductance M,
this is the primary leakage, secondary leakage or primary or secondary; so, the winding
resistances.
Now, this would be the equivalent of the transformer and all these impedance, if I take M
very large compare to all these impedances there will be a very small amount of current
magnetizing current only that will flow through this and all the load current flow through
in this fashion. So, all these impedances will act to limit the surge current, so that is the
logic. So, if we take M very large compared to LσP, plus LσS, then if you say the
Voltage across this as leakage inductance is VL then by Faraday’s law you have
VL = (LσP + LσS) dIin/dt.
So, assuming that this is very large and the current through this is very small compared
to the current load reflected current that flows through. So, the current is basically an
integral relationship with respect to the Voltage V L have developed across the leakage
inductances. So, initially when you switch it on capacitor here is not charged is like a
short coming across here and all the Voltage will drop across the inductance because that
will give the maximum impedance at that instant of switch on omega is very high, omega
L will be very large and all the Voltage will drop across thus or across the inductance
298
and gradually the current will start building up. So, it is like an integral filter and you
will see the current gradually building up and thereby inrush current limiting happens in
a natural sense.
So, this is one of the very good by-products or spin offs of using a transformer in the
circuit.
299
Prof. L. Umanand
Department of Electronics System Engineering
Lecture – 21
Inrush current limiting – MOSFET solution
There is another Inrush current limiting method that we will discuss, and that he is by
using an active semiconductor device like the BJT bipolar junction transistor or the metal
oxide semiconductor effect. The circuit description will be like this, we see that the
source through a switch is connected to the diode bridge rectifier output of which goes to
C filter like this and connected to the load then finally, is connected in this fashion. So,
this is our regular rectifier diode bridge capacitor filter circuit, the switch. So, we have t
equal to 0 the switch will be closing and a Voltage Vi will be applied, this is V naught.
So, this is our regular circuit.
So, what we will do is here we will remove this conductor and replace that with a
resistor; a variable resistor. So, here two you will get the same resistor effect where it
will limit the current. If you see the moment the switch is closed current flow in this path
through this, goes through this parallel paths comes through this protection resistor and
back again likewise in the other half cycle also.
300
So, this device which you are putting in here is a series device and it will limit the inrush
search current. So, if you make this is variable resistance wherein if I plot or versus time
this is time. So, with time we switch this resistance from high value to a low value. So,
that during time closer to 0 which means during this switch ON, turn ON you will see
that the value of the resistance is high and as time progresses the resistance of this device
will decrease ok. So, this device we will be emulating it with a semiconductor device like
a BJT or a MOSFET. You could either use the BJT or a MOSFET, I will demonstrate
this example with you using a MOSFET for example.
Now, let us replace this conceptual variable resistor with a MOSFET. So, let us remove
this portion and let me introduce a MOSFET here, n-channel MOSFET, is an n-channel
MOSFET arrow pointing towards the channel. So, let us introduce this MOSFET at this
point here.
Now, here I am going to place an attenuator. So, let us first put some number, so that the
understanding is clearer I am going to say that the grid Voltage is at 325 sin omega t
which will give you 230 Volt RMS and here let us put an attenuator comprising of two
resistors. These are low power low wattage resistors. And calculate the values of this
resistors in this following manner. Let us say if it is 325 Volts here at steady state let at
this point it be 15 Volts. So, that it drives the MOSFET to full on condition.
301
Ask the you will see that this potential is gradually rising; as it is rising if it is at 260
Volts, this will be at 12 Volts. So, the MOSFETs between 12 and 15 Volts it will be in
full enhancement which means low R DS on. So, across this here let us put a capacitance
and then I make a connection here and I take the gate drive for this MOSFET from here.
So, take to gate drive from here.
So now, the operation is pretty simple. Now, I have the x axis as time and I would like to
monitor this V gate to sources, this is Vgs V gate to source and let us have Vgs. Now,
what happens to this Vgs which is this point? Now, this is an attenuator and this is a
capacitor this RC filter so, this is gradually going to move up to this 15 Volts. So, you
will see a wave shape with time going on like that towards the 15 Volts line.
So, let us mark the 12 Volt flying, it will cross the 12 Volt at this point and at this point 2
Volt typically I am saying 2 Volt the threshold Voltage or the cut in Voltage for the
MOSFET. So, beyond 2 Volt the MOSFET will be in the linear region and up to 2 Volts
that is in the cut off region. So, let me mark these verticals and at the top here I am going
to draw one more marking this is as drain, this is source, this is gate. What is the
resistance between drain to source, R DS?
So, as this gate to source Voltage as it is increasing till 2 Volts till this point, this is in the
off-state R DS will be high, be very high. And as it raises as this raises to 2 Volts and
beyond, R DS will start reducing because MOSFET goes into the linear region here it
will be in the linear region greater than 2 Volts up to 12 Volts let us say; let me extend
the top.
So, you will see that the MOSFET starts going down the R DS starts going down, till at
around 12 Volts is it reaches saturation and enhancement and tape flattens out and that is
called R DS on. And R DS on of the MOSFET is very low. So, you will see that from the
off state which is very high value of impedance here, its starts going down and in the
linear region the R DS is going down and then finally, reaches a very low value of R DS.
So, the moment you switch this ON V gs is gradually increasing because, this is in the
off state there is no charge current flowing here, the charge control will flow this through
this path and charge of this capacitors. So, as this capacitor is increasing as its starts
going beyond 2 Volts this is going into the active region. So, there will be some charge
flow here. The current is automatically getting limited because this is a large value of
302
resistance. As this resistance decreases this R DS starts coming down more and more
current starts to flow on this side and finally, as the V gs reaches to around 12 Volts and
beyond the R DS of the MOSFET would have saturated.
So, this would be the full-on condition and which would mean that you are in the steady
state from here on, ok. So, in this way this is a very nice and interesting circuit which can
give you inrush current limiting simple circuit, active circuit and the MOSFETs can be
for even high-power ratings and high Voltage ratings. So, instead of the MOSFETs we
want to use IGBTs for higher current and Voltages that is also possible it could be an
IGBT, a MOSFET or a BJT depending upon the power levels.
There is one problem in this circuit which is similar to the problem that we saw in the
thermistor case one of time constants. Let us say this capacitor is charged up to 15 Volts,
the MOSFET is in full ON condition and this power goes off. The moment is power goes
this capacitor discharges through the load and this capacitor will discharge through this
resistance.
Now, if this time constant of this discharge if this RC time constant is much larger than
this time constant then there is a situation where you can have a problem. Now, let us say
after this power goes this has discharged and this has not fully discharge and the power
comes back. So, during that condition you will see that this is in the charged condition
303
and the power comes back and you can have an inrush current. So, that is a critical
situation which can cause a both of any of the components.
There is further problem if you consider most commercial MOSFETs there is a body
diode here. In itself the body diode is not an issue, but if you take the condition when this
power goes off; switch is OFF this power capacitor here discharges through this load.
Not only does it discharge through the load it can discharge through this path parallel
combination of this through this body diode and then complete this path.
So, therefore, you could see that this capacitor will always be charged up, the capacitor
discharge capacitor would be discharge and when this power comes up you will see that
the MOSFET is in full-on condition and inrush current. So, there is no decoupling of the
discharge of this capacitance and this capacitance. In order to decouple we can put as
diode here in this fashion as I will indicate.
I will start modifying this by erasing here, and I will make some space by pushing it right
and then make this connection, make this connection here. And now to this point instead
of taking it from here, I will tap through a diode, I will put another rectifier here and
from here I will tap through by putting another rectifier diode and then connect it here.
So, if you see these two are low-power rectifiers currents only to deliver to this gate
circuit which is the high impedance high input impedance circuit. Now, these two diodes
304
along with these two diodes form a full bridge rectifier; this form a full bridge rectifier,
supplying this capacitor load, but there is a large resistance here, so there won’t be input
surge current problem. These two diodes along with this are the regular power diode
supplying to the power portion of the circuit.
So, when you switch ON this switch, from Vi the power will be supplied to this
capacitor, the capacitor Voltage will increase Vgs will cross the threshold value and this
will go into the linear region when this will gradually start charging up. And once Vgs
goes to 12 Volts enhancement occurs and this will go to the full-ON condition and the
capacitor will be operating in the steady state. And when this goes OFF, when this Vi
goes off the capacitor will discharge through the load.
305
Prof. L. Umanand
Lecture - 22
Inrush current limiting – Relay, contactor
Another very popular method of Inrush current limiting is by use of the Relay and
Contactor. This is a very robust and reliable method and this is very well suited for high
power application where the power flow is very large and you will see that even though
the circuit may look a bit complicated it is a very very reliable.
306
Now, let us introduce this relay and contactor into the circuit. I will explain the operation
of the relay and contactor a bit later, let us first now introduce let us understand the
concept of introducing this relay into the circuit.
Now, I am getting a resistance here and connected a resistance like this. Now, this is the
typical first inrush current limiting circuit that we discuss that is putting a series
resistance, same circuit. So, here the moment the switch is closed you will see that the
inrush current is limited by this resistance here and as a consequence all these design
values are safe and after sometime it will go into the steady state the first cycle inrush
current limiting is done by this.
But we do not want this resistor to be present after it has done its job of limiting the
inrush current. So, after one or two cycles we want this to be removed from the picture.
So that is where the relay and contactors come into the picture. So, let me introduce these
two terminals. The idea is that for the first few cycle, first one or two cycles this resistor
is in the picture limits the inrush current, steady state is achieved, after that I would like
to short this terminals and bypass this resistor and the resistors out of the picture and
there is no loss, no dissipation.
Now, here let me introduce the contactor. Contactors and relays both are electromagnetic
devices they have a core and around the core you have a coil bound, so like a solenoid.
So, in the coil is energized the core is pulled attracted and because of this moment this
307
core which is conductive will make contact with this terminals and establish short circuit
between these two terminals. If the coil is not energised due to the presence of this spring
this will be released and it will not make contact with these two terminals and the
terminals of the open circuit.
Now, how to energize this coil? This coil comes in different Voltages available
commercially you get 24 Volt, 48 Volts, 230 Volts. I will right now choose a 230 Volt
coil which means if I connect 230 Volts across this coil, the coil is energized, contactor
is energized, it will attract the core the core will be pull down it will make contacts with
these two terminal and short circuited and there by bypass this resistor.
So, let us make that push to the circuit. I will take this 230 from here itself tap it and
connected to this coil through a switch. I will put a switch here let me explain that later
why I am putting the switch and finish the connection on that side. So, if this switch is
closed 230 comes directly across this coil. Coil is energized it will attract this, this will
short circuit these terminals and bypass the resistor, that is idea. So, how do I close the
switch? Now, this switch is actually coupled to the coil of the relay. So, there is a relay;
the relay basically is for low Voltages. So, these are low Voltage devices contact was
generally use for high Voltages these are high Voltages contact switches.
So, these relay can be operated from DC sources. So, I will connect a DC vcc maybe 15
Volts and use a semiconductor device like a BJT and connect it to the ground in this
fashion. And the BJT I will provide a simple resistive based drive circuit in this fashion.
Across the relay coil you need to put a diode, you will have to give a freewheeling path
for the relay coil is inductive and once the switch is off; switched off there is a current
flow and that should find a freewheeling path to dissipate its energized, its energy
through the l by r constant the r being the resistance of the coil itself.
So, you can go and buy relay coils having 100 ohm resistance, 200 ohm resistance, 12
Volt 100 ohm like that. So, there are very many number of different specs available in
the market you can look at the data sheet and purchase of coil accordingly, probably a 15
Volt 100 ohm coil can be used appropriately for this particular circuit.
Now, how do I energize this here? So, let me energize this by connecting it to an op-amp
and the op-amp output, output will drive this BJT. So, let me power of this op-amp
power supply plus and minus terminal to the minus terminal I will give the threshold
308
Voltage. How do I give the threshold Voltage here? Let me set up a threshold Voltage I
have a resistance I will use Zener diode and across the Zener diode let me connect a
potentiometer. So, by adjusting this potentiometer, I can get various, now let us say I use
a 8 volt Zener 8.2 Volt Zener. So, you have 15 Volts; 8.2 Volts come across this it is a
stiff reference and by adjusting the potentiometer I can get any particular Voltage from 0
to 8.2 Volts.
Now, this let me connect it here and that will form the V threshold for this circuit. So,
what does it mean? So, let me explain what we connect here. So, here let me connect the
feedback from the output that is V naught. So, the V naught let me step it down through
an attenuator and pick it off at this point you can choose the value of the resistor such
that you get any particular desired attenuation let us say a is attenuation, so that is
connected to this point. And let us say the attenuation here is and therefore, a V naught is
the potential that will come at this point here. So, V naught is here, a V naught comes in
here at this point.
Let me choose at V naught 260 Volts. So, a 260 value will be the potential I will set here
through pot. Now, this becomes the V threshold reference. What it means is that I close
the switch the input is charging the capacitor through this resistor. Why? Because this is
low, this is off, this is off, coil is not energize therefore, this switch is open and as this is
open this coil is not energize and therefore, this is in the release not connected position
this is open and therefore, the only path is through this resistance and into the
capacitance to charge it up and this Voltage is building up.
And as this Voltage is building up you will see this Voltage a V naught is raising so, as
this a V naught is a raising and reaches 260 Volts, as V naught reaches 260 Volt this
potential will cross this V threshold point at which point the out the op-amp becomes
high and this will drive; the base drive, base current through this BJT, and the BJT will
turn on. Once the BJT turns on there is a current flow through this relay coil.
Now, once there is a current flow through the relay coil this will switch on among the
switch is on the 230 Volts comes across this coil and energizes this coil which will
attract this core of the contactor and this contactor will make contacts with this two
terminals short circuited and thereby effectively bypasses this inrush surge limiting
resistor.
309
So, this way this circuit will operate and when you switch off this the Voltage here will
decrease and as a V naught decreases and comes below 260 this will go switch off. This
will free wheel through it and the energize, this will open and once it opens this coil is
the energized and you will see this will get released and this will get open and surge
limiting resistor again back in action immediately.
So, this is the circuit which is most popular and reliable and used in majority of the high-
power circuits. In order to complete this circuit, we should see how we get this DC
power supply here only then the circuit becomes complete. So, we will get this DC
power supply from this means 230 Volts itself. So, how do we do that?
So, let me draw this portion here, the circuit and let me tap it from this. Let me put a
transformer. Now, you see I do not want to put another surge limiting type of circuit
active surge limiting circuit. I will use a passive surge limiting circuit like a transformer
for a low power application because all these power put together will be very small
compared to the load power therefore, one is justified in using a transformer based
rectifier capacitor filter where isolation and inrush current limiting will happen.
So, connect it in this fashion we know how to connect this this is the full bridge diode
and output the full bridge diode you connect a capacitor and across the capacitor. The
load here will be a linear regulator and outer the linear regulator will give you the power
supply.
310
And say we have 15 Volt power supply, there are various linear regulator ICs which can
be used to generate the 15 Volt power supply. You can use the 7-8 series; you can use
the lm 317 lm 350. For example, here I can say that we can use a 7815 to generate a 15
Volt supply here and this will be connected across these two points, these are the
common points here and the ground, yeah. So, that way all this portion will get power.
Now, the circuit becomes completely operational and self contained.
311
Prof. L. Umanand
Lecture – 23
Three-phase rectifier capacitor filter
For three-phase AC to DC applications you have the three-phase rectifier capacitor filter
circuit. It is very similar to the single phase rectifier capacitor filter circuit, except that
we need to add one more arm to account for all the three-phases.
Now, let us take the sources three sources, va, vb, vc they are star connected in this
fashion the neutral and these three line voltages you bring it out and let us connect it to
the diode bridge. Now, you have a three-phase diode bridge one for each phase and
connect the phases in this manner, then output of the diode rectifier you connect the
capacitor and then the load resistor. So, this a b c this will be va, vb, vc, let us say
measured in this fashion, this v naught and i. Except for the addition of this extra diode
bridge and three-phase supply rest remains the same.
Let us look at the waveforms of this current and the output voltage to understand the
three-phase rectifier capacitor filter operation. Let us now draw the wave forms for the
three supply voltages and see how the ripple voltage waveform at the output would look
like for a three-phase rectifier capacitor filter circuit.
312
Now, let us draw the x axis, the x axis is t and let us now draw the three-phase wave
forms. I will draw in this fashion you will know that three-phase wave form drawing is
difficult, but doing it this way will make it at simpler. Yes, now this is the a b c, a b c, so
on and now when it passes through the rectifier at this node, across this node if capacitor
were not there you will see that this will rectifier for the this will be rectified wave form
fold back.
So, I will show that in dotted. So, you will see that these will get rectified this bottom
portions will get rectified in this fashion. And then look at this envelope. This is the
envelope that you would see here and this envelope you obtain even without the c filter.
So, this would give you 13.3 percent ripple.
Now, this if I zoom I should be able to I will take this as the reference 0, and then with
respect to that I will zoom this, and the output wave shape will be something like that.
Similar concept, where in steady state at this point let us say the diode conducts, the
output voltage tracks the input envelope and then discharges. At this point tracks the
input envelope and once the input envelope goes below the output diodes are reverse
biased and the capacitor discharges, so on, this continues.
Now, this would be the conduction times whenever they tract the input envelope, and if
you draw the verticals you will see the current wave shapes will look something like that.
The current wave shape, this current wave shape is at this point.
313
Prof. L. Umanand
Lecture – 24
Simulation – 3 phase rectifier capacitor filter
Let us now develop the spice simulation model for a Three-phase Full Bridge Rectifier
Capacitor Filter.
314
So, first let me change this title. I will make it as a three-phase full bridge rectifier
capacitor filter and place it there. We will remove this, remove this.
Now, we have to add three sources, let us redesign these arcs. So, let me take that out.
Now, let us copy paste because we need a third arm and I will put that there. Now, you
need to expand this a bit more. Let me pull it down, so that you can accommodate more
space here to accommodate a three-phase star connection, then make this connections
proper completed complete this. Let me center this. Yes, now we can move this together.
We will make duplicates copy, control paste, control paste. So, we have three sources
remove you can rearrange them accordingly so that does not clutter up.
315
You have to name them. Let say we name them as V a, V b, V c or V s a V v s, VSe. So,
V b, V c, then rename this diodes 5 and 6. So, you have to give unique names for the
components those that you have copied and then make this one as sin 0, sin omega t and
this one 13.3 milliseconds and 6.66 milliseconds. So, you will have that. You will have
to duplicate this and connect control copy, control paste, copy, control paste, control
paste. So, let us rearrange and connect this.
So, after rearranging and then properly connecting the entire three-phase circuit looks
like this. So, you have V a, V b, V c, these are phase voltages with respect to the neutral,
V b with respect to the neutral, V c with respect to neutral. So, the phase voltages are 325
volts peak ok, and they are all delayed 120 degrees apart and you have Lsa, Lsb, Lsc as
the lumped non-ideality in the AC line or Rsa, Rsb, Rsc are the surge limiting 0.5 0.5 0.5
ohm surge limiting resistors.
Then this is another non-ideality. This extra diode set that we have put here, these two
are the same. Now, we will simulate I have observed the node points that I have labelled
this is node a, node b, node c, node n. So, if I say V a comma n it will be between these
two, if say V a comma b it will be the line voltage between a and b, V b comma c would
be the line voltage between b and c, c and a, so on. So, between this and this it is line
voltage, this and this it is line voltages. So, what you expect here? It is the peak value of
the line voltage which will be which will be 560. The voltages V a, V b, V c are with
316
respect to the neutral. So, this node will be 325 volts peak, 325 volts peak, 325 volts peak
or 230 volt rms. Now, the line to line voltage is root three times which will be 400 volts
rms and root 2 of that around 560 volts will be the peak and you will be expecting 560
volts approximately about in that range across the output.
So, now, let us see how we go about simulating and let us see the various waveforms. I
like to see the current at this point current at this point, would like to see this voltage,
like to see all these three voltages. So, let us modify the rectifier dot c i r circuit, so that
automatically it calls ngspice, runs and then plots the first draft waveforms.
So, look at this. I have opened the rectifier dot cir. These are the standard tran statement
and include the net list statement. I have included between the dot control and dot end c.
The control statements to be operated within the ngspice environment automatically; I
am setting the foreground background colour as white and foreground as dark black.
Then asking ngspice to run the simulation on the net list and plot v a b that is the line
voltage a to b away from here to hear plot b c that is this to this and c to a. And I would
like to see the inductor current here, so that I will get this branch current and I have also
would like to see v naught. So, let us have a look at all these waveforms in ngspice after
the simulation has been concluded. So, for that let us open a terminal.
317
So, we have opened the terminal and let us run this simulation I am calling a runsim and
rectifier.
So, if you do that goes in the ngspice runs the simulation plots.
318
Let me maximize that. So, you see whole lot of plots here, you see. These are the line
voltage red this one, where my cursor is travelling through that is v a to b, blue that is v b
to c and this yellowish colour will v c to a. Green, green is the current in the branch just
before the capacitor. You see after one cycle steady state is reached see that at every at
every point where there is a ripple where the output conducts you will see the conduction
here. You will see this drop here is due to the resistive surge limiting otherwise this surge
would be very very high urge limiting circuits are similar, even in the case of three-phase
you can use the circuits that we discussed for surge in rush current limiting given here.
So, after the first cycle surge limiting you see that it goes in to the steady state you will
see that there are 1, 2, 3, 4, 5, 6, ripple 6 ripples in a fundamental cycle. So, the ripple
frequency is 6 times and this is the output wave shape. If you would like to see what
would be the inrush current limit when we remove or when we reduce the surge limiting
resistor value, then you will see that the inrush surge limit can go to very large current
values. Let us have a look at that also.
319
I will change these values to 0.1 ohms; almost no surge limiting resistor there. These 0.1
ohms would represent all the line resistances and no more.
So, let me save this, and let us go to the command line called rectifier, run that.
320
And you see the; you see the first cycle inrush current 2.6 kilo units, say the 2.6 kilo
amps and then immediately in the next cycle you see that it has achieved steady state.
And observe the drop because we have reduced r s the surge limiting resistor the drops
are much lesser, if you want to zoom to any particular place right click drag that and you
could see the zoomed values at that point.
So, this way I will recommend that you try to study the three-phase circuit and the
various waveforms branch currents node voltage waveforms at various points to
understand gather more insight into this three-phase full bridge rectifier c-filter circuit;
very similar to the single-phase full bridge rectifier c-filter circuit.
321
Prof. L. Umanand
Lecture - 25
Power factor - motivation
Power factor; what is power factor and why is power factor so important, let us see the
motivation for trying to find the power factor and also to improve the power factor of
any given load or circuit. So, we have here the regular rectifier capacitor filter circuit that
we have been discussing till now, this is input Vi source sinusoidal AC, i in is the diode
bridge you have the current i here and the v naught here.
Now, look at the input portion power factor is mainly aimed at improving the effect that
the grid c or the sources c because that we have put is non-linear rectifier capacitor filter
load. So, if we look at the waveforms there this is the x-axis kind t and you put in the
voltage sinusoidal voltage curve and the current i in that you would see here when the
circuit is active. So, this is Vi and this this is iin.
Now, let us try to see how the power curve looks like. So, let us draw bigger version of
that and let us put in the voltage wave shape, I am putting it dotted indicating sampling
and then let us put in the current wave shape for one cycle. Now, let us calculate the
322
instantaneous power. Instantaneous power is found by multiplying instantaneous value of
current with the instantaneous value of voltage at that time instant.
So, let us say here power is 0 which means what it basically means is that the current
from here to here is 0. So, the current into whatever the voltage will always be 0, so the
power value dots that I am putting here is the instantaneous power which is 0. From here
one current condensed current increases, so there is a finite value of current into the
voltage so that keeps coming up there. So, you have the power values and then again it
become 0 here all up to this point when current begins to flow both current and voltage
are negative here, power v into i will be positive. So, it is of this shape and 0 again. And
if I can I call the dots this will be the power curve.
Now, this power curve is actually occurring at a very narrow period in time and at other
times it is 0 and to the load it supplies an average value. So, let us say this is the
instantaneous power curve, and the average of that is the average value deliver to the
load here v naught, i naught and that will be naught or P load. Now, if you take the peak
value of this this is let say P i max which is v i max or v max here into imax here. Now,
that will be the peak power that is drawn from the source. Now, imagine let us say for an
example the average value of the load here is 100, this is 100 and this Pimax here is
1000.
So, for delivering 100 watts of power to the load here you will have to draw 1000 watts
at some instant in time from the input mains. What it means is that the input supply and
all the upstream, equipments, cables, connectors, switches, plug, sockets all will have to
be rated for the high power rating that you would be drawing, only for a short period of
time so as to supply the low average load. So, which means that you are under utilising
the supply side equipments on the supply. The power factor provides a measure of this
utilisation.
323
Prof. L. Umanand
Department of Electronic Systems Engineering
Lecture – 26
Power factor – discussion
Now, consider this simple circuit. You have a voltage source, and the voltage source is
connected across a resistance R and a current i flows through it. So, this is a resistive
load. Now, for a sinusoidal voltage waveshape V what is the current i waveshape under
steady state conditions?
324
So, you see that the voltage waveshape is like this sinusoidal and importantly the current
waveshape is in phase with the voltage waveshape as seen here.
Now, consider a pure inductance connected across the same voltage source. Now, for
this pure inductive load and also for a sinusoidal voltage waveshape Vm sinωt what is
the current waveshape expected under steady state conditions.
325
So, you will see that the current i in this case is delayed by pi by 2 or 90 degrees with
respect to the voltage waveshape.
Likewise, if you consider a pure capacitive load connected across a sinusoidal voltage
wave shape and then for a sinusoidal voltage waveshape V m sin omega t what is the
current that you will expect under steady state conditions?
326
Here in this case you will see that the current is leading the voltage waveshape as against
the inductive load where the current was lagging by 90 degrees. So, the current here is
leading by 90 degrees.
Putting them all together comparatively you will expect visually on the oscilloscope for a
resistive load V and i like this for a resistive load, V and i like this for an inductive load,
and this is for a capacitive load.
327
Now, take the case of the resistive load waveshape. You have voltage and current in this
fashion. Now, let us say you would like to plot the instantaneous power. So, when you
plot the instantaneous power which is instantaneous value of V into instantaneous value
of i, instantaneous value of V into instantaneous value of i, instantaneous value of V into
instantaneous value of i, at same time locations you will get this power curve as shown
here.
Observe that at this point here power is 0 instantaneous value of power is 0 because V 0 i
0, and here too power is 0 because V 0 i 0 and at around this region power peaks because
V m and i m both are at max value. And observe that the power curve is having double
the frequency of the voltage or current frequency. And another thing is that the power
curve is having an average value is shifted up. So, if you now, only consider as this
power curve and remove all the other waveforms.
You will see something like this, this is the power curve and there is an average value P
average. So, this is a visualization that you should expect for any resistive load where in
the source is a sinusoidal source.
For an inductive load where the waveshapes are like this, voltage waveshape, and current
waveshape, multiply instantaneous values of the current and instantaneous value of the
voltage at the same time points you will get the instantaneous power curve like this, w
waveform. Observe that this is having a 0 average, if you remove out the voltage and the
328
current waveshapes you will get only the power waveshape for the inductive load which
is like this, the average value is 0 whereas, the instantaneous value is double the sin
frequency as the fundamental.
In a similar manner if you look at the capacitive load reaction waveshape waveforms
voltage a leading current waveform like this. If you multiply the instantaneous values of
the current and the voltages you will get the power curve in this fashion. This is the p is
equal to vi instantaneous power curve. Here again if you remove the voltage and the
current waveforms becomes clear that the average power is 0.
329
Comparing all the three cases, you should visualize that on an any oscillogram for a few
resistive load this will be your power curve and the average power.
This would be the power curve for an inductive average power is equal to 0, average
power is equal to 0 for the capacitive pure capacitive load also.
330
Now, consider the same voltage source Vm sinωt applied across R and L, hybrid
impedance consisting of resistance and inductance. Now, the voltage waveshape is
shown like this. The current waveshape can vary with different lags. It can be 0 when it
is purely resistive and as the value of resistance keeps decreasing and inductance
increases you will see the phase shift starts moving like as towards when purely
inductive.
331
So, if you look at the power curve for the case when R is equal to 0 ohms it will look like
an inductive visualization, average power is 0 and this is the instantaneous power curve.
Now, keep increasing the value of R then it becomes an R L load.
So, as you start increasing the value of R you will see that the entire power waveshape
shifts up and there is an average power available and that is active power or the real
power that gets dissipated because of the resistor R.
332
So, as it starts moving up at a point you will see that as their value of inductance
becomes 0 the R L load tend towards the resistive load and you will see a waveshape like
this, where the average is given by V m i m by 2.
Likewise, for the R C load also where you are having R and C connected except for the
phase shift which changes from lag to lead.
You will see that the as the value of R increases the value of C decreases, you will see
that it starts moving in this fashion going more towards the pure resistive type of load.
333
Some remarks that we can make at this point in our discussion, is that the average power
is maximum only for pure resistive load. For any other type of load any other
combination of the loads the maximum power is less than Vm im/2.
Let us now discuss how to measure power factor. We would like to measure power
factor only using the waveshapes as seen on the oscilloscope, the voltage and current
waveshapes.
334
Let us begin with sinusoidal waveshapes, sinusoidal voltage and current and then move
on to the non-linear waveshapes of rectifier capacitor filter circuit. Now, on this graph let
us put in the voltage sinusoidal voltage waveshape and let us take an arbitrary delayed
current waveshape. So, there is an arbitrary phase delay of phi which means it is an R L
load
Now, this when you multiply the instantaneous values of the voltages and the currents
you would get the instantaneous power curve and this would look like this. Observe that
the power goes negative here. This is because the voltage is positive in this region
current is negative in this region the power is v into i and therefore, it has to go a
negative. At this point voltage is 0 current is negative at this point current is 0 voltage is
positive. So, therefore, the power crosses the 0 point at this point.
Likewise, here also voltage is 0 at this point power is 0 then voltage goes negative here,
current is positive and therefore, the power negative in this region till this point when
both current and voltage goes negative and from here onwards power becomes positive.
So, this is the power curve having double the frequency the voltage or the current. So,
this power curve will have an average value like this as shown. Now, this negative part
which I have showing shaded here this is the negative power which means that it is put
back into the source. It is taken from the source and then put back into the source and the
positive part is pushed into the output. So, the positive parts of this would be the active
335
power pushed into the output and this negative part is the part the reactive part which is
put back.
So, this is the power curve showing the positive and negative aspects, the active and the
reactive part.
Now, if you compare with an R-L load, R and an R-L load you will see that the current
waveform here is in phase and that is having a phase shift of phi, and the power
336
instantaneous power curve here and the instantaneous power curve for the R-L load is
like this.
Now, the averages. Look at the average the average for the R load is much higher than
the average for the R-L load because of this negative portion which get subtracted from
the area when you are integrating.
Let us remove the voltage and current waveforms (Refer Time: 12:33) that becomes
clear. Let us move this waveform arc. I am going to compare this R-L load with respect
to a pure resistor which implies that by comparison I am going to get the power factor
which is basically telling which is giving me a measure of this R-L load how close it is to
a pure resistance
So, let me call this as P R the average of a resistive load pure resistive load and this we
can call it as P load this is the R-L load. So, now, what is power factor for this load? We
define that power factor is a measure of the degree to which this R-L load matches to that
of the pure resistance.
337
Power factor let us say we define it as P load divided by the reference pure resistive load
power which is P R, all else being same the amplitude of the current and amplitude of the
voltage.
So, this is the definition that we will be using and which will give you the measure of the
closeness of any given load to that of a pure resistance. When PF is 1 that is P load and P
R are same, means these two values are same. Then we say, it is a resistive load or the
the load is equivalently same as that of the resistive load when PF is 0 that is when P
load is 0 then the load is farthest away from being a resistor it is a reactive load.
338
So, now this definition
PF = Pload / PR
Pload is the average power of any given load and this can be measured from the voltage
and current waveforms of the oscilloscopic itself. PR is the average power of a resistive
load for the same voltage and current magnitudes this is the important clause there. So,
as power is an average quantity, so this measure definition should be valid for any
voltage and current wave shape.
339
Prof. L. Umanand
Lecture - 27
Power factor – sinusoidal
Let us now try to find out the power factor for sinusoidal voltages and currents. The
power factor of sinusoidal voltages and currents are already defined, well defined in the
literature. Let us apply this measure definition and see if we get the same value and then
we shall apply it to the wave shapes of the rectifier capacitor filter voltage and currents.
340
Consider an oscillogram x axis is time and let us place a voltage waveform on this is a
voltage waveform and it has a peak amplitude of V m and superimpose a current
waveform with an arbitrary delay meaning an R-L load and that delay is phi. And let the
current have a maximum value of I m. So, the voltage wave shape will have the equation
Vm sinωt, the current wave shape will be having I m sin omega t minus phi.
So, if you measure, now let us say this is the unknown load and you have measured
something like this on the oscilloscope. How do you compare it with the pure resistive
341
load? So, let us say we would like to bench mark it with respect to pure resistive load for
the same magnitude of Vm Im. So, let us draw the voltage having the same magnitude
Vm, then draw the current in the case of a pure R-L load the current has to be in phase
with the voltage, now retain the same magnitude as Im. And you have this picture that
you would have got if it had been pure resistive load.
Now, let us find out the powers the instantaneous power curve we know is like this.
There is some negative portion here for the R-L load and for the R load there are no
negative, there is no negative portions. This is the average P R, this is the average P load
for the unknown R-L load.
Let us move it up. Observe this negative portion this is the reactive portion and P load
what goes to the load is the active portion. Now, this is the reference resistive load. Now,
this value is given as Vm Im / 2 this for a pure resistive load. Now, what is this Pload?
We have the reactive power component; Pload is the active power component.
342
And P load is less than P R, it is less than Vm Im / 2 because of this negative part.
And if it is a sinusoidal case we know that for sine waves Pload is (Vm Im / 2) cos phi
where this is phi. So, if we now apply our definition of PF is P load by P R any wave
shape and for this special case of a sine wave you will get cos phi. So, it verifies that the
definition that we have made as a measure of closeness to a pure resistive load any load,
any unknown load, measure the closes to a resistive load will provide you the power
factor value.
343
Prof. L. Umanand
Lecture – 28
Power factor for rectifier cap filter
Let us now find the power factor of a rectifier capacitor filter circuit, only by looking at
or only by using the measured voltage and the current wave shape on an oscilloscope.
344
So, let us say we have an oscillogram with the x axis, the time axis and let us put a
voltage wave shape. The voltage the input voltage to the rectifier capacitor filter is a sine
wave because we are taking it from the AC source and it has a peak maximum amplitude
of Vm. Superpose the typical capacitor, rectifier capacitor filter input current wave shape
which is like that and this is familiar to us, we have this here.
Now, the peak value, what is the peak value? This is the peak value Im, ok. Now, if you
want to compare it with the resistor a pure resistor, a pure resistor will be having the
same wave shape as the voltage and you will adjust the resistor value such that the peak
of that current should be same as this Im, like this.
So, we have two sides, one is the unknown rectifier capacitor filter load the other one is
the benchmarking R load, pure R load. Voltage wave shape Vm, and voltage wave shape
Vm same for both R. The capacitor filter load measured on the oscilloscope gives an I m
which is this value, use that same Im and draw the resistor current wave shape like this
having the same Im magnitude and having the same wave shape as that of the voltage,
then it is a pure resistance.
345
Now, the power, multiply the instantaneous values of the current and the instantaneous
value of the voltage for a given time point and place the power points. So, for the
resistive load we have seen that this is the power curve instantaneous power curve and
for the rectifier C-filter this is the power curve because the currents do not flow on any
other periods of time so power is 0 the power is going to flow only during the period of
the time when the diodes are conducting and there is a current flow.
Now, compute the average the average in the case of the C-filters P load or P naught or
V naught into I naught, DC output is DC and therefore, easy to compute P load this is the
average. In the case of the R load we know that P R is Vm Im / 2.
346
So, let us remove the voltage and current and retain only the power wave shapes with the
averages move up the wave shape a way bit and then we know this is Vm Im / 2. How do
you find P load? As I told you P load can be found from the output wave shape the
capacitor filter, V naught is a DC, I naught is a DC, V naught into I naught will be the
average power or you could on the oscilloscope compute the instantaneous; if you want
to include the ripple if the ripple is not negligible. Compute the instantaneous value of I
and P, I and V and find the value of P and then average it. I will I will just let you know
how to do that just few moments later.
So, we can find that of course, by measurement on the oscilloscope. So, P load can be
found out. And once you know that combine these two
PF = Pload / PR = 2 Pload / Vm Im
So, this would be the power factor equation for the rectifier C-filter.
347
Let me now briefly tell you how to measure average power. So, that you can measure P
naught or P load from the oscillogram and use it for computing the power factor.
Now, consider that this is an oscillogram you have the voltage wave shape and you have
the current wave shape. The inputs of your diode rectifier capacitor filter circuit. Now,
you mark points tick points tick marks along the x axis the time axis and number each of
the tick point 0, 1, 2, 3, so on till you have n. Now, at any point k at the k’th tick point if
you take this point on the voltage curve, you will get them value of voltage and that point
kth point we call it as v k and if you measure the value of current I at that point it will be
Ik at time instant k. Like that you have to measure for every time instant and tabulate.
So, let us make a tabulation.
Let us say you have one column for k, one column for v(k), one column for i(k) another
column for p(k), where p(k) is the product of these two columns, v(k) into i(k) this is the
instantaneous power. So, for the first point correspondingly take the i and the v value, v1
i1, in this case i1 would be 0 p(k), second value v2 i2 and you will get p2 so on, for the
nth value vn into vn in and pn which is vn into in, then take the sigma of the power
columns.
The P average or P naught or P load is equal to sigma of all this sum it up sigma of Pk, k
is equal to 1 to n divide by number of samples which is n. This will give you the average
power directly from measured wave shape on the oscilloscope. This P load value you
348
plug it into the power factor formula equation to get the power factor for the rectifier
capacitor filter circuit.
349
Prof. L. Umanand
Lecture – 29
Passive power improvement circuit
Let us look at Power Factor Improvement. Now, we will look at passive power factor
improvement, later on much later after we discuss the switched mode power converters
we will also discuss active power factor improvement. For now, let us improve the
power factor by using passive components. Let us, to do that let us get a hang of what we
should do.
We have the input voltage wave shape sinusoidal like this and the current wave shape for
the rectifier capacitor filter circuit is in this fashion. This will deliver an average load
power like this the red line. So, this would be the average power delivered to the load
and we can call that as P naught or P load.
Now, to deliver the same average power the same P naught if we had a pure resistive
load how will the current look like? Resistive load the current will be like this. Now,
observe that the peak is much reduced compared to this, that is because for the same load
if I had a resistive a resistance load you will see that the current is conducting for the
whole cycle because the current has to be in phase with the voltage and have the same
350
wave shape. And vm im / 2 is the power that is delivered to the load and therefore, I m is
fixed by that and you will see that the im is much lower than what you would get for the
case of the rectifier capacitor filter for this particular Pnaught.
The take away from here is that as I widen the base; as I widen the base of conduction
the peak will start reducing. Now, that is the take away that we are going to use for
improving the power factor.
To consolidate this, I am going to show 3 possibilities on the graph. Let me draw the
rectified input source and let me take the output voltage ripple in this fashion. Call this as
case 1. Now, case 2, again let me have the input voltage wave shape and let me consider
the output voltage ripple in this fashion. This is a slightly lesser ripple than this case and
call that as case 2. And the 3rd case I would like to give more ripple than previous two
and let us say that the output ripple is in this fashion larger ripple.
Now, let me mark the current conduction time. This is the current conduction time for
case 1, correct conduction current conduction time for case 2 and this is the current wave
shape for the rectifier capacitor filter circuit. Now, if you look at the current conduction
for case 2, it is now narrower because the ripple is smaller the conduction period is much
narrower. Now, let us say in case 1 some amount of power is being delivered and let us
say this is P load, consider for the same power average power delivered to the load you
351
will find that the current conduction as it is narrowed in order to keep the same area you
will have much higher value of peak, current peak will be much larger compared to this.
Likewise, in case 3 you will see that for the same average power we have increased
consciously the current conduction time and the current peaks will be much lower. So,
this is actually the concept that we will be using in passive power factor improvement.
We will be trying to increase the current conduction time at the expense of ripple, more
ripple and thereby reduce the current peaks and improve the power factor, making it
closer and closer to the full conduction as in the case of a pure resistive load.
So, that is what we would be like doing. I will take up an example passive power factor
improvement circuit for rectifier capacitor filter which will give conduction for a much
wider range and probably simulate that to understand its operation.
Let us now draw a passive power factor improved circuit. Let me put the source and this
diode bridge we are familiar with this diode bridge then connect the source through a
switch like this and across the bridge in this fashion. Now, this portion we know this P
sine wave source followed by a rectifier and let me draw the load R naught first. So, this
is the R naught that we connect across the rectifier capacitor.
Normally, we would have connected a capacitor like this; however, we want to modify
this into a improved power factor improvement configuration. I will now split the
352
capacitor into 2 parts into this fashion. So, you see that now I have 2 capacitors in series.
So, when you connect this when the capacitors are getting charged it will charge up in
this fashion.
However, there is no discharge path into the load because of the presence of this diode,
there is a reason we will see that. Charging is happening in this path, half the charge will
come into this capacitor, half the charge into this capacitance, so if you are having v
naught each will take vnaught / 2 and v naught/2. Now, when you want to discharge it
discharge each capacitor separately. So, provide the discharge path, you put a diode in
this fashion. So, this diode will give you chance for this capacitor to discharge through
the resistor in this fashion and if I put a diode here you will see that this capacitor can
independently discharge through the load in this fashion.
So, while charging you will see that the charge path is only in this way, the charge is
divided equally between the 2 capacitors if the 2 capacitors are equal and the 2 capacitors
will independently discharge into the load. So, the diode steering network takes care of
this issue. So, let us call this as C1 this is C2, x, y and you have v i and v naught. So, let
us look at the wave shapes at various nodes to understand this a bit better.
Let me draw the waveforms. Let me put in the rectified waveform at this point if it was
only a resistance connected. Let me draw two more time x axis, one will be vy, vy is the
voltage across C2. Note that voltage across C2 and voltage across C1 will be exactly
same if C1 and C2 are same values. Then vx is the voltage across this diode, voltage
across this diode will be same as voltage across this diode. So, you need to see just one
of these two waveforms. Now, I will paste these images here. Now, when you switch this
on the output voltage here will track, on the first cycle it will track the input wave shape
like this if I say this is the vm/2 line and this is also vm/2 line. So, this v m this will be
vm/2.
V y, I told you that when you switch this on the charging path is through C1 and C2 and
C1 and C2 will equally distribute the charge. So, if you look at the voltage across v y it
will go to the halfway mark and stop here. Likewise, C1 also will reach the halfway
mark and stop there. Now, once it is charged fully here, then the input voltage would
have gone low, these diodes would be reverse biased under the normal condition if it was
a normal circuit. But now, let us say it gets reverse biased reverse condition, then this
353
capacitor starts discharging during which time the diode conducts the capacitor is
charged only half the voltage vm. So, this node potential v naught will go to vm/2.
So, if it is vm/2 again the input will input will charge up the load the diodes will conduct
because the input is still higher than vm/2. So, till the point vm/2 is reached the input
output will track the input. So, let me draw that vm/2 point. So, till it reaches this point
the output will track the input and till that time voltage across the capacitor and voltage
of across C2 both will be fixed constant because there is no discharge path for the
capacitances. The load is serviced directly by the input during this time.
Then further on at this point when the input goes below this halfway mark vm/2, these
diodes are out of the picture and the capacitance is only servicing the load. So, this is
where the capacitor discharge happens to the load or not. Each of the capacitance C1 will
discharge in this fashion, C2 will discharge in this fashion. Till it reaches this point when
input means crosses this and this diodes will conduct and again the cycle will repeat. So,
this will conduct here and once it starts conducting and once the C 1 and C 2 charges are
replenished it will go and stay at the v m by 2 level and this continuous in this fashion.
Now, if you look at the envelope of the voltage output voltage it in the first cycle it goes
in this fashion up to the point it crosses over and comes to the vm/2 point it is tracking
the input. Now, at this point these diodes switch off and the output is just the capacitor
voltage which are discharging like this and then at this point again it will start tracking it
will start tracking the input and comes down and then again here you will see it will track
the capacitor voltages. Now, this will be the output voltage envelope look at the very
large ripple. But look at the conduction time the conduction time is all this time the
current conduction time.
So, this is v naught. So, let me draw these critical marker points like that. So, during this
time only the current is 0 here this current i is 0 here because at that time the capacitor is
discharging and during this time you will have the current this first cycle charge it is like
this, the second cycle onwards you will see that it will just be in this fashion. Now, you
see that the conduction period of this current has been widened so much and therefore,
the peak would be significantly lower compared to the normal capacitor rectifier filter
capacitor circuit.
354
Now, when you come to the v x waveform here, the v x will have an envelope. Now, this
is a constant fixed, let us say this is a constant fixed vm/2 this will be whatever the value
vnaught here, vnaught minus vm/2 will appear across this one. Vnaught is this waveform
minus this vm/2. So, I am just taking it from this point here let us say. So, it is this
envelope that v x will have. And this point will be the 0 point because it is during this
time this diode is conducting. When the capacitor is discharging during this time the
diode is conducting and then you will have a horizontal. So, this is the type of waveform
that you will see for v x and also across this diodes.
355
Prof. L. Umanand
Lecture – 30
Simulation – power factor improvement
Let us now have a look at the Simulation of the Passive Power Factor Improvement
Circuit. Observe that all parts of the rectifier filter circuit is same, except for this part that
we discuss the power factor improvement part. You have the capacitor 1 and capacitor 2,
you have this central diode and then you have to reverse diodes like this. So, when the
capacitors charging you will see the path go in this fashion, and when the capacitor
discharges it will discharge through this diode and resistor and this diode capacitor will
capacitor Cf 1 will discharge through the load resistor.
Now, I have a node here x 1 I have another node here y and this is the output node. Let
us look at the current through this inductor which will give you the current before the
power factor improvement portion and let us verify if the way from that we have
understood is correct and whether we can gain more insight into the power factor
improvement portion. So, now to simulate we need to go to the terminal, but before we
do that I want to make some changes in the rectifier dot cir file to see what we would like
to measure and let me change that open it and show it to you.
356
So, rectifier dot cir let us open that and see how it looks. These two statements are
familiar to you the Tran upon other the Tran directive. And here you have the control
statements to operate within the ngspice environment. I am setting the background as
white, foreground as black, run the simulation and then plot. So, what am I plotting? I
am plotting v naught, that is this node potential this node potential then v x, v y and the
current through L; v x, v y and the current through Ld. So, this is what we will do. So, let
us close open the terminal and see how the simulation goes through.
357
I now have the terminal open. Let us run the simulation runsim.
So, it will generate the net list, after this net list is generated portrait into ngspice
environment and we have the plots. So, if you look at the plots, observe this red
waveform is v naught, so as we said that v naught has around 50 percent ripple. Now, the
blue waveform is the v x waveform, again the v x waveform similar to the one that we
had discussed going up to v x is the voltage waveform across the diode.
Now, look at this v y here, the voltage across the capacitor. See that after it has followed
the input and then come to the 50 percent mark v m by 2, it then holds there till the input
voltage comes below the v m by 2 point when the rectifier diodes stop operating and the
capacitor is discharging into the load. Each of the capacitor discharging into the load, till
again the input comes up and then crosses over here the rectifier diode comes into
operation and then you have the output following the input.
Now, during this portion you will see that there is current drawn; the current drawn is
having 2 parts. This part is actually, this part the smooth part is actually the current direct
being given directly to the load and this part is now this is where the capacitor gets
recharged. So, this capacitor recharge point gets recharged at this, at this point and then
you see that small spike coming in. We did not discuss this point while calculating while
discussing the waveform in theory.
358
Just that it would confuse the issue, but you can explore much more while you are having
the simulation in front of you and you can explore the various branch currents on a node
potentials at great detail. So, I will enclose this circuit also as a resource for you, so that
you can work on it and try to understand understand and get as much in site as you can
about this circuit.
The basic take away is that we have introduced a current wave shape with a larger
conduction angle and thereby reducing significantly the peak current and thereby
improving the power factor. But all these at the expense of an output voltage ripple
which is very high. This is generally not a problem because the output of this is generally
given to DC DC converters or DC AC inverters and they have a very good closed loop
control mechanisms which will handle fluctuations in the DC link as provided by this
power factor improvement circuit. So, this fluctuation this very high ripple is generally
not a problem, but that is disadvantage of this particular circuit.
359
Prof. L. Umanand
Lecture – 31
Linear regulators - intro
We shall now discuss DC-DC converters. We have seen how to get DC from AC using
rectifier capacitor filter, so that DC is unregulated. So, we need to pass it through a DC-
DC converter to get a regulated DC output at the output of the DC-DC converter. There
are 2 methods broadly in which we can achieve this; one method is by linear circuits, the
other method is by using switched mode circuits. We shall discuss the switched mode
circuits later. Today we will take up DC-DC conversion using linear circuits, also known
as the linear regulators.
When you compare the functional performance of the linear regulator switched mode
regulators, the linear regulators are very superior in terms of line regulation, in terms of
load regulation, in terms of temperature regulation and all aspects of functional
performance. The only issue which is a disadvantage in the linear regulator is efficiency,
it is very lossy. It cannot match the switched mode power converters with respect to
efficiency as a consequence the linear regulators are bulkier and costlier.
360
However, linear regulators play a very important role in power electronics and in power
components and in equipments especially in applications like power supplies which,
where in precision output specks are needed to be provided to many of the critical and
sensitive loads. So, today we will discuss about the linear regulators.
The typical application for the linear regulators is power supplies. So, most of the power
supplies for equipments laboratory power supplies they are based on the linear regulators
especially if your requirement is for a very highly stringent output spec. We have seen
how to get DC from unregulated DC from AC source. So, if you see we have used a
transformer based rectifier diode rectifier capacitor filter like this, we have used this kind
of circuit and we have used transformer for inrush current limiting. We discuss this.
They not only does this transformer do inrush current limiting, it also provides galvanic
isolation between the AC side and the load side. And further it also provide scaling you
can now step down. So, you have 230 volts on this side 325 volts peak and here you may
want a 15 volt power supply, the transformer will do the job of scaling. And especially
for low power systems low power supplies the choice of using the transformer will be a
good one especially from the point of reliability and functional spec.
So, here you have unregulated DC. Now, this is given to a block and that will be the DC-
DC regulator block or the linear regulator block and the output of that now here you have
regulated DC that will be given to a load in this fashion. So, all this block is the input to
the DC-DC linear regulator and this portion is the output load v naught. Now, all this
portion the AC to DC portion we can replace it by this symbol here indicating that that is
Vi as far as the DC-DC linear regulator is concerned. So, this the unregulated voltage
that you get at this point we can name it as Vi and henceforth when we are discussing
DC-DC linear regulator. We will not be writing all this instead we will replace it just by
this battery symbol at this point.
361
Henceforth, we will be using this symbol to represent either a battery or an unregulated

DC obtained from an AC to DC rectifier capacitor filter circuit. Now, this is given to the
linear regulator and then further to the load and this will be the regulated the output.
So, this linear regulator the design the operation and the design of that is what we will be
looking at. We will call those as Vi for the linear regulator and this is the Vo. There are 2
methods in which to achieve the linear regulation; one is called shunt regulator and the
other is called the series regulator. Both are in use both are quite popular for different
applications. Generally, the shunt regulator is used for references, voltage references
much lower power compared to the series regulator which are used for most of the power
supplies. We will look at both and we will look at the circuits that govern each of these
principles.
362
Prof. L. Umanand
Lecture – 32
Shunt regulator
Let us now discuss the Shunt regulator circuit. So, let us draw the input unregulated
voltage Vi, it contains a series component here and a shunt component here, it’s a Zener
diode and the load. So, this is the structure of a shunt regulator. The hero of this is this
Zener diode. This is Vi, the unregulated Vi, we will call this as Rs, this is Vz and this is
the load resistance Ro and you have Vo and Vo is equal to Vz.
So, if I name this as Is or Iin and this is Iz and Io. So, these are the various parameters in a
shunt regulator. How does the regulation happen? Basically, Vz is a reference Zener, the
voltage across the Zener is constant and it tries to maintain Vo constant in spite of the
fluctuations in Vi or fluctuations in Ro. Let us see how the operation, how the regulation
operation functions.
Now, look at the Kirchhoff’s current law here at this node. You have
Iin = Iz + Io
363
Now, V i can vary from a Vi,min value to a Vi,max value and Ro is not under our
control again can vary from Ro,min value to Ro,max value and as a consequence when
Ro is minimum Io is maximum and Ro is max Io is minimum.
The operation of the shunt regulator hinges on this device which is the Zener diode. Let
us try to understand the characteristics of that Zener diode static characteristics of Zener
diode and how it affects the regulation performance here. So, let me draw the i v
characteristics, this is the v, the i and let me take only the reverse characteristic, Zener
reverse characteristic which will be something like this, where this line intercept will be
Vz on the voltage.
So now
Iin = (Vi – Vz)/Rs
So, this is that current flows in through here and this current is independent of the value
of Ro or value of Io. So, depending upon the value of Ro there is commutation between
Io and I z to maintain I in fix to this value.
So, the current through the Zener has flexibility. It can operate from a max value I z max
to a min value I z min and this in fact, gives you the regulation. If suppose Io is 0 then
the whole of I in current flows into I z, so the operating point will be somewhere here. If
Io is a maximum value then I z will decrease and it will probably be sitting at min value
the operating point will be here. So, it is I z here that actually is adjusting its value. The
regulation happens because of the capability of the Zener to take varying currents
through through it at a constant voltage. The Zener current can vary from I z min to I z
max depending upon the value of the load.
364
Consider now few operational cases. Let us say case 1. Now, here V i is at its max value
V i max and let us say Io is 0 which means this open circuited, then the entire I in current
flows through the Zener. So, I z will be equal to I in because Io is 0 and that is V i max
minus Vz divided by R s. Now, this is the max value of the current that can flow into the
Zener in this circuit. So, it will get, this will be the operating point it will get positioned
here and that will be I z max for the circuit and see to it that it is less than or equal to P z
by Vz, P z is this Zener’s power rating.
So, when you go and choose a select a Zener it will have a power rating in the data sheet.
So, you will have a 1 watt Zener, you have 400 milli watt Zener. So, whatever that power
rating divided by Vz Zener Vz will be the limiting or the max permissible value of I z.
So, this I z value which you calculate here should be less than that so that the Zener does
not blow.
The second case, the second limiting case happens when V i is V i min and Io is
maximum, that is when Ro is at a minimum value. So, you have Io max. Under this
condition this Io plus I z should be equal to I in, there should be a at least the minimum
value of I z to pass through the Zener in order that the Zener provides a constant voltage
Vz.
365
Let us now see what is the efficiency of this circuit? Efficiency is defined as
η = Po/Pi = VoIo / Vin Iin
Let us express I in in terms of input voltage and output voltage.
So, now therefore, we can rewrite efficiency as Vo Io Rs divided by V i into delta V io.
So, this will be the efficiency relationship. Observe that as delta V io input output
voltage differential increases the efficiency will decrease, which basically means that if I
am having a linear shunt regulator let us say with input output differential as some
voltage. For example, if I take Vi of 10 volts, Vo of 5 volts input output differential
would be 10 minus 5, 5 volts. For the same 5 if I am having V i of 15 volt then input
output differential would be 10. So, the 15 volt input would have a lower efficiency than
that 10 volt input linear regulator.
So, Vo Io is nothing but Po which is the which is which can be obtained from this spec,
R s is calculated by design, Vi is given from the input spec, delta Vio is also an input
spec and from here we can calculate the efficiency of the circuit. So, the important thing
to note here is the input output differential voltage, higher rate is lower the efficiency.
366
Prof. L. Umanand
Lecture – 33
Example on shunt regulator
Let us now look at an example of this shunt regulator circuit and try to gain more insight
into this by putting in some numbers. So, let us say that V i the unregulated voltage V i
varies from 9 volts to 13 volts, this is the spec. So, this is unregulated V i varies from 9
to 13 volts. Then this Zener, let us say the Zener is the 5 volt Zener, because we want 5
volt at the output so we need to take a 5 volt Zener and let us say the Zener has a spec of
400 milliwatts. What is the range of output Ro? So, let us try to work it out and find out.
So, let us say I z, I z is the current. What is Iz,max? Iz,max should be such that the max
power dissipation of the Zener 400 milliwatts is not exceeded. So, P z if I take P z 400
milliwatts divided by Vz, Vz into I z is P z, which is 400 milliwatts by 5 volts this will
give you 80 milliamps. So, Iz,max or I z the current through this branch should not
exceed 80 milliamps, if it exceeds 80 milliamps then the power dissipation limit of this
Zener is crossed and it will blow.
So, this is one limiting value that you would get and if you actually plot it on the Zener
reverse characteristic, as this is a Zener’s reverse characteristic this is 5 volts here and let
367
us say this is the Iz,max operating point this is Iz,max and that is 80 milliamps as
calculated like this here.
So, now, R s let us try to find what is this R s. V i max take V i max which is 13 volts
minus this node potential which is Vz divided by I in. Now, I in I will take it as Iz,max
assuming that I naught is 0. This is one of the extreme conditions that we saw earlier. So,
that would be 13 minus 5 divided by 80 milliamps and which is 100 ohms.
So, what it means is that under worst case condition when there is no load and Ro is
infinity all the current I in flows into I z, and that would be Iz,max. Now, that current I in
is maximum when V i is max, so V i max minus Vz by Iz,max will give you the value of
R s and what should be the power rating of R s. So, the power rating of R s is given by V
i max minus Vz that is voltage potential difference across R s square by the value of R s
which is 13 minus 5 square by R s which is 100 and this turns out to be 0.64 Watts. So,
we have selected R s.
Next let us look at Ro. So, first what is the maximum value of Ro I naught that can be
allowed through Ro? So, I naught max is even under minimum input voltage condition V
i min minus Vz divided by R s will be the I in that flows here under minimum input
condition. And at that condition what is the minimum Ro that you can apply. 9 minus 5
divided by 100 and that is 40 milliamps, that is the maximum I naught that you can allow
through the output and therefore, Ro min if it is max I naught you write down the Ro min
which is Vz V naught potential is Vz divided by I naught max which is 5 volts divided
by 40 milliamps this one and that comes to 125 ohms.
So, therefore the range of Ro is 125 ohms on one side, on the other side it is open circuit
or infinite. So, this implies that Ro can take a minimum value of 125 ohms and a
maximum value of open circuit. So, this is the range of Ro for this particular circuit for
this kind of specification.
368
What is the efficiency? You can calculate the efficiency of this to complete the
understanding. So, efficiency is we have found out that V naught into R s by V i into
delta V io or the delta input output differential.
Now, let us say efficiency max. What is the max efficiency? P naught is max. So, we will
make P naught max, R s ,R s is a fixed value calculated for this particular circuit, V i min
take V i min and delta V io. When you take V i min delta V io will naturally be smaller
minimum of and therefore the efficiency would be the max possible that you would get.
So, if you calculate V naught max which is V naught, 5 volts into I naught max, 40
milliamp into R s which is 100 ohms divided by V i min which is 9 volts. Delta V io,
when it is 9 volts and the output is 5 volts V io will be 9 minus 5, 4 volts and this turns
out to be 0.55 or 55 percent. So, the efficiency of this circuit for these specifications is 55
percent at max load condition.
369
Prof. L. Umanand
Lecture - 34
Non-ideality and solution
Let us now consider and discuss the same shunt regulator, but a practical one with a non-
ideal Zener. Till now in the discussions we had used Zener which was more or less ideal,
wherein the Zener did not have any series resistance or the series resistance was 0 in the
breakdown region. Consider the static characteristic of the Zener to the forward, but let
us consider the reverse characteristic.
We have till now considered the reverse characteristics which is going vertical down at
V z. But a more practical Zener will have a follow a slope like this and the slope of that
is 1 by Rz or the series impedance or the series resistance of the Zener. So, this can
equivalently be shown symbolically in this fashion where you have the Zener with an
ideal Zener, with an ideal zener having a V z reference. And in series you have a small
resistance Rz. Now, this is the practical Zener. Let us try to include that in the shunt
regulator circuits. So, you have the V i, you have a R s and you have this Zener and
connected across the load as shown like this here. So, this is V z just like here and this
will be the Zener series resistance.
370
Now, this whole thing together this block is the Zener device and we can say this whole
thing is actually the Zener, practical Zener that you would have placed at this point in the
circuit. So, let us mark V i this is Vo, R naught, R s and this is I in and this is Iz current
flowing through this branch, this is Io current flowing through the output branch. Now,
what is Vo? Vo before it was equivalent to just V z when Rz was 0, but now it is V z
plus Vo is equal to V z and in the earlier case when you use the ideal Zener Rz was 0 and
Vo was equal to was equal to V z.
Now, you have an additional term due to the drop across Rz, so that drop is Iz into Rz
which is Iz into Rz. Now, this is the extra portion due to the non-ideality and this is the
cause for deregulation or an improper regulation or loss of regulation. So, this results in
the loss of regulation. Now, if you want to improve the regulation for the circuit you
have to reduce this term. So, reduce Iz, Rz term to improve this regulator.
In order to improve the regulation of the practical shunt regulator we now modify it and
see how the modified shunt regulator behaves and improves the regulation. Now, this is
the regular practical shunt regulator, let us do some modifications, let me push the low to
the right, let me join this and here let me include a BJT, an npn BJT. And connect the
emitter here to the load and I will connect the collector to the supply and the base is
connected to this point this junction point.
371
So, if this is Q the emitter current is now Io. So, the base current will be Io by beta plus
1, where beta is the collector to base ratio I c by I b. So, you have Io by beta plus 1, I
mean in here this is V be base to emitter voltage is around 0.6 volts. So, now, what is
Vo? Vo here is equal to V z plus Iz rz, then this V be drop. So, plus and minus in this
fashion therefore, minus V be is what will come to the output here.
Now, here look at this Iz before without putting this transistor BJT the entire Io was
being commutated by Iz. Iz was used to commutate the entire Io, but now after having
put this BJT Iz has to commutate only Io by beta plus 1 that is the virtual load for this
Zener shunt regulator is Io by beta plus 1 and as a result Iz swing is much reduced and it
is reduced by beta plus 1. And beta being high if your BJT is chosen where beta is
around 100 say then the Iz swing is reduced by 100 and which means that Iz Rz would
be reduced by 100. So, therefore, you see that Iz rz is very much reduced and as a
consequence Vo regulation is greatly improved.
One other thing that you need to note is that V be comes into the picture here as such V
be is constant around 0.6 volts, but there is there is a variation with temperature. You
know that the P n junction here, this V be will change at minus 2 millivolt per centigrade
for every degree centigrade rise in temperature, there is minus 2 millivolt drop. Likewise,
Zener is also P-N junction. So, there is a possibility for one to match this P-N junction
with the base emitter junction such that temperature compensation can be achieved.
So, by using this modified shunt regulator one can achieve much better regulation than
the practical shunt regulator, the one without BJT.
372
Prof. L. Umanand
Lecture – 35
Applications of shunt regulator
Let us now discuss where we use shunt regulators, what are the applications of shunt
regulators. We will look at couple of important ones, one is as a voltage reference. This
is a very popular circuit used at many places in bigger circuits as a voltage reference. So,
we have the shunt regulator, I am writing it in this fashion a V cc here and the Zener
diode and we have a load connected like this. Now, this is a shunt regulator.
This V cc here is unregulated DC voltage, have an R s, a Zener voltage V z and across

this you get a regulated constant V z voltage. And many a times we would like to get a
variable voltage reference, so you replace this resistance with the potentiometer and V
reference is given by a times V z, where a is the attenuation ratio of this potentiometer at
this tap of point. So, this is a very common and popular circuit that you would use in
many bigger circuits where this will come in as a voltage reference circuit.
373
Another important application of the shunt regulator is in making a constant current

regulator. Let us see how we make a constant current regulator. Let us first draw a Zene,r
I will now put the Zener up and the series resistance R s down below is V cc. And at this
point I will introduce a transistor I am using a PNP transistor in this case. One can make
a constant current regulator with NPN transistor also. It is easy for explanation now.
So, I will put a resistor in the emitter like this, and here across these terminals we will
take the load resistance Ro. So, constant current is supposed to flow through the load
resistance. Whatever may be the variation in Ro the current that flows through Ro, I
naught will be a constant. So, let us see how this is a constant current regulator, this is V
cc let us name this as R e, this is R s this is V z and there is a drop V be across the base
emitter of the PNP transistor.
Let me mark this and take the potential across V Re in this fashion V Re, this is I E and
this is the base current, ok. So, now, V Re is V z constant minus V be drop will be the
voltage. So, if you take this loop and apply the Kirchhoff’s voltage equation you will get
V z minus V be. And I E is V Re divided by R e and there for you have V R e which is V
z minus V be V by R e. Now, this is a constant, V z is a constant V be is a constant and R
e is a fixed value therefore, I E is a constant.
Now, I E is nothing but I B plus I naught and I B is nothing but I naught by beta, beta is
the current gain of the transistor. So, I naught by beta plus I naught, so which is I naught
374
into beta plus 1 by beta. So, therefore, we can write I naught as I E into beta divided by
beta plus 1. Now, here I E is a constant as we saw here beta is a constant value for the
transistor, so therefore, you can say I naught is a constant. So, I naught is constant
irrespective of power supply variations V cc irrespective of variation in Ro as we see that
Ro does not figure in this equation. Therefore, this is a constant current regulator. So, a
very nice circuit has lot of applications.
Now, if you see this is Vo and let us say the voltage across the collector emitter of the
BJT, VCE. What is the limiting value of Ro? On one end I can short circuit these Ro can
be 0. Now, Ro is 0, Vo is 0, there is constant current that is going to flow through it. It
will operate perfectly well. What is the max value? Can I open circuit it? If I open circuit
it there is no way current can flow through and therefore, I naught is no longer constant.
So, there is a upper limit on Ro. So, what is the upper limit on Ro?
Now, let us write this voltage equation for this loop. So, VCE, VCE this is given by V cc
minus this drop minus I naught so which is V cc minus V Re minus Vo. Now, VCE for
the circuit to operate this drop has to be greater than 0 in the ideal case or it should be
greater than VCE sat in the ideal case. So, therefore, VCE should be as shown here it
should be positive, the positive, negative should be greater than VCE sat and therefore,
you will see that Vo should be less than V cc minus V Re minus VCE sat.
375
If you apply this relationship here at this value should be greater than VCE sat, Vo will
be less than. What is Vo? Vo is nothing but I naught into Ro that should be less than V
cc minus V z plus V be, because V Re is nothing but V z minus V be and minus VCE
sat.
And therefore,
Ro <= (Vcc – Vz + Vbe – Vcesat) / Io
376
Prof. L. Umanand
Lecture - 36
Series regulator
Till now we have been discussing about the shunt regulator, shunt linear DC-DC
converter. Now, we will discuss the other type that is the Series Regulator. The Series
Regulator is much more popular and then is useful for higher powers. The structure is
like this, I have an input unregulated input V i connected through a series element and
you have the output load. So, this is V i unregulated, this is Vo, this is Ro and there is a
series element R s.
Now, the principle of operation is that we try to control the value of R s is we regulate or
we change R s value based on the output feedback. So, let us say we have a controller
and the job of the controller is to check a reference value. So, let us say we have a
reference value Vo reference and checks also with the output Vo and then uses that to
control the value of R s. Idea is like this if R s value is increased then the drop across that
one will increase and the drop across this will decrease. If the R s value is decrease then
the drop across that one will decrease and that drop across the output will increase. So, in
this way by adjusting the value of R s we can bring about regulation in the output.
377
So, if you see the regulation action it is something like this. We have Vo, now let us say
for some reason it may be due to V i increasing or change in the Ro. Let us say Vo
increases for some reason. So, if the potential here increases then the potential at this
point increases Vo ref is a reference constant therefore, e has to decrease. So, the error
here will decrease and if the error decreases, we would like to increase R s such that the
drop across R s VRs will increase. So, once the drop across VRs increases in this outer
loop applying Kirchhoff’s law, as this increases Vo will reduce.
So, if you look at the regulation action because of an increase in Vo finally, it has
resulted in a decrease in Vo bringing it back to its original position by this regulation
action. So, in the practical circuit we replace R s with BJT, so that this change in the
impedance is automatic.
Now, let me include the BJT a semiconductor impedance variable impedance rather than
fixed R s. So, I am going to put a BJT there in place of R s and you have the load
connected in this fashion. Now, here I will try to put some kind of biasing circuit for
biasing the BJT such that it operates automatically.
Now, at this point I will include Vo reference and to include Vo reference I will use the
shunt regulator. We are familiar now with the shunt regulator. See this is the resistance
and that is connected here. This blue portion is nothing but the shunt regulator which will
378
define the potential here in that we can call it as Vo reference, related directly to the
Zener voltage.
Next let us obtain the feedback from the output. I will use the potential divider resistive
divider tapped from the midpoint and then use it to drive this BJT here. So, let us then
we include one more resistor here to bias, to give bias current for this BJT like this. So, I
will call that one as R b, I will this an R z I will call this one as R1, R2, Ro. This is V i
unregulated, Vo which we want to regulate this is Q1 and this is Q 2. So, this is a Vo, a is
nothing but the step down or attenuation ratio which is R2 divided by R1 plus R2.
So, under steady state condition let me look at this current here. This will give you the
crucial operating operation of this linear regulator. This current IRb what is it equal to?
On one end we have V i potential V i at this end. Let me put it at V I, minus at this end
what is the potential? Here it is Vo Vo plus V be minus Vo plus V be. So, V be is
nothing but this drop across transistor Q1 V be 1. So, that is what I have used here. Now,
that divided by this value R b will give you the current through this. Now, this current
here if you see V i is fixed value constant, Vo is constant, V be is constant R b is a fixed
value therefore, IRb is approximately a constant value here.
The only variation is due to variations in V i because V i is unregulated. So, if IRb there
is constant, this IRb divides into two parts, one part goes to goes to Q1 as a base current
base drive and another part goes down through Q 2 as collector current of Q 2. So, let us
say IRb equals this is IC2, we will call this one IC2, I f collector C 2, then this is IB1 .
Now, as IC2 increases IB1 has to decrease because the sum has to same as IRb which is
a constant. If IC2 decreases then IRb has to increase, so that the sum is the fixed value
IRb. Now, this herein lies the regulation action.
Another thing that you may have to notice let us say you have the voltage across VCEof
Q1 will call that VCE1. And as V CE, now we need to control V CE, so as IB1 increases
as the base drive to that is increased more drive, it will go more towards saturation VCE1
will decrease. So, VCE1 decreases and as IB1 decreases you will see that the drive is
lesser as comes out of saturation moves more towards cut off VCE1 will increase. So,
this is our basic control or regulation action.
The control action is like this. As Vo increases, as Vo increases here due to any reason it
could be change in Ro load resistance or it could be change in V i, you will see that this
379
point the tap off point a Vo will also increase a is a constant. So, a Vo increases the V be
of this one will increase, because you are increasing this point and this point is fixed at V
v z reference voltage. So, as this voltage increases V be increases the drive increases and
therefore, IC2 will increase and therefore, IC2 increase.
If IC2 increases you see that IRb is a constant current, if IC2 increases IB1 has to reduce.
So, IB1 will reduce. If IB1 reduces you see that VCE will increase, VCE will increase.
And as VCE increases in this loop apply Kirchhoff’s law V i is fixed as VCE increases
Vo has to come down. Therefore, Vo comes down. So, this is the regulation action. Any
increase where the regulation action comes into picture and brings back Vo to the
original value, if there is a decrease in Vo you will see that Vo will again increase and
will be brought back to the original set value and thereby the regulation action. So, this is
the operation of a series regulator.
380
Prof. L. Umanand
Lecture – 37
Efficiency of series
What is the efficiency of the series regulator? So, you see the circuit you have in the
power loop just three components, you have the input source V i, you have R s and you
have R naught. So, efficiency is given by
η = Po/Pi = VoIo / ViIo = Vo/Vi
So, if the input output differential is very large, then you will see that the efficiency is
low. If the output is very close to the input comparable to the value of V i, then you will
see that the efficiency is high. Let me rewrite the efficiency as Vo divided by V i can be
replaced with Vo plus delta V io that is a input output differential. Now, as you see that
as the input output differential is very large if it starts becoming larger, then the
efficiency goes down.
So, an example say Vo is 5 volts, and V i is 15 volts input output differentially is 10. So,
you have 5 by 15, which is efficiency one-third 33 percent. For the same 5 volts, if V i is
10 volts, then efficiency is 50 percent 1 by 2. So, if you see that the difference between
381
Vo and V i if it starts becoming smaller and smaller efficiency is better and in fact that is
the reason the low dropout regulators, where the input output differential is small how
much higher efficiency.
382
Prof. L. Umanand
Lecture - 38
Negative and dual voltage regulators
Till now we have discussed the series voltage regulator which is positive input positive
output, but there are many applications where you will need a negative voltage output
negative voltage regulation. So, how do we get a negative voltage regulator? So, let us
start with this positive voltage regulator and modify it to give you the negative voltage
regulator. First of all v i which is positive here and negative here, we will reverse that.
So, we want plus and minus like this. Now, this should finally, result in an output with
plus and minus here. So, this will give you a negative output.
However, the current flow is in this fashion before we have to reverse the transistors. So,
to keep the circuit same, I will what I will do is convert these transistors from n-p-n to p-
n-p. So, you will see a reversal in the current flow without affecting the topology of the
circuit. I will change this from n-p-n to p-n-p still maintains the same topology.
So, I should also reverse the zener diode direction in this fashion. Now, you have your
negative voltage regulator, I will remove that and this current directions will reverse. So,
you will have the base current flowing this way and the collector current flowing this
383
way and I b flowing in this fashion. Still the same concept holds this potential, and this
potential are fixed and the current through Rb is more or less constant. And these two
sum up add up to make I Rb.
And this is the direction of I naught. So, the operation is exactly same as we discussed
for the positive voltage regulator, now this is a negative voltage regulator where I can
give a negative input to get a negative output. If you combine the positive voltage
regulator and the negative voltage regulator, you will get the dual power supply. Let us
just integrate that these two circuits and see how a dual voltage power supply looks like.
Let us now see how we get the dual power supply. First let me insert here the positive
voltage regulator. So, you have this as the ground line, and this is positive with respect to
the ground line, and the output voltage also this node is positive with respect to the
ground line, all are n-p-n transistors. Now, this is a positive voltage regulator. So, giving
a positive input, I get a positive output here. Now, let me place here the negative voltage
regulator, so we, we saw we just now discussed the negative voltage regulator. The
negative voltage here and I have replaced all the n-p-n’s by p-n-p compare with this.
384
Now, I will use this as the common ground point so that as the ground place that ground
there. Now, I will common these two ground points which means that I will have to flip
this circuit such that this ground comes in there.
So, let me redraw the flipped negative regulator. So, let me redraw that. So, I am
redrawing with the ground flipped compare, now this is the p-n-p I am replacing the n-p-
n with a p-n-p, I am replacing this n-p-n with p-n-p, but now the regulator is flipped
compare with what we have drawn drawn previously this is R naught, these are the
385
output terminals. And I am having an attenuation and taking the attenuation point here I
need to have the bias resistance for this transistor, and I have to give the series resistance
for the zener.
So, this becomes our dual power supply regulator common ground. So, for example, let
us say I give 15 volts for this positive regulator, 15 volts plus and minus. And for this
also 15 volts plus and minus like this, but this is the ground point I will pull it out let us
say that is 0. So, with respect to this ground, this node is positive so plus 15. With
respect to this ground this node is negative 15, so it is minus 15. So, we have 15, 0,
minus 15 at the input.
And at the output, this is again ground ok. This is the measurement direction. Now, let us
measure like this with respect to the ground always with respect to the ground. So, you
are measuring this node potential with respect to the ground that is this, this node
potential with respect to the ground that is this. So, let us say a 10 volt supply we want at
the output with a input output differential of 5 volts in each of the regulators. So, at this
point, let us say we have 10 volts plus 10 volts with respect to this ground this is 0. And
with respect to that same ground if I measure here, I will get minus 10 volts.
So, this gives gives me a plus 10, 0, minus 10 volt dual output power supply. So, this 10
volts, plus 10, minus 10 volts is the regulated plus 15, minus 15 is the input unregulated.
So, this is the integrated schematic of a dual power supply linear regulator both positive
and negative together.
386
Prof. L. Umanand
Lecture – 39
Over current limiting circuits
Let us now discuss how to provide current limiting for the series regulator circuit. Now,
consider the series regulator circuit and let us introduce some components here such that
we limit the current under, overloading or short circuit conditions.
387
Let us create some space here to introduce some components here. And let me introduce
a resistance in this fashion. And across the resistance let me connect the base emitter of a
transistor, and the collector of the transistor is connected to this point. .
Let us call this transistor Q3. This is R s and Io flows here. You can say Io is the current
flowing here is approximately equal to Io, because the current flowing through this
feedback tap point is very, very negligible.
So, let us say this Io value which is flowing here is dropping across this and you have Io
into R s drop. So, this Io into R s drop is the one that will start cutting in this transistor
Q3. So, once Io into R s drop reaches 0.7, so that is where you can design the value of R
s for a given Io rated value. Once it crosses 0.7 at say the Io rated value flowing through
this. The transistor will be almost reaching the saturation stage, and it will be drawing a
large amount of current reducing the base drive here. And thereby making this more
towards the off region, and it will be withstanding more and more voltage.
Imagine the condition where you short circuited the output, when you short circuit the
output the voltage here is 0, a large current is expected to flow. But I R is drop and drive
this into saturation and take away all the current IRb current into this and Q1 will be in
off state and withstand all of V i. So, this is the constant current limiting. So, if we take
V R, this is VRs. So, Vbe3 across this is nothing but Io into R s and R s can be selected
by Vbe3 divided by Io rated, whatever is the rated value for which you would like to
388
provide the limit. So, around 0.6 volts for V b e divided by Io rated will be the value that
you select for R s.
So, if you plot the, I versusV, Io versus Vo, so at Vo whateveRbe the load, Vo is going
to be constant till it reaches a rated Io rated current value. Beyond Io rated current value,
this would be draining out the drive for Q1, and Q1 will support the entire voltage.
So, even if you short circuit it this is the short circuit with 0 Vo point that is here. So, it
will limit the current to Io rated. So, this is the constant current limiting type of a circuits,
very nice circuit, it can be used in most of the series regulator type of configurations.
Let us look at another interesting current limiting technique called the fold back current
limiting. We had seen this constant current limiting by putting these set of components
one BJT Q3 and R s. Now, let me modify this slightly, and you will get this fold back
current limiting.
389
Now, let me explain this. Let me remove this portion make a bit more space and I will
put the resistance R s here. And from this point, I will tap a voltage divider to ground.
And to the centre point of that, I will connect to the base of the BJT Q3. So, the collector
I connect it in the same way and the emitter also I connect in the same way to this point
output point. .
So, by making this small modification you will find an interesting way of current limit
we will call this Q3, we will call this Ra, Rb, and this is R s.
Vbe3 = (IoRs + Vo) (Rb/Ra+Rb) - Vo
Now, let us look at the potential of base emitter of Q3. So, we will call that one as V b e;
V b e is equal to this potential which is whatever this potential Io R s plus Vo with
voltage divider action of R a, Rb..
So, which will be Io plus Io into R s plus Vo multiplied by this ratio Rb by R a plus Rb
will be the potential at this point minus the potential of this point which is Vo . So, we
will get that as the voltage across Vbe3. .
Now, I will define the symbol alpha for this Rb by R a plus Rb, so that equation becomes
simpler to write. And we know that alpha is less than one varies is between 0 and 1. .
Now, Vbe3 can now be written as alpha Io R s plus Vo minus Vo ok. .
390
Now, you can simplify it a bit further alpha Io R s I will take it and this Vo and Vo, I will
take it common and write it as 1 minus alpha Vo note the sign change here accordingly
to take care of this.
So, now on short circuiting, let us say I short circuit this. So, Vo becomes 0. On short
circuiting Vo becomes 0; and Io we will call it as I s c short circuit current ok. Here
again assume that the current flowing through R 1, R 2 is negligible and therefore this
value is Io itself whatever is flowing through the load. .
So, under short circuit condition look at this equation Vo is 0. So, this part this term goes
off and you have only this term. So, therefore, I s c this I replace by I s c which is equal
to Vbe3 divided by alpha R s. So, this is a design equation for choice of R s and
appropriate choice of the design parameter alpha. You can also design what should be
the short circuit current.
Now, this importance will become clear as we go to the next steps. Now, consider this
equation. So, I can now write Io equals Vbe3 divided by alpha R s plus I am bringing
this term on to this side 1 minus alpha Vo by alpha R s.
391
Io = Isc = (1-α)Vo/αRs
This circuit will operate in the normal mode, regulation mode for all the loads which
appear here till Io rated. Now, let us say I apply a short across the output Ro is made 0.
So, the moment Ro is made 0, Vo becomes 0, this term vanishes. So, Io reduces to I s c
actually Io when Io when it was operating in the normal operating mode, a value of Io
here is I s c plus something. And now Vo is made zero under short circuit condition, you
will see that this would be the operating point I s c because Vo is now 0. And this is the
fold back.
So, when it was operating at Io rated and something happens a short circuit happens, you
will see that the current that flows will be lesser than the rated thereby protecting the
devices. And this is called fold back current limiting or current protection, a interesting
circuit which can be used by you for any series regulator.
392
Prof. L. Umanand
Lecture – 40
Improvements to series regulator
Let us now discuss some improvements that can be made to the series regulator circuit,
so that the precision of the regulation can be improved. So, first let me take up the case
of Rz, and then next the case of Rb. See both these through both these components the
current that is supposed to flow through or supposed to be constant for effective
regulation. However, Vi here is unregulated, and therefore it is a varying voltage.
So, therefore, we cannot strictly say that the current that were flowing through Rz and Rb
are constant. Now, let us take the case of Rz. If we shift this Rz from here to here, Vo is
a regulated voltage node point. Therefore, if we shift this to here the current through Rz
can be effectively said to be constant. So, so let me remove this Rz from there, and then
reconnect it here in this fashion Rz.
And you see that here the voltage is Vo, and Vo is the regulated voltage. And, therefore,
it is constant and fixed; therefore the current through Rz is fixed. And as a consequence
the the regulation the Vo reference is much more stable in this arrangement.
393
However, the question arises will this arrangement allow start up? Means at the time of
starting the output voltage is 0, what will happen? At the time of starting Vo is 0, zener
breakdown has not happened here also it is 0, a Vo. Q2 is how to the picture it is off. So,
all the input bias currents through Rb flows through the base of Q1 and tries to turn that
on.
And the voltage starts getting developed here. Once the voltage starts getting developed
here, zener comes into the picture Vo ref is set, and the output voltage will track this Vo
ref. So, therefore, we can say that this arrangement will allow start up, and also improve
regulation by virtue or some more constant current flowing through Rz.
Let us discuss another important improvement in the series regulator circuit to improve
the regulation precision, and that is in this current I Rb. We saw earlier that for effective
regulation I Rb supposed to be constant and I Rb splits up into IB1 base current of Q1
and I C 2 collector current of Q2. I C 2 and IB1 the base current of Q1 and collector
current of Q2, they commutate among themselves so as to achieve effective regulation of
output voltage Vo.
Now, we see here that Vi is unregulated voltage which means that it can fluctuate. If it
can fluctuate the voltage here,
IRb = (Vi – Vo – Vbe1)/ Rb
394
So, as a consequence, IRb is not strictly a constant current value, which is flowing
through Rb, and therefore, this is a serious cause for deteriorated regulation.
So, how do we solve this? One way to solve this is to remove this Rb and putting there a
current source like this. And, repeating the current source in irrespective of variations in
V i, the current here will be constant and there while their regulation will be very precise
for the circuit for any variation in load and line. We also know that a current regulator
can be made using the zener shunt regulator.
Let us make that and introduce at this point. So, let us make some space. Let me put in
the current regulator circuit a resistance a pnp, and the output of that connected here, and
then I have a zener, and the resistance here. So, look at this, this is a zener which give me
constant voltage that is applying between these two points I am showing. V b is constant,
so essentially Vz -Vb will be come across this resistance.
And this resistance being fixed Vz - Vbe by this resistance will be a constant current
flowing through the emitter, and thereby a current which is alpha times the emitter will
come into the collector, and then flowing here. And the collector current will be
essentially constant irrespective of the variations in V i here. And now that is a crucial
improvement where this current being constant, the regulation will be very, very precise
and effective.
395
Now, let me discuss another important significant improvement to the series regulator
circuit. Here I am talking or replacing this transistor Q2 itself. The transistor Q2 ws
supposed to provide amplification, current gain and also a high input impedance as seen
from this side. The same thing can be provided in a much better way by using an op-amp
which is supposed to have an infinite gain and almost infinite input impedance. So, let us
try to use an op-amp and improve the regulation performance of the circuit much, much,
much better.
So, remove this portion this portion which is providing the Vo references still retained,
and this attenuator which is actually tapping off the output providing the feedback a Vo
is also retained. The transistor portion will be replaced with an op-amp. So, let me have a
drive resistor to drive Ib into the Q1 base. And the I B drive is be obtained from the
output of an op-amp like this. So, op-amp is a three terminal device. There are two
terminals for power supply. Let us take the power supply for the op-amp from the
unregulated voltage itself.
And I will give the feedback tap of point to the minus terminal of the op-amp, and the
voltage Vo reference to this point. Observe that the op-amp has infinite infinite input
impedance, which means that the V plus voltage and the V minus volt are almost same
meaning that there is a virtual ground in the op-amp, even though there is no current
flowing through the high input impedance. This means that whatever value of Vo
reference you are setting here, the a-Vo will track this Vo reference almost exactly.
Further, there is almost zero current flowing into V plus, and therefore the zener the
zener impedance that is Rz also does not cause deregulation, because this is essentially a
fixed value of current. Now, the regulation action, how does it regulate? Now, let us say
for example, this is IB1 and for example, Vo increases due to some reason, the Vo may
increase due to change in V i change in Ro. So, if Vo increases, then a Vo will increase.
And as a consequence a Vo which is given to the minus will be higher than will increase
more than the plus V plus terminal voltage which is being held constant. Therefore, the
error decreases and the output of the op-amp reduces op-amp output reduces.
And as a consequence IB1 drive will reduce. And if IB1 drive reduces VCE will
increase, because the transistor Q1 is pushed more towards the cut off, and as a
consequence VCE will increase. And if VCE increases, then the outer loop Kirchhoff
396
Kirchhoff loop will sit with that Vo decreases and is brought back to its original value.
Now, this is the regulation action, the closed loop regulation action. Likewise even if Vo
decreases, the same sort of regulation action happens Vo decreases a Vo decreases, error
increases, IB1 increases, VCE1 decreases, and therefore Vo increases and is brought
back to its original state.
So, to this improved regulator that is the one with the op-amp connected like this, you
can improve it still furtheRby placing the zener such that it is tap from the output side
rather from the input side. So, let us say we tap the zener input instead of from V i, you
tap it from Vo. Vo is a regulated, and therefore this will be well regulated absolutely stiff
Vo reference and the one of the most precise regulator that you can have. However, will
they start up? So, when you are starting up Vo is 0, this is 0, this is 0, output of the op-
amp is not defined.
And, it may turn on this or it may not turn on Q1, but we cannot a designer cannot leave
that one to chance; therefore we should ensure or guarantee turn on. So, what we shall
put is we shall put a resistance division like this. And at this point, I will tap off put a
diode normal diode and connect it there. So, how does this operate? Let me take an
example. Let us say we have a 5 volt zener. So, this Vo reference is 5 volts. This is a Vo,
let us say this is R, and R equal which means a is 0.5, and therefore, Vo will be 10 volts.
So, if this is 10 volts, input of course will be much greater than 10 volts.
397
Now, let me set this value, this potential division to 4 volts, then initially on start up 4
volts comes across at this point, and here 4 volts will be available to you if this is an
ideal diode or 4 volts minus the diode drop. Now, this will act as the reference and this
will push this transistor and the regulator will come into being. And at this point, it will
be 2 times 4, 8 volts. The moment this potential here crosses 5 volts, then the zener starts
to go into the zener breakdown, this will be 5 volts. The moment this is 5 volts, this
diode will get reverse bias, because it is 4 volts there and 5 volts here, and covers out of
picture, and the zener will set the reference and the operation will be as usual as normal.
Additionally you can also give current limiting for this circuit also at this point. You can
include a constant current limiter like this. Put a resistance there then put a BJT and
connected in this fashion. As the voltage across this resistance increases, V b will
increase. And it will try to drive remove the drive from the base drive from this main
transistor, divert, it will divert the base drive and see that this transistor blocks a higher
voltage. You can also replace this with a fold back current limiting circuit too. It will
also operate equally well in this case.
398
Prof. L. Umanand
Lecture - 41
Regulator performance parameters
Regulator performance parameters, these are important parameters that actually

characterize the various regulators, linear regulator that are available in the market. You
can find them in the data sheet in one form or the other. First one of the important
parameters the line regulation parameter, it is actually given as (∂Vo/∂Vi)*∆Vi
∆What is basically means is the variation in V naught with respect to variation in V i for
a given absolute variation difference variation in delta V i. So, this parameter is called
the line regulation parameter or the line regulation coefficient.
The second important one is the load regulation. What is the effect of load on V naught?
So, that is what is given by (∂Vo/∂Io)*∆Io
And the third one is the thermal or the temperature regulation, which is (∂Vo/∂T)*∆T
399
Overall, if you put all of these together, you can get the overall output voltage regulation
∆Vo = (∂Vo/∂Vi)*∆Vi + (∂Vo/∂Io)*∆Io + (∂Vo/∂T)*∆T
Next you have the effect of change in temperature on V naught due to change in
temperature keeping input line and the load constant. So, this becomes the important
regulation equation for the for the output voltage for any regulator. And this can also be a
good comparison or a bench mark bench can be used for as benchmarking for various
regulators. You will find these regulation coefficients in the data sheets.
400
Prof. L. Umanand
Lecture – 42
Datasheet of few IC regulators
Let us look at some commercial linear regulators. The mu A78 series regulator is a very
popular and commonly available linear regulator. It is a 3-pin you see here 3-pin, 3-
terminal linear regulator. You have a input the output and the common pin very popular,
and it comes in various fixed outputs 78xx or 00 series. If you get 7805, it is a 5 volt
regulator. If it is 7808, it is a 8 volt regulator; 7810 7812, 12 volt; 7815, 15 volt regulator
so on.
401
Very popular output currents up to 1.5 amps, and it is available in TO-220 package.
See here different 78x x series this is a five volt regulator this a 8 volt regulator, a 12 volt
regulator, it is the 15 volt regulator, 24 volt regulator so on. So, these are constant
voltage regulators just 3-terminals, you put this 3-terminals no other extra components
and you are in business.
402
In the data sheet of the 78xx series you will see some of the input important parameters
output voltage minimum 4.8 to 5.2. Now, the input voltage regulation ok, so for a
variation from 7 to 25 volts, 3 to 100, you should map it into the regulation coefficient
that we just now discussed. Ripple rejection, output voltage regulation that is with
respect to the load. There is a temperature coefficient of output voltage with respect to
variation in temperature ok.
403
These are the regulation coefficients that would matter, and you will have to use these in
effectively designing your regulators.
Another versatile precision regulator available in the market is the 7803.7 Another
versatile precision regulator available in the market is the muA723. It comes in a d
package and few other packages as well.
404
It is a 14 pinned IC. Of course, many of them are not connected, but you will see that the
internal block diagram of this regulator is very similar to the discrete regulator, series
regulator that we discussed.
And you will see that many of the features that we have discussed is already
implemented in IC form in this regulator. And you can make the best use of it.
Here are some data sheet values, input voltage 9.5 to 40 volts max. So, let us stay within
that for an output voltage. See 723 is a variable precision regulator, you can set for
405
different output voltage from 2 to 37 volts. Input-output differential voltage this is an
important parameter, so not exceed input-output differential voltage of 38 volts, and you
should keep a minimum of 3 volts. And you have a 150 milliamp output current
possibility.
Here to look at the regulation coefficients input regulation. One typically 1 and
maximum of 5 millivolt per volt, ripple rejection, output regulation, and you have the
temperature coefficient output voltage. Now, these are the important parameters that you
should consider in doing your design.
406
Look at this, this is a typical circuit given in the data sheet for the 723 regulator. So, this
is where unregulated V i is given. And internally reference is generated; you do not need
to put an external zener. So, internal reference can be used and given to the plus terminal
of the op amp internal op amp. Even the op amp is internal, and to the minus terminal
you will give the output that is fed back. And there is a facility to add RSC short circuit
protection or the current limit through this across these terminals when you have the
current SENS. So, this is a typical case what we discussed in the discrete version.
407
LM 723 also provides you with facility to provide an output transistor, transistor on the
output side to provide a current boost. Let us say this internal transistor is capable of
taking 150 milliamps, and you want 1 amp at the output load current you can put a
current boost and then increase the current capability.
Observe again from the data sheet of muA 723, they have implemented the fold back
current limiting. The transistors are internal. The resistance you have to put external, and
you see the fold back current limiting operation.
408
LM317 is another 3-terminal adjustable regulator. I will discuss this a bit more in detail
later. Another version of LM317 called the LM350 is also a 3-terminal regulator. LM317
has 1.5 amp output current rating, whereas LM350 has 3 amp rating. So, they are
powerful regulators for quite reasonable amount of power.
Here too you have just 3-terminals, the input, output and the adjust. There is a fixed,
there is a fixed voltage between the output and the adjust and that is used as the reference
voltage and that is what is used for at achieving adjustable output voltage regulator.
409
In the electrical characteristics of the data sheet you see for the LM317 line regulation
aspect, it is given here, terms of percent per volt, you have the load regulation given
here, thermal regulation. So, all these aspects are given, and you will have to map it to
the regulation equation that we discussed.
Another important aspect that you need to note that in most of the data sheets you will
find R theta JA junction to ambient thermal resistance.
410
So, from the core, the core of the regulator should not cross 150 degree centigrade; and
to be on the safe side we set it at 120 degree centigrade and pass on the heat to the
ambient. How do you pass on the heat to the ambient? You can use thermal resistance is
given here in degree centigrade per watt. If they do not give reasonable thermal flow,
you may have to put extra additional heat sinks now that is very important if any of your
regulators or semiconductor devices has to function properly.
411
Prof. L. Umanand
Lecture – 43
Common IC regulator circuits
Let us now discuss some common IC regulator circuits IC regulators that are available
in the market popularly available in the market. One of them of course, we just now saw
in the data sheet which is 7-8 xx series, let us see how we will incorporate them in the
circuit. Consider that we have 230 volt mains this is an AC voltage and we will use a
transformer and the output of the transformer is a step down transformer. So, that we get
output 5 volts.
Then the step down transformer is connected to a rectifier circuit like this and output of
the rectifier is connected to a capacitor filter and output of the capacitor filter is
connected to a 3 terminal regulator like this and the output this is the load r naught.
So, let us have if I use 7805, 7805 which means a 5 volt regulator this has the in terminal
out terminal and the common terminal and here you will have 5 volts coming at the
output of 7805. You should have a minimum differential law of delta vio of at least, 3
volt should be greater than 3 volts which means that here the voltage of the greater than 8
volts. So, when you are designing the turns ratio, you should see that at least under vi
412
min condition you should get 8 volts at least. So, that then only this 3 terminal regulator
will provide regulated output.
Let us say now that you have you need a multi output power supply; let us say you need
a 5 volts and you have another you need anotheR15 volts which is isolated you do not
need to put another transformer, you can wind one more secondary on the existing
transformer. So, on the same transformer you have one more secondary have one more
rectifier diode bridge on the secondary side, only this turns ratio and this turns ratio will
be different and put a capacitance and output of the capacitance put a 3 terminal
regulator let us say 7815 and connected to the load we have the IN OUT COMMON.
So, if you need 5 15 volts here you lead a minimum differential of at least 3 volts and at
least 18 volts is needed at this point. So, accordingly you can adjust the turns ratio
between the primary and the secondary. Now this is a way you will get a multiple output
power supply. Now let us say that we would like to have a dual power supply with a
common ground plus 5 and minus 5 volts how would you go about doing that? The input
side remains the same you will have a transformer with multiple secondary rectifier and
the capacitors make some modifications here on the output side, let me remove this and
then put a 7805 I will set this to 8 volts like this, this will be 5 volts which means I will
replace this by 7805.
So, both these are giving 5 volt isolated outputs no common ground. So, what you do
you short the minus of this to the plus of this and indicate that has the ground or the 0
potential. Now all your further circuits should consider this as the 0 potential or the
circuit ground, then with respect to this 0 potential with respect to this node this point
will be at plus 5 volts and with respect to this 0 ground, this point will be minus 5 volts.
So, this way this power supply together shown like this acts like a dual positive negative
power supply.
413
Let us discuss another example this time an example of LM 317 where LM 317 is a high
power variable output voltage regulator you can set to any output voltage just like LM
723 also both LM 723 and 317 or variable output voltage regulators I will take LM 137
as example.
Another (Refer Time: 05:08) sister of LM 317 is LM 350 which is a much higher power
regulator. So, let us construct the regulator structure I have an input unregulated voltage
and that is connected to the 3 pin regulator the IN pin the ADJUST pin OUT pin and
output connected to the load R naught like this.
Now, the ADJUST pin is brought down connect a resistor here like this and from here
you connect another resistor in this fashion name this as R1 name this as R2 how do we
choose R1 and R2? Observe that look into the data sheet there is a fixed voltage
difference between the out and the ADJUST pin out paying more positive than the adjust
and that is 1.25 volts for the 3 1 7 regulator and also 350.
Now if this is V naught then we can say V naught equal to if you look at this network
1.25 by R1 is the current that flows through R1 this is a high input impedance pin
therefore, all the current flows in through R2. So, you will have 1.25 by R1 current
flowing through R2 into R2 will be the voltage drop across R2 plus 1.25 will be V
naught. So, therefore, V naught is equal to 1.2 into 1 plus R2 by R1. So, this is the
voltage relationship and then accordingly you can choose the value of R1 and R2.
414
If you make R2 adjustable like this, you will see that R2 in this equation is a variable
and you can get an output voltage setting as different values for different values of R2.
Apart from this there are few other protection components that are needed to complete
the circuit. First of all of course, we need to put a 0.1 to 1 microfarad capacitor here this
is just in case to be to make sure that there is decoupling at the input side and there is
sinking capability also.
Likewise in the output side also you put a capacitoR10 to 20 microfarads then important
you have to put a protection diode like this. The protection diode does not come into
operation normally, but this comes into operation during switch off condition let say you
switch on, when you switch on there is some voltage here and then there is a drop across
this and then there is a voltage at the V naught some set value.
Now when you switch off this let us say the input goes to 0 quickly and the output is not
yet discharged; which means the out terminal will have some finite positive voltage in
terminal be 0 and then the input output differential voltage will be negative which will
damage the IC. Therefore, if you put a diode like this whenever the output terminal is
more positive than the input terminal the diode will conduct and sink it to the capacitance
thereby protect the IC.
So, there is an important protection diode that you need to put. So, that the input output
differential voltage does not go negative. Another important component we can have is
add a capacitance here 10 microfarad. This will act like a filter and stabilize the potential
here once it stabilize the potential here output regulation is much better.
However, you will see that when you switch off there is no discharge path for this 10
microfarad capacitance it has to be through R2 or it has to be through R1 and the output
capacitor or the input capacitor. This can be a long time constant in order to speed up we
can put a diode like this, which in this way this capacitor can directly go and sink into the
output or sink into the input. So, these are some of these protection features you add and
this will make it workable circuit.
415
Prof. L. Umanand
Lecture - 44
Practicals 1
Here are some tasks that you can do for Practical work. Of course, it is not in the scope
of this course, but to gain further insight into the circuit nothing like implementing it on a
lab table. So, if you can take the help of your mentor and instructors in the colleges and
if you can find a lab table with appropriate components, do please try to implement them
by yourself so, that you get a much better understanding of the circuit. So, the linear
regulator circuit we have discussed this circuit.
416
We have here the 230 volt AC and followed by a transformer; the transformer does 3
jobs here scaling the input output scaling it gives galvanic isolation and third importantly
it also gives inrush current limiting. And you have a rectifier diode and this is followed
by a capacitor filter, and this is the point which acts as V in for the regulator this is the
unregulated DC.
And you have the series pass element here you have a constant current limiting current
protection, you may modify it into fold back current limiting protection also and try it
out and then this is the feedback for the output voltage brought to the minus terminal,
this is the zener which is filtered output brought as reference this is the start up circuit
and the output of the op amp is driving the transistor Q 1.
So, this would be the entire current limited linear discrete voltage regulator, please try to
rig them up you can use 2N3055 for Q 1 2N3019 or 2N2222 for Q 2 these are low power
1 amp devices this is 10 amp device uA741 you are already familiar with op amp or any
other equivalent op amp. Do the design using octave or matlab m files so, that you can
do iterations perform the load regulation test tabulate and plot V-naught versus I naught.
And set the current limit at 1 amp and test the over current protection by adjusting the
value of R e. Short circuit the output terminal bravely and check if the current protection
works. Then finally, if you it is preferable recommended I would recommend that you
document the results take the waveforms, design equations put them into a word file and
417
convert them into pdf and zip them along with the design file and keep it for records or
submit it to your instructor.
There is another exercise that you can do and that is linear regulator using LM 317.
So, this is this is IC regulator, it is a variable output voltage regulator download and read
the data sheet very important and design select the components based on the using a dot
m file.
418
in octave or matlab this is the circuit similar circuit we have the 230 volts AC. And then
you have the transformer and you have the rectifier and the C filter then after that you
have the LM 317 now the in, out, adjust pin and you have R2 and R1 and I have made R
2 adjustable. So, that you can adjust that and get different voltage settings.
I have explained in the class on the need to use C3, D2 and D1 recall that and
understand what they do and what the functions here are ok. So, do all these tasks to get
better insight and then document them, tabulate the results wave shapes and then put
419
them in the word file zip it along with your design file, submit it to your instructor or
keep it for records.
420
Prof. L. Umanand
Indian Institute of Science Bengaluru
Lecture – 45
Switched mode DC-DC converter intro
We shall now discuss Switched mode DC-DC converters; I mentioned earlier that DC-
DC converters can be executed either using the linear circuits called the linear regulated
circuits which we discussed in the last week or using switched method by switching the
semiconductor switches.
In the case of the switched mode DC-DC converters all the semiconductor switches that
is the diodes, the BJT and the MOSFETS they are all switched and they are either in the
cut off region or in the fully saturated region, which means all the switches are either
fully OFF or fully ON. The advantage in such mode of operation is that the switches
which are coming in series will either be ON or OFF when they are on the current
through them will be flowing through the ON switch, however the voltage across the
switch is 0.
When they are OFF the voltage across the switch will be finite and the current through
them will be 0 in both the cases the power decapitated across the switches will be 0.
Therefore the efficiency will be very high in an ideal case the efficiency will go towards
100 percent, however in a practical case there will be switching transitions while going
from ON to OFF or OFF to ON the switch has to transit through the linear region and
therefore there will be some dissipation. Due to switching losses and therefore the
efficiency will not 100 percent will not be 100 percent but will be slightly lower than
that.
So, we will take up now DC-DC converters using the switched mode technique wherein
the semiconductor switches will be switched, we will take simple converters and then
gradually go to isolated converters and more realistic power supply cases.
421
Let us now consider an unregulated DC voltage and this DC voltage suppose to supply
power to a load or not, let us say this input unregulated DC voltage is V i output load is
Ro load resistance is Ro and output voltage Vo is demanded is required. V i and Vo are
different values and therefore they cannot be just connected directly with the conductor,
so they need to have some series element.
In the linear regulator case what did we do? We put a variable resistance component
which could be a resistance or it could be a semiconductor switch operated in the linear
region. So, basically this Rs resistance was varied. So, when this Rs is varied accordingly
the drop across that varies and therefore we could regulate Vo, but unfortunately there is
a voltage drop across this Rs there is a current through Rs and therefore there is a power
dissipation VRs* io happening in this resistor Rs and it is not available as useful load,
therefore the efficiency is very low.
Now, if we modify the constraint that Rs can take only 2 values it can either take 0 or Rs
can be infinite there is open circuit. So, let us say let us put the constraint that Rs is either
0 or infinity, in such a case now let us plot Vo with respect to time what happens.
Let me make time divisions like this let me divide the time in this fashion, now let us say
first in this time let us make Rs is equal to zero, so this is zero V i comes completely
directly to the output and we will see an output like this with V i . Now in this time region
let us make Rs infinity means open this, so when you open this the entire V i dropper
422
drop drops across this and there is no load current and therefore the voltage Vo across
the load is zero this what we have drawn there.
Now, let us make again Rs is equal to zero this is a sorted, again V i comes across Vo and
therefore Vo is V i and Rs infinity again it repeats in this fashion if we take just only this
two possible values of Rs. Now if you see here V i , zero, V i , zero this or the possible
outputs voltage states, one thing that you observe from here the output voltage is
pulsating not constant fixed at DC.
Another thing is we have restricted Rs to only these 2 possible values and when Rs is
zero the voltage drop across Rs is zero and therefore there is no dissipation across Rs
when Rs is infinite the current through Rs is zero and therefore there is no dissipation
across Rs. In both these conditions the power dissipation across Rs is zero has no power
dissipation, so the entire power goes only to the load the only drawback is that the output
voltage is pulsating and not regulated DC. So, one thing that we have achieved by doing
this is to remove the power dissipation and improve the efficiency.
Now, to take care of this pulsating or chopped DC that is coming across the output we
need to put a filters, so we will put a filter which is not dissipative so we have to put a L
C filter, so that is what we will do to obtain a DC again at the output. Let us now
introduce a block where which is suppose to do a filtering, so let us say that there are
these components the filter consists of three important components one is the L or the
inductor which is the kinetic storage, C or the potential storage capacitance and there can
be switch elements like diodes and other semiconductor switches. Now the filter
comprises of these components.
Now, if I take before the filter V A considers it as V A now that will be chopped like that
and after the filter the chopped waveform will be smooth and out and that will be the
smooth DC which will be applied to the load Ro. So, therefore, in a DC-DC converter
switched mode DC-DC converter you have an input V i unregulated and output Vo
regulated different value. Now in between there is a switch which chops the waveform
there are filters which will do the job of smoothening the chopped waveform to obtain
filtered DC, so these are the main components of a DC-DC converter. Now, let us take
the simplest DC-DC converter with one switch, one inductor and one capacitance.
423
Let us now draw the circuit of a DC-DC converter with one switch one inductor and one
capacitance this is the load and this is the input unregulated input L C Ro, we have V i
unregulated and this is Vo. Now this is as single pole double throw switch because this is
a pole p this is throw 1 this is throw 2 double throw. So, the most fundamental switch
would be having a single pole and 2 throws if 1 throw is not connected then we call it as
a single pole single throw, but more fundamental is a double throw switch.
So, let us also make the connection for this throw T2 and let me connect this to ground.
So, which means throw T1 is connected to V i throw T2 is connected to ground and the
pole P will alternately connect to T1 and T2. So, this is a full-fledged DC-DC converter
it has a single pole double throw switch and inductor L capacitor C, so this portion will
be a switched mode power converter.
Let us now have some clarity on when the pole will be connected to T1 and at what
times the pole will be connected to T2. So, let us make some definitions and
clarifications on that. So, for that let us shift make some space and I am going to have
graph of time verses amplitude here. But I am not going to use the y axis in the time axis
alone I want to split it up into different periods of time. So, let me make these kind of
divisions in time and let us see how this time divisions or defines split and define. So, let
us say this period this is called Ts switching period and this is called again the next
switching period Ts.
424
So, this width period time period is switching, now the time period and switching period
is divided into two parts this switching period is also divided into two parts. So, in one in
part 1 let us say pole is connected to T1 and in part 2 here pole is connected to T2,
likewise the next switching cycle pole is connected to T1 during this period and during
this period pole is connected to T2.
So, let us say when pole is connected to T1 we will call that one as the ON time. So, I
will say on when pole is connected to T2 T1 is off so I will call that one as OFF time
pole is connected to T1 we will call that one as the ON time, when pole is connected to
T2 T1 is off therefore we say that is the off time for T1. So, with reference to T1 I am
saying this is the ON time OFF time ON time OFF time or symbolically we can say this
is T ON period this is T off period.
So, look at this and this and this Ts can be written as T ON + T off . So, let us divide by Ts
we have 1 = (T ON /Ts )+ (T off /Ts). So now, this T ON /Ts we will define it as d which is
the duty ratio or duty cycle it is called the duty ratio or the duty cycle and this T off /Ts we
will call it as (1-d) just (1-d), because the some should be equal to d there are only 2
states p can be connected either here or here no in between states.
So, T ON /Ts is d so therefore Ts = dTs+(1-d)Ts. So, we have now converter this time into
an equation based on d that is the duty ratio, so duty ratio is the value that varies from 0
to 1. So, this portion here will be dTs this is dTs and this portion will be (1-d)Ts.
So, what we have now achieved is that when we can say for a period dTs time p pole is
connected to T1 for a period time (1-d)Ts pole is connected to T2. So, this is the way we
will be describing the switch ON OFF without confusion and d is an important variable
called the duty cycle or the duty ratio please carefully understand this particular variable.
425
Let us now mark the period for which P T1 will be ON as dTs and the period for which P
T2 will be ON will be (1-d)Ts or is the period for which T1 will be OFF. So, now with
this definition of operation of this single pole double throw switch let us look at some
waveform. So, let me now divide the time in this fashion this is dTs period this is (1-d)Ts
period this is neck cycle dTs period and this is (1-d)Ts period so on its keeps going, so
this is 1 period dTs+ (1-d)Ts is 1 period Ts.
So, during this time let us see what is the pole voltage with respect to the ground so let us
plot vp. So, during the time when during the time d Ts P is connected to T1 so it has to
be V i . So, V i comes in there now during the time (1-d)Ts p is connected T2 and T2 is
connected to ground zero, so that is that point then dT s period again V i and zero.
So, this is the voltage that you will see the chopped voltage that you will see at the pole
of the switch. Now what is the average value of this particular wave shape we will call
that one as vp average and from this relationship waveform we can find vp average is
this is V i in a cycle V i into dTs and here it is zero. Therefore, 0 *(1-dTs)/Ts so you have
Vi*d this is the average value.
Now, what is the function of this L C; L C is a filter which is nothing, but an averaging
network. So, this is an averaging network or averaging circuit and therefore Vo is the
average of vp and therefore Vo is equal to vp average which is equal to V i *d.
426
So, the relationship between Vo and V i is V i *d duty ratio and note that d is a value
between 0 and 1 and therefore Vo is always less than V i for the circuit topology,
therefore it is called a step down or up buck converter. So, this is the first primitive DC-
DC converter circuit that we have seen.
427
Prof. L. Umanand
Lecture – 46
Volt-sec and Amp-sec balance
For simple switched mode power converter circuits like the Buck converter that we
discussed it is easy to get the input output voltage relationship. we saw that we had at the
pole the pole voltage was chopped and taking the average of it, because it is going
through an averaging network we got the relationship between Vo and V i which was
V i *d the duty cycle.
But we would like to have a formal method of getting the input output voltage
relationship or input output current relationship, because as circuits switched mode
converter circuits become complex it may not be as simple and as evident as the case
was for the simple buck converter. Therefore, we would like to use the method of Volt
second balance or the Amp second balance the dual of this one. So, what is it?
So, if you take an inductor any inductor whichever circuit have an inductor L there is a
voltage across the inductor we call that one as v L and there is a current through the
inductor and called that as i L . In the steady state; in the steady state should be noted that
428
the inductor cannot sustain or cannot support an average voltage. So, we can put that
down like this cannot support an average voltage across the inductor.
So, it may be for any coil, so what it basically means that in the steady state the average
voltage across the inductor is always zero. Of course, during dynamics there will be
voltage fluctuations, but in the steady state you will see whatever be the voltages across
that the average will be zero. So, let us take graph of across voltage across the inductor
v L with respect to the time and let us say it is connected on network where there is an
arbitrary voltage across the inductor could be any arbitrary voltage.
But one fundamental character of the coil of a coil or inductor, so in this waveform if
you take the average of this waveform over a period it should be resulting in a zero
which means that zero average, which means that the area under the top portion of the
curve area under the bottom portion of the envelope they should match and then cancel
each other and result in an average which is zero.
If you take the area under the top portion it basically means that you are taking an
integral of the V i envelope to dt over the whole period and that should result in zero.
Now this is very fundamental and it is called the Volt Second Balance equation. So, this
volt second balance equation has to be satisfied for any coil or any inductor in any circuit
and under steady state it will always be zero.
Now, this is the guaranteed principle which is existing fundamental by nature and we
will use that for developing or deriving the input output voltage relationship of any
converter circuit.
429
Exactly, the dual of the volt second balance across the inductor is the capacitors amp
second balance condition. So, consider a capacitance and the capacitance has a voltage
across it called v c and the current through the capacitance is called i c . In the case of the
capacitance the steady in the steady state the current the average current through the
capacitor has to be zero always and that is fundamental and that is the relationship that
we will be using.
So, which means i c average is equal to zero, if we now draw a graph of the i c current
with respect to time and take an arbitrary current that is pushed into the capacitance, in
order now to have charged build up this area the positive area under the i c envelope and
the negative area of the i c envelope must match and cancel each other out and therefore,
have average current equal to zero in the steady state.
So, this boils down to integral of the current with respect to time and that should be equal
to zero in the steady state and that is called the amp second balance. So, the amp second
balance condition is also a fundamental relationship and that is used for finding the input
output current relationships of any given circuit or converter and the volt second balance
relationship is used for finding the input output voltage relationship for the converter
circuits. Now these two important fundamental relations we will use for finding the
relationship between voltages and currents in various circuits.
430
Prof. L. Umanand
Lecture – 47
Input-output relationship
Now, I will demonstrate to you for the circuit that we had discussed till now, how to get
the Input output relationship by application of the volt second balance. So, let us take this
buck converter circuit which we have discussed and let me mark this current Io; Io is the
output current that flows through the output load resistor Ro, the voltage across the
output is Vo, Vi is the voltage of the unregulated input. Now the inductor, so when you
want to apply the volt second balance you have to look for the inductor or the coil in the
circuit and take the voltage across that coil.
So, you go to the inductor take the voltage across that and mark that v L look at the way I
have put the arrow what it basically means is that when you are making a measurement
using a physical device or virtually you are to mentally keep the common of the probe
here and the measurement probe on the side that is what the arrow mark indicates. So, I
could as well now use the arrow mark on this fashion which means that, if the arrow is
pointing on to the right then you keep the common of the probe on the left and make
431
them use the measurement probe on the right. But here what I am doing is I am keeping
the common probe here and the measurement probe here.
So, let us also mark the time periods for which PT1 is on and time period for which PT2
is on or PT1 is off which is (1-d)Ts. Let us also put a graph here indicating that will
indicate the wave shape of the voltage waveform the inductance and let us divide the
time scale in this fashion I will call this period as dTs time period during which PT1 is
ON (1-d)Ts time period during which PT2 is ON dTs again the second cycle starts and
(1-d)Ts, so these are the time periods. Now let us look at the voltage waveform across
the inductor. So, during the dTs period the pole is connected to T1 and VP is equal to Vi
on this side of the inductor in the steady state this is Vo.
So, in the steady state the capacitor would have charged up we are talking of steady state
conditions only because we want to have the steady state input output relationship and
during that time we know the volt second balance occurs across the inductor voltage. So,
this is Vo here and this is Vi here during the dTs period, so therefore I will say it is
constant Vi-Vo during that time the voltage across the inductor.
Then during the time (1-d)Ts P is connected to T2, so this is grounded the inductor is
freewheeling so let us say the voltage here is zero VP is zero. But Vo is there Vo does
not discharge so quickly we are talking of the switching cycle, so -Vo appears across the
inductance. So, it has to be minus voltage here because then only volt second balance
will occur, so you will see –Vo appearing across that then again during the next cycle
dTs P is connected to T1 you have Vi-Vo and during (1-d)Ts time when P is connected
to T2 you have -Vo appearing across the inductor.
So, this is the inductor current wave shape and during this time what is the volt second
this is a rectangular voltage no need to do integral you can find out the area just by the
height into the time width, so we have (Vi-Vo)*dTs. So, let us calculate the volt second
balance across the inductor, so during this time (Vi –Vo)*dTs is the area of this
rectangle. Now the other rectangle is this which is -Vo(1-d)Ts.
So, that is -Vo (1-d)Ts now that should be equal to zero these two should add up to zero
or basically what it means is these two should cancel each other. So, here I can remove
this Ts so the remaining portions if I simplify you see Vo*d and you have a –Vo*(1-d) if
you take it to the other side they become Vo*(d +1-d). So, basically it becomes 1 and in
432
the side you are having Vo*d, so if you do the algebra in simplification you will get Vo=
Vi*d and that is the input output relationship.
Now, to get the input output current relationship you can use the amp second balance
average current flowing through C is equal to zero, but for that you have to understand
the waveforms the currents that go through which we have not yet come to that point.
However, you can do the amp second balance. Later on I will show you how the amp
second balance also work, but I have an easier method for you right now.
So, let us say I will define this as I and let us find the input output current relationship.
So, to find the input output current relationship I will make use of the fact that the
efficiency of the circuit is 100% meaning there is no lossy component within. So,
Pi=Vi*Ii and Po=Vo*Io and I must being efficiency is 100% no loss within. So, in such a
case Po = Pi or Vi*Ii= Vo*Io.
Now, I know the input output voltage relationship here, so let me apply it here Vo I will
replace it by Vi*d. So, you have Vi*d* Io= Vi*Ii, so remove Vi they cancel of and you
have Ii= d*Io. So, this is the input output current relationship you can get it through the
capacitances amp second balance also, but that is sometime later after we know how the
waveform current waveform is through the capacitor.
433
But for now you can use the power relationship or the power balance relationship
knowing that there are no dissipative components switch is either OFF or ON no
dissipation here the ideal sense inductance is a energy storage device, no dissipation
capacitance is a storage device potential storage device no dissipation again.
So, all these components or dissipation less devices in the ideal sense, therefore if I take
100 % efficiency I will get this input output current relationship knowing the input
output voltage relationship. So, this is the formal way of using the volt second balance
across the inductor to get the input output voltage relationship you can use it for any
circuit however complex it is.
434
Prof. L. Umanand
Lecture – 48
Buck converter – operation and waveforms
Let us now try to understand the Buck Converter. So, there is this unregulated voltage
source V i here and this is the single pole double throw switch, an inductor followed by
capacitance. Now this blue portions are the inner inside components of the buck
converter, the black portion are external not under your control for design and this is the
load. So, let us complete now this is the buck converter circuit we have seen and
discussed.
Let us try to understand little bit more in detail. This is the pole, this is throw 1, this is
throw 2, L, C, Ro, V i unregulated, this is dTs meaning that pole and throw 1 are
connected for dTs period of time, pole and throw 2 are connected for (1-dTs) period of
time, this is Vo, this is Io, this is i L inductor current and this is the i c the capacitor
current, let us also indicate the voltage V L across the inductor.
Now, let us try to understand the circuit visually graphically by looking at the
waveforms. So, I am going to have the time on the x axis and the time is split into dTs,
(1-dTs. dTs, (1-dTs) are two switching periods. So, let us first draw a waveform Vp with
435
respect to time. So, during dTs, p is connected to T 1 , so it is V i, in then during (1-dTs), p
is connected to T 2 it is 0 then during dTs connected to V i and so the waveform had the
pole P V P is given here so it is V i and V i .
Next let us try to draw the voltage V L this also we have seen, so during the time when
during the time dTs when the pole p is connected to T 1 Vp is V i and on the other side of
the inductor it is Vo, V i -Vo is appearing here and here when p is connected to T 2 Vp is
0 and on other side you have Vo and you have -Vo coming across the inductor. So, it
repeats every cycle in this fashion.
Next let us have a look at i L the current through the inductor, now current through the
inductor has 2 components one component the Io component and other component i c
component. See the Io component is the DC current average current. So, it has to flow
through Ro because capacitor current has an average. So, entire DC current or the
average current goes through Ro and the average zero current flows through i C .
So, if you look at these 2 parts let us say this DC part is Io and the AC part the part that
flows through i C is given in this form. let us say we draw this line how did I get this line.
So, that is obtained from the Faradays law voltage across the inductor is given by
L*di L /dt. So, this means that di L /dt has a slope of V L /L, now if you see here the voltage
V L during the dTs period is a fixed constant value V i -V o so it is constant. So, V i /L; L is
a constant so V i /L is a constant, so therefore the slope of di L /dt has to be fixed therefore
a straight line having a slope (V i –Vo)/V L .
Now, during this time the voltage applied across the inductor is negative and the slope
will be -Vo/L and because it is -Vo/L, so negative slope. So, -Vo/L positive slope
negative slope so on it keeps going. Next let us have a look at i C the capacitor current see
Io is the fixed DC value remove the Io component from iL what remains is the AC part,
so the ripple part so you will see the i C are the a capacitor current contains the zero
average portion or just only the ripple portion of the current of the inductor current.
So, it will have the same ripple envelope wave shape average is zero, next let us look at
current through T 1 P when pole is connected to T 1 there will a current flow when the pole
is not connected to T 1 current flow is zero. So, let us say T 1 P in this direction, so you
have let me draw this Io level because it is the inductor current that has to flow either
through throw T 1 or through throw T 2 . So, the level the inductor current has to be level
436
as to the P the same, so this portion of the inductor current which I am indicating will
flow through T 1 P.
So, let us indicate that and this portion of the inductor current will not flow through T 1 P
because during that time (1-dTs) time T 1 is cannot connected. So, T 2 P if you see is
mutually exclusive T 1 P you will see that the current flows during the other portion (1-d)
portion. So, this portion of the inductor current flows through T 2 P that is here so the
inductor freewheels in this fashion.
So, when the T 1 P is connected inductor is charging up like this it is going up inductor is
charging up and then when it moves in this fashion and when pole is connected to T 2
inductor is discharging it is falling. So, it is freewheeling through this so the current flow
is in the direction T 2 P current flows in the direction T 2 P. So, if you take the inductor
current peak to peak ripple we will call that as i L . So, if you take the peak value of the
current flowing in T 1 P it is Io level (Io+Δi L /2) this from here to here it is Δi L . So, from
level Io to peak it is Δi L /2.
So, this can be used for designing the switches rating the switches, then let us look at
voltage across T 1 P these are all useful waveforms to rate the switches. So, during the
time when dTs, T 1 P is connected 0 voltage when it is not connected, P is connected to
T 2 . So, p is at 0 potential T 1 is at V i potential it has to support V i during the off
condition.
437
So, this is the voltage across T 1 P voltage across T 2 P, so voltage across T 1 P so actually P
voltage P is more positive compared to T 2 . So, during the time and T 1 is connected to P
is at a positive potential compared to T 2 because T 2 is always at ground. Therefore, P
being more positive I am putting it as V P T 2 voltage is from P to T 2 . So, it is supposed to
withstand V i here 0 V i and 0 in this fashion, so these are the important waveforms that
you should obtain even on the oscilloscope when you use the buck converter circuit.
438
Prof. L.Umanand
Department Electronics Systems Engineering
Lecture - 49
Buck converter – component selection
The Buck Converter has 3 important components to be designed or rated, one is the
single pole double throw switch the other one is the inductor and the third component is
the capacitance. Now let us take one by one let us take the inductor. So inductor L value
how to find the value of L? We will use the Faraday equation V L = L* ΔI L /ΔT. Now we
see that in the inductor current waveform it is a straight line linear because the rate is
constant.
So, therefore, I can take it as ΔI L /ΔT and ΔT is nothing but dTs here or (1-dTs)
depending upon which period you want to take. So, let us say I am talking about dTs
period, so in the dTs period the slope is the current traverse through ΔiL that is this in a
period of dTs in a period of dTs. So, for that the voltage that is seen across inductance is
Vi-Vo which will put it down, other way I can look at the other portion 1-dTs portion
also. So, L*ΔiL the current traverses through ΔiL in time (1-dTs) and during that the
magnitude of the voltage is Vo, you can put only the magnitude which is Vo because, the
439
traversal is negative you can put negative Vo negative negative can be removed and you
have Vo.
So, either I can use this equation of this equation because let us say for example Vi; Vi
can be expressed in terms of Vo as Vo/d substitute here and you will get all in terms of
Vo. So, the value of L is Vo*(1-d)*Ts/ΔiL will bring it down call it fs switching
frequency Δ iL switching frequency. Now this is the basis for the design equation for L
value Vo is a design spec given for you based on the output load side spec d is obtained
designed from Vi and Vo both are specs and therefore d is calculatable fs is the
switching frequency. The designer chooses whether you want to switch at 20 kilo Hertz
50 kilo Hertz 100 kilo Hertz.
Then ΔiL again is a design designer spec and normality it is taken as 10 % of Iomax Io.
So, with this you can get the value of L.
Let us now obtain the capacitor design equation so capacitor C. So, the capacitor current
is the important reform parameter that we are to look at just like in the case of the
inductor design we look at the inductor voltage; voltage across inductor here you look at
the capacitor current. So, let us remove some parts of the waveform focus on the
capacitor current.
440
Now, let me shade this, this is the area under the capacitor current curve on the top and
this is the change in charge we will call that one as ΔQ. So, the change in the charge that
is charge up and the area here would be charge down discharge. So, the charge up and
the discharge should be equal and should cancel out for amp second to balance. So, ΔQ
ΔQ we know is equal to C into ΔV change in the voltage across the capacitance.
So, let us try to find what this area is? This area is nothing this is a triangle, therefore half
base into height. Let us find out the height; height is the distance between these 2 from
zero to peak of the triangle and that we know is that ΔiL/2, because this height this whole
height from here to here is ΔiL. So, this is ΔiL/2 and what is the base of the triangle, so
that is the distance from here to here. So, what is this distance? So, here if you see that
this is from -ΔiL/2 to +ΔiL/2.
Now, this is linear therefore this has to be a midpoint, so therefore this must be dTs/2.
So, this is dTs/2 likewise this is the midpoint in the (1-d)*Ts portion, so therefore you
have (1-d)Ts/2. So, together you add up (dTs/2)+(1-d)Ts/2 which will be Ts/2. So, this is
Ts/2 irrespective of the value of duty cycle this is Ts/2 which is the base.
So, now we can calculate the charge this is half base into height, so half base Ts/2 into
height ΔiL /2 and that should be equal to C*Δ V. Now see let us calculate C=ΔiL/Ts*I
will take it down as fs switching frequency this 2*2 *2 =8 . fs*ΔV comes down that is
the ripple across the output. So, this is the design equation for basis of design equation
for C, ΔiL, design parameter while designing V L we took it as 10%of the max Io fs
switching we know what switching frequency we are designing for ΔV the output
voltage ripple due to capacitor action is output spec.
So, we know that so based on that we can calculate what would be the value of C. Of
course, in a practical capacitor there will be ESR drop across ESR also will come into the
ripple, but anyway we take here the ripple across the capacitor due to that capacitor
action which is equal to I=C*dVc/dt.
Now, let us look at how we will implement this single pole double throw switch, off
course it cannot be implemented by a single semiconductor device it needs at least 2
semiconductor devices and let us see how we select the appropriate semi conductor
device for an for appropriate throw.
441
So, let us make some space here, now consider for example, a single pole single throw
switch single pole a single pole single throw switch like this like that of a relay contacts
how it is iv static characteristics so draw the v versus i. So, v is the voltage across the
SPST switch, and I is the current through the switch. So now let us say that this is throw
and this is the pole and let say that positive direction is current flow from throw to pole.
Now does this switch allow positive direction current flow yes.
So therefore, we will mark this like this, will it allow current to flow from pole to throw
this SPST switch like the relay contact will definitely allow. So, which means this is also
a potential operating point. Now coming to voltage can throw to pole support a voltage
the throw as positive and pole as negative, when the switch is off can the voltage that
needs to be supported by the switch be positive like this throw positive pole negative yes
in this case this is possible therefore, you can say that these are possible operating points.
Likewise we can also find out whether the switch can withstand sustain voltage in the
negative direction yes, so this also become an operating point.
So, for the SPST switch static iv characteristic though entire y axis current axis and the
entire voltage axis or operating points, which means the switch can support positive
voltage during OFF condition it can support negative voltage during the OFF condition
and during on condition it can support positive direction current flow it can support
negative direction current flow.
442
Now take for example, another switch a diode; diode is a switch with the anode cathode
like this, so is it static iv characteristic as a switch. So now, let us consider does this
device allow current flow from anode to cathode YES, so that if we call as positive
direction then we say this is the set of operating points. Will it allow negative direction
of current flow that is from cathode to anode NO that will not be allowed by this device
therefore, there are no operating point the negative access of i.
Will the device the diode support voltage from cathode to anode YES, but cathode to
anode is negative voltage yes can you support positive voltage and anode to cathode
during which time will be forward bias NO, therefore there is no operating point in
positive direction of the voltage axis. So, this means that diode can have operating points
on these marked segments, where it can support negative voltage when it is OFF and
when it is ON it can support positive current to flow through it which is an anode to
cathode. Take for example, another semiconductors switch like the BJT or a MOSFET.
So, you have the collector base emitter gate drain and source. So, how do the iv static
characteristics for switch appear like if these 2 are operating as switches. So, when they
are ON does these device allow current to flow from collector to emitter if you consider
that as positive direction drain to source if we consider that as positive direction then yes
all these are operating points. Does it allow emitter to collector current flow or source to
drain current flow NO, so therefore, the negative portions negative portion of the i axis is
are not operating points.
Likewise, now for the voltage can a positive voltage collector to emitter be supported by
BJT during OFF condition or drain to source voltage positive drain to source voltage be
supported by the MOSFET during the off condition yes it can do that, therefore all these
are potential operating points. On the other hand negative voltage for both BJT and
MOSFET cannot be supported by this and therefore the negative axis cannot be
operating points.
This indicates that doing OFF condition this device can support only positive voltage and
during ON condition it can support only positive current. Now with this background let
us get back to our buck converter single pole double throw switch, now there are 2
throws T1 and T2 PT1 is one semiconductors switch PT 2 is another semiconductor
443
switch. So, let us take one static characteristic for PT1, so we will call this voltage T1 P I
am taking T1 as the first note T as the second note i T1 P.
So, the positive direction of current flow is T1 to P and positive for the voltage support is
T more positive than P. Likewise for T to P I am taking V T to P throw to pole and i
throw to pole. Now let us see how they how this characteristics get filled up and then we
will map up them to the available device, now take for example, PT1; in the case PT1
you can see the; you can see the current through T1 P the current through T1 P is
positive only positive there are no negative direction currents.
So, all the current that flow through d T1 P is positive and if we go back to this page you
will see that the positive axis of current are all possible operating points, because the
current in this pole throw switch is always positive. Likewise the current through the
pole throw to pole T 2 to P is always positive, observe that T 2 to P; T 2 to P the current
flow is always positive. So, here to you will see that the possible potential operating
points lie on positive current axis, likewise when you talk of the voltage supporting when
PT1 is off T1 is more positive than P because when T1 is OFF P is connected to ground,
so T1 is more positive then P.
So, only the positive side of the voltage axis are potential operating points for PT1 and in
the case of P T 2, P is more positive than T 2 is grounded 0 potential P is either 0 or V i.
So, when P T 2 is OFF P is connected to Vi positive potential, therefore VT to P is
negative. So, only the negative portion of the axis are possible operating points. So, from
this set of requirements of the switch i v we can say that this matches with the BJT or
MOSFET or IGBT type of iv switch characteristics.
So, we can choose a BJT or MOSFET for P T1 switch or even an IGBT and this matches
with the characteristic of a diode. So therefore, we can say diode is the best suited for T 2
P, so we can place a diode somewhere like this here and we can place a BJT or a
MOSFET or an IGBT in the T1 P place and these 2 together will form the single pole
double throw switch the equivalent function.
444
So, let us place these components in place of single pole double throw switch and see
how the circuit will look like. So, let us erase those portion we will place the BJT here in
this direction T1 to P this is the pole diode in this direction this is the pole this Q and this
is D. So this is the buck converter circuit with semiconductor switches in place. So, when
this Q 1 is made ON you will see that current flows in the direction charges of the
inductor and then back, when you switch OFF Q1 the inductor current will pass through
this and this and free wheel throw the diode automatically and the energy will start
discharging and therefore the following slope in the current.
So, this is how the switches are selected for a converter for any general converter. Now
let us look at the waveform let us see how you will rate Q and D. Now Q is T1P so T1 P
the voltage V T1 P look at this when this Q is ON it carries current i T1 P is one and
when the Q is OFF the voltage is withstanding capability should be at least Vi comes
across this. So therefore, if we take Q 1 the V CEO rating for Q 1 should be greater than
V i. So, it should be greater than this height and Vi max because this unregulated this can
swing from Vi min to Vi max it should be able to withstand Vi max.
The current rating current max rating flowing through this BJT or MOSFET should be
capable of withstanding this much. So, it should be greater than I naught plus Δ iL /2,
likewise for the diode the peak inverse voltage rating for the diode the negative voltage
that diode should be able to with stand is P T 2 and this is Vi and therefore it should be
445
greater than Vi max and the current IDm max current that flows through the diode is here
shown here T 2 P and this again is i naught plus Δ iL /2, therefore IDm should be greater
than i naught plus Δ iL /2.
Other currents like average and RMS currents can be easily calculated from this wave
shape and there be we get the ratings of the 2 semiconductor switches. So, now we know
how to calculate value of the L, how to calculate the value of C how to choose the
semiconductor switch and the semiconductor switch ratings Q and D for the buck
converter.
446
Prof. L.Umanand
Indian Institute of Science Bengalore
Lecture - 50
Primary configurations
Consider this Buck Converter topology here we used one SPTD switch single pole
double throw switch 1 inductor and 1 capacitor, using one of each of these components
we can have 2 other configuration. So, let us just take only those parts see Vi will be on
the inputs side the outputs side V naught we need to have a capacitor on the output side
to filter to get filtered V naught, so the capacitor will always be on the output side.
So, the position of the single pole double throw and the position of the inductor can
change, however note that the inductor current cannot be discontinuous. So, the inductor
cannot be on the throw side of the single pole double throw switch. If suppose the
inductor was on throw 1 or throw 2 and whenever the pole was not connected to those
throws, the current there would be 0 and the inductor current cannot go in a
discontinuous way to 0. Therefore, it is also constraint that L has to be connected to the
pole.
So, now with these constraints we have 2 other positions where we can connect the
single pole double throw switch and the inductor. So, let us look at these 2 possibilities.
447
So one possibility is to how the inductor on the input side flip this single pole double
throw and the inductor. So, if you do the flip so you have the inductor first and then
comes the single pole double throw switch, like this you have the input and the output
capacitor is connected to the output and to the input we have the unregulated DC input
voltage and to the output to connect the load like this and complete this single pole
double throw switch connection.
Now, this is the single pole double throw switch L this is the pole this throw 1 throw 2,
C, Vi, R naught, V naught measured in this fashion Vi measured in this fashion this is iL.
Now such a converter where inductor is on the input side and the single pole double
throw switch is configured in this fashion is called a Boost converter.
Why it is called a boost converter will become evident shortly when we discuss the input
output relationship V naught will always be greater than Vi therefore, it is called a boost
converter. Then the other possibility see in the buck converter the inductor is on to the
output side in the boost converter inductor is on input side one other possibilities is to
have the inductor connected to the ground.
So, let us do that possibility also you have the single pole double throw switch in this
fashion we have the pole here, inductor connected to the pole and other end connected to
the ground and the throw L end of the throw connected to the input and other end of the
throw other throw connected to the output. So, complete the circuit in this fashion and
you have the output Vi, R naught, v naught, and Vi, L, C, this is pole throw 1 and throw
2.
So, this is called a Buck Boost Converter, why it is called a buck boost converter again
this will become evident that we derive the input output relationship V naught can be less
then Vi or V naught can also be greater than Vi depending upon the value of the duty
cycle. Therefore, this is this can do the buck operation or even boost operation therefore
this converter is called buck boost converter.
So, you see in the case of the buck converter you have the inductor on the output side, in
case the boost converter you have the inductor input side and the case of the buck boost
converter the inductor is neither on the output nor output is in between connected to
ground. In all the cases inductor is connected to the pole inductor is connect to the pole
in the boost convertor also, inductor connect to the pole in the buck boost also and the
448
capacitor is to the output connected across the output to filter the output voltage. So,
these are the 3 primary converters the buck, the boost and the buck boost converters they
are called the primary converters.
449
Prof. L. Umanand
Lecture – 51
Boost converter
Consider the boost converter circuit. We have this inductor, single pole double throw
switch and the capacitor. Let us see how it operates and the various waveforms. Let me
mark this I naught current flowing through the output load resistor, I in current flowing
through the input source V i. Let me mark the voltage across inductor is the voltage
across the inductor V L with the common point the probe placed here and the measuring
point the probe placed here.
Now, let us discuss its operation and wave forms visually and get some insight into this
converter topology. Now I am going to display the wave forms here and I have split the
time axis into four parts as before. Now let us say that P T 1 is going be on during d T s
period and P T 2 is going to be connected during 1 minus d T s period. So, let me mark
this time periods here on the graph d T s and 1 minus d T s; d T s and 1 minus d T s
together form a switching period T s.
450
So, let us plot first the voltage V L across the inductor and during d T s, the pole P is
connected to T 1 pole P is at 0 voltage and the other side of the inductor is at V i voltage.
So, a voltage V i measured positive like this.
Now, during this period 1 minus d T s 1 minus d T s T is connected to T2. So, the
potential here is V naught V P or the pole potential is V naught and this side the inductor
is till V i. So, voltage across the inductor is V i minus V naught. Now is V i minus V
naught positive or negative? We know that the inductor cannot support an average
voltage. Therefore, the average voltage has to be 0. In steady state which means in steady
state during the one minus d T s portion, there should be a negative voltage so that there
is volt second balance.
So, that can happen because the voltage across the inductor during that period is V i
minus V naught that can happen only V naught a greater than V i. So, it is automatic that
V naught will be greater than V i in order to have the volt second balance happening. So,
this will become negative like this and so on to the next cycle. V i and V naught V i
minus V naught. So, this is the voltage waveform across the inductance.
So, now, let us now see the current waveform through the inductance. I am drawing this
average current here; this average current flowing through the input flowing through the
same inductor and the inductor has two parts the average part and the ripple part.
So, let me call this I in average and let me draw the ripple. I have drawing the ripple as a
straight line and it has a slope V is equal to L d i by d t. So, d i by d T slope have placed
it as a straight line linear fixed slope. The voltage across the inductor is V i during that
time constant L is a constant; V i by L is the slope which is the constant. And during this
time V i minus V naught is negative. So, it is a falling slope V i minus V naught by L is
the slope. So, it repeats for the next cycle and so on. So, this is the inductor current
waveform.
Next let us see the waveform of the switch P T 1. So, let me draw this I average level
position I average and i PT 1 P T 1 current flowing from the inductor into this and when
this open the inductor current cannot be discontinuous its start flowing into P T 2. So, the
inductor current inductor current flows through P T 1 during the time d T s when P T 1 is
on. And only during d T s time you will see the inductor current flowing through P T 1,
otherwise P T 1 is 0.
451
Next let us see what is the current through P T 2. P T 2 So, during P T 1 the inductor
current flows through like that and during P T 2 during 1 minus d T s the current flows
through P T 2 and that is same as the inductor current portion during that time. So, this is
the current through this P T 2.
Now, what is the average of the P T 2? Average of the current through P T 2 flows
through capacitance and through the output load resistance. We know that the average
the current through the capacitor has to be 0. Therefore, the average has to be I naught.
So, average is I naught. What is the current through the capacitance? The current through
the capacitance is this waveform P T 2 waveform shifted down by the average because
the average cannot be any finite value here, it has to be 0. So, we have a 0 average
current flowing here in this fashion where this amount, but the amount by which it has
shifted down is I naught.
So, these are the various waveforms that we see for the boost converter and using this
waveforms, let us find the input output relationship and the values of the inductor
capacitance and what goes on to form this single pole double throw switch. Let us now
try to find the input output voltage relationship for the boost converter. So, for that we
have to use the volt second balance equation. So, let us apply this volt second balance to
what voltage you will be applying it to the inductor voltage. So, look for the inductor;
look for the voltage across the inductor which is this waveform and let us apply the volt
second balance.
Now, during d T s this is the volt second area, we can use, we can use a simple
relationship of a rectangular area because this is a constant voltage would. So, let us
apply that. So, this is V i into d T s V i is the amplitude d T i d T s is the width and this is
the area the rectangle during the d T s period. Now during the 1 minus d T s period the
voltage across the inductance is V i minus V naught. So, this area we will calculate V i
minus V naught into 1 minus d into T s.
Now this area positive area negative area should balance out. So, therefore, that should
be equal to 0. I will remove these 2 variables for the picture. Now the remaining
variables if you simplify you have a V i into d and you have V i and V minus V i into d.
So, this V i into d and this V i into d will cancel out, you have a V naught into 1 minus d
push it to the other side. So, you will end up with V naught equals V i by 1 minus d.
452
Now this is the input output voltage relationship for the boost converter, observe 1 minus
d comes to the denominator here; d takes on values from 0 to 1. If d is 0, V naught will
be V i if d is one V naught will tend to infinity, but it will not actually tend to infinity d
cannot take value 1 because then this will be permanently on and there would not be
switching.
So, it should be less than d should be less than 1. So, V naught greater than or equal to V
i and d takes on values between 0 and 1 in this fashion. So, this would be the input output
relationship the boost converter and because V naught is always greater than V i we call
this as the boost converter. Let us also obtain the current relationship.
We could use the amps second balance across the capacitance or like we did last time we
could use the power balance
Pi = Po
But it is not a straight forward because a trapezium and find the trapezium area, we need
to know this heights finding this area is easy because we this is a rectangle with height I
naught, but may be difficult to find this in a straight forward manner. So, we use this
current relationship we obtain a current relationship using the power balance which is
much easier.
453
Now, let us find the values of the components L C and then identify the switch elements
this of this single pole double throw switch. So, first the L consider the inductor L, what
is its value? Take the inductor current wave form and we know that the height of this
current ripple; let us set it as delta i L. Now this is the change in the inductor current. So,
we will use a Faraday’s equation V is equal to L d i by d T and we know the slope of this
V i by L we have written here V i minus V naught by L for this part of the slope.
L = Vi d /∆iL fs
Here you know V i; V i is a input spec of the un regulated dc source; d how do you find
d? We know we know V naught; V naught is again output spec. So, knowing V naught V
i, we can use input output relationship and find d, f s is the designer parameter at what
switching frequency are you switching these switches could be 20 kilohertz, 50 kilo hertz
or 100 kilo hertz.
So, whatever the designer fixes specifies. So, that will be the switching frequency here
and delta i L. Now here i L unlike in the case of the bug converter, the average value the
inductor current was I naught in the case of the boost converter the average value the
inductor current is the input current average. So, 10 percent of this input current average
will be delta iL. How do you know the input current we know V i we know P naught P
454
naught and P i are same and therefore, P i by V i will give you I i; so, you 10 percent of
that will give you delta i L. So, all these parameters are determinable from the specs and
you can find out the value of L.
Next let us find out the value of the capacitor C, for that we have to look at the capacitor
current waveform and we know from amp second balance that this area the top portion of
it above the 0 line and the bottom portion of the area below the 0 line, they should match.
Now the area under this i C curve will be the charge Q. So, in every cycle this amount of
area will be delta Q. So, let us use like in the both buck converter case, we will try to find
what is delta Q and delta Q we know from physics is C into delta V. So, what is delta Q?
It is the area covered under the i C curve i C envelope. In this case between one minus d
in T s period and d T s period, it is much easier to find the delta Q of this because this is
a rectangle of height I naught and width time width d T s.
So, therefore, delta Q i will replace it by I naught, I naught is a height of this rectangle
and the width of that the time in the time scale is d T s. So, that is equal to C into delta V.
Now C we can write it down as I naught into d T s. I will take it to the denominator as f
s; f s is 1 by T s and delta will come down delta V will come down into f s. So, this is the
relationship for the capacitor value. Do you know I naught? I naught is obtained from the
output spec. Whenever a power supply is designed P naught is given P naught P naught
and V naught are given the voltage of the output and the power that needs to be delivered
455
to the output, d can be found in the input output voltage relationship, f s is a designer
parameter, delta V is again an output spec what should be the output voltage ripple and
using that find out the value of C. So, this is the basis for the design the basis for the
design of the C for the boost converter.
Now, let us select the switches for the single pole double throw switch and find out the
ratings. So, the switch elements like we did for the bug convertor, we will use the static
characteristic for the switches, the i V characteristic. So, consider first P T 1; let us
consider this switch P T 1 and then afterwards P T 2. So, let us say we draw the i V
characteristic of v P T 1 and i P T 1. And then let us match the requirement with a
practical switch characteristic and then select the component.
So, now, if you take i P T 1 current flowing from P to T 1 is i P T 1 positive and if you if
you look at the operation of the circuit when P and T 1 are closed, the current i L flows
in this direction from P to T 1. You can also observe i P T 1 wave shape here; all are
positive there is no negative component of i P T 1 which means there is no current flow
from throw 1 to P all current flows are P to throw 1. So, only the positive current access
or potential operating points. So, let us mark that.
So, these are potential operating points meaning that P T 1 should allow positive
direction of current flow from P to T 1. Now for the voltage, when the switch is OFF
when the switch is OFF what should it support what is the voltage of P pole P? So, when
456
T 1 is off, P and T 2 are connected. So, T and P and T 2 are connected the output voltage
will come at this pole point. So, the pole voltage will become V naught; V naught is
positive T 1 is at ground potential.
So, P T 1 is a positive voltage and the switch P T 1 during OFF condition should support
or withstand a positive voltage which means all the positive voltage points on this axis
become possible operating potential operating points. So, we expect a switch to have this
kind of a characteristic and we saw while we were discussing the buck converter that it is
similar to a BJT or MOSFET characteristic or even an IGBT. So, I can use a BJT or a
MOSFET here in place of P T 1. And for P T 2, let us draw the i v characteristic v P T 2
and i P T 2.
So, looking at the inductor current we know that the current flows current used to flow
from P to T 1 and then when you switch off P T 1 and connect to P T 2, the current has
to flow from P to T 2 in this direction. And you can also look at P T 2 current wave form
all are positive there is no negative, meaning there is no current flow from through 2 to
P. So, therefore, we can say that all the positive current access can be potential operating
points.
Now, when the switch P T 2 is off meaning when P is connected to T 1 during that time
P T 2 is off, what is the voltage it will with stand? T 2 is at V naught potential, P is at
ground potential when P is connected to T 1. So, P T 2 is negative and therefore, you will
see that it has to withstand all negative possible negative potential. So, this character is
provided by a diode. So, therefore, P T 2 should be a diode. Now let us replace the single
pole double throw switch with the selected types of the switches; let us say BJT or
MOSFET and the diode at the respective places.
So, let us erase some portion of this P T 1 and then introduce a BJT wherever I have
introduced BJT, you can as well introduce a MOSFET or also an IGBT all are valid
switches in those places; they are controlled switches. So, I have put this BJT in here to
match for this characteristic requirement, then this P T 2 we said matches a diode. So, we
will make some space there and introduce this diode here.
457
So, let us clear up this is called Q and this is D. So, this becomes the complete topology
of a boost converter with the semiconductor switches in place. What are the ratings of Q
and D?
So, let us say Q the ratings of that so, first the voltage rating V CEO rating. So, during
the OFF condition, what is the voltage that Q has to with stand? During the OFF
condition, the inductor current is passing through the diode to the output. So, the diode is
conducting. So, V naught comes as the pole voltage here. So, Q should with stand at
least the V naught amount of voltage. So, V CEO should be greater than V naught
definitely, then what should be the max current rating I c m rating collector current max
rating should be greater than. Here you see in i P T 1 this point here is the max current
that occurs and that is i in average plus delta i L by 2.
So, this is I in average plus delta i L by 2 and I in average can be related to the I naught
current we know, we develop the input output current relationship where I i is equal I
naught by 1 by 1 by I naught by 1 minus d. So, therefore, I will say I naught by 1 minus
d is this I in average plus delta i L by 2. So, in terms of known items d is known I naught
from specs delta i L by 2 is known you can say what should be the current max spec for
Q, I naught by 1 minus d plus delta i L by 2.
So, for the diode rating, you have the peak inverse voltage. Peak inverse voltage should
be greater than what? So, when Q is conducting the diode D is OFF and it is seeing the
458
peak the inverse voltage. So, during the time when Q is ON, the pole point is connected
to the ground. So, pole voltage is 0 and the diode cathode is connected to V naught. So,
therefore, at least V naught amount of reverse voltage the diode has to withstand. So, it
should be greater than V naught and I d maximum is like here you have i P T 2 the max
point is same I naught by 1 minus d plus delta i L by 2.
So, all other currents like average and rms currents can be calculated from the wave
shapes. So, this would be the way you choose the ratings for the semiconductors
switches of the boost converter.
459
Prof. L. Umanand
Lecture – 52
Buck-Boost converter
The third of the primary converter is the Buck Boost Converter it can do both operations,
it can do buck operation and also boost operation depending upon the value of the duty
cycle or the duty ratio. Now let us draw the circuit, you have the input voltage source
single pole double throw switch connected like this, in the case the buck converter the
inductor was on the output side, in the case the boost converter the inductor was in input
side, the case the buck boost converter the inductor is in the middle connected down to
ground. Then you have a capacitor connected across the output in this fashion and R
naught.
So, this is V i this is R naught this is V naught measured in this fashion and V i measured
in this fashion you have the pole throw 1, throw 2 C and L. Now, during dT S period let
us say the pole is connected to T 1, so let us say we connected like that. So, which means
the current will flow in this path, current flows in this path and charges of the inductor.
Now during the 1 minus dT S period the pole is connected to T 2 in this fashion. The
460
inductor current should continue to flow, but it cannot flow in the input path because T 1
throw is open now.
So, it has to continue to flow in this path, take this direction and come back and join and
free wheel in this fashion. So, the inductor current continues to flow, but when it is
flowing to the capacitor look at the direction the charge on the capacitor will be plus
minus in this fashion which means you will have a negative voltage V naught will be
negative and I naught will be flowing in this path in this direction. So, that is one
difference that you would find in the buck boost converter with respect to the buck and
the boost topology.
So, this freewheeling action which causes the output voltage to develop is called the fly
back action and in the isolated version we call the buck boost converter; isolated buck
boost converter as the fly back converter. So, anyway let us now try to understand this
buck boost converter with waveforms and try to design find out what are the values of
the L and C and also the type of switches that you will use for the single pole double
throw switch.
Let us now try to understand the operation of the buck boost converter with the help of
waveforms, typical and important wave forms. We have here the single pole double
throw switch P is connected to T 1 for a period of time dT S p is connected to T 2 for a
period of time 1 minus dT S, we will measure the voltage across the inductor v L in this
fashion, common point here measuring pro point here. I naught is as indicated here in
this direction i C is in this direction as we explain, i L current waveform and now we
shall plot the important parameters and try to understand the operation of the buck boost
converter.
Like before I will divide the time axis into four segments dT S, 1 minus d T S dT S and 1
minus d T S together they makeup T s that is one switching period, likewise I have two
switching periods here for the waveform and let us plot first the voltage across the
inductor. So, we will plot V L and during dT S time what is the voltage across During dT
S the pole is connected to throw 1 and the entire vi is coming across the inductance pole
is having the same voltage as v i, so this will be V i. And during the time 1 minus dT S,
the pole is connected to T 2 and therefore, T 2 potential which is actually minus V
naught here minus V naught comes directly across L.
461
So, therefore, you will see V naught coming directly across L and we have seen that the
charge on the capacitor is reverse plus minus in this direction because the current is
flowing like that to complete the inductor freewheeling current path and therefore, V
naught measured in this fashion with the common here and the measuring probe here we
will show a negative voltage value, so that is V naught which is minus.
And then again next cycle V i here, V naught here and so on and so forth it continuous.
So, this is the inductor voltage waveform remember to use this for the volt second
balance equation to obtain the input output voltage relationship. Next let us look at the
current through the inductor i L, so that is another next important wave form that we
need to look and understand.
So, the inductor current will have an average value, so this is the average value which I
am indicating and for now I will call this as i L average because it is neither connected to
the input nor to the output infact it is connected to both, for parts of the time and
therefore, we will see what it is later or now call it as i L average, this is the study dc
portion, on top of it there will be the ripple portion.
Now, the ripple portion as before I am trying a straight line because the voltage across
the inductor is fixed constant V i during the dT S period. So, there is going to be at delta
i L change with fixed slope, fixed rate and that is V i by L. And then, during the 1 minus
dT S period it is going to fall because the inductor is going to discharge into the load at a
constant voltage being applied across the inductor which is minus V naught. We see that
there is a minus V naught applied here, so there is a minus V naught by L and that is that
falling slope.
And so on it continues again for the next positive and the next negative discharge portion
of the inductor current. Next let us have a look at current through T 1 P; T 1 P current,
the current through T 1 P; the current through T 1 p should be that part of the inductor
current that flows through the input source when P is connected to T 1 during dT S
period and that will be the dT S portion of the inductor current. So, we can draw the dT S
portion inductor current 1 minus dT S portion will not flow through the inputs source
because T 1 is open.
So, there is no current flow during 1 minus dT S period again the next dT S period you
will have the inductor current flowing through the input source, so T 1 P will have a
462
current like the inductor current, but only the dT S period part. Now this has an average,
this is the current this i T 1 P is the current that is flowing here and we will call that one
as I i and I i, the average I i current will be will be the average of this wave shape, so this
average is I i average can call it as that.
Next let us look at current T 2 P, the direction of the current through inductor cannot
change, so, what was T 1 P now becomes T 2 P the inductor freewheeling in this fashion
and T 2 is connected to P during 1 minus the d T S period. So, the that portion of the
inductor current that occurs during 1 minus d T S period will flow through T 2 P. So,
during dT S period T 2 P will show 0 current, then it will show the inductor portion of
the current passing through T to P during 1 minus dT S period again dT S period 0 and
so on.
So, this will have an average. Now this current the same current flowing here also splits
in two parts i C and I naught i C is the 0 average part and the average part flows through
the load. So, the average part is nothing, but I naught looking at the figure you can see
that the average part has to be I naught and the non average part flows through i c. So, in
effect if I remove this average part bring down this wave form to 0 average I will get the
current through C that is i C.
So, if I do that exercise I will see that I will be bringing down removing the average part
bringing out the wave form by I naught. So, this amount which it has come down is equal
to I naught, now this current wave shape wave from will have 0 average because the
current through the capacitance has to be 0 average under steady state conditions.
So, these are the various wave forms, various voltages and currents that you will see in
the buck boost converter. Let us now find the input output voltage relationship and the
input output current relationship for the buck boost converter. So, a first the input output
voltage relation let us use the voltage across the inductor wave shape, the voltage across
the inductor and apply the volt second balance for the voltage waveform.
So, this is the portion of the waveform inductor voltage waveform above the 0 line and
this is the portion of the inductor wave form below the 0 line, the area under this portion
of the curve and area under that this portion of the curve should cancel out and balance
each other out in the steady state. So, that is what we will be doing, applying the volt
second balance, so let us consider this area the volt is V i second is dT S plus take this
463
portion of the area during 1 minus dT S period is V naught into 1 minus dT S should be
0.
So this is the volt second balance equation. I will remove this variable and rearranging
you have V naught is equal to minus V i d by 1 minus d, so this is the input output
voltage relationship. Observe the negative sign because of the direction of the
measurement here and because the capacitor charges and reverse here you will see the
negative sign appearing. If you choose chose the direction of measurement in the
opposite direction this will become a positive one, any way we are bothered about the
magnitude of the conversion, so V naught is equal to V I d by 1 minus d.
So, if you look here the absolute value the only magnitude V naught I am considering is
less than V i it is the buck operation for the value of d 0 to 0.5. So, between 0 and 0.5
and the value of d is less than 0.5 you will see that this is less than 1 and therefore, it is
attenuation or buck operation. V naught is equal to V i for d is equal to 0.5 0; 0.5 you
will see the attenuation factor here is 1 and therefore, V naught is V i by magnitude.
And V naught is greater than V i for d between 0.5 and 1, during which time this ratio
becomes greater than 1 and therefore, it is boost operation. Therefore, you see that this
converter operates in buck mode for d less than 0.5 and in boost mode for d greater than
0.5 and therefore, it is called the buck boost converter.
464
Let us now find out the input output current relationship, so let me move this up a bit. So,
the input output current relationship you can use amp second balance, but much simpler
to use the power balance
Pi = Po
And now this portion let me replace it with this equation I will use that equation from the
volt second balance relation V i into I i average is equal to I will only bother about the
magnitude, so V i d by 1 minus d into I naught. So, let me remove these two these two
parts V i and V i and simplifying you get I i average is equal to I naught into d by 1
minus d and this is the input output current relationship for the boost converter.
Thus
L = Vo (1-d)/∆iL fs
465
The i L average is equal to i T 1 P current waveform average which is I i average plus i T

2 p average which is I naught because these two sum up to give you i L. I i average from
the input output current relationship we know is I naught into d by 1 minus d plus I
naught which workouts to I naught by 1 minus d.
So, this is i L average and therefore, is i L average is I naught by 1 minus d, so this also
you can estimate from the specified parameters, I naught you know from the max power
that needs to be delivered that is a specification, output specification d is obtained from
the input output voltage specifications and therefore, i L average is known and therefore,
10 percent of i L average is known and you can find out the value of inductor.
466
Next let us find out the capacitor value, we go again by the change in the charge in the
capacitor with in a cycle,
C = Io d/∆v fs
I naught is known from the output spec, d is calculatable from the input output voltage
relationship spec, delta V is a spec of the output voltage ripple, f s is the designer
parameter which is chosen by the designer depending upon the frequency at which the
single pole double throw switch is being operated. So, using this equation you can arrive
at the value of C.
467
Let us now select the switch elements for the single pole, double throw switch.
Considering the switch elements now, let us find out what is the static switch
requirements? Static characteristics of the switch; now we will take T 1 P and T 2 P one
after the other and see how the static characteristics for the switch looks like v T 1 P on
the x axis, i T 1 p on the y axis.
Now if we take this switch during the time dT S when pole is connected to T 1 there is a
current flow, the current flow is from T 1 to P; T 1 to P being positive and you see that i
T 1 P is always positive above 0 and there is no current flow from P to T 1. So, therefore,
we can say that all the positive axis of the current are possible operating points, then
when P T 1 is open P T 2 is connected, the pole voltage is at minus V naught and T 1
throw voltage is at plus V i.
So, throw T 1 is at higher positive potential as compared to P and therefore, T 1 P should

support positive voltage during OFF condition. So, therefore, all these are potential
operating points. Now this if you look at the characteristic and characteristic of typical
devices you will see that they map to BJT and MOSFETs and even IGBTs. So, T 1 P can
be replaced with a BJT, now when you come to T 2 P V T 2 P and i T 2 P are the
important variables here.
So, during the time when T 2 and P are connected there is a current flow from T 2 to P
because inductor current should not change direction. So, all the positive axis of current
468
becomes potential operating points and during the time and P T 2 is OFF P is connected
to T 1, so P is a positive potential T 2 is connected to the output, so it is a minus V
naught, T 2 is minus V naught P is positive potential, so T 2 P is a negative voltage.
So, the switch T 2 P should support negative voltage. So, all these negative axis of the
voltage become possible operating points and this has a typical nature of a diode switch
characteristic. So, now we know that for this we need to use a diode. let us now use the
BJT for T 1 P and diode for T 2 P replace it in place of this single pole double throw
switch and see how the topology evolves.
So, let us erase some portion making space we will put a BJT here like this, observe the
direction of the current flow is from T 1 to P direction and then we will make some space
here and then put the diode here again the direction of current flow is from T 2 to P
keeping that in mind, so this is T 1 and this is T 2. So, now, you have the BJT and the
diode interposed inside the circuit and this gives you the entire buck boost converter
topology with semiconductor switches.
So, now this is Q and this is d let us write down the ratings for Q and ratings for D, the V
CEO rating of this when the switch is off. Collector is at V i potential and the emitter the
diode is free, the diode is conducting freewheeling during that time the emitter will be
connected to minus V naught.
469
So, V i and on this i minus V naught, so the Q V CEO should be greater than V i minus
of minus V naught V i plus V naught. I Cm the current through that V C i T 1 P this is
the peak current and we know this is i L average, the peak value is plus delta i L by 2, i L
average we know is I naught by 1 minus d and plus delta l by 2.
So, therefore, I C m is I naught by 1 minus d plus delta i L by 2. The d rating similarly

we can find out peak inverse voltage across this when diode is OFF T 2 is at minus V
naught pole because Q is conducting pole is at V i. So, the peak inverse voltage is V i
plus V naught should be greater than that and I D m the peak voltage is same peak
current is same as this I naught by 1 minus d plus delta l by 2 and you can find out from
the wave shapes current wave shapes here, the average and the rms values also if needed.
470
Prof. L.Umanand
Lecture – 53
Simulating the primary converters
Let us now see how we will simulate the primary converters. I will supply in the resource
potion of the course website, the simulation files for the buck the boost and the buck
boost converters. Here what we have; we have discussed everything about the primary
converters accept that how to give the information signals the gate pulse signal to the
controlled switch. So, what I have done here is that I have passed it; I am passing it
through a PWM generator block during simulation and giving a voltages which is
proportional to do recycle.
So, if I want lower duty cycle, I will decrease the voltage; higher duty cycle, I will
increase the voltage and this will compare with the triangle waveform to generate the
PWM and this signals will turn on this BJT. So, inside this PWM block what you may
find is that there is a triangular career. So, this is let us say 10 kilo hertz; I will set it 10
kilo hertz for now, but you can experiment with different frequencies and this carrier is
swinging from minus 1 to plus 1 and then you can have a control voltage a voltage is
471
proportional to duty cycle. And then you will see that after comparison of these two
voltage signal the output that you will get will be something like this.
So, let us say all the at the comparison point you will see that the voltage proportional to
control voltage upon duty cycle is higher. So, let us say we will have high there and then
here the control voltage is lower than the triangle and its higher than the triangle so on.
So, as this control voltage moves up and down, you will see a modulation here
happening in the pulses and these are pulses that will be given to the control will be used
as control signals for Q and this is what you will be seeing at these point. Later on I will
discuss in detail what goes into making the PWM so that you can build other PWM
blocks and techniques.
Let me now show you some simulation examples. In the resources folder I have three
folders; one on buck converter, one boost converter and another on the buck boost
converter. And within them I have the schematic files, net list files in this ar files along
with the edt 01 dot sub. This is subroutine sub circuit file. Now when I open the
schematic file, you will see something like that buck dot sch and this is the circuit. So,
you see that I have here a switch a semiconductor switch controlled semiconductors
switch like a BJT or IGBT. It is a generic switch a Vin a diode here L inductor capacitor
and or not and see here that I have a PWM module which is basically V in a diode here
L; inductor capacitor and r naught and see here I have a PWM module which is basically
472
built of a triangle wave form and I have a source here to represent the voltage
proportional to duty cycle.
So, I have set it at 0.5. So, 0.5 in range of minus 1 to plus 1 approximately it is now 75
percent duty cycle. So, let us see how the circuit simulates. I have put various node
points; I have node point here for the input a node point, here for the pole node point for
the output and node point for the gate voltage. So, to simulate let us go to the command
line; command terminal.
I have opened the command terminal and here we will run the simulation. I will run, I
will use this and run buck. So, it will create the netlist and then run the simulation and
show the waveforms. Now what are the waveform that I have here?
473
I will indicate to you before that let me open the buck dot cir circuit. So, apart from the
dot run statement and includes statement, I have the control statements where the
background of the figure windows is said to y foreground is black and then plotting the
pole voltage; pole the output voltage that is the inductor voltage output voltage and the
inductor current.
These are the things that are being plotted. So, now, back again to the plot, you will see
that the plot goes through transient phase here and then starts to reach the steady state.
So, you would like to look at the steady state waveform. So, you can right click on the
mouse and drag a very small portion on the steady state portion so that you will be able
to see that portion only.
474
So zoom that up. So, this is the steady state portion and you have green color is the i L
waveform pole to output is the inductor waveform pole voltage waveform, V naught
waveform. So, you can study them in detail here take for example, the green colored il
wave form very small here if you want to see you can right click here and then amplify
them by taking a zoomed small portion of that. It will zoom in and you will see that the
inductor current wave shape as we had drawn while discussing the buck converter.
So, I will leave it you to explore change value change duty cycle, change frequency
change the values of the L C’s and explore inside out and try it out and try to learn about
the buck converter. In a similar manner I also have the boost converter circuit schematic
here. So, observe the boost converter circuit, you have the switch here the inductor on
the input side diode here and similarly I have the node labels here, you can also look at
the inductor current likewise I am having a control signal generation based on assign
based on a voltage triangle comparison principal. So, this let us simulate, we will go to
the terminal window and in the terminal window let us run the simulation for the boost
converter schematic circuit. It will generate the net list, it will run execute. There will be
a portion of dynamics and then reaches the steady state.
475
So, what you could do is right click near the steady state. This has still not reached the
steady state, but anyway you can observe. So, take a small portion out of that.
And you will be able to see it much better, green is the inductor current you could
probably have a look at the inductor current by zooming in like this.
476
So, it is just at the boundary of discontinuous conduction the waveform as expected. So,
you can study this further, get greater inside by playing with the values and understand
the boost converter. In a similar manner I am going to show you the buck boost converter
schematic.
So, the buck boost converter schematic is like this, you see the throw one to pole is BJT
or an IGBT as you may consider fit and throw two to pole you have this diode D. Look
at the placement of the capacitance of course, this is PSpice symbol showing plus here,
but you will get a negative voltage the buck boost will give you a negative voltage for as
we had discussed.
So, let us stimulate this PWM portion is like I had told for the other two converts, you
generate the pulses by giving a voltage here proportional to duty cycle. Let us go to the
terminal and let us run the simulation for the buck boost. It will generate the net list and
the waveforms.
477
And you will see that the steady state is around the right side. Let me take a small portion
of that right click and drag an area and you will get a zoom portion of that and this will
give you a much better detail.
So, observe that output voltage the blue line here negative; it is in the negative side.
Change the duty cycle observe that, observe the values, observe the various waveforms.
If you want to have a closer look at the inductance waveform, this is the orange line here.
You will see the inductor current waveform like that.
478
Slightly more amplified zone and then you see as expected, you will see the waveform of
the inductor current in this fashion. So, again I will recommend that you try to
understand as much as you can by changing the values and changing the frequency and
probing at various points of the currents and voltages of the circuit.
479
Prof. L. Umanand
Lecture - 54
Forward Converter
Today we will begin the discussion on Isolated Converters. Till now we have been
discussing on non-isolated converters mainly the buck converter, the boost converter and
the buck boost converter which are called other three primary converters. We will use the
buck converter and the buck boost converter and derive the isolated versions of these
isolation has many benefits.
Firstly, there is no physical connection between the primary side and the secondary side
of the galvanic isolation, the primary side which is the input source side. And, the
secondary side which is the load side if it was isolated there will not be any surges
passing on to the load side and damaging components on the load side. There will be
reduced conducted emission both ways from the source side to the load side and from the
load side to the source side.
Not only that the galvanic isolation gives us one more degree of freedom by means of the
turns ratio. By using trans ratio I can use the buck derived forward converter or the other
versions of the converters to also provide you boost the output voltage can be higher than
the input voltage because, of the trans ratio freedom and you can have multiple outputs
that is the you can have the same converter 5 volts plus minus 15 volts and so on.
So, these are the advantages that you gain by putting in a transformer and most of the
pack practical power supplies will use the transformer as isolated version of the
converter for most of the practical power electronic equipments.
480
We shall now try to look at and study some of the isolated converters. Firstly, we will
begin with buck derived. There are many buck derived isolated converters. We will be
studying the following. One is the forward converter issues involved with that and then
we have the push pull version or the push pull converter.
Then the half bridge converter followed by the full bridge converter. These 4 are the
important converters which are very popular and most power electronic equipments and
we will discuss these. Then on the buck boost derived isolated converters there is one
very important converter which is one of the most popular converters in use and that is
the fly back converter. So, we will discuss the fly back converter also.
481
Let us now see how the forward converter circuit apology is. This forward converter is a
buck derived isolated converter. So, we shall start from the buck converter topology. So,
let us have this unregulated source V i and we know the buck topology another BJT you
have this diode the inductor capacitance and the load. So, this is the buck converter
circuit, you have V i Ro Vo i and c. Now let me do one thin. First let me shift this
transistor which is on the top rail to the ground rail because then we need to drive this
transistor with respect to the emitter. I lifted up floating emitter is much more difficult to
drive than a grounded emitter. So, let me make some space here and let me place the BJT
here. Now the emitter is on the ground side same direction of the current flow to see the
current flows in the same direction.
So, only way it is on the return rail now this emitter is grounded. If I take this is ground
and driving this BJT or MOSFET with respect to the ground is much easier. So, normally
we will do that. So, we will put the BJT or the transistor that is the control switch on the
return rail rather than on the top right. Now we can remove this transistor. So, we will
remove that and close that. Another modification that we will do in the representation is
let us put label symbols for the ground. This is the ground, I will put a label symbol for
the V cc or V i. The positive of the source now, I can remove this without loss of
generality, so that it does not clutter up the circuit.
482
Now that we have removed this, this still represents the same, but converter circuit I will
allow put this switch down vertical like this in this fashion. So, this is still the same, but
converter circuit and I will shorten this. Let me put the label here and remove this
portion. So, this becomes the buck converter circuits, still the same buck converter
circuit only with some repositioned switch positions. Next I would like to insert the
isolation.
So, at that point let me put in the primary winding, then I will put the secondary winding
and join them up. This still naught complete I cannot just join them up here because let
us say when the inductor is freewheeling, when the indicator is freewheeling, this diode
is conducting and it is effectively shortening the secondary. When it shorts the
secondary, the primary will draw a huge current and blow up this transistor.
Therefore, we cannot allow shortening of the secondary when the inductor is

freewheeling through this diode. So, we will put a diode here, so that you do not have a
reverse current flow. I will make some space and put the diode. So, at least now it is safe.
Even if this is prevailing the voltage, this diode will not allow short circuit to happen
across secondary and thereby reflected on to the primary. Next let me mark this V p V s
and this is turns ratio 1 is to 1.
This voltage here let me call it as V A, this is inductor current i L, this is output I naught
current and this is I c, this portion from V A onwards is nothing, but pure buck converter.
So, observe here that the circuit is still incomplete. Let us see what happened now. Let us
say this transistor is on. The moment the transistor is on, there is going to be a
magnetizing current that is going to flow here and the reflected load current also will
flow through here.
When you switch it off, when you switch it off inductor freewheel there is a reversal of
polarity. The dot will become negative, non dot end will be positive, dot will be negative,
non dot end will be positive, this diode will be reverse biased because this diode is
conducting what will come across that will be a negative voltage. This diode will be
reverse biased and off. So, there is no path for the flux, internal flux of this core to reset
meaning when there is a current flow here, you have switched it off. There is a id L by
dt, the voltage across this di by dt has been cut off.
483
There will be a huge voltage because of the negative di by dt and huge voltage will
appear across this and blow up this transistor. Therefore, you have to provide a flux a
path for the flux in the core to reset. So, we will see that there is a path for the inductor
current to flow in this direction here. So, we will make a path in this fashion. We will put
a resistance and the diode like this. So, we will call this as Rm and Dm diode
corresponding to the magnetizing flux reset path. Rm is corresponding to the
magnetizing current reset path.
So, now when you switch this on, there is a current flow through this magnetizing
current and the load reflected current flowing through this and when you switch off,
there is no load reflected component. Only the magnetizing power portion of the current
will freewheel like this. It will freewheel and Dk with Lm by Rm time constant Lm is the
mutual inductions and Rm is this resistance. So, Ly r time constant It will start decaying
and go to 0 and when the magnetizing current goes to 0, the flux in the core is reset and
brought back to its original state, so that you may next you can start the next cycle by
switching this on. Now this is the complete forward converter circuit or working forward
converter circuit.
Let us now try to understand this forward converter circuit. Let us try to understand its
operation and the waveforms at various points. Now if you see this portion of the circuit
the one that I am indicating with the cursor, this portion of the circuit on the secondary
side is exactly like a typical buck converter. So, you have two switches here, two diodes
if you consider this as throw 1 and throw 2 and this is the pole, this is the pole voltage
VA would be the pole voltage have the inductor current capacitor current I naught. All
this portion of the circuit, the waveforms will be very typical like the buck converter
circuits that we discussed.
On the primary side is nothing, but a chopper or a switcher and we will discuss the
waveforms of the chopper and the switcher and also this portion of the circuit where you
are doing core resetting, these are the extra portions that will come apart from the regular
buck converter waveforms. Now, let us relate the regular buck converter waveforms to
the forward converter. Let us look at the buck converter waveforms for now.
484
And revisit these buck converter along with the waveforms here, so that we again
remember these waveforms. This portion the buck converter portion is very much similar
to the secondary portion of the forward converter. All these waveforms are valid. Only
we will allow to look at the notations and change the notations to the forward converter
notations. So, now what I will do is copy these waveforms which we had drawn for the
buck converter and place it along with the forward converter and in comparison change
the values on the notations.
485
So, let us place the waveforms move it. So, here you see that these are the same buck
converter waveforms during dts time we have the switch on 1 minus dts times switch is
off two cycles; so, again dTs and 1 minus dTs. Now V p the pole voltage will map To
VA. So, let us change that. So, I will remove that and put it as VA. So, VA here and
when during dTs the voltage here what will happen when the switch is on? What will
appear is V i and this V i look at the dots will translate to the secondary voltage as n
times V i the n times V i will appear across at VA. So, therefore this V i will change to n
times V i. So, let us make the change n times V i and then during the time when this is
off, the inductor will freewheel through this diode and the potential of VA with respect
to hear the output ground could be 0 and 0.
So, this will be the VA wave form and from the VA wave form all these waveforms are
typical, but converter waveforms for example, the voltage across the inductor V L will
remain unchanged except that here V i minus Vo will become n V i minus Vo. So, let us
make that small change. So, this will become n V i minus Vo n V i minus Vo. This will
remain Vo. When the indicator is freewheeling this potential is 0, V A potential is 0 and
the other side of the inductor is at minus Vo. So, that is what we have indicated here
remains unchanged.
Now let us take the inductor current i L. The inductor current i L also will be similar we
will see the inductor current rises. Now it rises for a change of delta i L in dTs time, it
will be n V i minus Vo by L. So, instead of V i minus Vo by L, it will become the n V i
minus Vo by L. Let us make that change. So, we have n V i minus Vo by L.
On the falling side during the freewheeling time the voltage across the inductor is minus
Vo. So, this will remain minus Vo by L. Look at the ic current waveform i c there is no
change. It will be this ripple component of the inductor current only that will flow
through i c zero average current because second balance has to be maintained. The
average value of the inductor current is I naught, that still remains I naught.
Now, let us mark some symbols here. Q for this BJT and the current that flows through
Q is i Q and let me mark this as d 1 diode this is d 2 this is i d 1 and this is i d 2. So, these
are the currents that I have marked. So, now let us look at some of these waveforms.
Now, look at the waveform here i t one p it one p for the buck converter was the bgt the
controlled switch current. So, when during dts time when the control switch was on this
486
was the current that was this part of the inductor current was flowing through that switch.
So, you still have the inductor current flowing here. So, what will flow through i d 1 will
be equal and to i T 1 p. So let us change this it 1 p to i D 1.
So, it will be that current that is flowing. So, i D 1 this diode will be on whenever Q 1 is
on and only that part of the inductor current will flow through it during dTs time period.
Now if you look at this wave form, I T 2 p it should be same as i D 2. So, I can just
change it i D 2. Nothing else will change in these two waveforms. So, now what is the
current i Q through this primary side switch.
So, when the switch is on the current flow through the primary side switch i Q and i
primary are same and i primary is the reflected current whatever i d 1 n times i d 1 gets
reflector on the primary and i Q is the same as that. So, i Q is equal to n times i d 1 we
can say that. So, therefore if I change this to n times i d 1 it becomes i Q current. So, if I
say n times i d 1 all the values here will be multiplied by scaling factor n. So, let us say
this becomes n and the average value will be n times i naught. So, this wave shape is the
wave shape that you can expect to see through the BJT Q or i Q.
With respect to the primary side switches with respect to V t 1 p, the voltage or the V t 1
p, I will discuss that separately. V p 2 t V p t 2 is the voltage across the diode d 2. There
is no change except that V i becomes n V i. So, I will remove the voltage portions of the
wave form now and I will get back to that in the meantime. Before that I would like to
give the input-output relationship.
487
So, let us write down the input output relationship and for that we need to look at the
inductance voltage wave form that is the voltage waveform across the inductor. We
know that we have to look at the old second balance. So, what is the voltage during d T s
period multiplied by the time period will give the area of this triangle and during the 1
minus d T s period we have minus Vo and multiplied by 1 minus d T s will give the area
of this rectangle.
These two should be equal for volt second balance and that will give us the input output
voltage relationship.
Vo = n Vi d
So, you the designer has the flexibility in the choice of n and therefore, the output can
even be higher than the input. So, in a typical buck converter, the output is always Vo is
always less than V i, but here even though d varies from 0 to 1 0 is less than 1. In this
case the duty cycle cannot go more than 50 percent because if the transistor is on for 50
percent, there needs to be an equal and amount of time needed to give for the core to
reset.
So, therefore normally the duty cycle we keep it at 50 percent and use the freedom of n
to scale it to any level. So, if you have n is equal to 10, you can have Vo much greater
than V i. So therefore, buck and boost everything is possible because of the isolation
488
because of the scaling provided by the isolation, however this n is fixed variable which
cannot be changed dynamically, d is available which can be changed dynamically. So,
this can be taken care at the end at the design time once and d can be used as a control
input to take care of changes in a dynamic fashion. So, d becomes a control input.
So, this is the input output relationship for the voltage. In a similar manner the same
manner we can also get the input output current relationship. We can use the power
balance thing P i is equal to P naught which means V i i i average because the input
current is chopped switched and therefore, the average of that will be V i into ii average
will be equal to P naught which is Vo into I naught.
So, now we can replace this Vo with the input output relationship that we have just
derived and that means, that we can say equal to n V i into d into I naught, this will be
the P naught output power. So, this implies that i average will be equal to n i naught into
d. So, this will be the input output current relationship. So, this value is the average of
this i Q average of this i Q wave form will be this in terms of n i naught and d.
489
Prof. L. Umanand
Lecture - 55
Core Reset in Forward Converter
Till now we discussed on the wave forms on the secondary side of the forward converter
which was mostly like the buck converter. Now let us look at the primary side, this
switcher and the core reset issue. Let us draw the wave from the related to that and then
try to understand that. So, I am going to divide the time again into four parts where you
have d T s d T s the time during which the switch Q is on 1 minus d T s the time during
which switch Q is off d T s again Q is on 1 minus d T s Q is off.
So, let us first consider the voltage wave the current waveform through this diode and
that is what will actually get reflected on to the primary. So, let us mark here the primary
current i p. Remember that primary current and i Q are same only during the time when
Q is on when you how i p and i Q same. The moment Q is off, i Q becomes 0, but ip
does not become 0 because there is a freewheeling current that will happen in this
direction. So, therefore, ip and i Q are naught strictly same during 1 minus d time ip has
some finite current i Q is 0 during that time.
490
So, what is ip? Ip is the reflected secondary current and at the secondary what is flowing
when the transfer Q is on is i D1. So, n times i D1 is what is supposed to be reflected on
to the primary the load reflected part. So, let us take n i D1 which is i D1 into the load
reflected path flowing through the primary. This is one component. So, we know that the
load reflected part during that time inductor current is rising and this is how the current
wave shape look like and this we have seen before and if you take the level at this point,
you know that that is the n times I naught again load reflected part.
Now, next we should understand that this is a transformer, then there is some current
needed to magnetize the transformer and that is the magnetization current. So, a
magnetization current also will flow through this primary and that should be added to
that to get the total ip. So, what will be the magnetizing current? I will say i p mag when
Q is on V i is applied across the primary winding, it is a constant voltage. So, the
magnetizing current from the Faradays law V i is equal to L m di by dt. The magnetizing
current will have a slope of V i divided by L m. So, it will start from 0 like that it will
have a slope of V i by L m, L m being the magnetizing inductance during 1 minus d i
will explain later.
For now we will leave this part and again it will start from 0. During this time we have to
reset the core. We are to reset the core and bring it to 0, so that it will again start back
again. In the meantime let us also draw. Before we draw this portion of the magnetizing
current let us draw what will be ip. Ip is the primary current. The primary current will be
having two components. One is the load reflected part which is n i D1 and other one is
the magnetizing part which is ip mag. These two will add up together to form ip.
So, this part is magnetizing part. I am actually showing this bit zoomed. It will actually
be much smaller as compared to the value of n i D1. So, let us say this is the n i D1 part
which I am writing down here. This part you can add it up here, starts from this value
and then gradually it will go up like that same as this d T s period also.
So, this will be the ip, this green portion will be the total added up ip value during the
time when Q is on. When Q is off, you will see that i Q becomes 0, but ip will start
freewheeling in this path. So, it will freewheel in this path because inductor current
cannot become 0. There is this resistance R m and there is a mutual conductance L m. L
491
m by R m will be the time constant. So, at that time constant exponentially i ip will
decay. Likewise here also you will see that it will decay in this fashion.
So, now what will be the ip? we will in the ip waveform the same thing will repeat. You
will see a drop because the moment Q is off, the reflected component vanishes i D1
component vanishes that becomes 0 and only the ip mag component will come into the
picture and that is what we will find in this.
So, the total ip current we will look like this drawn thick in a thick green line. This will
be the ip current that will be flowing through the primary winding. Let us now see also
what is the voltage across V across Q V CEQ. So, let us draw on more access here and
see V CEQ during the time when Q is on, you know that this is 0 and during the at the
moment when Q is turned off, you will see that there will be a sharp rise. What is that
value we will see that and then afterwards post that rise, you will see that this is V cc
level the V ca will start decaying down to V cc level. Why does this happen?
You see that moment this is switched off, this is prevailing in this fashion. So, there is a
drop across the diode forward drop across the diode, there is a drop across this R m ip
into R m. Current is flowing in this direction ip into R m and there is V i. So, V i drop
across R m and the diode drop will all be reflecting here when it when this shoots up and
then ip current is decaying as you see here it is decaying exponentially. So, the drop
across R m should also decay exponentially, therefore you see a decay exponentially and
once the decay is complete, it will settle at V i because when once this is off, it is only V
i that will come at this point.
So, this is how the voltage across the VJT will look like and then, it will repeat again the
next cycle and decay in this fashion. So, this point I would like to say is the V CEQ
which is equal to V i plus this diode drop V Dm forward diode drop around 0.7 volts plus
the current flowing this direction plus minus here you have I m into R m if I say this
value I m peek at that point I m that I m into R m will be appearing across. So, this will
be the voltage that will come across at this across this transistor Q during the time at the
time at the moment when it is being switched off and then further on during the time on
minus d T s, it will start decaying down like this because I m is decaying down.
Now this decaying time constant is equal to magnetizing inductors L m by R m. Now

what should be the value of R m? So, let us say 5 times time constant 5 times L m by R
492
m, e is the time constant which says that this would have reached steady state, it would
have reached 99.9 percent of the final value. So, by that time this would have core would
have reset if I m comes down magnetizing, current comes down to 0, then you can say
that the core is reset. So, 5 in 5 times time constant the core would have reset, so that it
can take up the next cycle. So, this should be less than this time 1 minus d T s. So, before
the off time is over, the core would have been reset. So, this is one relation.
Another relation is the voltage across Q V CEQ. So, the voltage across that is as we saw
here V i plus V Dm plus I m Rm. This is the peak voltage that the device will see and
that it will see at the point of switch off. Now that should be less than V CEO rating of
the device is the second relationship. Now these two inequalities using these two
inequalities from this you can see that
Rm > 5 Lm / (1-d) Ts
Rm < (VCEO – Vi - VDM)/Im
So, this is the value of Rm which will be the upper limit, this is the value of R m which
will be lower limit. R m should be a value which is in between in this range. So, this is
the design equation for R m selection. Now what is the power rating of R m? To see that
when this is switched off, there is a current I m starts freewheeling through R m. The
energy in the magnetizing inductance at this point is half L m I m square. Now that much
493
amount of energy has to dissipate, then only the core will reset. So, that has to dissipate
only in R m because R m only dissipating element. So, let us say power rating of the R m
can be found out from half L m I m square amount of energy has to be dissipated every
cycle, every switching cycle.
So, you can say this is equal half L m I m square into F s and this is the power that will
get dissipated in R m whatever may be the value of R m which you are chosen in this
range because this energy, the energy here is independent of the value of R m. So, this is
the power rating choice that you have to do for R m. So, with this you can now say that
whole components of this forward converter circuit with core resetting feature can be
now implemented in this fashion and also simulated there is one important variable that
we will not be able to measure and see on the oscilloscope.
And that is the flux inside this core. We have been talking of core resetting the core in
the flux which is to be reset, but we will not be able to see the flux waveform, but
indirectly we can know how it looks like. So, let me try to plot an estimate of the flux.
See the flux in the core with the symbol phi and let us try to plot that one.
Now, we know from Faraday's equation that V p the voltage across the primary is equal
to L m, the magnetizing inductance d L m by dt or demagnetizing current by dt it is also
equal to n or np d 5 by dt. So, from this basic fundamental equation we see that L m d I
m by dt is equal to n d phi by dt. L m is a constant, n is a constant, V p is fixed for the
period of the time and Q is on, therefore d i mag by dt demagnetizing current by dt and d
phi by dt will have the same wave shape.
And therefore, same slopes and therefore, I i magnetizing current and flux in the core and
also have the same wave shape. Therefore, the flux within the core will take the same
wave shape as ip mag that we have drawn here. So, let us draw that the same linear rising
and fall because when V p is fixed content when Q is on V p by n V p by n will be the
rate at which the flux will rise and it will fall exponentially because the current I m is
falling exponentially and again the next cycle it will go in a linear fashion to a max value
and then, fall exponentially as maximum value is phi m.
So, this is how the flux wave from within the core may look like if you were able to
probe it and see it on the oscilloscope, but we are able to get this wave shape indirectly
through the Faradays law. So, you see that the flux rises and then the flux is reset. This is
494
the core resetting action. The flux in the core it is the reset to 0 and it is ready for action.
The next cycle when it again rises and then during 1 minus d T s period, the flux is reset
again; so, this is core reset in forward converter. We have been discussing about this
forward converter with this type of core reset.
Now, this is a lossless core reset. The reason being that there is some dissipation
occurring in R m and we calculated that the dissipation that happens in R m is half L m I
m squire into F s meaning half L m by I m square is the energy store stored in the
magnetization and that half L m squared must be removed every cycle and therefore, half
L m I m square F s is the amount of power dissipation in this R m. So, this will bring
down the overall efficiency of the forward converter.
So, let us discuss another method of core reset where it is lossless. So, lossless core reset,
so what we do is first we remove this and then let me make some space.
495
And I will introduce one more winding, one more winding which is tightly wound to the
primary 1 is to 1. In fact, bicular wound you take the primary and this winding which I
will call as demagnetizing winding because it is used for resetting or reset winding you
take them together and wind them, so that there is minimum leakage as possible.
And then you attach a diode like this and other end you connected to V i or V cc and you
place the dot polarity in the spanner opposite dot polarity ok, then let me put this as 1 is
to 1 and let me push this back here. So, now this is the modified forward converter where
you have removed the lossless core reset circuit and you have interposed another
winding, no resistance only the winding, but still there is a prevailing diode.
Now, how does this become lossless? Let us see how the core reset action is performed
and how it is lossless. So, this winding is normally turn the literature as the
demagnetizing winding or simply the reset winding because of the job that it does. Now
just like your measuring the voltage V p here in this fashion V s here in this fashion will
measure the voltage across the demagnetizing winding and let me call that 1 as V d. So,
now it will be appropriate to look at some wave forms to understand the operation of the
circuit. Again I am dividing the time into four parts d T s 1 minus d T s d T s and 1
minus d T s two cycles two switching cycles d T s, the time period when Q is on. So,
now first let us see the two waveforms V p and V d simultaneously V p and V d.
496
Now, during the time when during the timed T s during the timed T s Q is on. So, when
Q is on the entire V i is coming across the primary coil with dot end positive, non-dot
end negative. So, if that is the case in the other winding you will see the dot end positive
non-dot end negative for the reading this fashion and dot end positive non-dot end
negative in the secondary and the secondary operation we know it is the regular buck
convertor operation. We will not be discussing it now. Let us focus only on these two
coils the primary and the demagnetizing coil.
So, during that time V i comes across V i comes across V p for that V i and as I am
measuring in this fashion the dot end being positive, it will appear minus for V d minus
V i. So, because it is minus V i, the diode is reversed biased because there is a minus V i
and there is also V i of the supply. So, minus 2 V i comes across the diode. Diode is
reversed biased no current flows through it, but it is also interesting for us to look at what
is this diode current id.
So, let us also look how one access here for the diode current id I would also like to look
at the magnetizing part of the ip. We will come to that later. So, the diode current id
because the diode is reversed biased, it will be zero current at that moment of time. Now
if I take the magnetizing part of ip remember recall that ip is consisting of two parts. One
is the load reflected part that is the n times i D1 and the magnetizing part ip mag together
that form ip. So, consider only the magnetizing part ip mag which is a kind of a ramp in
this fashion which is having a slope of V i by L m the slope.
Now during 1 minus d T s Q is switched off the moment, Q is switched off. The current
here is immediately brought to 0. So, there is a large di by dt negative d i by dt and
therefore, they will be reversal of potential across the coils. So, let us mark that reversal
of potential non-dot end becomes positive dot end is negative, non-naught end is
positive; dot end is negative likewise on the secondary side the secondary side now here
inductor is freewheeling. This is plus and minus in this fashion this there is a reverse
biased. So, the entire secondary is like open circuited. We will not talk about that. We
will talk about these two.
Now here if you look at the non-dot end is positive and this will keep this will become as
positive as it can pump against this V i and make this diode forward biased, so that the
diode is conducting for freewheeling action. So, if I draw the voltage source V i portion
497
also, here you will see that there will be conduction of current in this path like that and
comes through this against the source V i and reaches here from this plus it goes in
against the source. So, if the diode is ideal, the drop just a bit more than V i is what this
will be needed to pump current into the pump current for resetting the core. So, there is
freewheeling action that will happen.
So, what are the voltage across V d and V p? So, across V d because it is this a diode
clumping plus and minus this is coming directly across V i, it will be V i in this
measured in this direction nor non-dot end positive. So, in the case of V p you will see
that it is negative minus V i because this is 1 is to 1 times ratio, this is 1 is to 1 and what
happens to id see that ip had charged the flux up to that point and now it will take over
from there and discharge during that time with a slope of minus V i by L m. This charged
with V i by L m, this will discharge with minus V i by L m.
So, therefore if this is taken x seconds, this should also take x seconds. That is why we
say that if this has taken T s by 2, this would be T s by 2. You should give the same
amount of time for it to reset. Therefore, you cannot have a duty cycle greater than 50
percent in these type of converters, but is not a disadvantage because you have n has a
degree of freedom whichever you can adjust the voltage.
So, next again next cycle next cycle behaves in the same way. You have the repeat and
during this portion there is no id. There will be ip mag and then id comes down in this
fashion. So, ip ip mag magnetizing close only during d T s portions; id flows doing 1
minus d T s portion as the freewheeling path. If you look at the flux within the core, the
flux within the core will have the same profile as id and ip mag. It will take the profile of
ip mag during the d T s period because that is a magnetizing current that is flowing and
during the 1 minus d T s period. The magnetizing part is flowing in the freewheeling
portion of the demagnetizing winding and that will be this part.
So, this is V i by n, this is also is V i by n negative because it is falling. So, this repeats.
So, this will be the nature of the Flux. The flux builds up and the flux resets during 1
minus d flux builds up, flux resets during 1 minus d period because the slopes are same.
You need to at least that much same amount of time for it to reset therefore, limiting the
duty cycle for this type of converter max to 50 percent.
498
So, this is the way the core reset functions in this. Note two points this windings here the
primary winding and the demagnetizing winding should be 1 is to 1 and they are to be
tightly woven together, wind them together. That is called bifilar winding, so that the
leakage inductions is minimized. The reason is that if there is a leakage here that portion
is not covered for mutual inductance for demagnetizing, so that can cause a large spike
across the device.
When it is switching off that can also be avoided by having this number circuit which I
will discuss later when I am discussing devices. But, I could also provide you with a
quick solution provide even the leakage inductance shortly when I am talking of using
dual switch or two switches for the forward converter and remove this magnetizing
portion, then we can still get lossless core reset and avoid the leakage inductance issues
also.
We can do lossless core reset with dual switch forward converter also. So, let me briefly
describe the dual switch forward converter and explain to you how this lossless
demagnetizing can be done. So, this is the circuit of a forward converter without any
demagnetizing; without any core reset and let us introduce core reset now. So, let I
remove the V cc for now and then here I will introduce another switch another BJT this
is V i. Let me call this Q 1 and this is Q 2 ok.
499
Now, these two switches are switched on simultaneously ok. Now I will add two
freewheeling diodes; one diode freewheeling from here to V i to V cc and another diode
from here down to ground in this fashion. I will explain to you how the current flows.
Let us also visualize that there is a supply here. So, let us place that supply because we
have placed the label. So, let us visualize to understand the current flow.
Now, let us consider the situation d T s when Q 1 and Q 2 are both on together. So, they
are both on together. You will see that this will be this is on. So, V i this is plus and this
is minus, on the secondary side this is plus and minus, secondary side operation we know
is buck converter operation, but let us see how the current flows here on the primary. So,
it starts from here, goes in this direction, comes like that through the ground and back
again to the power supply. So, this is the path of the current flow during the period d T s
and where the voltage across the primary is plus minus dot and positive V s dot end
positive buck operation follows.
Now, what happens when you switch off Q 1 and Q 2? So, when you switch off Q 1 and
Q 2 there is a cut in the primary current meaning the primary current is cut off to 0, but
we cannot allow cut off to 0. There will be a reversal of polarity. So, first these polarities
will reverse because of the minus di by dt and there has to be a path for ip current flow in
the same direction in the coil.
So, let us start from here ip flows here now as Q 2 is off current flow in this direction
through this diode and then into the supply. So, I am indicating the supply here goes into
the supply and this comes in this direction through this diode up there again and then,
down and maintaining the same direction of current flow in the primary only. Now the
dot end is negative; non dot end is positive. So, it is demagnetizing.
So, this is this coil is acting like a demagnetizing winding during the 1 minus d period
because the secondary side is open, this diode is off. So, this because the same winding is
operating as a demagnetizing winding, there is no leakage inductance problem because
everything is included here and primary reflected and you will not have leakage
inductance spikes due to due to the leakage inductance spikes on the two switches when
they are off. So, this is the path the this is the path where the demagnetizing current will
flow and if you will consider diodes as ideal, this whole V i is coming across the primary
plus minus in this direction.
500
So, you have to allow the same time because again it is 1 is to 1. The magnetizing is
happening at V i by n the demagnetizing is also happening at the rate V i by n.
Therefore, you have to allow the same amount of time for the magnetizing as you are
allowed for magnetizing and therefore, this is also limited to 50 max duty cycle. This is
not actually a serious limitation because of the n here, you cannot just the value of the n
to get whatever required output voltage. So, this is the dual switch forward converter.
You have removed the windings, but you have included one extra diode and one extra
switch and an extra controls switch means extra get right for that.
501
Prof. L. Umanand
Lecture - 56
Simulating with lossy core reset
Let us now see how we go about simulating the forward converter circuit. Forward
converter with lossy core flux reset calling it lossy, because this is the core reset circuit
which we studied and this is the lossy resistor which is going to dissipate the
magnetizing energy out of the core.
So, this is essentially the forward converter circuit that we have been discussing till now.
This is the power semiconductor switch which is going to turn on and off and it is getting
the gate drive signal from PWM, block and the PWM block is given with a control
reference of minus 0.1.
Remember last time when we were discussing the PWM block we, we said that this
triangle within the PWM block is transiting from a minus 1 to plus 1. So, at 0 if the
control signal is 0, you will get 50 percent duty cycle. If the control signal is minus 1
then you will get 0 percent duty cycle, if the control signal is plus 1 then it is 100 percent
duty cycle. So, at minus 0.1, it will be around 40 percent duty cycle.
502
So, this is, the gate drive signal which is given to the BJT switch. Now, if we come to the
transformer portion the transformer is not the regular transformer provided by spice that
is a linear transformer model, we have included a sub circuit model here so, that even
saturation effects can be taken care of. One could also use the core based model of spice
here, but this is a generic model, any good model of the transformer can be used without
loss of generality. Now, this portion of the circuit as you already know is the core reset
freewheeling circuit, we have a resistor and the time and this is of course, the input and
we are applying 15 volts DC here.
On the secondary side, we have the diode D 1 and D 2 and LC and you can recognize the
buck converter topology in the output side. Now after that for the display of the results I
have been telling you to label the nodes like this. So, you have the node A and this is
node Q, if you want to see the voltage across the device here there is a node S, here for at
the dot end of the transformer. Secondary there is a node P indicating the pole voltage of
the buck converter and the node O; you should remember that the output side is not at the
ground potential, this is not at the ground potential. So, you should because it is isolated
from the primary side. So, whenever you want to measure a signal here, it should be with
respect to some point on the output side. So, let us say we have a reference node R here.
So, will causes the reference node and anything on the secondary side, we will measure it
with respect to this reference node.
So, when you want to see the secondary voltage, you will say VS with respect to R VP
with respect to R that is V P , R VO , R then the voltages are with respect to this
reference node. So, I have put in some typical values here.
In fact, these are the same values which we used for the buck converter. And you can
use the similar thing for the forward converter example here too. Using the PWM with a
switching frequency of 10 kilohertz and this is the spice include which we have been
doing for all the simulations and I have included the EDT 0 1 dot sub file.
So, this file, let us simulate. Now going back to the folder the forward dot CIR the
forward dot CIR contains dot TRAN attention analysis statement 0.1 micro second step
and up to 10 milliseconds use initial conditions.
503
Then dot include the forward dot net forward dot net has to be generated which we will
do now shortly. Now coming to the EDT, let me close this. Now the RDT dot sub I have
already included the model for the electromagnetic transformer. So, there is a model for
the electromagnetic transformer, the primary dot point primary non dot point secondary
dot point, secondary non dot point there is a default primary number of TRAN secondary
number of turns and C, here C is not to be thought of as capacitance , but as permeance.
So, within the magnetic domain the permeance in the core behaves very much like the
capacitance or capacitor. So, in spice the equivalent component would be A C. And
therefore, we have used C, I will explain more about this capacitance permeance
relationship. Later on when we are discussing the magnetics so, we have used this kind
of the model here. We have also put some kind of saturation effect here, if the value the
permeance if the value of the flux is greater than a particular value then we will not, we
will take the same value of C as is given here, but if it is greater than the value a
particular set value, then it will start exponentially decaying down to 0.
So, this will emulate the BH curve of a core. So, this is a saturating type of a magnetic
circuit and therefore, this transform a transformer has is a generic transformer which
behaves much more closer to reality. Now this transformer uses to further models called
gyrators and the gyrator is modeled here.
504
Anyway this is the model of electromagnetic transformer and you will find many models
of transformersin the internet, you can use anything that is suitable to you.
And you can also probably use the spice provided transformer and even the linear
transformer to at least get the concepts, only if you want to look at saturation then you
will have to include something to make the core saturate, ok.
So, we have the EDT 0 1 sub dot sub file, also in place now we have to go forward and
do the simulation using the ngspice engine which is what we will be doing.
505
Now, so, let me go to the terminal and inside the terminal I will go to the ok. So, we are
in the directory and we will generate the netlist and also call ngspice after we have
generated the netlist, let us now generate the netlist and go forward.
So, generating the netlist is with this. Now familiar ,nd G netlist dash G spice dash S DB
dash O forward dot net has to be generated from forward dot SCH of the schematic file.
So, from the schematic file, you give this ,nd and now you would see that there is a net
file generated.
Now we can use this net file this net file is being called in the forward dot CIR and
ngspice can perform the simulation. Now all the ngspice engine, ngspice forward dot
CIR. Now this will load the circuit schematic with the netlist and now run to perform the
simulation.
So, this will take some time I will later on tell you methods in which you can reduce the
simulation time. The first time simulation will take quite some time basically, because
you would have given an extended range till it reaches steady state. After it reaches
steady state, you can then note down the steady state value and give it as initial
conditions. And I do the simulation with those initial conditions then you will see that it
in just a few cycles, you will quickly get the wave shapes and wave forms that you
actually would like to see at various parts of the circuit.
506
So, now, here you have some time for 80 percent of the simulation done, you just have
few more moments to wait and you have the complete simulation done now. So, you
could plot. So, when you plot let say I want to plot out do not do not just put V in
brackets O.
Now with respect to R with respect to the R node, because observe that we want to see
the output with respect to this R node. So, it will be V O , R. So, coming back here, you
will see that this is the output ratio.
Now let us, before we further investigate few more aspects of the waveforms and
waveforms of the other parts of the circuit one tip here would be, ok. Let me put that
back here, you have a black background and on the black background, you have the
white grid and the waveforms showing up in color here. In some cases, you would prefer
to have a white background and have black grid lines and then the waveforms in color.
This is especially useful when you want to document it or cut I will take a screenshot of
the plot out result and then put it into your document
507
So, to change the background what you could do? You could use this following ,nds.
You can always go back to the ngspice manual and check out these ,nds color 0. So,
color 0, I will now set it to white. So, this basically sets the background color 0
represents background color set, these are fix it reserved words color one is the
foreground which we use set it to black.
So, now, same plot ,nd would appear in much deserved black a white background with a
black foreground, ok. Now with this let us see few more waveforms of interest to us. So,
we would like to see the, inductor current waveform here. So, let us have a look at the
inductor current waveform plot i l.
508
And see as expected the inductor current triangular wave shape; this is what we are
expecting. And you observe that later on around here after the 9.5 millisecond almost it
is trying to reach stable point, you could probably give the initial condition at around this
value. So, that quickly, you will see your circuit coming to a stable state from the initial
condition that is a very nice approach do to shorten the simulation time, if you are not
interested in the transients.
Now another the point that we would like to see is Vq here, we have been discussing.
509
So, much about v q, you will take across the transistor. So, let us say v q, v q let us
expand it to just a few cycles and you would see.
Now, you see this is exactly what we have been discussing this is the on time. So,
therefore, the on state then the moment the BJT switches off. So, you see a high spike
and then the current decays exponentially and goes towards vvn value or vcc value.
So, this is typical of the lossy flux reset and this exponential decay is the LM by R time
constant and this is due to the decay in the magnetizing current. So, this is one important
510
waveform which I wanted to show you., other another point important point of waveform
that we would like to see is the currents that are flowing through this. But unfortunately
here there are no see if it is an inductor, inductor being a state energy storing variable
component the current is a state variable and then it will so, give the branch current of
that.
If it is a resistor or any other branch, you will not see automatically with the branch
currents available here. You, you observe to the branch currents are available in the list
of plotouts only for L branch. If there is a source vin and vc branch. So, you see that
there will be currents available only for this vc a source branch another source branch
and NL branch. So, how to see and measure currents at various other points?
So, I will give you one more trip, one more tip here we could put 0 sources wherever we
want to measure the currents and then on simulation, you will get those branch currents
and then you could see those branch currents.
Let us see we can do such a thing and the meantime first what I will do is quit this plot
quit and clear the screen.
And now, we will do the modification here. So, what I am going to do? I am going to
source like this I am going to introduce here, here maybe here. And here no what we
could do, we need not have two here, we will have one here and one here. So, these two
511
will give you all the information of the current which flows here and one here which will
give you the secondary current. So, three voltage sources with 0 voltage value we will
interpose inside in this circuit schematic.
So, now you see, I have made the changes which I said I would do which is to put a
source here a source here and the source here. Now observe also that I have put 0 voltage
value at all these source. So, these are just now behaving like current sensors, now I have
put the positive at in such a way that the current flows into the positive. So, here we are
expecting the current to flow in this fashion. So, it flows into the positive, we are
expecting the current to flow in this direction here also it flows into the positive and
likewise from the dot end into the positive. So, if you put the source in such a way that
the positive is in such a direction that the current flows into it then you will get the
proper direction of the current ok.
So, now, this otherwise there is no other change that we have done, we have output
current sensors. And with this current sensors, we will be able to get more insight into
the circuit. So, let me see this and do the simulation once again.
512
So, let us generate the netlist, netlist is generated. And now we will go into NG spay of
spice and run the simulation once again.
So, it should take some time then after that we could we could change the background
color to white and the foreground color to black. So, this would make a much more
presentable display and a display which you can later on use it for documentation
purposes. Now, while the simulation is going I would like to also show that in the
transformer, when I double click on the transformer.
513
It is sub circuit we, I have put here the primary number of turns 10 secondary number of
turns 20 and the permeance is 100 micro. So, this is the parameters that is passed on to
the transformer. So, therefore, you see that NP 10 and NS 20 means there is a trans ratio
of 1 is to 2.
Now coming back to here the simulation is over. And let me set color background equals
white set color foreground equals black. Now let me plot, now what a plot. Now I would
like to see the switch current here. Now the switch current remember is nothing, but the
514
reflected part here of the secondary reflected part and there will also be a magnetizing
part. So, these two components together form the switch current.
And during the time of the switch is off the reflected part is not there, there is only the
magnetizing part which will decay. Now this effect, we will be able to see clearly. So, I
will view this current this current and this current. So, I of vid i of viq and i of vis. So, let
us i of you could also see here you have the vid branch current vis branch current, viq
branch current. So, voltage of the demagnetizing winding diode winding vid i of the
current flowing through the switch viq and the current flowing through the second wave
vis. So, let us plot that let me expand this and you will see that and let me also probably
see just few.
515
Now, you see here, viq blue viq, this is the switch current. So, the switch current is like
this, it is the blue portion clipped. Now this is actually N times I naught recall that this
point here where the cursor is would be N times I naught. Now this is actually the I
naught current. So, if I plot the inductor current, it will flow all along through like this.
So, therefore, it may be a good idea to plot that also I will do that later. So, this is
actually the secondary current and you see the red one is the D magnetizing current see
the it is so, small in magnitude. So, actually the magnetizing part will take a linear wave
shape here and goes in an exponential manner and from here again linear and so on.
So, let us say if I say N times I naught, this will is supposed to match here exactly except
for the magnetizing. So, you will see that this will more or less come in this like this here
like a small triangular gap would be left here that would be the magnetizing part.
516
So, let us see that what I will do is now I will give the this is actually the secondary
current into N and N is to in this case. And now plot that and let me expand that.
Now, you observe now you see this, the primary or the secondary exactly same except
for the difference which is the magnetizing part. So, the primary switch is composed of
the reflected part and the magnetizing part and the magnetizing part actually is an
amount equivalent to this which will come and match exactly here and then this will start
exponentially decaying.
517
So, that is one aspect I would also like to compare with the inductance, inductance
waveform inductors waveform. So, let us say we also have I will remove this and also
put i of l the inductor waveform the interesting to see that let me zoom in.
Now you see the green is the inductor waveform, you see that this inductor waveform the
part when the switch is on is the same which we are measuring as the secondary current.
And that is actually what is reflected to the primary. Now the central the average of the
inductor waveform is I naught and that will come as the flattop equivalent of the primary
518
switch current which is N times I naught. So, the primary switch currents flattop
equivalent equivalent flattop current will be NI naught in this case.
Now I would like to show you one more thing before closing the simulation and that is if
you want to do the immediately some changes and check the waveform that various
points, you would not like to wait long. And then wait for all the simulation to run
through and then wait to see what is happening here. First off one thing what you could
do is you could take the value of the state variables. If you see there are two state
variables the inputs and the state variables define the system and inputs are DC in this
case and the state variables are il and vc.
So, let us take the value of il and vc at the end of the simulation plug it in as initial
conditions here and then you will see the simulation can be much faster. So, that is one
exercise I will just show you.
I will just quit from here oh no not quit, I will just plot plot first IL 1 state variable.
519
So, let us take the value at the end of the simulation here click on that and then you will
see the values one point had at that final value, it is around 1.19444 so, that I will use it
as the initial condition for the inductance. So, when I click on the value for the
inductance millinery, I will also put initial condition equals 1.19444 .
520
So, this will set the initial conditions see, it will be valid only for only for this value of
inductance and state values.
And next, let me see the value for the output voltage for the voltage across the
capacitance plot V O , R.
521
Now, let me take the end value here click on that. So, it is on 11.8871 there could be
some minor difference does not matter, it will reach stable state quickly. So, that is
11.887. So, go to the value of the capacitance and give initial condition equals 11.887.
522
So, this is for the capacitance and let us save this and let us close and redo the
simulation. So, now, before redoing the simulation I do not want to do the simulation for
10 milliseconds.
523
Now, that I have almost reached the stable state I will do it for let us say 0.5 milliseconds
it is supposed to be pretty quick. Now rather than waiting it is actually 120th. So,
therefore, let me save it and go to (Refer Time: 29:45).
524
So, now let me redo the netlist. Now after doing the netlist Ngspice forward dot CIR,
yes. Now, run the simulation.
You see it is so quick and let us plot, plot I will plot the currents v vid plot the current
switch current plot the current of the secondary. This is, ok.
525
Now this is again the black background I do not like it, I will set color to white set color
of foreground to black.
526
And look at the waveforms once again. And you see that you, you are already in the
stable state. And you could also see the output waveforms V O , R and probably you
could see Vq switch voltage across the switch and distinctly you will see this is the
output voltage V naught this is a voltage across the switch.
527
Now, this first one is a bit off, because our initial condition it will take some time for it to
stabilize. It has taken just one switching cycle to stabilize and you see you get stable
output waveforms. So, now, you can do lot of changes in your output load and the such
accept any change in the state values which you have to run the simulation right from 0
to 10 millisecond. Once again and then re edit the initial conditions, but this stick this
trick would be very useful. When you want to do repeated simulation of the circuit for
changes other than the state values and you can quickly see the output results in this
fashion.
528
So, I will leave it to you to explore the circuit and try to get as much insight as you can
into this forward converter operation and the waveforms.
529
Fundamentals of Power Electronics.
Prof. L. Umanand
Lecture - 57
Simulating with lossless core reset
Let me now show you the simulation or the forward converter with lossless core flux
reset. This is the entire circuit diagram of the forward converter with lossless core reset.
Observe this transformer part so, in the transformer part you have 3 windings; one is the
primary, other the secondary and in between you have the demagnetizing winding. So,
this demagnetizing winding is the one which will perform the core reset. I have this
demagnetizing diode flowing into the source.
I have here current sensors in the form of 0 voltage sources I am sensing the i q current
by using V i q 0 voltage source V i d 0 voltage source for the demagnetizing current and
V i s another 0 voltage source or measuring the i D 1 current. Another change you will
notice is in the values of L 10 millinery have also provide initial condition of 1.1944.
Now, this value I took from the simulation after it has reached steady state so that the
simulation can be quicker and you can see the outputs in less number of cycles. I have
also given a voltage initial condition for the capacitors these 2 are the states of the
system and therefore, you need to provide the initial conditions for these 2 states and I
530
given a voltage initial condition value here. Again taken from the steady state value of
the previous simulation that I did without initial conditions. Now, in giving this initial
condition the simulation can be very very quick and you can directly look at the steady
state and see the simulation for just a few cycles 5 or 6 cycles.
This transformer here forward converter transformer when you double click you will see
that it is a sub circuit x N p the primary turns is 10 N d that the demagnetize number of;
demagnetizing winding number of turns is 10 N s 20 and the permiance I have put as
150, I have put it in electrical parameters of microfarad, but it is micro, 10 to the power
of minus 6 in terms of the magnetic parameters. So, we can now simulate it in Ngspice
before we simulate it in Ngspice let us have a look at the dot cir file.
531
So, double click on the forward dot cir file and you will see that you have the trans
statement for the transient analysis, the include statement including the net file which
will get generated from after having run the schematic to net generation command. Then
in the control statements I have here set background color to white, foreground color to
black, run the simulation and plot. Now, in plot I am using output voltage v naught with
respect; with reference to or if we just put v naught it will be with respect to ground, and
ground is on the primary side and you need to see actually the v naught with respect to
its secondary ground which is the reference here in this case is r.
So, therefore, you need to see v naught with respect to r, i L the inductor current and i v
is is nothing but the diode D 1 current. So, you can put in many other variables; other
variables to see the various other wave forms, but I just put these two so that you get
started and then you include an explorer on your own.
532
Now, let us open a terminal window command window and run the simulation. So, I
have here the command window opened in front of me, I will run the simulation like
before I will use a runsim and forward. So, this will run the simulation in the forward
lossless and generate the output.
As you can see I have given it for very small number of cycles, you will see this as the
output voltage v o, r ok. So, that’s close to around 12 volts and calculate it with their
input and then find out if by the input-output voltage relationship you get approximately
533
this value. Then you have this blue and this other one the diode current and the inductor
correct. So, let me probably expand it so, that you get I will expand it here and see. So,
this is your inductor current; so, this is inductor current and the diode current here ok.
And this portion diode current is 0 and inductor current is falling. You can separately see
these waveforms also just to confirm whether your theory is proper.
Anyway I will allow you to explore the circuit and you can try to change values and
change the initial conditions and also the change the value redesign for different
frequencies of the PWM and try to get more insight into the workings of these forward
converter with lossless core reset.
534
Prof. L. Umanand
Lecture - 58
Flyback converter
Let us now discuss another converter, very popular converter; one of the most popular
converters that you will find in most power electronic equipment that is the Flyback
converter. It is a buck-boost derived converter. The forward converter was a buck
derived converter, the flyback converter is the buck-boost derived converter and let us
see how the isolated buck-boost derived converter looks like.
So, let us begin by drawing the circuit of the buck-boost converter and let us derive the
isolation from that. So, we have the source, you have the switch, controlled switch,
inductor, uncontrolled switch, diode there, the capacitance and the load. So, this is the
buck-boost converter which we know and which we have discussed and we know its
operation function.
Let us now name the parts; you have V I measured in this fashion, R naught, Vo, L, C,
Q, control switch D diode. Observe that when the switch Q is on the inductor current
flows in this direction when the switch Q is off the inductor current will flow in this
fashion charging of the capacitor in reverse so, this is i c, Io.
535
Now I will make one modification just like we did in the forward converter. This
controlled switch is on the rail side. Now, this switch has to be driven to saturation on
and off by controlling the voltage across its base and emitter. The emitter is lifted up,
floating it is much better to have the switch ground based and the ground based drive is
much easier to make. So, without loss of generality we can move this switch from the top
rail to the return rail, maintaining the flow of the current path in the same manner.
So, we can place that Q here, you see the current flow is in this path when the switch is
on and the return path for the current is in this path, in this direction. So, we are still
maintaining the same direction of the current flow and we can remove that switch. So,
we have shifted the controlled switch to the ground rail, we can mark the ground this as
the circuit ground and we can label the positive of the input source V i like this, and in
order not to clutter up the circuit you can remove that.
We can do next step of the modification make this portion vertical. Let me, let me draw
the controls switch vertical ground and that becomes Q erase this portion. And you can
shorten this, in this fashion and, so this becomes the non-isolated buck-boost converter.
Now let us introduce the isolation this portion of the circuit we will push it out and
introduce isolation here. Observe that this is an inductor and we are winding one more
winding on the inductor so, that it provides the isolation, but the behavior of this
magnetic will still be inductor energy storage.
536
So, let us put in one more coil there, let me make some space arrangements. Put in the
core and put in the secondary coil and then connect it in this fashion. This is still
behaving as an inductor L, it is the behavior is in like an inductor not like a transformer.
Let me put the dot polarity here, this is dot polarity. So, operation is as before, so we
have the Q here, when the Q is on during d t s time you will have current flow like this
and the voltage across this is V i, the voltage across this is V i here, n times V i and
reverse biases the diode.
Because here it is plus minus reverse so, this the Vo and the secondary voltage will
reverse bias the diode. And then when the switches off, there is a reversal of polarity the
non dot end becomes positive, non dot end becomes positive and that will drive free
wheel through this diode and charge of the capacitor. So, the buck derive buck of buck-
boost operation is still maintained.
We can do some more modification in the circuit to make it a bit more attractive. So,
what we can do is remove, first remove these markings this diode and the dot polarity.
Let us reverse the diode, if you reverse the diode I can reverse the I should reverse dot
polarity. So, that when the dot end; non dot end is positive the diode should be on and if
you reverse the diode the capacitor will get charged in this direction plus and minus and
we will have as measured in this fashion a positive voltage.
So, just to make it more pleasing in that fashion we have reverse the position of the
diodes and the dot polarities without loss of generality. The operation still remains the
same. So, this is i c and this is Io, you have i Q here, this is the primary voltage V P
across the inductor measured in this fashion. And the secondary voltage V S, this is D
and this is i D ok. We will push it out and then let us say that there is a turns ratio
between the primary to the secondary as 1 is to n.
Now, in order to understand this circuit, now this is the complete isolated buck-boost
converter also called as the flyback converter; very popularly called as the flyback
converter. And look at the simplicity of it, you have one switch, you have one magnetic
core operating like an inductor and you have one diode and one capacitor simplest least
number of component count and that is why it is so popular. Now we know how the
buck-boost converter operates, this is entirely similar to the operations of the buck-boost
537
converter not withstanding that we should draw the waveforms and try to understand its
operation, but let us revisit, relook at the buck boost converter.
So, this is a buck-boost converter that we had discussed some time before and the
waveforms here. So, V L this is the voltage across the inductor remember this the current
through the inductor i L. Now the current through the inductor will be split into two
parts; current through the primary of the inductor and current through the secondary of
the inductor, I will explain and the currents through the switches here. So, let us copy
this portion of the waveforms and paste it alongside the flyback converter and try to
understand the operation of the flyback converter with these waveforms and try to
modify this waveforms to fit the variables in the flyback converter.
So, let us now paste the waveform here move it around make some space and this is the
waveform that we will try to map here. So, v L here will be v P primary of the inductor
of the flyback converter will be the V L waveform d Ts is the period of time and Q is on,
1 minus d Ts is the period of time and Q is off. Again d Ts when Q is on, 1 minus d Ts
when Q is off. i L is the inductor current we will split that into the primary indicator
current and the secondary inductor current this portion flows through the primary and
this portion the falling slow portion will flow through the secondary and let us rename
some of these variables to map into this circuit.
538
V L here the voltage across the inductor is indicated as V P here voltage across the
primary inductor and V S voltage across the secondary inductor. So, let us replace this V
L with V P voltage across the primary inductor. Now. the voltage across the primary
inductor during the time d Ts when the Q transistor Q is on the entire V i comes across
the primary inductor. So, let us indicate that with the plus, dot end is plus, non dot end is
minus.
On the secondary side you have the dot end plus, non dot end minus. Observe that the
voltage across V S is n times V i and that is that plus Vo is coming across this diode and
reverse biasing it and switching it off. So, what is that voltage which is coming across
the diode reverse biasing it, n times V i plus Vo comes across that and this will reverse
biases switch it out. Therefore, during the time n Q is on, during d Ts time only the
primary current is flowing the secondary is off. Of course, the capacitance is discharging
to the load.
Observe also that during d Ts the current that is flowing through the primary i Q is only
for the part d Ts and during 1 minus d Ts there is no current flowing through that. The
only link between the primary and the secondary side of the inductor is the flux. So, let
us indicate that flux here and i L we will replace it with flux. So, this is the flux
waveform or we have to change these values. So, let us say here flux is phi average, this
is the average phi and the slope is not V i by L, V i by L is for the current waveform and
here it is V i by. Now if I say this is the N P and the secondary is N S number of turns.
So, it is V i by N P the slope the rate at which flux changes.
Now, what is the waveform of the current that is flowing through the primary or the
current that is flowing through Q, i Q. i Q is nothing but the current that was flowing
through 1 to pole current that is i T 1 P. So, i T 1 P it becomes the i Q. So, see that during
the d T s period of time i Q there is a current which is raising linearly like this and the
slope where is V i divided by L and that is the slope of this.
So, observe that i Q you will have a finite current only when Q is on and during the
period 1 minus d Ts Q is off and there is no i Q i Q current is 0. Now what is the current
through the diode D, i D? Let us change this T to P it is from throw 2 to P of the buck-
boost topology that becomes i D and as you can see from this operation when Q is on dot
is plus, dot is plus, diode is reverse biased, diode is off there cannot be any i D current.
539
However, the capacitance is charging into the load. So, you have a negative Io flowing
through at that time, discharging into the load.
Now let us take the time period 1 minus d T s. During 1 minus d T s Q is off. So, the Q is
switched off i Q will become 0 and what will happen to the polarities of the primary and
the secondary of this inductor? So, you will see that the non dot end will become plus,
dot end will become minus, non dot end is plus, dot end is minus and you will see that V
S the secondary voltage will rise up to that level just greater than Vo so, that diode D is
forward bias and diode D will conduct and charge up the output capacitance.
So, V S will be Vo because the diode is conducting, Vo will come across V S and what
will be V P? V P will be Vo by N because there is a 1 is to N turns ratio so, V P will be
Vo by N. And how does it reflect on to the waveforms? We will see that V P waveform
during 1 minus d Ts will go negative, but we know; but the amplitude of that voltage
across V P will be Vo by N.
So, let us mark that Vo by n and in the flux waveform it will be Vo by N S because Vo is
coming directly across N S, Vo by N S will be the downward slope of the flux waveform
downward rate, i Q is 0, Q is off i Q is 0 at the time and i D; i D is the current through
the diode. Now the current through the diode observe the flux minus Vo by N S because
the non dot end is positive it is going to be a downswing current flow and also the flux.
So, the current will have rate of minus Vo by N square L. L is on the primary side on the
secondary side will be N square L and that will be the slope of the current there.
The average of i D will be Io like before in the case of the buck-boost converter. The
current wave shape of the capacitance will remain unchanged and it will be of the same
shape and amplitude was output portion we are not changed anything. Now, let us try to
find the input-output voltage relationship. So, input-output relationship first a voltage we
apply volt second balance for the voltage waveform across the inductor you can choose
either primary or secondary. So, let us say we take the primary voltage across the
inductor.
So, during d Ts we have V I, during 1 minus d Ts we have Vo by N. We use that so, we

have V i into d Ts plus Vo by N into 1 minus d Ts should be equal to 0 volt second
balance. You can remove this variable and then on simplification and if we take only d
magnitude of Vo because you do not need to worry about the sign because you can take
540
anything as ground because is isolated it will be the n times V i d by 1 minus d. So, this
becomes the input output voltage relationship.
Note that the output voltage is having n as 1 degree of freedom and d as the other. So, is
the variable it can continuously vary on the fly during operation so, you can use that as a
control input, n is again a variable which you can change only once at design time. So,
this can be a designer variable. So, by design choice you can fix the value of n. So, you
have more flexibility in choosing the output.
Now likewise for the current also you can get the input output current relationship, can
use the power balance relation
Vi Pi = Po
Iin = n Io d/(1-d)
When the transistor Q is off, we saw that when the diode D was off it the peak inverse
voltage that it need to support was N times V i plus Vo and for the case of the transistor
when it is off it has to support a voltage of V i. So, you have supply voltage V i coming
across that plus also the reflected voltage from the secondary at the time when this is off
diode is on there is Vo coming across this one and on the primary side it is Vo by n with
plus minus in this fashion plus minus this one, plus this V i. So, you have Vo by n
together coming across the collector emitter of Q.
541
So, you will have to rate the vco rating of Q to be greater than V i plus Vo by n. So, in
this fashion we have this flyback converter which is this is the functional schematic of
that and you can use that and very very popular and you can also have multiple outputs
you can have multiple secondary windings across the; wound across the inductor and you
can take out isolated multiple outputs from this flyback type converter.
542
Prof. L. Umanand
Lecture - 59
Simulating the flyback converter
Let us see now the Simulation of a flyback converter. So, this is the circuit of a flyback
converter. You have the switch here; you have a flyback transformer. The reason I am
saying this is the flyback transformer is that it is actually inductor with an extra winding,
which gives isolation to the secondary part. You have the diode, the capacitor I have put,
a small resistance here R s 0.1 ohm, this is basically because the simulation engine can
run into numerical instability if you do not put this resistance. You will be connecting a
capacitance directly across a voltage source and that can give issues while simulating.
So, good practice to decouple that by putting 0.1 ohm resistance, and you have the load
resistance here. And, I have the reference node r, output node o. And, you have the
transformer this is the switch Q and the PWM block. I have set the control signal to
minus 0.5. Remember that the triangle is going from minus 1 to plus 1, minus 0.5 would
mean around 25 percent duty cycle.
543
So, if it is 25 percent duty cycle, N here is you see N D is 100; N s is 100. Therefore, N
is 1 is to 1. So, because it is 1 is to 1, you have V naught is equal to V i to d by 1 minus
D, 0.25 by 0.75, which is one-third of this which is 5 volts one-third of 15 – 5 volt. So, it
should settle somewhere around 5, 5 volts, but it can be lesser than 5, because you have
various other drops you have the diode drops, and you have the drop across this
resistance R s. So, now, let us simulate it in ngspice before simulating it have a look at
the dot cir file.
544
So, let us open the dot cir file. What does it contain? So, it has a dot option statement, I
have put reltol 0.01, and abstol tolerance for voltage and tolerance for current 0.01. You
sometimes when you end up in numerical instability issues, you need to limit the
tolerance of voltage and current, I have done that one. Dot transit transition as a
statement and you are including the flyback netlist, which is generated from the
schematic.
And, within the control statements again setting the background white, foreground black,
running the plot, running the simulation and plotting. What do I plot v o, r, there is
output with respect to r; v i, p which is the input current flowing through the switch,
flowing through the switch. And, i v i s which is the secondary current have seen here.
Now, we open a terminal and simulate the circuit.
545
So, I have the terminal open here, let me simulate it. I am in the folder flyback, it is
available in the resources section. And, you can you are welcome to use it, let me run the
simulation using runsim flyback.
546
So, when you execute that simulation goes through. And you will see that there are some
noise spike spikes coming in here again due to numerical instability. What you see this is
the waveform. You can zoom that. Let me zoom just that portion of the waveform. Let
me zoom it a bit more, so that you can see it clearly.
So, you see this red waveform is v o, r output with reference to with respect to r and here
the blue and these way from the primary in the secondary currents. I could zoom the
primary secondary current a bit more, and have a look at that.
547
So, blue one that is the primary current ok, and this falling one is the secondary current.
So, these are the currents that are flowing through the q switch i q and i d. And, I will
leave it to you to explore the rest and experiment with it.
548
Prof. L. Umanand
Lecture - 60
Octave mfile for design
I have here on the screen on the left side, a text editor and on the right side I am having
the octave workspace screen. You could either use octave or MATLAB both are
compatible and on this text editor, I have here m file, forward design dot m file. This is
basically an m file a text ASCII file, where I am trying to summarize, the equations in to
make some kind of a semblance for design.
Now, designer first I am using clearing the screen and clearing the variables. in a
systematic way let us first make the specifications. So, the input voltage specification, 48
volts nominal, DC link voltage we call that , the input VI can take a minimum value or a
maximum value, you should design for worst case. So, this example takes care of
variations, you have a minimum value of the DC input and the maximum value of the
DC input.
V naught, let us say you want it for 12 volts. There is a delta V, there is output voltage
ripple peak to peak ripple of 0.05 times 0.05 and I naught is 1 amp, this is the output
current. So, designer variables delta I is 10 percent of I naught fs is 20 kilohertz, Dmax
549
do not go beyond 0.544 you do not go beyond 0.5. So, to be on the safer side I have kept
it at 0.45 and you can calculate R0
Now, power calculation P naught is V naught 1.1 to account for winding losses plus the
Vd diode drop will be the output, on the secondary side the voltage into I naught. The
turns ratio, you can calculate the turns ratio, I will use D max value and Vcmin when D
max occurs, input voltage is minimum and when the input voltage is maximum D is
minimum. So, D minimum you can calculate using that v v i min into D max is equal to
v i max into Dmin and find out the value.
Inductor design like we discussed, I have here the equation and I naught plus delta i by 2
will be the peak inductor current.
Likewise, you can design the output capacitor C the voltage across the capacitor voltage
rating and the Esr values also. Output, the diode selections the currents through the
demagnetizing diode, average current, peak current and the voltage peak inverse voltage.
Likewise the blocking diodes, peak current, average current, peak inverse voltage. As we
discussed put in the formulas there and the power switch rating also can be selected in
this fashion putting the equations have a look at it.
550
After the equations we need to have one more important program section, which is the
display of the results. So, I am using f print f, both MATLAB octave will permit that.
So, FARWARD CONVERTER SPECIFICATIONS, start printing the spec, designer

variables, then the switch ratings, diode ratings, blocking diode ratings, inductor value
and the capacitor value use the variables and then give meaningful statements.
Now, this if you run executed, it will design the values for these set of specifications.
Now, let us run that and see I am already in that resource file, resource folder. You can
551
see that in octave workspace forward design, let me run that. So, on typing forward
design do not put dot m just without the dot m. So, you execute it, you will see that the,
all the commands are executed, you get the specs, the designer variables and then the
power switcher ratings, diode ratings, all are calculated and you get the entire design of
the forward convertor.
We have not include the transformer here, because we have not studied the transformer.
Once, we do the study of the transformer and inductor magnetics, which I will do in the
next session, then you will be able to use the magnetics portion of the design also and
then we will update this file to include the magnetics design. I will request that you do a
similar such m file for flyback converter also and try it out.
552
Prof. L. Umanand
Lecture - 61
Magnetics Design Intro
In this session we will discuss one of the most important aspects in switch mode
converters and that is design of magnetics. Earlier in the rectifier portion of our
discussion we discussed power supply where, the main voltage is stepped down by using
a transformer and a transformer can act as isolation and inrush surge limiter. So, we saw
a transformer come in there in a symbolic fashion in the circuit. How do we make such a
transformer? How do we fabricate that?
Later when we discussed the non-isolated converters i.e. the primary converters: the
buck, boost and the buck-boost there was a magnetic component that came in all this
convert and that is the inductor. We used the inductor symbolically and calculated its
value, but how do you physically realize it such that you obtain that calculated or the
designed value such that it gives you a designed value of “L” as we required. So, how to
design such inductors?
Then later we discussed isolated converters like the fly-back converter and the forward
converter and both these converters had inductors as well as transformer isolation. So,
553
how to design the magnetics of these transformers and inductors which will give you the
desired electrical effect? So, this is the aspect of the discussion that we will be carrying
forth now, so that we will be able to design inductors and transformers for our
converters, but the focus of the class, in this session is the design of magnetics.
I have here a transformer. You can see this is a low frequency transformer and these are
laminations, this is Cold Rolled Grain Oriented (CRGO) steel, silicon steel and you see
the former here and on the former the winding, the winding is done and then after that it
is insulated. You will be able to see windings coming out and then terminating on to the
terminals on top.
So, if I show it in this direction, you will see the terminations marked. See here you have
0 to 230 Volt, this is the 180 Volt tap, it is 230 Volt and on the secondary side you have
two 18 volt and 18 volt tap and then there is another tap here. So, this is the transformer
that we normally use for stepping down and using it along with here.
You will you will connect it with rectifier capacitor filter and then the linear ICs to get
the ±15 volt supplies. So, this is a 50 hertz transformer. Directly from 50 Hz, 230 Volt
mains.
554
This is another transformer. This is made of ferries. I will just remove this, this is an EE
core. The shape is like E and therefore, it is called EE core, this is the ferrite material and
I will remove that, there are two E-E cores. This is the former and on the former you
make the windings and the former will be something like that. I have this, this is an
empty former and on the former you will be winding the copper.
And the wound one will be something like that. You will insert it into the EE core like
this and then take the other half of the core and then place it.
555
So, this will be another transformer or an inductor, whether transformer or inductor I will
tell you later how to do that, but the core will be still the same. So, this is a ferrite based
EE core high frequency transformer or inductor. This is another bigger size EE core
where you have again two half’s and makes up the complete EE core and this can be
used for either transformer or inductor.
And we will discuss how to go about doing the design. You will get various sizes of this
type of EE cores, but there are many shapes not only EE cores, you have UU cores and
you have pot cores. I will show you an example of a pot core.
You see this is a pot core. It is totally enclosed, best for electromagnetic interferences
where it rejects all electromagnetic interference going out and coming in and it has two
halves replace them like that, and there is a former here. You see this former, it has a
multi section; you get single section former, multi section former, you can have the
primary, one secondary or two secondary like that and then wind it and place it into this.
These cores can also be used as high frequency transformers or high frequency inductors.
So, like this there are various shapes. You have RM cores and many other shapes even in
the EE cores. With circular section the central section can be circular. So, all those
possibilities are there. Look at the Siemens data sheets, the Siemens cores are pretty
popular. Look at the Siemens core data sheets to understand the different types of shapes
556
and materials that are available. These are all ferrites on the high frequency side we
normally use ferrites and on the low frequency side we use the CRGO laminations.
This is a power supply board and you see of course there are the power switches
mounted on the heat sinks. I want to bring here focus on to these components here; these
are the pot cores in this particular power supply board. There they have used the pot
cores for inductors and transformers high frequency switching at 50 kHz. So, this is how
it will get integrated into the PCB.
557
Another example of the PCB you see this, these are high frequency again ferrite EE
cores, this is EE core, there is another EE core here. So, both transformers and inductors
are done here in this board with EE cores. So, you see there are different cores, different
core materials, different core shapes, different core types, different frequencies for which
you can operate. You have ferries, you have silicon steel, you have move metal, you
have amorphous materials like met glass amorphous glass and powdered iron. Many
materials have different magnetic properties. So, different shapes and sizes all these are
available in the market which you can effectively use for making these inductors and
transformers.
So, I have given you an idea of how it looks like and where it will be used and how it is
used just to give you a visualization. And, now we will go into these inductors and
transformers in detail and see the equations leading to fabricating these inductors, such
that you get an electrical effect as we required.
558
Prof. L. Umanand
Lecture - 62
Magnetics review
Let us now discuss inductor design and before going into the inductor design let us review
some of the magnetic principles. Assume that you have a magnetic core material like this
which has a cross section, when I take a cross section it has a rectangular section like this.
And over this core material, let us wind some copper conductor turns so, these are copper
conductor turns.
So, let a current I be pass through this, it will go and finally come out through that. So, in
the copper turn coils you will see the current flowing in this fashion. Now, you will see
that all these N coils have the current flowing in the same anti-clockwise direction now,
this is going to produce a resultant flux rate of change. So, let us say there are N turns and
then there will be a resultant rate of change of flux which is perpendicular to the cross
section area so, this is the cross section area Ac.
Now, all these turns N*i is called the MMF or the Magneto Motive Force. So, this magneto
motive force Ni, the current flowing through all the turns included is the forcing function
to produce this rate of change of flux or these changing flux lines.
559
Now, let us see, what is the fundamental graph between the power variables in the
magnetic domain? So, the magnetic domain there are two variables, I will say one is the
H or the forcing field and the other is B the flux-density here. Now, how these are related
to the power variables; in the electrical domain the power variables are V voltage and i
current. So, let us bring about some kind of a connection between them so, that we are able
to visualize and understand them a bit better.
So, let us say we have the core so, this thing is a core almost a square core and we have
the turns wound just like here. Now, let us say on the insid you have the current flowing
in and on the outside the current coming out. So, let me just take a section. So, the copper
conductors will look like this circular sections and here the current is coming out so I will
put a dot here the current is going in, I will put mark like that this coming out, going in
coming out and going in.
So, if you look at this particular you have three turns here. So, if I take a path any path
which encloses these which encloses the current some other current carrying turns here
part of them. So, if I take this along this path, now I have three such turns; so, three times
i will be the mmf (magneto motive force). Now, this is encapsulated in a more generic
form by the ampere’s law. So, what does ampere’s law state? It states that mmf is given
by if you take the integral over a path, a closed path or the force field H over the path, now
that will be equal to Ni, Ni is the mmf.
Now, in this case if I take for example, the path is the mean magnetic path for this core lm
let us say, right through the centre in this fashion. Now, it is enclosing three sections where
the current is going in. So, therefore, it will be 3*i in this case and if there are N turns it
will be Ni. So, H* lm=N* I, and H =Ni / lm. Now, this is what you have on the x-axis which
is Ni / lm you see that, N is fixed lm is fixed constant, i is the electrical parameter to which
H is linked and you have the current or one part of the power variable coming in picture
here.
Now, B let us take another law which is operative with the magnetic domain which is the
faraday law. The faraday law talks about the EMF, there is induced across the coil ec, is
the induced EMF across the coil and this ec whatever is the rate of change of flux within
the coil. So, this is another fundamental relationship that we will be using very much
especially during design of this inductor and transformers. So, if you take this flux Ф. Flux
560
𝑙 Ф
Ф is nothing but, ∫ 𝑒𝑐 𝑑𝑡. So, this is the flux relationship and the flux density B is ,
𝑁 𝐴𝑐
flux per unit area or weber per unit area, Ac is the core cross sectional area.
So, you see that flux here is proportional to the integral of the voltage with time, there is a
Ф 𝑙
voltage function coming into the picture; B is 𝐴𝑐
and flux itself is 𝑁 ∫ 𝑒𝑐 𝑑𝑡. So, you see
the voltage coming into this axis. So, the voltage and the current and this would be the
fundamental B-H curve or the fundamental B-H axis that one will be using in most
magnetic design and analysis.
So, if you take this B-H curve; take this B-H axis and take a typical magnetic material let
us say it is ideal then, it will look something like this a straight line and it has a slope and
the slope is called µ, µ is the permeability. So, 𝐵 = µH. So, this is one fundamental
relationship 𝐵 = µH, where µ is the permeability which relates the flux density and the
forcing field which is Ni / lm.
In practice in a real core, the B-H curve is not a straight line like this there are non-
idealities. So, the first one is that you will see that there is knee and then it flattens out into
a saturation like this, the positive side and also on the negative side there is a knee and
flattens out in this fashion. So, if I take what does it mean, what does this kind of saturation
mean. If you take an operating point here and going down there is some value of force
561
field, there is Ni / l that is applied at this point. If you take another neighbourhood point
you will have some value of H field here.
Likewise, another neighbourhood point you will have H field here so, there is a difference
of ΔH field between this point and this point, but how does it reflect or project onto the
B axis. So, on the B axis they all appear to be at the same point; we call that one, let us say
as the saturation flux density because there is no change in B; the ΔB =0 in all these
regions in this flattened out region. Therefore, you say ΔB =0 there is no induction, there
𝑑Ф
is no action of there, because of that here in this region µ = 0, permeability is 0 i.e. no
𝑑𝑡
longer acting as an inductance or as a transformer in this region.
So, we should not operate in the saturation region and most of our discussions in designing
L and the transformer is in the linear region, but you should know that there exist the
saturation zone and then you should not go near that. Likewise on the negative side also
you have a symmetrical value -Bsat. So, not only that you have the saturation non-linearity
there is also history or memory in this most of the course they seem to remember. So, you
will see that when you magnetize it, when you apply Ni mmf, it will reach this operating
point and then when you remove the current it will go and settle here, there is some kind
of flux; resident flux density in the core.
So, you have to apply a negative if you take Ni / l, negative mmf so, that it becomes 0 here,
then starts going negative. Likewise in the other direction also at this point there is a even
though you have removed the mmf, there is some amount of resident flux density at this
point which can be removed only by application of a positive mmf and so on. So, this is
called the hysteresis loop which is coming because of the memory present in every
magnetic core.
Now, if you see take the an area within this; the shaded area within this, every cycle the
amount of energy that is enclose in this shaded area because if I take one axis as the
potential variable, other axis as the kinetic variable. The product of them will be the power
variable and integral of that over a period will give you the power. So, within a cycle the
area under this is a power variable and that is amount of loss that goes within the core and
that is called the hysteresis loss. So, this loop is the hysteresis loop, significance is that the
amount will be the number of times this hysteresis loop is travelled will results in covering
562
that much amount of area which will result in that much amount of energy being lost and
that is called the hysteresis loss.
But for the purposes of design, let us not complicate with this complex non-linear curve;
B-H curve we will reduce the complication in this fashion. Let us take a simple, linear B-
H curve and flatten it out at the ends to account for saturation. So, we will call the top
flattening as +Bsat, the bottom negative side has minus -Bsat. The operating point well
below the saturation will be the maximum operating point and an operating point below
on the negative side just above the -Bsat will be the negative max operating point. To ensure
that the system or the core is of having an operating point well within the saturation in this
green portion of the curve.
So, there are many core materials available the most popular is the ferrites for the high
frequency switched mode converter applications. So, the ferrites have a Bsat of 0.3 Tesla
and we normally operate at a maximum flux of 0.25 Tesla for inductors and 0.2 Tesla for
transformers. Likewise on the negative side, you do not go below minus 0.25 Tesla and
these are negative sat is minus 0.3 Tesla.
So, there is another core material called powdered iron. Powdered iron is also used for
high frequency applications, it has 1.5 Tesla as Bsat and we normally do not go beyond 1.4
Tesla; likewise on the negative side - 1.4 Tesla and -1.5 Tesla or the limits.
563
We have another common material called the CRGO. CRGO is Cold Rolled Grain
Oriented; cold rolled grain oriented silicon steel. This is one of the most common material
used in all low frequency transformers where you see the laminated transformers that is
the laminations it is generally CRGO cold rolled grain oriented silicon steel and it has 1.2
Tesla as Bsat and we normally limited to ±1.1 Tesla, saturation is ±1.2 Tesla.
Another cousin of the cold rolled grain oriented silicon steel is the CRNGO which is Cold
Rolled Non Grain Oriented silicon steel. It is also used for transformer; low quality
564
transformers, but mostly CRNGO is used for making machines; induction machines. It has
1.1 Tesla saturation and we limit the operation of the maximum flux density to ±1 Tesla.
But generally, for good quality low frequency transformer cores, it is CRGO that is used
and not CRNGO.
But for our discussion and of the switch mode converters we will mostly use ferrites which
has ±0.3 Tesla as the saturation flux density and we limit for the inductors between ±0.25
Tesla for the transformers ±0.2 Tesla.
565
Now, this has in order to understand the operation of the inductor and the transformers, we
should understand the energy or the power flow from the electric to magnetic domain and
from magnetic domain to electric domain how it reflects back. So, this understanding of
power or energy flow across these domains electric - magnetic; across an electric magnetic
interface is very important.
So, let us consider that, on the one side you have an electric domain and on the other side
you have a magnetic domain within the core. So, electric domain is measurable with
oscilloscopes the parameters like voltages and currents are measurable whereas, in the case
𝑑Ф
of the magnetic domain it is within the core and the parameters and the mmf are
𝑑𝑡
invisible.
So, let us say you have an interface boundary between the electric and magnetic domain
and we would like to see how the energy exchanges across these domains. So, let us create
a port. So, this is a port and it is through this port there is an exchange of energy between
these two domains. So, the energy can flow from the electric to magnetic and it can flow
from magnetic to electric both ways.
So, let me create an invisible membrane across this electric magnetic interface and let me
mark two terminals here on the electric side, you have an inflow of current and current that
comes out from the other terminal, there are two terminals here. So, we say there is a
current input, a current part of the power variable. Power has two variables, one is the
566
current and the voltage in electric domain, current is the kinetic variable and the potential
variable is the voltage. And across the terminals there is a potential and we will say it is
the potential of across the coil ec voltage across the coil.
Now, likewise on the magnetic domain let us say there are invisible terminals and there is
potential across invisible terminals and that is the mmf (magneto motive force), very
similar to the electro motive force on the electrical side. Now, from the invisible terminal,
there is some flow kinetic part of the power which goes out and comes back in and that is
𝑑Ф 𝑑Ф
and completes the circuit in this fashion. So, is equivalent to the current i, mmf is
𝑑𝑡 𝑑𝑡
equivalent to the electromotive force.
Now to just verify that energy or power is same across the domains so, basically their units
should be same and power should be conserved. So, let us say the power on the electric
side what is it? And what is it on the magnetic side? let us try to find out. Now, how is
𝑑Ф
mmf related to the electrical parameters, it is Ni that we saw, ampere-turns and how is 𝑑𝑡
𝑑Ф 𝑑Ф 𝑒𝑐
related to the electrical from the Faraday’s law we know, ec = N . So, = 𝑁.
𝑑𝑡 𝑑𝑡
So, if you take power on the electrical side which is nothing but ec* i we know is vats, on
the imaginary invisible magnetic domain side power is equal to its potential variable mmf
𝑑Ф
into its flow variable which is kinetic variable which is and if we substitute back you
𝑑𝑡
see that Ni*ec/ N = ec* i. So, power is conserved it is one on the same and has the same
units flowing across the electrical to the magnetic domain. And therefore, mmf has to
𝑑Ф
represent the potential variable and has to represent the kinetic variable. So having
𝑑𝑡
made this equivalence, let us symbolically represent the interface between the electric and
the magnetic domain.
567
So, let us draw a block like this, now let us say this is the interface between the electric
and the magnetic domain we will give it a name later. And on the primary side I am putting
indicating a bond and I am indicating a half arrow to indicate energy flow direction. Energy
is flowing from the side into the magnetic domain in this fashion. So, this is the energy
flow direction this is just to indicate the energy flow direction.
And on this electrical side there are two variables energy variables or power variables one
is the potential variable called ec the other is the kinetic variable called the current. On the
𝑑Ф
magnetic side also we have a potential variable called mmf and the kinetic variable 𝑑𝑡
always the product of potential and kinetic variable in any domain has to be power, that
𝑑Ф
we have verified ec * i is the power in the electrical domain, potential into kinetic mmf * 𝑑𝑡
is power in the magnetic domain.
So, we will say that these two domains are related by one parameter that is called the turns
or the number of turns which relates these two the variables on both the sides. So, ec is the
𝑑Ф
potential variable current is the kinetic variable likewise, is the kinetic variable mmf is
𝑑𝑡
the potential variable.
Now, the relationship between these variables a very unique in the sense that, the potential
variable of the electric side voltage is related to the kinetic variable of the magnetic side
𝑑Ф 𝑑Ф
which is by the Faraday’s law which is ec=N 𝑑𝑡 . And the other variable, kinetic variable
𝑑𝑡
568
of electric side is related to the potential variable of the electric side, the cross linkage and
there mmf = Ni; again, with the parameter N, N coming in both the cases therefore, N is
the parameter.
Such a device, where you have a port, a power port separating two different energy
domains, one is electric and one is magnetic in this case and the relationship between the
power variables on both sides of the domains. Potential variable linked to the kinetic and
kinetic variable linked to the potential cross linkage is called a gyrator. So, the gyrator is
the port element that links the two domains magnetic and the electric by this cross linking
mechanism. So, keep this in mind that this will help in doing the derivations later.
569
Prof. L. Umanand
Lecture - 63
Permeance
There are another couple of terms that you will come across whenever you are working
with magnetics and that is Permeance is one of the term, another is Reluctance they are
very well connected. And let us see what this permeance and reluctance means. Let us start
with the basic equation B=µH, this is coming from the B-H characteristic for the material.
Flux density is equal to µH i.e. Ni/lm. So, let us replace flux density, 𝐵 =
𝑓𝑙𝑢𝑥 Ф 𝑁𝑖
= = µ𝐻 = µ .
𝑎𝑟𝑒𝑎 𝑜𝑓 𝑐𝑟𝑜𝑠𝑠 𝑠𝑒𝑐𝑡𝑖𝑜𝑛 𝐴𝑐 𝑙𝑚
µ𝐴𝑐 µ𝐴𝑐
So, flux can be written as Ф = 𝑁𝑖 . This term here is called permeance, Λ. So, it
𝑙𝑚 𝑙𝑚
is called permeance and therefore, flux is equal to permeance* mmf this is one important
relationship, it is as important as Ohm’s law in the electrical domain. This is also written
slightly in a different manner, there is another term called reluctance which is frequently
used and it is defined basically as 1/ Λ, inverse of the permeance. And therefore, this
𝑙 𝑁𝑖
becomes ℜ=µ𝐴𝑚 . Now, if you reframe this you will see that, Ф = ℜ
. So, phi is equal to
𝑐
mmf by reluctance this is also a popular way in which Ф flux and mmf are related.
570
In fact, many of the literatures sometimes people confuse this with the Ohm’s law in the
electrical domain and call Ф as current and mmf is the potential and ℜ equivalent to the
resistance. Reluctance is not equal or equivalent to the resistor, reluctance is not dissipative
𝑑Ф
quantity and Ф here is not a kinetic term, is the kinetic term as we saw to have to
𝑑𝑡
𝑑Ф
conserve both the units and power it is which is the kinetic term in the magnetic domain
𝑑𝑡
and mmf is the potential term.
𝑑Ф
So, let us try to understand this a bit further, we know is the kinetic term or the kinetic
𝑑𝑡
variable in the magnetic domain and it is similar to the kinetic term in the electric domain
𝑑Ф
which is current i. So, if I integrate both the kinetic terms, ∫( )dt and ∫ 𝑖dt, what do we
𝑑𝑡
get. We know ∫ 𝑖dt is nothing but charge, Amp-second. So, this term is nothing but charge
𝑑Ф
Q and what is this term ∫( 𝑑𝑡 )dt is, nothing but ∫ dФ or Ф. So, this is Ф or the flux.
So, if you see the flux (Ф) is linked more with the charge in the electrical domain. You
substitute there in this equation, you see how well they fall in place. Now for Ф, I will
replace it with charge Q and with permeance I will put it as C and mmf is V; Q =CV. This
is a well-founded relationship in the electrical domain and you see the equivalent
relationship in the magnetic domain Ф= Λ(Ni).
So, permeance is actually behaving like equivalent to the capacitance and in fact, it is so.
571
The energy storage within the magnetic domain is happening in the permeance it is
happening in the equivalent capacitance in the magnetic domain. So, what is reluctance?
Reluctance is 1/C and nothing but elastance. Normally the permeance is a much more
visualizable term than the elastance; so, elastance is 1/ C. We will be using permeance and
the term reluctance in our design of the magnetics frequently therefore, let us just expand
this terms a bit more.
Let me draw the B-H curve, straight line like this and the saturation parts; now, this is the
B-H curve of a core. Now, what will be the B-H curve of free space, no core in it. So, the
slope will be very very low, almost horizontal, but I will show it zoomed with a finite slope
like this.
Now, this is the character of free space and it has a value µ𝑜 , we will call that one as mu
naught the permeability of free space and it is a constant, it is 4*pi* 10 ^(- 7) it has units
Henry/ meter. Now, this is a constant, universal constant, it is the permeability of free
space.
Now the moment I introduce a core, we will see that the slope jumps to a very high value
and that is µ. Now this µ is having µ𝑜 slope of free space into some correction factor µ𝑟 ,
µ𝑟 is the relative permeability it is an unit less quantity it is just a number. So, this for
ferrites varies normally between 2000 to 2500.
572
What it basically means is that, if this is the slope of free space; the slope of the B-H curve
in ferrites will jump by a factor of 2000 to 2500 and therefore, it becomes almost vertical.
So, this is the relative permeability. So, µ is having two parts, the permeability of free
space and that of and that constant part are unit less part which is the scaling factor which
scales the free space permeability to get you the permeability of that in the core.
µ𝐴
So now if you take this equation Ф= Λ(Ni), permeance is given by, Λ= 𝑙 . Permeance is
𝑚
µ𝐴𝑐
nothing but and µ gets expanded into µ𝑜 µ𝑟 , µ𝑜 that of free space is 4*pi*10^(-7) and
𝑙𝑚
µ𝑟 , relative permeability of that particular core which I have going to use. Ac: Area of
cross section of the core by the lm: mean magnetic path length for the core.
𝑙𝑚 𝑙𝑚
Likewise reluctance is 1/ permeability which is µ𝐴𝑐
, and µ can be expanded . So,
µ𝑜 µ𝑟 ,𝐴𝑐
this is the relationship between reluctance and the core cross section area, permeability
and the path mean path length and also between permeance and the core cross section
permeability and the mean path length.
573
Prof. L. Umanand
Lecture - 64
Inductor value and energy storage
Consider an inductor which is carrying a current 𝑖𝐿 through it, inductor of value L and there
is a voltage across the coil 𝑒𝑐 and we have discussed while discussing converters how to
find the value of L using the Faradays law. The faradays law related to electric currents
𝑑𝑖
𝑒𝑐 =𝐿 𝑑𝑡 . So, the rate of change of current into the inductance will give you the voltage
across the coil.
Now is it possible to get the value of inductance L from the other parameters, physical
parameters and magnetic parameters of the inductor? Now that is what we will try to find.
See typically this is a circuit symbol, what we have written here. In reality you will see
that there is a core as I said ferrite core or CRGO core or any of the magnetic core materials
on which you will wind some turns.
And it will have these two terminals just like here and then a current will go in through
and come out; this is 𝑖𝐿 and voltage across the terminal will be the voltage across the coil
𝑑Ф
𝑒𝑐 and this coil may have N turns and within the core it will set up a flux and this core
𝑑𝑡
574
will have some predesigned value of permeance (Λ) shown here. Now with respect to
such a physical fabricated inductor what is the value of the L or how is L related to these
core parameters and physical parameters?
So, let us back go back to the fundamental equation Ф=Λ(N𝑖𝐿 ). Recall for as a comparison
you can jog your memory using charged Q = CV; Ф flux is equivalent to charge Q,
permeance is equivalent to C, mmf is equivalent to the potential and there is what this
equation is. Now Ф=Λ(N𝑖𝐿 ), can be written as (N𝑖𝐿 ). Let me now differentiate this
𝑑Ф 𝑑𝑖𝐿
equation; =𝛬𝑁 .
𝑑𝑡 𝑑𝑡
𝑑Ф 𝑑𝑖𝐿
Now, let me multiply both sides with N, so 𝑁 =𝛬𝑁 2 . Now if you look at this is
𝑑𝑡 𝑑𝑡
𝑑Ф
nothing but Faradays law 𝑒𝑐 = 𝑁 𝑑𝑡 , voltage across the coil and this should also be equal
𝑑𝑖𝐿
to 𝑒𝑐 and we see here the Faradays law from the electrical perspective which is, 𝐿 .
𝑑𝑡
𝑑𝑖𝐿
So, this is the form 𝐿 and therefore, this L must be equal to this 𝛬𝑁 2 . So, the inductance,
𝑑𝑡
L= 𝛬𝑁 2 . See that inductance is related to the permeance, magnetic properties and the
physical properties have important relationship, will be using it in design of the inductor
remember this.
µ𝑜 µ𝑟 𝐴𝑐 𝑁 2
Now, let us expand permeance; 𝐿 = , 𝑙𝑚 is the mean magnetic path length.
𝑙𝑚
575
So, you see that the inductance is related to the permeability, the core cross section
area, the number of turns, mean magnetic path length all the physical and the magnetic
properties.
So, this is one way of getting the L from the physical magnetic properties in the physical
𝑑𝑖𝐿
properties. The other way is to get from 𝐿 𝑑𝑡
which we obtained from the electrical
requirements of the convertor. So, these two we have to map and match to fabricate the
inductor. Of course let us see how we will go about do it, some more fundamentals and
then we can design the inductor.
Let us now discuss the energy that an inductor can handle. From the electrical perspective
1
we know that EL, let us say EL is the energy in the inductor is 2 𝐿𝑖 2 , i is the current at a
given instant, so this is the instantaneous energy at an instance when the current is taking
1
some value i. So, 2 𝐿𝑖 2 is the energy in the inductor and this is coming from the electrical
perspective by virtue of the motion of the electrons.
Now, let us see we can get the energy in the inductor from the magnetic perspective. Now
1 1 𝑁𝑖 2
𝐿𝑖 2 can be written as 2 𝐿 ( 𝑁 ) ; N i is nothing but mmf, so mmf-square by N-square .
2
𝐿
Now what is L by N-square? is nothing but permeance because we know that 𝐿 = 𝛬𝑁 2 .
𝑁2
1
So, this nothing but permeance, so we can write 𝛬(𝑁𝑖)2. So, this is the magnetic
2
576
1 1
perspective for the energy in the inductor2 𝛬(𝑁𝑖)2, this is 2 𝐿(𝑖)2, L in the electrical angle.
Consider the electrical equivalent potential storage, how does it look like? We know
permeance is equivalent to capacitance. So, potential mmf is nothing but voltage. So, this
1
will translate into an equivalent as 2 𝐶𝑉 2 , so we can be satisfied that the equivalence works.
Now, let me further introduce the small modification, permeance* mmf; I will take that
1
together into mmf. Now permeance into mmf is nothing but phi Ф, 2 Ф(𝑚𝑚𝑓), this again
1
from the equivalence you can say it is 2 𝑄𝑉, so here again equivalence works. So, you see
1
that 2 Ф(𝑚𝑚𝑓) is still energy, now Ф I want to express it in terms of flux density B.
1 Ф Ф
So, let us say 𝐴𝑐 𝐴 (𝑚𝑚𝑓), : the flux density B. So, I can write that one as
2 𝑐 𝐴𝑐
1
𝐴𝑐 𝐵(𝑚𝑚𝑓). So, this becomes the energy that is stored in the inductor from the magnetic
2
perspective.
Now, let us say Ac for a given core; Ac is fixed and for a given core let us say ferrite, Bm:
the maximum operating flux density, Bm can be fixed, as I told for ferrite we can fix it at
around 0.25 Tesla. So, like that if I know the core type, core material you can fix this Bm,
so these two are fixed constants. So, then energy is directly proportional to mmf or current
because mmf is N* i; N is again a fixed quantity, energy is directly proportional to the
577
current. So, greater the current greater will be the energy that can be stored in the inductor.
So, how do I increase the storage capacity of the inductor? Let us see that one.
It is of interest for us to see that for a given size of the core and flux density we would like
the inductor to handle more and more energy. So, let us understand this by drawing the B-
H axis and let us take a typical core where you have the B-H curve like this; this is the
saturation, negative saturation and this has a slope of permeability mu or mu1 in this case.
So, let me fix the operating flux density, so this is the operating flux density on the positive
side we will call that one as Bm, likewise operating maximum flux density on the negative
side -Bm. So, we know that the flux density in the core will swing only between ±Bm,
within that and never crossing that.
So, if we limit the flux density to this value, what is the maximum energy that this inductor
𝑁𝑖1
can store? So, if you take the intercept here; here the field will be , this is the H field at
𝑙𝑚
this point. So, when a current 𝑖𝐿1 is flowing then we know that the energy stored in the
inductor EL1 is proportional to 𝑖𝐿1 . So, this particular core having this µ1 as the
permeability can store this particular amount of current for a current 𝑖𝐿1 flowing through
that.
Now, can I increase the energy handling capability of this core which means that, I should
be capable of handling higher currents, higher mmf which means it should go further on
578
into the x-axis. But if you start going further on into the x axis, the core will reach
saturation it will flatten out and there is no induction and therefore, permeability is 0. So,
how do you still keep the flux in operation without saturating the core and still be able to
store larger energy?
So, let us see what happens when you reduce the permeability mu? So, let us do that. Let
me draw another line here, so which is having a lower slope and this intercept here. You
see for the same Bm, the intercept here for this lower permeability core where µ1 is greater
𝑁𝑖2
than µ2 , this is and here also let say the energy EL2 is proportional to 𝑖𝐿2 .
𝑙𝑚
So, we see here that this the energy at this point for this particular current being handled
for this curve, for this core the green line is EL2 is greater than the energy that can be
maximum energy that can be handled by this core indicated by the blue line. So, if I reduce
the permeability for the same flux Bm this will go and meet the saturation much later, in
the much later in the field axis as shown here.
So which means that it is capable of handling higher currents, larger currents and therefore
larger energy. So, how do you achieve that? You have to decrease permeability. How do
you decrease permeability? By decreasing the permeance and what is permeance?
µ𝑜 µ𝑟 𝐴𝑐
Permeance we saw is .
𝑙𝑚
Look at this µ𝑟 ; µ𝑟 is the relative permeability which is in the order of 2000 to 2500 in the
µ𝑜 𝐴𝑐
case of ferrites, let us say you make it equal to 1. What is it µ𝑜 , permeance becomes ,
𝑙𝑚
which is for free space, so µ𝑟 is equal to 1 for air.
So, if you start introducing air gaps, then my relative permeability value will start coming
down for the same core and the effective permeability mu-naught into mu-r will start
reducing and the slope decreases and as the slope decreases the intercept at the Bm
maximum operating flux density will move farther onwards the farther on away into the
x-axis and the core will be able to handle higher currents and therefore, the inductor is
capable of storing higher energy.
So, this way we try to maximise the energy storing capability of the inductor. See one
primary important feature of the inductor that it should be capable of storing energy. So,
that is why we try to design the inductor which is capable of carrying higher and higher
579
currents.
In the discussions still now we have made some equivalence between electrical and
magnetic parameters let me summarise them. So, that we understand the electrical and
magnetic properties well equally well.
So, the parameters potential, potential in electrical is voltage v or emf electromotive force,
in the magnetic domain it is the magneto motive force. The kinetic parameter; the kinetic
parameter or the flow parameter in the electrical domain is current i Amps and in the
𝑑Ф
magnetic domain it is . Power in the electrical domain is emf (voltage) * current or
𝑑𝑡
potential * the kinetic variable, here also it is the same thing potential variable mmf into
𝑑Ф
𝑑𝑡
and both have units of power which is watts in both the cases.
Capacitance in electrical domain is C and the magnetic domain it is permeance. Elastance

in the electrical domain it is 1/C and the magnetic domain it is called the reluctance,
reluctance R which is 1/permeance. Charge in the electrical domain Q= ∫ 𝑖𝑑𝑡 here it is Ф
1 1
flux. In energy you have 2 𝐶𝑉 2 is replaced by 2 𝛬(𝑁𝑖)2 or you can club CV and call it as
1 1
QV and here also 2 Ф(𝑚𝑚𝑓).
2
580
(Refer Slide Time: 16:20mmf)
So, these are some of the equivalence that we saw. You can write many more equivalence
you will see that because the energy is the same across the domains, there is a very nice
strong equivalence between the electrical and the magnetic parameters. Use this
equivalence to remember the equations on the magnetic domain.
581
Prof. L. Umanand
Indian Institute of Science, Bangaluru
Lecture - 65
Inductor area product
We will be using the area product approach for designing the magnetics both inductor and
the transformer. So, let us look at what area product means. So, there are two important
areas for any core, one is the core cross section area Ac meter-square core cross section
area. What is the core cross section area? You see this part here, the central, this is the E
core you have seen this E core I have shown it to you before. On the extreme size there are
thin limbs and on the inner the central arm is the thicker one. So, this area here which I am
showing with the pencil is the core cross section area. And perpendicular to that orthogonal
to that there will be flux flow.
Next let us see another area called the window area. Window area symbol is Aw meter-
square and what is the window area? In the case of the E core, you see there is this one E
core, I will take the other one and place it on top of this like this. This will be the completed
EE core as you see it here.
Now, this, this gap here what you see here is the window area. This area here is window
area, not both together. Any one of the area only either it is this area, we have to take or
582
that, because a turn entering here also come out there. So, it is only one area that we need
to consider. So, this is the window area. So, that is what we call as window area.
And now area product Ap is Aw*Ac or Ac*Aw and this is 𝑚4 . Now Ac * Aw is called the
area product. When we do the core design for inductors and transformers we will be
calculating this area product Ap and checking against the table of core course for the area
product that matches or is greater than the value calculated. And we will select such a core
which has a greater datasheet value Ap then what we have calculated and then use it for
design. This is an iteration process of course we will now discuss how we will use this
area product to design the course for the inductors and the transformers.
Now, consider a core, let us take an EE core and we saw just now how the window, window
area looks like this the window area and this is where you will see the cross section of the
wire. So, first let me draw the former, you will have to place the former. There will be a
former section which is going to come in here, it will eat up some portion of the window
area this is the former section.
Then let us say we put in the turns. So, these are the wire cross section. So, this is one
single turn, the cross section will appear on this side of the window area and also on this
side of the window area. So, I will just put one layer, there may be many layers which will
fill this entire window area. So, this is an EE core.
583
Now, this portion here which I am showing marking on the curve with the cursor is the
window area. So, this is the window area Aw; however, you should understand that the
entire window area is not available for you to fill it up with the windings. So, there is some
loss in the available window area. What is the loss? One is the curve we just saw that there
is this coil former which is eating away a portion of the area.
Then when you are winding the cross section of the copper wire is circular in nature and
there will be air gaps in between the windings. So, that is going to take away some window
area. Thirdly, the cross section of the wire has enamel on enamel covering on top of the
copper. So, that there is insulation between the windings. So, that there is no short across
the windings.
So, this enamel over copper. So, if this the copper cross section wire on over that there is
the enamel such that there is an insulation between two neighbouring coils. There will not
be a short circuit, but this enamel cover is going to eat up some portion of the window
area. So, now, all these are going to reduce the window area by a factor called the window
factor Kw. So, this window factor Kw has the value less than 1. And the available window
area can be written as Kw*Aw. And for the inductor Kw is between 0.4 to 0.6 , should
understand that it is not easy to deterministically evaluate all these loss and available
window area, because many a times the air gap between the windings depends on the skill
with which one winds.
So, therefore, we have a range of Kw. So, if someone is the no wise who is winding it for
the first time who is the new beginner the Kw factor will be lower and someone who is
very experienced and has very good skills on winding the coils Kw will be around 0.6.
So, now I have shaded one wire cross section. So, that area which is shown black is called
𝑎𝑤 lower case w which is the wire cross section area. It is the cross section area of the
copper wire, it is only of the cross section area of the copper wire inside excluding the
enamel, because the enamel portion is taken within Kw.
So, if there are N such turns N such cross sections will be there. So, Kw* Aw should be
equal to N * (area of cross section of the wire).
584
Now area of cross section of the wire. We will write Irms, there is a rms current that flows
through the cross section of the wire and a permissible current density. So, there is a
permissible current density J and this for copper is at a default value of around 3 𝐴/𝑚𝑚2
or 3 ∗ 106 𝐴/𝑚23. So, this is an Amps, this is 𝐴/𝑚2 and therefore, the units of Kw Aw will
become 𝑚2 .
So, this is the starting value of the default value the J varies in practice from 2 𝐴/𝑚𝑚2 to
even 5 𝐴/𝑚𝑚2 in very low power systems. As the systems become higher and higher in
power the J value becomes 2 𝐴/𝑚𝑚2. 3 𝐴/𝑚𝑚2 is a reasonable value to take for most of
our design purposes.
Let me represent the rms current here in terms of max current. So, let me introduce a term
crest factor. Crest factor is Kc which is (Imax or Ipeak) / Irms. And therefore, using this relation
𝑁𝐼𝑚 𝐾𝑤 𝐴𝑤 𝐾𝑐 𝐽
here you have 𝐾𝑤 𝐴𝑤 = or Im can be written as 𝐼𝑚 = .
𝐾𝑐 𝐽 𝑁
So, this is now is relationship we will mark it as one the current the max current through
the inductor is expressed in terms of a number of turns the window area or the window
factor and the current density J.
585
𝑑𝑖
Next we go back to faradays law, the voltage across the coil induced is equal to 𝐿 𝑑𝑡 which
𝑑Ф 𝑑𝐵
is also is equal to 𝑁 𝑑𝑡 . Now Ф can be written as Ac * B, Ac is a constant. So, 𝑁𝐴𝑐 .
𝑑𝑡
So, consider these two, if the voltage is constant the rate of rise of the current is linear.
Therefore, rate of rise of the flux, flux density should also be linear. In a given time, let us
say the current rises to Im from 0 to Im, L*Im and that same time the flux density would
have reached Bm; so, N Ac Bm. So, this is another relationship where L*Im current is related
to the magnetic properties and the core cross section area (𝐿𝐼𝑚 = 𝑁𝐴𝑐 𝐵𝑚 ). So, this is
relation 2.
586
Now, let us evaluate the energy stored in the inductor. So, what is the maximum energy
1 2 1
stored in the inductor half 𝐸𝐿 = 2 𝐿𝐼𝑚 . So, this I can write it as 2 𝐿𝐼𝑚 , I will group this
1
together as 2 𝐿𝐼𝑚 ∗ 𝐼𝑚 . And we have developed relation for L*Im wherein L*Im is related
to Ac and we have also relation for Im which is related to A w, we will substitute these two
relations here and rewrite. So, L*Im, I am writing it as N*Ac*Bm and then for Im here Kw
Aw Kc J /N.
So, let us remove these two N’s and I will keep Ac and Kw to one side their product, this
is the area product Ac* Aw which is equal to 2* EL / (Kw*Kc*J* Bm). So, this is area
product Ap 𝑚4 and this is the area product equation. And this is the equation that you will
be calculating and checking it in table of cores and comparing it with the area products of
the cores. And you will pick out that core which is having an area product greater than that
which is calculated.
And that is what we will be using for the inductor design. And later on for the transformer
design too, but therefore, the transformer design the area product equation will be different,
we will come to that later at that point in time.
587
Prof. L. Umanand
Indian Institute of Science, Bengalure
Lecture – 66
Inductor Design
Let us down write down the steps for designing the inductor. Step 1. Step 1: Find the value
of the inductor now this is something that we have done when we studied the various DC-
DC converters based on the converters volt second balance relationship we were able to
find the value of L. Step 2 onwards we proceed calculate the area product. So, area product
Ap is given by 2 EL for inductor 2 *energy, max energy that the inductor can store
Kw*Kc*J*Bm.
12
The energy EL is given by 2 𝐿𝐼𝑚 . So, you know what the current flowing through the
inductor L. Values obtained from step 1, Im is known from the converter operation what is
the maximum value of current, you can calculate EL. Kw as I said is an empirical value 0.4
(Refer Time: 01:42) for novice, Kc is equal to 1 for square wave and for various other ways
you can evaluate, J is 3 𝐴/𝑚𝑚2 or 3 ∗ 10−6 𝐴/𝑚2 and Bm is the flux density 0.25 Tesla
for ferrite core materials. So, we can evaluate the value of Ap.
After evaluating the value of the area product choose from the core table. So, you have to
588
look into the datasheet of cores and go through the table of area products. I will show a
typical example of the core table.
Here is a typical core table I have taken it from the appendix of a book. So, you see cores
without air gap, ok. These are high permeability cores and you need to introduce air gap if
you want to use it for inductor purposes, for transformer purposes you can use it directly.
So, there high permeability transformer cores, but you can use it for inductors too if you
provide the proper air gap and reduce the permeability.
So, this is this column, is the mean length per turn, the mean magnetic path length lm that
we have been using in various equations is this. See, there are various types of cores part
cores this is the first the first top view rows of a pot cores the core cross section area you
see this is the core cross section area, the window area. These are numbers for the window
areas where given as mm-square and convert it to meter-square.
The area product now this is what we need to be concerned about interested in after you
calculated, you need to you need to go down this and then choose an appropriate area
product.
589
You also have EE cores you can decide on the type on the shape of the core and then
appropriately choose the core of a specific area product.
So, whatever you have calculated and let us say your calculated some 18000 you go down
and choose a core which is greater than 18000 like that and once you have chosen that note
down the A c cross section area, Aw window area. So, like that there are few other cores
you have the UU cores, you have the toroids. So, the data so, look at the datasheet to get a
list of cores there are many many cores available and mainly look for the area product.
After you have chosen a particular core from the core table note down the cross section
area Ac and the window area Aw of the selected core and subsequently after this step you
will use this value of Ac and Aw for all other next calculations.
Step 3: permeance, let us find out the permeance for the core this is the next important step
that you need to do.
590
So, consider a core, like this. I have taken a circular core, toroidal core and then I have cut
open like that and I have introduced an air gap. Now, there is an air gap lg and the remaining
is lm mean magnetic path length. What is the reluctance? Reluctance is having two parts,
𝑙𝑚
one is this part. So, I will take approximately and reluctance of this part which is
µ 𝑜 µ𝑟 𝐴𝑐
𝑙𝑔 𝑙𝑚 +µ𝑜𝑙𝑔
and you can combine them you will get .
µ𝑜 𝐴𝑐 µ𝑜 µ𝑟 𝐴𝑐
What is permeance? 1 by reluctance which is 𝑙 . So, this is the permeance of the core.
𝑚 +µ𝑜𝑙𝑔
Now, if µ𝑟 is much much greater than one like in ferrite 2000 to 2500 or even 3000, then
you will see that permeance reduces further permeance is equal to µ𝑜 𝐴𝑐 , I will divide by
numerator and denominator by µ𝑟 , lm by µ𝑟 plus lg, µ𝑟 being very large this term will
µ𝑜 𝐴𝑐
vanish becomes insignificantly small and the permeance becomes . So, in this high
𝑙𝑔
permeance cores when you want to introduce lg that is the air gap you could use this
equation to quickly get the value of the permeance.
591
Now, take the example of the E core that we saw permeance is given Henry per turns
square. So, if we take the EE core. So, let us say you want to introduce air gap. In the
centre there you can cut a piece of paper or cut a piece of miller and then place it there on
the central arm and then place the former and then on top you will place the other half of
the core. So, this will be your air gap introduced core.
Now, the air gap here that you need to introduce should be half the lg that you would like
1
to have finally, because there is 2lg here if I introduce then if you go along the core the flux
goes along the core there is an air gap which automatically comes in here air gap that
automatically comes in air, but this is a thinner cross section. So, you will see that when
1
you give 2lg here, then the overall lg throughout the core will land up as lg. So, give only
1
lg.
2
So, if you have calculated, let us say you want a particular permeance and you have
calculated that knowing µ𝑜 𝐴𝑐 and let us say for a given permeance you want this particular
1
lg or for a given lg this is the permeance, then you introduce 2lg here and use lg for
permeance calculation.
In the case of high permeability transformer cores you will introduce lg , but sometimes in
the market cores are available specifically for inductors without the air gap, but with lower
permeability. So, they have been categorised with the factor call the AL factor and AL
592
𝑛𝐻
factors are units of . In fact, AL factor is nothing, but permeance.
𝑡𝑢𝑟𝑛𝑠2
We know the inductor value L is equal to permeance * N-square. So, from this relationship
the number of turns is equal to L by permeance L by permeance square-root. So, using this
relationship you can find out the number of turns that are required for the inductor for that
particular selected core.
Next, step 4: we have to find out the wire thickness or the wire gauge. So, let us calculate
𝐼𝑟𝑚𝑠
the cross section area of the wire aw, which is , Irms the rms value of the current flowing
𝐽
𝑎𝑚𝑝
through the inductor wires divided by J the current density it is 3 𝑚𝑚2 . Then after having
calculated go to the wire table and from the wire table list choose a wire gauge from the
wire table which is having aw greater than aw calculated and then choose the wire gauge
which is normally given in terms of standard wire gauges, SWG – standard wire gauges.
For example, SWG 26.
593
Let me show you a typical wire table. This is a typical wire table wire size table. You see
the first column SWG standard wire gauges given in terms of these numbers the higher the
value of this SWG number thinner the wire.
594
So, if you go down you will see that SWG 19 or SWG 8 will be a very very thick copper
wire SWG 45 will be a hairline thin wire. So, you have various informations about the
wire the diameter with enamel area of the bare conductor this is the area of the bare
conductor is aw that we have been using in the equations and the way it so on so forth.
So, go down the wire table look at this particular column area of the wire cross section of
the bare conductor without the enamel. So, go and pick that particular row which is having
aw greater than the one calculated and choose that particular wire gauge standard wire
gauge. For example, if it is coming greater than this value you can say SWG 40 or SWG
39. So, in this way you will pick the standard wire gauges from the wire gauge table so
that you can buy it from the market.
So, after you choose you note down the value of aw of the selected wire gauge and use that
to do a window area cross check what it means is that, you have to ensure that this
inequality Kw Aw is greater than N*aw means that the window area is available window
area is able to accommodate all the turns having this wire cross section then your design
is complete for the inductor.
595
Prof. L. Umanand
Lecture - 67
Inductor example
Let us now take an example and see how to design an Inductor. Consider a buck converter,
we have discussed and studied the buck converter at some length, we know how it operates.
This is Vo, let say Vo we want is 3.3 volts and Io is 5 Amp load requirement. The switching
frequency is 20 kilo Hertz; 20000 Hertz, Vi input voltage 10 Volts ±10 percent varies from
10 Volts -10 percent to 10 volt +10 percent. This is Q D L C and Ro and our job is to design
L inductor.
So, let us try to find the duty cycle Vi max* dmin is equal to Vi min * dmax should be equal to
Vo, buck converter relationship. Now, Vi max=10 Volts + 10 percent of 10 Volt is 1 Volt.
So, 11 Volts*dmin or 9 volts * dmax is equal to 3.3 Volts; dmin from that relationship will
turn out to be 0.3, dmax is 1.1 by 3.
596
𝑉𝑜 (1−𝑑𝑚𝑖𝑛 )
We will need the dmin value, L= . While, we discussed the buck converter we
∆𝑖𝐿 𝑓𝑠
𝑉𝑜 (1−𝑑𝑚𝑖𝑛 )
had use this equation; we had developed this equation L= . I am now including
∆𝑖𝐿 𝑓𝑠
it has dmin because, dmin is worst case when d is minimum this value in the numerator within
the brackets is maximum and I will get the larger value of L.
So, that would be the worst case. So, let us plug-in these values as 10 percent of 5 Amps
which is ∆𝑖𝐿 =0.5 Amps. Here fs is 20 kilo Hertz, Vo is 3.3 Volts dmin=0.3 and you will land
up with L=0.23 mHenry. So, this is the value of L, that is the first step in the design of
inductor, find the value of L which we have done from the understanding of how the buck
converter operates.
597
Next let us try to find the area product. So, for finding the area product we need these
2 1
values, what is Im because in the energy equation we have L𝐼𝑚 . What is Im? Im is you
2
know that the current in the inductor the average value is Io and there is superimposed on
that Δi, peak to peak ripple. So, the peak value will be Im plus delta iL by 2. This we
know, we have studied 5 plus 0.25, 5.25 amps.
2 1
Then the energy in the inductor at peak current will be 2L𝐼𝑚 which will turn out to be
3.1834*10−8joules. And, now the area product can be calculated 2*EL divided by
Kw*Kc*J*Bm. We know all the values here, I am going to take Kw as 0.6 and experienced
winder will wind this, Kc is the creast factor which is Im by Io, Io is the rms value.
𝐴
You can find that out J is 33 ∗ 106 , Bm is 0.25 Tesla for ferrites we are using ferrites.
𝑚𝑚2
And, this will work out to be 1.34747*10−8 𝑚4 . Or, it works out to be 1.34747*10−8 ∗
1012 𝑚𝑚4, you will get 13474.7 𝑚𝑚4 . Now, look into the core table and try to select a
core which will give a value greater than this.
598
Now, let us look at the core table. Now, here is the core table I am now looking at the area
product column, this is the area product column. Then look at this 13474 you are having
10200 for P 30/19 greater than that 20301. Now, this is greater than the one that we have
calculated. So, it is good to choose P 36/22 pot core of course, you can choose any other
core shapes, but let us say for now for example, we choose this pot core. So, select pot
core P 36/ 22. So, this is the core that we have selected.
Now, what are the values for this core, the selected core? We have the core cross section
area of 201 mm-square. We have the window area of 101 mm-square and we have the
mean magnetic length of 53.2. So, let us write that down here; core cross section area of
201 mm-square, window area 101 mm-square, total area product which is the
multiplication of this 20100 𝑚𝑚4 and lm that is the mean magnetic length 53.2 mm. So,
these are the parameters for the selected core which is the pot core P 36/22.
599
Next let us calculate the permeance. See from the datasheet of CEL HP3C grade ferrite, I
have obtained the value of relative permeability as 2000±25%, where the worst case
permeability minimum value is 2000±25%=1500. So, I will use this minimum
permeability for the calculations in evaluating permeance.
So, in the permeance; if you look at the permeance equation , this is what we had
𝑙𝑚 +𝑙𝑔 µ𝑟
developed. We know Ac, we know µ𝑟 , we know µ𝑜 , we know lm, µ𝑟 ,lg; lg the air gap. You
pick up a piece of paper or Mylar sheet, measure the thickness of that and then you can
plug that in here. For now, I am going to take a paper of thickness 0.5 mm.
So, plug in the values; let us say µ𝑜 is 4 ∗ 𝜋 ∗ 10−7 , µ𝑟 is 1500 as given here, Ac is
201*10−6 meter-square, lm is 53.2*10−3 , µ𝑟 is 1500 again and lg is 0.5 mm so, 0.5*10−3
meters. So, this will work out to 4.717 * 10−7 Henry per turn square. So, this is the value
of the permeance.
600
So, once we have obtained the permeance, next it is easy for us to obtain the turns; number
of turns that are needed to be wound on the inductor core. This is given by the relation
𝐿
N=√𝛬. So, this is 23.13, choose after rounding it off to the upper next upper integer 23
turns and this is N. Next we have to find the wire gauge, the area of cross section of the
5
wire be calculated Irms/ J which is Io/ J. This case which is 3∗106
which turns out 1.67 ∗
10−6 meters-square or 1.67 mm-square.
Now, look into the wire table. So, look into the wire table; we need to look at a wire bare
conductor area having greater than 1.67 mm-square.
601
Let us go down this and you see that here SWG-16. Now, here in this row we will find
SWG-16 which has bare conductor area of 2.075 mm-square greater than this. So, pick
that value.
So, now let us write down select SWG-16 which has a wire cross section area of 2.075
mm-square. SWG-16 is selected and this has to be wound on to the inductor core which
we have chosen P 36/ 22. There is a last check that we need to do what is known as the
window area check, to see if the windings will fit into the available window area.
602
Aw*Kw is 101 (mm-square) *0.6 which 60.60 and N* aw that is wire cross section as we
have read out from the wire table is 23* 2.075 which is 47.75. And, you see that Aw*Kw is
greater than N*aw and therefore, the inequalities satisfied. And, you know that this number
of turns or this gauge will fit into the available window area. And therefore, the design is
successfully completed.
603
Prof. L. Umanand
Lecture - 68
Transformer design
Let us now apply our magnetic design knowledge to the design of transformers. Essentially
the transformer and the inductor will appear very similar physically it will also have a core,
it will have winding, former all those things. What essentially is the difference is in the
BH characteristics. So, let me just point out to you the main differences in the BH
characteristic of the transformer and that of the inductor.
So, let us say we have a BH graph here for the transformer and let me have another graph
here BH graph here for the inductor. So, in the inductor the main focus main job is to store
energy. So, energy storage is the main issue in the inductor we want to store as much
energy as possible. In the case of the transformer it is power transfer instantaneous power
transform of transfer from the primary side to the secondary side. So, this is the main job
in the case of the transformer, no energy storage.
So, in the case of the inductor, in order to have as high energy storage as possible we
discussed that we would like to have the mu the permeability as low as possible. So, the
core is designed to have as low µ as possible. So, that you have very high magnetisation
604
current here. The magnetisation current Ni/ lm, the i part here should be very high. So, that
2 1
the energy that is stored in the core maximum. Energy that is stored in the core is 𝑙𝑖𝑚 .
2
So, more energy get stored because the energy is proportional to the magnetisation current.
Whereas in the case of the transformer we are not talking in terms of energy storage; there
is no energy, we do not want energy storage. It is only instantaneous power transfer. So,
we would like the µ permeability to be as high as possible. So, if you look at a transformer
core BH curve it will be like this very high slope; for the inductor very low slope and
because of this high slope you will see that the magnetisation to reach the same Bm you
will need much lesser current the Ni/ lm, less magneto motive force or less electrical current
i or magnetising current i is needed to bring the transformer core to a level operating level
Bm or even Bsat.
So, we have to design the transformer to have low magnetisation current. So, that is the
essential difference. A transformer is a power transfer device whereas, an inductor is an
energy storage device and when you design we would like to design transformer with low
magnetisation current whereas, in the case of inductor we have high magnetisation current
so that the energy that is stored within the core is high.
So, with this fundamental and main difference let us look at the transformer again in terms
of the area of products and arrive at the design equations.
605
Let us now discuss the area product relation for the transformers. Now, consider a generic
transformer which has a primary coil and the secondary coil like this as shown, we have
the dot polarities. Primary, the voltage ep measured in this fashion, secondary potential
emf measured es in this fashion. Primary has Np number of turns, secondary as Ns number
of turns, primary current ip flows in this fashion and secondary current is as shown like
this.
𝑑Ф
So, if I take the primary induced voltage ep, so, it is given by Faraday’s law Np∗ , Ф is
𝑑𝑡
the flux in the core here. Now, the voltage here it is not only the voltage, it is the volt-
second that we have to be bothered about. So, let us take some arbitrary wave shape. Let
us say we have sine wave, this is the sine wave transformer. Let us say the voltage is
sinusoidal in nature. So, this is the halfway mark here. This is Ts/ 2 and this is Ts/ 2. So, if
it is supply frequency 10 milli-second and 10 milli-second if it is switching frequency,
whatever the switching frequency that switching period by 2 and switching period by 2.
Now, we are not that much interested in the wave shape as we are interested in the area
under the curve. Especially, for design of the transformers we are bothered about the volt-
second the area under the voltage profile this one. So, this volt-second we need to be
worried about. Now, this volt-second and equivalent square wave volt-second here let us
say we have equivalent square wave flat top both these will have the same volt- second let
us say. Then what is this flat top value that will be the average of this particular wave
shape.
So, whatever be the wave shape whether be the sine wave, whether it be a square wave
whether it be any other or be triangle wave or any other arbitrary wave shape take the
average of that and for the half period it will be equivalent to square wave with the flat top
and where the height is nothing but e primary average because the volt-second is the same
in all these cases. And this average remember is not the average of the entire cycle the
average of the entire cycle is zero because the average of the entire cycle in the steady state
for across any coil will always be zero because volt- second balance will be there you are
taking the average for half the cycle. This is the half cycle average keep that in mind.
Now, let us see how we are going to get that dt, we will integrate ep with respect to dt and
this also with respect to dt which becomes dФ. Then ep with respect to dt and Np, Ф can
be written as Ac * flux density. So, therefore, (area of cross section of the core) * dB.;now,
606
integrate for half the period only 0 to Ts/ 2; 0 to Ts by 2. So, in Ts/ 2 time how much will
the flux density vary? It will vary from -Bm to +Bm in most cases, but there are cases like
forward transformer where it will vary from 0 to Bm.
We will say 0 to ΔB and then we can make ΔB has 2Bm or Bm accordingly depending
upon the topology use. So, let us say some ΔBm which is the variable quantity. So, Ts/ 2
in half a period there is a change in the flux from 0 to ΔB. Now, on integration you will
see that this is a square ep average*Ts/ 2 and this is Np* Ac*ΔB.
So, now rearranging you will get Np is nothing but ep average, half cycle average remember
that 2 then you have Ac here core cross section area coming down ΔB and the Ts will be
written as fs; Ts is in the numerator will make it as fs.
Now, this is the relationship between the turns and the half cycle average of the voltage
𝑒𝑝𝑎𝑣𝑒
across the primary or any coil : 𝑁𝑝 = 2𝐴 .
𝑐 ∆𝐵𝑓𝑠
𝑒𝑝𝑎𝑣𝑒
Likewise for the secondary also you can write it down Ns is nothing but , 𝑁𝑠 = 2𝐴 . So,
𝑐 ∆𝐵𝑓𝑠
you see that there is now a relationship between Ac the core cross section area and the volt-
second or the voltage. So, Ac gives an idea of the voltage supporting capability or the volt-
second supporting capability for the transformer.
607
Next, let us talk about the window area Kw*Aw is the available window area; just like in
the case of the inductor we have loss of window area due to coil former thickness. We
have to insert coil former into the core and then the thickness of the coil former that amount
will go unavailable for the windings. So, that much amount decrease in the window area
is there. Not only that when you have circular cross section coils one there is always air
gap in between the coil sections and that will eat away some area.
Apart from that we have seen that enamel over the copper; copper cannot be just pure
copper because copper will come in contact with the next turn and there will be short
circuiting of the turn therefore, you need to insulate and therefore, copper is normally
available as enamelled copper there is insulation which acts as an insulation and avoid
short circuit between coils.
Now, this enamel has a finite area and that will also eat into the available window area.
Now, in the case of the transformer there is another additional loss of area and that is due
to inter winding insulation. Normally transformers at least minimum two windings are
there; one primary, one secondary and in some other applications you may have multiple
secondaries in which case for every winding there has to be an insulation between that
winding and the one previous previously wound.
So, this inter winding insulation can eat up a significant amount of area. So, the available
window area is much lesser than that for an inductor.
608
So, Kw is will vary between 0.2 to 0.4. In the case of the inductor Kw varied between 0.4
and 0.6, but in the case of transformer because of the additional loss of area the Kw or the
window factor vary between 0.2 to 0.4 and no wise unskilled winder should use 0.2 and
very skilled winder will use 0.4.
So, therefore, the available window area has to be used to accommodate Np primary
number of turns into the wire cross section of the primary wire plus Ns number of
secondary turns plus wire cross section of the secondary wires; if it is just a primary and
the secondary, but in the general case there can be multiple secondaries. You can have a
primary, secondary 1, secondary 2 so on to secondary m. In such a case Kw*Aw should
accommodate Np wire cross section area of primary plus Ns1 wire cross section of
secondary 1 plus Nsm into wire cross section area of secondary m and so on.
So, that is something that let us now take just these two primary and secondary and we see
that Np and wire cross section of the wire can be calculated from the rms current flowing
through that let us say Iprms / (the current density J), just like as we discussed in the case of
the inductor windings. And, secondary you can calculate Ns * Isrms/ J.
So, now we saw Np and Ns we have a relationships here with respect to the half cycle
primary average voltage and half cycle primary secondary average voltage with respect to
the other parameters core cross section area. We will replace it here accordingly.
609
So, (ep-average /2*Ac*ΔB*fs)*( Iprms/ J) plus for Ns I will put (es-average half cycle average/2
Ac*ΔB*fs) * (Isecondary-rms /J).
So, now you see here Aw, Ac; now let me bring Aw and Ac as a product to one side and
then put all the remaining terms. So, you see that the denominator here (2*Ac*ΔB*fs),
take it out common. Ac and Aw we keep it on one side, Kw bring it down. So, you will
see 2*Kw*ΔB*J*fs all this come in the denominator and in the numerator summation of
the half cycle averages and the rms currents.
So, ep-average is the half cycle average* Iprms plus es-average half cycle average * Isrms. So, this
is the generic equation for the area product. If there are multiple secondaries like here so
on it will the half cycle average of that across that particular winding and the rms current
through that winding will come depending upon as many windings that are there.
Now, this is an important relationship a general area product relationship for any given
transformer and we just have to calculate ep-average and Iprms and ΔB for specific application
of the transformers and you will get the area product for that specific application.
Let us find out what is the area product for sine transformers from the generic area product
relationship that we just developed. The area product generic area product is given in this
fashion we have here on the primary side ep-average half cycle average Iprms, es_average
secondary half cycle average, Isrms and the denominator portion given by the winding factor
610
ΔB*J*fs. We need to find out for sine transformers what is the ep-average es-average and ΔB.
So, let us look at ΔB first. If you take the BH characteristic and let me say that a typical
BH characteristic something like that one this is plus Bsat, minus Bsat and let us say the
transformer flux. The flux in the transformer cores swings from positive Bm value to a
negative Bm value minus Bm and that is plus Bm.
So, which means total swing, so, for every cycle it will go minus Bm to plus Bm and then
back minus Bm which means in a Ts/ 2 half the cycle it would traverse minus Bm to plus
Bm or plus Bm to minus Bm. So, at ΔB of 2Bm, 2Bm swing it will undergo in Ts/ 2 time.
So, this portion gets replaced with 2Bm.
Next for the averages here if you take a sine wave, typically if you think of the half cycle;
the half cycle average if this is Vm; Vm*sin( 𝑤𝑇) average is ep-average for the half cycle will
be 2Vm/ pi and we know the rms value is Vm/ sqr-root (2) and the relationship between
rms and the half cycle average will be 1.11 times the average. So, therefore, ep-average will
replaced by Vrms/ 1.11. Likewise the secondary average will be Vsecondary-rms /1.11, so, we
can take that out common.
So, let us make that substitutions. So, Ap is Vprimary-rms* I rimary-rms plus Vsecondary-
rms* Isecondary-rms divided by 4.44. How does 4.44 comes is 1.11 because this is Vrms
by 1.11. This is another Vsecondary-rms by 1.11 that will get taken common come down
611
which will come 2.22. ΔB is having 2Bm, so, it becomes 4.44* Kw* Bm*J*fs.
Now, Vprimary-rms*Iprimary-rms is the power on the primary side. Vsecondary-

rms*Isecondary-rms is the power on the secondary side, we will call this power as Po. So,
for the transformer this is the Po=Vsecondary-rms*Isecondary-rms all sinusoidal
quantities and Pi is Vprimary-rms* Iprimary-rms, Pi we can say is Po/ efficiency.
And, generally efficiency of transformers are much higher than 80 percent; if we take a
conservative value worst case value of 80 percent then we will be rating the primary for
Po/ 0.8 and that is very very conservative. Generally the efficiency of the transformers are
much higher and therefore, if we design for a primary side Vprimary-rms* Iprimary-rms
of Po/ 0.8 we will definitely be on the safe side.
So, therefore, Ap will become (Po/ 0.8+ Po) divided by the denominator Kw*Bm*J*fs and
if you simplify that becomes 2.25* Po/(4.44* Kw*Bm*fs). Now, that becomes the area
product for sinusoidal transformers. Can be used for low frequency operation for where
you are using for rectifier capacitor filter power supplies, where you have to step down
230 Volts to 18 Volts. You can use this relationship for the design of those transformers
because they are sinusoidal operated where fs will become 50 Hertz you have to use CRGO
– Cold Rolled Grain Oriented silicon steel as the core material which has a Bm of 1.1
Tesla.
612
Now, let us take another example and that is of a switched mode DC-DC converter. We
have studied and discussed the forward converter. Let us try to find the area product of the
forward converter transformer. We use the generic Ap formula: eprimary-average*half cycle
Iprms plus esecondary-average *half cycle Isecondary-rms by (2*Kw*ΔB*J*fs).
First ΔB; in the case of the forward transformer we know from the BH, I will write the
BH curve. So, you have the typical BH curve of the core. The swing in the B is from 0 to
Bm only because the case of the forward converter, the forcing voltage is applied only in
one direction, unidirectional and therefore, it will swing from 0 to Bm and Bm to 0. So, in
Ts/ 2 time maximum it will swing from 0 to Bm and therefore, ΔB is nothing, but Bm.
So, this will be Bm.
So, let me show you again the way from that we had drawn for the forward converter, let
us recall that. Observe this is the forward converter, observe the voltage across the coils of
forward converter primary and the secondary. When the switch is on the voltage across
Vp is Vi and across the secondary it will be n* Vi and when the switch is off the
freewheeling power portion will come and the area will be same as Vi* dTs during the
secondary during the portion of the time 1-dTs.
So, I will use that equation during the time when the switch is on and let us say for the
forward converter the maximum on time or maximum dTs or d can be 0.5. So, during that
613
time Ts/ 2 when this is on you will see the voltage across the primary as Vi, voltage across
the secondary will be n* Vi and that would be the half cycle average whether it will be
during the on time or during the off time.
So, going back here let us write it down esecondary half cycle average n*Vi during Ts/2;
Isecondary-rms the current that is flowing through is Io during that time and rms value is
into √d. So, esecondary-average*Is-rms is nothing, but n*Vi*Io*√d. Let me multiply numerator
and denominator by d then n*Vi*Io. N*Vi* d is nothing but Vo. Vo*Io is Po and √d and
this √d will cancel. So, you will have in the denominator √d.
And, if you consider the maximum operating duty cycle d will be 0.5 and therefore, you
will have it as Po*√2 for d-max as 0.5. Likewise for the eprimary half cycle average
*Iprimary-rms current you will find that it is Po*√2/efficiency and if I take worst case
efficiency 80 percent (Po/0.8) *√2 will be the eprimary-average * Iprimary-average.
Let us replace it there and see what is that we get. So, P this would be; Po*√2, I will take
it out common both the cases. We will have Po*√2 in both the cases, let us take that out
common and it will be 1/ efficiency plus 1 which is 2.25 and the denominator 2*Kw*Δ
B*J*fs. So, this portion (1 plus 1/0.8) is 2.25* Po*√2 / 2, so, I will replace in the
denominator √2* Kw*Bm*J*fs. So, this would be the area product of a forward converter.
So, let us use this area product and do an example.
614
Before we workout any example for transformer design let us write down the design steps.
There were many equations that we discussed and developed. Let us sequentially list down
all the equations relations that are necessary to design the transformer and step by step
write them down.
Step 1: First let us estimate the load power Po; estimation of Po this is the load power for
the transformer with all the secondaries put together. So, Po should reflect all the power
in all the secondaries if it is the multiple secondary transformer. So, let us say in a more
generic sense if there are m secondaries, i is equal to 1 to m, 1/ T integral of 0 to T, voltage
across every secondary i th secondary, current through i th secondary*dt. So, this will give
you the average power of every winding, i.e. i th winding and iterating i from 1 to m all
the winding’s average power is obtained and they are summed up to give you Po.
Next, step 2: Area product because the area product needs the Po value. We need to use
the Po of step-1 to find out the area product for the particular application. Then after you
find the area product using the area product for the particular application select a core from
the core table just like we did for the inductor after the area product you have to go to the
core table and then select a core which is having an area product higher than the calculated
area product then after having selected note down the core cross section area Ac and
window area Aw of the selected core.
Next, step 3: So, in step 3 we have to evaluate the number of turns.
615
So, NP we know we have this relationship e p-average half cycle average voltage divided by
2*Ac*ΔB*fs. Secondary windings Ns the i th winding is given by the voltage of the
secondary of the i th winding average half cycle divided by 2*Ac*ΔB*fs and the turns
ratio of the i th winding or i th secondary winding with respect to the primary is given by
Nsi/ NP. So, this is the turns ratio of the i th secondary winding with respect to the primary.
Step 4: Next, we find the wire gauge, we have the turns ratio, we have the number of turns
to wind, we have the core, we now need to choose the wire; what is the thickness of the
wire. So, wire gauge selection like in the inductor the area of cross section of the wire the
primary is given by the Ip-rms; rms current flowing through the primary by J the current
density.
616
And, the area of the wire cross section of the wire of the secondary of the i th winding will
be the rms current of the i th secondary winding divided by J. So, like that for every
winding you can find it is rms current and divide by J you will get calculated value of aw
wire cross section area. J what you are going to use just like in the inductor we start with
3 Amp/ mm square or 3* 10^6 Amp/meter square.
After having calculated awp and aw of all the secondaries wire cross section area, go to the
wire table and choose the wire gauge from the wire table. After having chosen the
particular SWG Standard Wire Gauge; note down the values of the actual wire cross
section area of the primary and wire cross section area of the secondaries. This would
complete the design, but there is one more step needed to complete one cross check step
just like we did in the inductor design.
Window area cross check: Kw*Aw that is available window area or should be greater than
NP primary number of turns * the wire cross section area of the primary plus all apply for
every secondary winding secondary number of turns of the i th winding * area of cross
section, wire cross section of the i th secondary wire and add them all up. All these turns
* area should add up and fit into the available Kw*Aw, only then the design will be
successful.
If this inequality is not satisfied you go back to step 2 select a core another core having a
higher area product and then repeat all these calculations for step 3, step 4 again, till step
617
5 inequality is satisfied, then the design is successfully completed.
618
Prof. L. Umanand
Lecture - 69
Transformer example
.
Let us now take an example and design a transformer for a DC-DC converter. I have taken
here an example of a forward converter where we have discussed this operation and we
know how it works. So, let us say we have an output voltage spec of 12 Volts, output load
current as 3 Amps, switching frequency of 20 kilo-Hertz, input Vi as 24 Volts ±10% and
then we need to design this transformer.
So, let us see how we go about doing that. So, step 1: we need to find what the output
power is? Output power is Vo*Io, but we need to find what is the power of the secondary
here. So, if you see the secondary because Po here refers to when you are talking of design
of the transformer, all the power right at the point of the secondary terminals.
So, you see this is the power here Vo*Io and in principle the components are not ideal,
there will be a diode drop especially at higher powers the diode drop here will be around
1.2 to 1.5 Volts * Io that will come in here and then there will be winding resistance and
there will be some drop in the winding resistance. So, put all that together you can increase
Vo by 10% and add the VD drop.
619
So, what we can do is we can say 1.1 Vo plus VD will be the voltage that will be needed
here to take care of a output Vo of 12 Volts. So, this will give us and slightly higher voltage
at the secondary so that all these drops are taken care of into Io will be the power at these
terminals of the secondary. So, apply Vo 12 Volts, VD of 1.5 Volts and 3 Amps you will
get around 44.1 Watts.
So, for this secondary output power we need to design the transformer.
Next step is find the area product. So, the area product for a forward converter we have
derived 2.25*Po/(√2 Kw*Bm*J*fs). Po is what we have calculated here 44.1 Watts, Kw
we will take it as 0.4 a conservative value, actually the 0.4 is actually a value where the
winder is quite experienced and skilful. Bm for transformers we reduce the Bm for
inductors we used to take 0.25 Tesla, but for transformers do not swing beyond 0.2 Tesla
because we want to limit the magnetising current and also to keep the swing well within
the linear region of the BH curve.
Then J, same as before as you used for inductors 3* 10^ 6 Amp/ meter2 or 3 Amp/mm2.
Frequency switching frequency 20 kilo-Hertz as given as a spec. Now, if you apply all
these and calculate you get 1.46172*10^ -8 meter4 multiply this by 10^ 12you will get it
in mm4, 14617.2 mm4.
Go to the wire table as before and then search for a core which has a higher AP than this.
620
In our case here we cancel again select P 36/ 22 pot core which has an area product of
around 20000 mm4; Ac upto 201 mm square, Aw of 101 mm square, Ap area product of
20100 mm4, mean magnetic path length 53.2 mm. So, these are the characters of the core
that we have selected pot core P 36/ 22.
Next, we go to the turns: primary number of turns Np is given by Vimax /( 2*Ac*Bm*fs).

Vimax is 24±10% which is 26.4 Volts; Ac we have found out 201* 10^-6 meter2, Bm is 0.2
Tesla and fs is 20 kilo-Hertz on calculation you will land up with 16.4, set Np as 17 turns.
Next let us try to find out what is the turns ratio. Set d-max as 0.45. Remember that this is
a forward converter you cannot have a duty cycle more than 0.5. Therefore, set d-max as
0.45 to be on the safer side and Vimax*d-min should be equal to Vimin*d-max should be
equal to whatever the output voltage; here the upgraded output voltage 1.1*Vo+ VD.
Now, using these two you can find d-min because d-max is set. So, d-min is equal to
Vimin*d-max/ Vi max; you know Vimin is 24,d-max is 0.45, Vimax is 26.4 you can find out d-
min which is 0.41.
Now, this 1.1*Vo+ VD should be equal to our forward converter input output relationship,
n* Vimax* d-min and here n is the only unknown (1.1*Vo+ VD)/ (Vimax* d-min) apply the
values and n=1.35.
621
And, Ns is equal to n* Np which will turn out to be 1.35* 17 which is 23 turns. So, Ns is
23 turn. If you are using a demagnetising winding; remember, demagnetising winding will
be very thin wire because it is only the magnetising current that will flow through it
probably will be SWG 45, but you need to use the same number of turns as Np which is
17.
Next step is to find the wire gauge. So, Is secondary rms is Io*√ d is the rms current that
is flowing through the secondary. So, Io*√ d-max you can calculate around 2 Amps, 3
622
Amps*√0.45. Ip-rms is nothing, but n* I secondary-rms which is 2.7 Amps. And,
calculate the wire cross section of the primary Ip-rms/ J; J you take it as 3 Amp/mm2. You
will land up with 0.9 mm square. Area of the cross section of the secondary winding wire
Is-rms/J you have 0.67 mm square. For the primary go to the wire table and you will see
that SWG 18 will fit it as a wire cross section of 1.167 mm square and SWG 19 will fit for
the secondary aws will be 0.8107 mm square.
Finally, come and do the crosscheck window area crosscheck to ensure that the windings
will fit into the core. Kw*Aw can calculate 40.4 mm-square and Np*awp + Ns*aws will be
equal to 39.1 mm square. So, you see that the window area check is satisfied and the design
of the forward converter transformer is complete.
623
Prof. L. Umanand
Lecture - 70
Forward converter design mfile
Now here, I am going to show you an m-file wherein I have designed a forward converter.
Recall that in the last week resource I had included a forward converter design example,
now I am extending that to include inductor design with the core and also the transformer
design with the core.
So, for that some extra files are needed I have included cores dot m, ferrite dot m which
gives you the data for the ferrite cores like the pot cores EE cores and so on. I have included
some for metglas cores, powdered iron cores I have also included the wire table here.
624
Just to show you, if I go into ferrite dot m you see here POT CORES, EE CORES, UU
CORES, RM CORES and TOROID CORES. These are the various varieties that I have
included.
And I have put in the database Ap that is the Area product, Ac the core Cross section Area,
window area in the form of matrices and I have also included the strings and for all these
cores.
625
It is simple for you to add as much as you want into this database.
626
And then you say, in the wires table I am going to show you that I have included the
standard wire gauges details, these are the standard wire gauges going down from SWG
45 number toward SWG 7.
With all their bare conductor area and enamel thickness, all these things.
627
So, you have the wire gauge table and also the cores table and in forward design dot m I
have included the actual design calculations. So, what is there in forward design? Look at
it first the specifications, we saw the specifications we see 48, we see minimum, we see
maximum calculation, you can calculate it if you have tolerance otherwise list the
minimum and the maximum
Now, the change that I have done, 𝑉𝑜 is indicated as a vector 12 Volts and dv as another
vector the Δv and 𝐼𝑜 is also a vector one Volts, these are for a single secondary, but if you
have multi secondary you just have to include comma and put in let us say, you can put in
a 15 Volts, you can put in a minus 15 Volts so on. And accordingly the ripple for the output
corresponding that vector will change and then 𝐼𝑜 vector also will change, but now I will
just bring it back into a single. So, this can be used for multi output forward converter.
628
Then the designer variables ∆I, I have set it at 30 percent ripple inductor current ripple can
change it did that so that to reduce the core size. 20 kilo hertz switching frequency, I have
given two different flux densities; one for the flux density or transformer where I am
setting it as 0.2 Tesla, another Bml the flux density for inductors as 0.25 Tesla.
The efficiency of the transformer worst case minimum is 80 percent generally, when you
design it will be much more than 80 percent. So, this is a safe lower limit to calculate the
input power. Dmax are set at 0.45, J is 3 amp per mm square or 3* 10 to the power of 6
ampere meter square, Kw 0.44 for transformer 0.6 for inductor, Vd and all these other
calculations.
629
You see first the power calculation 𝑃𝑜 naught is ∑ 1.1 ∗ (𝑉𝑜 + ∆𝑉) ∗ 𝐼𝑜 , summing it over
all the windings sigma. TURNS RATIO; find out the TURNS RATIO, you have 1.1 ∗
(𝑉𝑜 + ∆𝑉) divided by Dmax*Vc-min or D min*Vc-max which will give you N of all the
all the secondaries with respect to the primary. N is a vector because I have used
MATLABs vectorization strength.
Then the FORWARD TRANSFORMER DESIGN or this is where the new thing comes
in, find the Ap; Ap using the equations and then going to the while loop you call the cores
function the cores function which I have indicated here and that will go into the cores
function and try to search for cores. So, it will ask which core you want whether you want
to have ferrite core, metglas different core materials. And once you choose the material it
will go into that and then choose the specific core according to the area product calculator.
Then it will give out the selected core, the selected cores Ac, selected cores window area
and the area product.
630
And using that Np, Nd, Ns are calculated and then you are calculating the wire gauges here
and for that I am using the wires using the wires table and the wires table will give you the
aw wire cross section actual for the wires and the and the SWG, Standard Wire Gauge.
Then you do the window area check if it is not fitting it will go into the next loop it will
ask you which again to choose a core and so on.
631
And then the INDUCTOR DESIGN comes in finding out the value of L for every
secondary you have an inductor. So, L here will be a vector and find out the energy of the
inductor and finally, you do a area product design just like we discussed.
And do the core you I am using cores here again, to select the cores and then wires here to
select the wire gauge for the inductor and only here you also calculate the air gap and goes
into an iteration.
632
If it is not fitting and then the OUTPUT CAPACITOR FILTER DESIGN; the OUTPUT
FILTER CAPACITOR DESIGN, the DIODE DESIGNS. The various diodes
demagnetising diode, the secondary blocking diode in the secondary freewheeling diode
POWER SWITCH all those things.
633
Then comes the display section where you display all the area details. So, let me just run
this for you for this specification.
So, I will go into octave, I am already in that folder, you see that you see forward design
and the other m files
634
So, let me type in forward convert FORWARD DESIGN and then you execute that it will
ask you to choose a core material, I will choose Ferrite which is 1. I will choose ferrite and
then within ferrite it will ask you to choose POT CORES, EE CORES this is the data which
I have entered. A minimal set of data you can always increase it you can add your own set
of new cores and cores shapes.
Let me choose EE CORE now, second and then see that it may then it is asking for the
INDUCTORS and I will again choose this one again EE CORE. And that is it the design
is done and it will get printed you see the specifications and the DESIGN VARIABLES,
FORWARD TRANSFORMER see, choice of E25/13/07, the primary windings, the
Gauges, Demagnetizing Winding Gauge so on, the POWER RATING of the switch.
635
And the DIODE RATINGS, Demagnetizing, Blocking, Freewheeling, INDUCTOR

DESIGN all these things are calculated and you see the output. I will allow you to explore
this and leave it to you to fine tune it and try to understand this forward design m file.
636
Prof. L. Umanand
Lecture – 71
Pushpull converter
In this session we shall talk on a 3 important converters. The Push-pull converter which I
have written here, the second one is the Half Bridge converter and the third one is the Full
Bridge converter.
These three converters are generally used for higher powers in order in which I told push-
pull converter, then followed by Half Bridge converters still higher power followed by
Full Bridge converter is for still further higher powers all very high power converters are
generally the Full Bridge converters.
Now these three converters are also Buck derived converters only thing is that it is special
in the sense that, they have a kind of an oring mechanism where during the 1 minus dTs
period also power can be delivered to the load, I will explain to that how this oring
mechanism comes into play.
637
This is done by having a centre tap on the secondary side. So, all these three converters
have secondary as centre tapped. So, let us these three very important converters which
are very very popular and used in many applications.
So, let us now consider the push pull converter. Let us begin by writing down the typical
forward converter with demagnetizing winding; this is the forward converter forward
converter primary is Vi or Vcc. This is the demagnetizing winding which we discussed,
the diode which is used during the demagnetizing time. The demagnetizing diode and then
you have the secondary.
The secondary you have the secondary blocking diode. And then you have the secondary
side, the freewheeling diode where the inductor freewheels during the time when this
primary side BJT is off. Now we have L, C and the load Ro connected in this fashion.
So, this is our typical forward converter, let me place the dot polarities on the primary side
I have placed the dot here. Remember that the dot polarity is down below here for the
demagnetizing winding. Well we want this demagnetized winding to come during the time
and this switch is off.
The secondary side will have similar kind of a dot polarity like the primary, because we
want the power to flow to the secondary at the time when the BJT is on. This regular
forward converter which we know and you know how it operates. So, let me put a diode
here, this is actually not an extra component, most of the switches the MOSFET’s, IGBT’s
will have this internal body diode coming to the picture.
So, I have drawn this for the sake of completeness and also symmetry. Now this diode is
there, let me also make that into a controlled switch. So, an IGBT and this diode is actually
the body diode of an IGBT. So, imagine it in that fashion, so now you have two controlled
switches. So, let us play around placing of the switches to see how the push pull converter
evolves.
Now the concept is that during the time when the BJT is on the primary is energized and
it will pump power to the secondary and then so on by buck converter action to the load.
And during that time when the switch is off the flux in the core demagnetizes by a current
flow through this diode through this demonetizing winding and into the supply.
638
Now, instead of pumping this power into the supply, let us pump the power into another
winding and then root it in some manner into the output. So, then we can have this we can
have an output which is having a higher power than as compared to the one where is oring
effect does not come into picture.
Now, how to bring about this oring effect? Let us see that. And this demagnetizing winding
in the forward converter is a very low power winding, it has very very thin wires and it
can handle only the magnetizing current cannot handle the power currents, like the primary
and the secondary. So, that is one change that you have to do, you will have to change this
demagnetizing winding into a power winding and then reposition.
So, let us see how we can bring about using this demagnetizing as a power winding, and
also try to steer the power through another winding into the output, rather than putting the
power back into the source. So, let us first mark this, this is Q1, this is Q2 and this Q1 let
me shift the position in this fashion without altering the circuit in any way.
So, this is I am instead of putting it vertically down to ground here, I am going to place
this transistor Q1 in this fashion horizontal and make space here. So, next let me make
space and let me bring the supply Vi again in a horizontal fashion like this here, remove
that, and now I have made space here.
Now this winding is demagnetizing winding what was supposed to be demagnetizing

winding I am going to flip and pull it up like this from this common points. So, from this
common point first the non dot end and then after the non dot end the dot end will come
at the remote side here.
So, from that you see this diode and transistor combination switch coming in here the diode
and this is Q2 and I place the dot here. So, now I can remove this winding and this part.
So, now it is that same forward converter transformer with demagnetizing winding but the
windings placed in such a fashion that it now looks like the primary is a centre tapped
common winding.
Both are on the same core, 1:1. Now, let me for rooting power to the secondary I will
replace this with another winding here, I am centre taping the secondary. And I will place
the dot here, and on the non dot end the diode. So, let me erase this, the non dot end diode
639
I am not going to use any extra diode; I will use the freewheeling diode and connect it to
the non dot end, and then I will make this connection.
Now, this would be the secondary side all these winding the secondary two winding centre
tap, primary two windings centre tap all will reside on the same core. So, all this is one
single core, now this is called the push-pull transformer, now this power supply, I can
complete it and ground that one.
So, now we will call this, name the parts this is Vi, this is Vo as we have normally named,
Vo is measured in this fashion. Then let us push it to one side because we need to draw
some wave forms here to understand the operation of the circuit. Let us call this D1, let us
call this D2, L, C and this is Io, this is IQ1. And here current flowing in through this switch
Q2 is IQ2.
Now to understand the operation of the circuit, let us draw the waveforms alongside here
and get more inside into this operation. Let, me now draw series of these lines I will explain
to you what these time divisions are.
Now, during this time I will say this during this time only Q1 is ON and I leave this space.
I will tell you what this other spaces are, Q2 is ON only during this time, Q1 is ON only
during this time and again Q2 is ON only during this time.
And all other times Q1 and Q2 both are off. So, during this time when I say Q1 is on Q2
is off during the time when I say Q2 is on Q1 is off, but in all the places where I have not
written anything it means both Q1 and Q2 are off. So, that is the meaning of these thing.
Now, let me draw one axis time and this is the pole voltage Vp and let us start by looking
at the pole voltage Vp.
Consider the period when Q1 is ON only Q1 is ON. So, when this BJT transistor Q1 is ON
this is connected to the ground here. The dot end is connected to the plus of Vi. So, Vi
comes directly across this coil. So, you are having plus minus across the coil like this.
So, because all the coils are coupled on the same core, you will see that all the dot ends
are positive non dot end is negative. On the secondary side also dot end is positive, non
dot end is negative, the other coil plus on the dot end, minus on the non dot end.
640
So, let us see what is the voltage, the voltage measured in the fashion from non dot end
keeping the common and the measuring probe on the dot end, you will measure Vi across
this coil and by induction it is Vi across the other coil, even though Q2 is off by induction.
In the secondary side if you take a turns ratio 1:n, you will have n* Vi. And the other coil
also we will have n* Vi. Now this will be the way the voltage will be appearing across
each of the coils. Now Q1 is ON, Q2 is off there is no IQ2 current, only IQ1 current is
flowing.
So, the only active portion on the primary side is this, what I am showing through this
cursor moment. So, this coil of the primary is active giving you a positive dot end positive
Vi. On the secondary side dot end positive diode D1 is forward biased. The moment D1 is
conducting D2 is reverse biased, because you see that this is positive and the other end of
the D2 is negative at the reverse bias over to the picture.
So, only this coil, the top portion of the centre tap secondary is active, because D1 is
forward biased D2 out of the picture. So, D1 forward bias will charge up the inductor and
this portion will act like a regular buck converter. So, what is the voltage at Vp; n*Vi will
appear at Vp with respect to the centre tap.
So, it will be n*Vi, likewise whenever Q1 is ON; you will have a same voltage appearing
n*Vi here. So, now consider the situation when Q1 is off only Q2 is ON only Q2 is ON.
So, let us write down when Q2 is ON, the dot end is now connected through Q2 to ground,
non dot end is positive Vi. Likewise non dot end of the other coil is plus, dot end minus
and the secondary side non dot end plus, dot end minus; non dot end plus, dot end minus.
And the voltages that you would see measured the same direction in minus Vi on the
primary side across each of the coils. And on the secondary side minus n*Vi. And now,
you see that D2 forward biased, D2 is forward biased D1 is reverse biased.
So, there will be a current flow through D2 through the inductor charge up and there will
be buck converter operation with respect to this portion of the coil. So, you will see that
this is the active coil, on the secondary side this is the active coil; there will be power
transfer in through this coil to the output.
641
So, at the pole you will see when Q2 is ON, you will have plus coming in here Vp. We
have n*Vi, and it will repeat for all Q2 ON periods. So, if I draw in this fashion complete
the waveform, this is how Vp will look like. Why do we say it is 0, why do Vp is 0 during
the time when Q1 and Q2 both are off.
So, during that time you will see that the inductor is freewheeling; the freewheeling path
will go through like that. Look at the curser it reaches the centre tap and then at the centre
tap it equally divides into this path and this path. And back again through the diode ors up
at this point and comes.
So, because the current is going one entering the non dot end, the same equal current
entering that dot end in the core they are all coupled one is having N*i mmf in one
direction. In the other N*i mmf in the opposite direction equal and opposite, because the
currents are getting divided by 2 exactly. So, the mmf in the core cancels, there is no
induction the voltage across the coil is 0. Therefore, potential at Vp both the diode are
conducting potential at Vp is 0, therefore, I have indicated as 0 here.
Let us now discuss how the inductor current waveform looks like? Let me mark the
inductor current waveform here, iL and let us draw its waveform. This is the time axis and
let me mark this label this as iL. Now, let us take this interval when we say that Q1 is ON.
642
So, when Q1 is ON we saw that this winding was active, this winding was active on the
secondary side, D1 is conducting Vp is at nVi. And we have marked that here in the Vp
wave form. So, what is appearing across the inductance? On one side you have Vp which
is it at n*Vi, on the other side you have Vo voltage across inductor is n*Vi minus Vo.
And at the rate of n*Vi minus Vo by L you will have a current rise, because Vo is lesser
than n*Vi you have a current rise, inductance is charging. Let me mark the slope (n*Vi
minus Vo) /L.
Now, during this period both Q1 and Q2 are off. So, when both Q1 and Q2 are off inductor
current is freewheeling, it freewheels through capacitance and the load, goes through this
line through the centre tap. And at the centre tap it divides equally into 2 halves, one going
in the direction, other going in this direction they are or up at the pole point and then go
through the inductor.
So, because exactly half the current is flowing through one part of the coil which is going
through the non dot end, and another parts of the coil which is the dot end. The mmfs N*i
𝑑Ф
cancel each other and their the flux remains at wherever it was. And there is no . As a
𝑑𝑡
consequence there is no voltage induced across any of the coils; and then therefore, Vp is
0 and we have indicated Vp 0.
But, there is Vo on the other end of the inductor coil; and therefore, the voltage across the
inductor is minus Vo. So, it will have a falling slope of minus Vo/L. And then, during this
period Q2 is ON and when Q2 is ON this coil is active and on the secondary side this coil
is active, D2 is active Vp we saw is again n*Vi.
So, consequence inductance will charge up again at (n*Vi minus Vo) /L rate, and it will
fall down at minus Vo/ L and this repeats. So, if you take the average of this, draw an
average right through is the DC part. And this DC part is nothing but Io, the reason is that
Io is the part that is going through the load and it is the average part.
The part that is going through the capacitance is the 0 average part, there is no average
current through C; and therefore, the entire average has to go through the output load.
Therefore, we have indicated is it as Io. To get a better inside of the operation of the push
pull converter, we need to look at the flux within the core, how does it look like?
643
So, let us deduce the wave form of the flux within the core, you cannot see the flux
waveform on the oscilloscope, but, let us deduce it using the Faraday’s law. So, let us draw
an axis time axis t and we would like to draw waveform of the flux; flux within the core.
Now, let us say during this period, when Q1 is supposed to be ON; when Q1 is supposed
to be ON this point is grounded this point of the coil is grounded. The dot end is connected
to Vi, so the dot end is more positive with respect to the non dot end.
So, therefore you can expect the flux to increase, you have the positive mmf being applied
flux will increase. So, in this period the flux will go from minus Фm to plus Фm at a rate,
what rate? It will go at the rate of this voltage whatever the voltage which is being applied;
divided by the number of turns of that particular winding. So, if I take the primary winding
half this from the centre tap to this end the voltage across this coil is Vi, that is the Np is
the number of turns, then Vi/ Np is the rate at which it will rise.
If I take the secondary winding, Ns is the number of turns and the voltage across the
secondary winding is n*Vi. And therefore, the rate at which the flux will change is
n*Vi/Ns, but because Ns is a related to n, you see that both these rate of same. So, let me
write that down, this will be the way the flux will change, constant rate voltage divided by
𝑑Ф
n. Because we are using V = n 𝑑𝑡 .
Now, importantly during this time this period when Q1 and Q2 both are off, when both
Q1 and Q2 are off, I said that the inductor current is freewheeling, how does it freewheel?
It goes in this path this path I am just shown as a short path but actually it will go through
the capacitance and it will also go through the load and then or up at this point.
So, I am just indicating it that it goes through these two impedances, and it comes in this
direction. And then it splits here equally exactly half because this windings are exactly
equal.
So, whatever the inductor current iL/ 2 will go there in this direction and then you have
iL/2 coming from this diode D1 and D2 also iL/ 2, both these will go down or up and then
flow through inductor again. But importantly if you take this part here iL/ 2 flowing into
the dot in the lower half of the secondary centre tap winding. And here iL/ 2 flowing into
the non dot end part into the upper half the centre tap winding.
644
So, the upper half and the bottom half they are having opposing mmfs and then mmfs
cancel there is no magneto motive potential. And therefore, flux wherever it is will not
change, it will remain. So, the flux was here and it will remain at that point.
𝑑Ф 𝑑Ф
So, if it remains at that point what is during this period, is 0, there is no scope. So,
𝑑𝑡 𝑑𝑡
𝑑Ф
if is 0 for all the coils coupled to the core. The primary top, primary bottom, secondary
𝑑𝑡
𝑑Ф
top, secondary bottom, all will see 0 voltage. Because is 0 and because this voltages are
𝑑𝑡
0, Vp will also be 0.
So, therefore you do you see the voltage Vp is 0. Next during Q2 ON, when the Q2 is ON
dot end is negative, non dot end is positive. And therefore, you have a negative magneto
motive force, and it will start falling like this. And then during this period both Q1 and Q2
are off flux wherever it is will stay.
So, it will stay there, and then from there again it has a positive rise Q1 ON stays and then
Q2 negative, negative rate falls and stays so on. So, you will have this plus Фm this is
minus Фm, you will see the flux swinging from minus Фm to plus Фm and in between
𝑑Ф
when both the switches are off. The flux just remains 𝑑𝑡
is 0 and voltages across all the
coils are connected all the coils on the same core will be 0.
Let us now look at two other important waveforms in the push pull topology. One is the
current that is flowing through D1 and the other one is the current that is flowing through
D2. These two currents are pretty unique in shape you would not have seen this current
waveforms, this type of current waveforms in other buck derived converter like the
forward converter and the non isolated buck converter. In the push pull, half bridge and
full bridge, you will see this type of currents flowing through the diodes.
So, let us have a look at them, let us draw the t-axis and look at the iD1 current here, and
another t-axis and let us look at the iD2 current here. Let us first draw the level because it
is the inductor current that has flows through iD1 and iD2 parts of the inductor current.
We will see which part of the inductor currents will flow through.
But, first let us mark this average level of the inductor current which is Io, and for iD2 also
I will mark the level. Now, when I take iD1, when is iD1 active? Whenever Q1 is ON this
winding on the primary side is active dot end. Whenever the dot end is positive you have
645
the dot end positive and this portion of the winding top half of the secondary will be active
and D1 will be active during the time when Q1 is on.
So, it will just follow the inductor current whatever the inductor current was will be iD1
during that time and we can write that down. And you see Q1 is ON again here, here also
it would be the same thing. Now what is happening during the time, when both Q1 and Q2
are off? Here also Q1,Q2 off here also Q1,Q2 off here so on.
Now, during the time when both Q1,Q2 are off, the inductor current is freewheeling and
we see that at the centre tap point it is dividing into two exactly two halves iL/ 2 and iL/2
through D2. So, D1 will be carrying iL/ 2 during the period when Q1 and Q2 are off.
So, this is the iL wave form by 2 half of that will flow through iD1. So, like this half of
that, then you see that it is having a down slope and negative rate. So, the iL wave form
would have been like that, but half of that would be like this.
So, like that we can draw there also and the other side here also you see that it is both Q1
and Q2 are off. The inductor current divides exactly into half and that is how it would look.
So, this is the type of waveform that you would see for the D1 current. And for the D2
current also similar during Q2, you will see that the D2 is active during the time when Q2
is ON.
So, D2 is active and you will see during this time it is exactly the same as the same as
inductor current during this time also it is exactly same as the inductor current, and during
the time when both the switches are off, half of the inductor current. So, half is flowing
for iD1, half is flowing through iD2. And here also you see that half flowing through iD1,
half flowing through iD2 likewise here also, so, completing the wave shape you will see
that.
So, during this time only iD1 is flowing iD2 is 0, during this time only iD2 is flowing iD1
is 0. So, you will see that the diode currents on the secondary side will be like this. And if
you actually probe the currents in an actual circuit with the help of an oscilloscope, you
will be able to see currents like this.
Let me remove the wave from iD2 and make some space, and in this space I will include
another wave from another very important waveform the voltage across the switch. The
646
voltage across the switch Q1 and Q2 very important we will be discussing that now. The
voltage across both these switches are very similar accept that they are phase shifted. I will
look at only one of the voltage waveform across the switch either Q1 or Q2, I will take Q1
for now.
And let me mark the way I am going to measure in this fashion vCE1 is the label. And let
us bring that label down here and see look at the wave from here of VCE1 across switch
Q1. We know that when the switch Q1 is ON during periods and during the period, the
voltage across the switch Q1 VCE1 will be 0 almost 0, and we will put that down.
Now, the another state is when Q2 is ON. So, when Q2 is ON Q1 is off, when Q2 is ON,
this dot end here is at ground potential. And this point here at the non dot end see that each
coil is having Vi with the positive side on the non dot end, Vi here also positive side on
the non dot end.
So, overall from this point to this point you have two Vi coming across the switch. So,
therefore, when the switch Q1 is off and Q2 is ON, Q2 is ON, these two periods here. We
will see two times Vi coming across the switch Q1, because of the voltage induced across
both these coils.
Now during the time when both Q1 and Q2 are off, we saw that the inductor current is
freewheeling in the secondary side and the inductor current is splitting into two equal
647
halves. And you have opposing mmfs Ni on one side in the top due to the current flowing
in the top half the secondary centre tap into the non dot end. And due to another equal and
opposite current flowing through that dot end of the bottom half of the secondary centre
half causing mmf cancellation and no flux.
And therefore, the voltage across all the coils is 0 as we see here even across Vp. And
therefore, if this voltage is 0; this centre tap point which is at Vi will come directly to the
collector point of Q1. And therefore, they will all be at Vi in between this. So, this Vi this
is the voltage waveform that you can expect to see on the oscilloscope across switch Q1,
VCE1.
Similarly, you would see across the switch Q2, CE2, only thing is phase shifted with the
0s or the low voltage 0 almost 0 voltage occurring during this period Q2 period. So, now
let us try to calculate the voltage equation, input output relationship. First let us see what
the time period is.
Now you see that this time switch Q1 is turning ON and it is ON for this period. I will call
this one as dTs. This whole period is Ts till again the cycle repeats. And this period when
the switch Q1 is ON, we call that one as dTs as the duty cycle of the switch Q1. Likewise
the duty cycle switch Q2 also is dTs only it is turning on here during this period.
So, now this is the switching pattern for the device, but what is the duty cycle for the
voltage profile at Vp? You see that the voltage profile at Vp is slightly different, in the
sense that it is having double pulse in that same Ts period.
So, Vp is actually having double the frequency as compared to the switching frequency of
these two devices. So, therefore, this arrow or dimensioning that I have indicated here will
be the frequency of Vp and this time is Ts/ 2.
So, therefore, what is the duty cycle for the Vp, the Vp duty cycle will be this period which
is dTs, we know it is dTs divided by Ts/ 2. So, this is nothing, but 2* d. So, the duty cycle
for the Vp or the pole voltage waveform is 2 times the switching duty cycle of the
individual switches off the push pull converter.
648
So, therefore Vo, Vo here is nothing about Vp into the duty cycle and duty cycle is 2*d.
And Vp we know is n* Vi we have marked it here, we saw how that came about. Therefore,
we can use n*Vi for Vp* 2*d.
So, if you write it down here, 2*n* Vi* d, d here is the switching duty cycle of the
independence switches and this is the input output relationship for the push pull converter.
649
Prof. L. Umanand
Lecture - 72
Flux Walking in pushpull
In a practical push-pull converter there is one issue called the flux walking. We will try to
discuss and try to understand what flux walking means and how to take care of this. Let
me first draw the push-pull circuit, I am going to focus myself only in the primary side.
So, you have the two transistors, the battery here, on the common core you have the
secondary, and a secondary centre tap and you have the further portions of the circuits here
that is the two diode the inductor capacitor the complete buck converter. But I am not
going to draw this here, we know that, because I want to restrict myself to this portion of
the circuit here on the primary side and see what this flux walking is.
It is a serious problem in practical circuits and which can cause damage to the topology
itself. It could blow up the transistors and it could cause saturation of the transformer. So,
let us see how this happens and how we can solve this problem.
650
So, let me mark these dot polarities here for the transformer, this is Q1, this is Q2. Let me
mark Vi here and let me also put in this waveform that we need to understand. And here I
am going to draw only one waveform and that is the primary voltage here Vprimary.
I am picking the entire primary coil, one could pick any one of the coils either a primary
or secondary it does not matter, because the flux inside is the common one linking all the
coils. But it is easier to draw the waveform for the primary. So, therefore, I will draw that
all the waveforms are similar, but easier to understand just with a complete primary.
So, let me draw the time axis, and let me mark the label here Vprimary the voltage across
this entire primary. Now, we know that during the time periods when Q1 and Q2 are off
the voltage across the coils are 0, because the output side inductor the buck converter
inductor is freewheeling and it is splitting across the secondary centre tap, secondary coils
and causing opposing mmfs to come up.
𝑑Ф
And therefore, is 0, no voltage during the time when Q1 and Q2 are both off together.
𝑑𝑡
Therefore, during all those periods you will see zero voltage across the coils. So, that is
what I am marking here and during the time when Q1 is ON, you will see that the non-dot
end will be connected to the ground zero potential. The dot end is connected to Vi and
therefore, the voltage here would be Vi with dot end positive.
So, all the dot ends will be positive with respect to its non-dot end, here also dot end will
be positive with respect to the non-dot end even on the secondary side. So, this we know,
the voltage will be Vi and Vi this is what will be induced across the primary coils. So,
therefore, if I take the entire primary coil it is Vi plus Vi which is 2*Vi.
So, let us mark that here during the time when Q1 is ON the primary voltage with the dot
ends being positive you will see a voltage of 2*Vi. And during the time when Q2 is ON,
during the time when the Q2 is ON the dot end is actually grounded the non-dot end is
plus Vi and this minus Vi is induced on all the coils.
So, therefore, let me place the polarities in this fashion non dot end is positive with respect
to the dot end, same with all the other coils too. And on the primary side you are seeing Vi
with the non-dot end positive. So, it is a minus voltage that you will see across the primary
and that will be minus 2*Vi. So, this will repeat cycle by cycle second cycle it is 2*Vi
positive, 2*Vi negative when Q2 is ON and so on.
651
So, if we connect them all together you see a waveform like this and this is the type the
waveform that you would see on an oscilloscope when you probe the entire primary. You
will see that it is also you will get similar waveforms across each of these secondaries, but
they will be scaled by the turns ratios.
Now, let us see the problem what is the problem? Now in a practical circuit; when all the
devices are consider ideal; we are expecting here the positive volt-second and the negative
voltage-second should exactly balance so that there is no average. So, this is the
fundamental physical nature of the coil where you cannot have an average voltage across
the coil..Now in the case of a practical situation if you look at this voltage here, this is the
time during which Q1 is ON. Now when Q1 is ON, an ideal case we say that the ground
is connected to this point which means that VCEsat is equal to 0. But in a practical case
the VCEsat is not 0. So, therefore, it will be Vi minus VCEsat1 which will appear across
this, and that is what will be induced with proper scale.
So, Vi minus VCEsat1 is appearing across Q; is appearing across the primary coil
connected to Q1 and 2 times that will be appearing across the entire primary coil. Likewise
during the time when Q2 is ON, you will see 2*Vi minus VCEsat2 coming across the
primary coil.
So, if you look at this there is volt-second balance, perfect volt-second balance only when
VCEsat1 and VCEsat2 are equal. But in a practical situation it will not, you should also
understand that the time period is dTs when Q1 is ON you would also be given control
signal dTs for Q2 to be on during the same time period.
So, a VCEsat1 and VCEsat2 are miss matched you can have a difference in the area. So,
if VCEsat1 is not equal to VCEsat2, then there can be a small average here, can be a very
small average we will call that one as Vprimary-average, it could be as much as a micro-
volt, but does not matter.
So, even if it is a micro-volt you will have a problem, that is because from the Faraday’s
𝑑Ф
equation Vprimary-average which is equal to 𝑁𝑝 𝑑𝑡 , If I consider the number of turns Np
𝑉𝑝𝑟𝑖𝑚𝑎𝑟𝑦−𝑎𝑣𝑒𝑟𝑎𝑔𝑒
for the whole primary. So, what is flux? Flux is 𝑁𝑝
∫ 𝑑𝑡, where Vprimary-
average which is a constant dc, which means that Ф is proportional to t. This is constant,
652
integral of dt is t, Ф becomes proportional to t which means that the flux is increasing
continuously with time.
Whatever may be the scaling Vprimary-average may be 1 micro seconds which means that
Ф will continuously increase and probably reach saturation after few months. But still you
will have a problem which means that initially these circuit will be working and after a
few months, after the flux saturates. You will see misbehaviour, it maybe catastrophically
damage by blowing up some of the components.
So, what happens is that the operating point keeps drifting depending upon whether
Vprimary-average is positive or negative, it will drift to plus Фm or minus Фm and
ultimately saturate the core in one direction. So, once the core saturates when you give a
duty cycle Ton there will not be induction and the current will increase and blow up the
switches.
So, this is the problem in a practical case, you cannot make the transistors perfectly
matched it is not possible. So, you will land up with VCEsat1 not equal to VCEsat2 and
you will have this problem of flux walking. So, how to solve this problem electronically,
so that is what we will be looking at next.
Let us now try to visualise the effect of flux walking on the primary voltage waveform.
So, let me draw this BH curve, here a typical BH curve of the core using the push-pull
653
transformer. And normally the average flux value operating point will be in the centre
here, average flux would be 0. And about that average flux you will have the swing going
±Фm.
Now, if you look at the average flux in a practical push-pull converter, we know that the
flux was proportional to time t which means that it starts growing. It grows from 0 in this
fashion, and at somewhere here you will see that the average flux point at this point is
away from 0. I have shown it positive of course, it could also be negative you could have
the flux decreasing.
So, you see that the swing will still be about that average of flux operating point. And you
see that it is near the the top end at the positive side you will see that it is near the knee of
the BH curve, still there is no damage done as at.
Now, as time progresses somewhere here, let us say at this point in time you have still the
much higher average flux somewhere here. And when you take the swing about the
average flux operating point you see that a portion of the curve on the positive side is
𝑑Ф 𝑑Ф
moving on the knee which means there is lower and because the there is lower, you
𝑑𝑡 𝑑𝑡
will see that the induced voltage, the back emf across a primary winding will be lower,
and as a result there will be a drop in the voltage.
So, let us have a look at that, voltage across the primary. So, the voltage across the primary
𝑑Ф
is like this Vprimary and as it is swinging in the linear region there is no change the 𝑑𝑡
𝑑Ф
rate here or even in this case there is no change in rate. So, the primary voltage
𝑑𝑡
waveform will look essentially like this.
Now, in the case of this last case where a portion of the flux operating region is moving
𝑑Ф
towards the knee the there is decreasing. And you will see a drop in the induced voltage
𝑑𝑡
across the primary. So, this drop in the induced voltage primary can cause a high inrush
current because Vi is fixed constant applied as a input.
The primary is showing a lower back emf and the difference in the voltage Vi minus this
lower back emf will cause a large inrush current and which can blow up the devices and it
is against this problem that we have to protect.
654
So, as the flux working continues you will see that there is a drop in the induced emf across
the primary due to loss of induction, and as a result a high inrush current and damage to
the transistors. So, this is the problem of the flux working, and let us see how we can avoid
that.
Let us now see if we can have a solution for the flux walking problem. So, let me again
draw the push-pull circuit topology, and I am focusing mainly on the primary side, because
we have to control the pulse widths of Q1 and Q2. So, let me draw that portion of the
circuit and bringing your attention to Q1, Q2 switches.
And we have a common core on which we have the centre tap secondary, let me draw the
dot polarities like this. Now here we have the diodes oring, inductance, capacitance, load.
So, the buck portion is there and we know the how that works. Now, I am going to put on
the same core, on the same core two more windings one winding these are all very low
power windings very thin wire gauge they are used only for sensing.
So, dot polarity like this, note that I am going to put another winding there dot polarity in
the other direction. So, this is going to sense this dot polarity would be positive when Q1
is ON, so, this will be positive. And when Q2 is ON the dot is negative the non dot end is
positive, so, the non dot end. So, I am basically trying to use sense coils and sensing the
dot side and the non dot side positive under different switch on condition.
655
So, let me go further in the circuit, I will put a diode. So, which means that I just need to
have the positive it is a clipping circuit, it clips and takes only the positive part of the pulse
across the primary and finds an average. So, basically I am trying to find the volt second.
So, this will find the average; this is an averaging circuit R, C, I am putting a simple
averager, I could also use an op-amp averager to get better precision.
And then for the other winding also I am clipping and then putting the R, C averaging
circuit. This is just to indicate conceptually that I am trying to take the positive part of this
pulse, Vprimary pulse and then averaging it to get the volt second. And the non-dot end
during the time and Q2 is ON that part I am taking it out and finding the average. So, that
I have the volt-second of the positive part of the voltage Vprimary voltage and the negative
part of the Vprimary voltage.
So, this I will call volt-second of Q1, volt-second of Q2. So, these 2 should in actuality be
same under normal operating conditions. But if they are not same we have to do corrective
action. So, let us say I subtract them plus and minus and take the error. So, the error if both
are same should be zero, but if one is saturating and the other is not saturating then this
error would be finite the error can swing positive or negative. So, you pass it through a PI
controller.
So, the job the PI controller is to make the error zero by giving a correct corrective voltage
V correction. This correction voltage should be added to the duty cycle patterns of Q1 and
Q2 in such a manner that error here become zero. Error here become zero means Vs volt-
second of Q1 ON and volt-second Q2 ON will both match.
So, this will take care of the differential VCEsats of the devices, now where do you go and
give the correction. So, normally under normal condition you will have a control a scheme
like this where you have a Vo reference and you have a Vo feedback coming from the
output of the converter. Then the error is given to a PI controller and the output to the PI
controller is compared with a triangular carrier plus minus in this fashion.
So, after the comparison of the control voltage the triangular carrier you will get a PWM
voltage here. And this PWM is steered alternatively to go and switch Q1 and then Q2. So,
this steering logic will act in such a way that Q1 and Q2 are mutually exclusive when they
are on turning on. So, you will have the drive for Q1 and the drive for Q2 and goes to the
base drives or the gate drives of Q1 and Q 2.
656
So, this is how the normal PWM switching pattern will work, I will discuss this in greater
detail shortly, but for now we will consider it as a block scheme like this. Now we need to
add somewhere here this correction, because Q1 is getting a drive pulse dTs, for a dTs
period of time Q1 is ON for mutually exclusive dTs period of time Q2 is ON.
So, let us say here I introduce something there and then take this correction and then the
correction is passed through some logic. Let us say, and then it is connected to this points
basically to in; let us say it subtracts from one and adds with other the pulse widths.
So, which means that here in this case, let us say if the volt-second when Q1 is ON that
that part is increasing that part is saturating. Now you need to decrease that volt-second,
so, you are subtracting some amount of pulse width from here and then adding some
amount of pulse width to the Q2 drive. So, that Q1 is turned on for slightly lesser time than
Q2 and then there will be a balance achieved, e will go to zero.
Now, this is the nature of the correction logic that we would use for correcting the flux
walking. So, as the flux tends to walk positive, let us say; whenever Q1 is ON there will
be a tendency for the back emf induced to reduce slightly when it reaches the knee and
that will be detected.
So, you will see that value will be lesser than this value and then the correction value will
go positive in a way such that you can add and subtract here such that Q1 and Q2 will
automatically correct each other. So, they will not be given same dTs in the earlier case
open loop case the dTs of both would have been same.
But dTs will get corrected by the PI controller in such in such a fashion that Q1 will be
again brought back below the knee and the average flux will tend towards the 0 point. So,
this is the effect of the closed loop which will be happening continuously and then prevent
the flux walking. Of course, I have shown all this in a kind of a block schematic way, I
will in detail show what is actually the logic, so that you will understand that.
657
Prof. L. Umanand
Lecture - 73
PWM Generation
Let us now discuss how we will generate PWM for the push-pull topology. Let me draw
an op-amp comparator you have plus and minus, to the minus I will give a triangular
carrier. And to the plus I will give the control signal or the control voltage like that which
is supposed to modulate the triangular carrier. So, let us draw the waveform respect to t. I
have this triangular carrier in this fashion 0 to some positive value and the modulation
signal Vc is shown here in green.
Let us see what the output is like, so, let us mark the intersection points and bring it down
and bring it down in time and project it on to the time scale. So, whenever Vc is greater
than triangle, because triangle we get the minus and Vc control voltage to the plus.
Whenever the control voltage is higher than the triangle the output will be positive
otherwise 0, positive 0 it goes like that, depends upon what type of op-amp you are using,
if you are using a bipolar op-amp then then it go positive minus, positive minus like that.
So, let us say that we are having a unipolar op-amp with only Vcc and Vss grounded.
Therefore, you will see output going positive to 0 in this fashion just thereby giving a pulse
658
width modulated waveform. So, we will call this one as Va. There are many PWM
integrated circuits that use saw tooth carrier instead of a triangular carrier, much easier to
generate.
So, they use that one, this is also fine you will super impose you can superimpose Vc that
is the control signal like this. And then mark the intersection points in this fashion,
whenever the saw tooth reset happens that would be the start of the period.
So, we can say period Ts is of that duration here, and whenever the Vc control signal is
greater than the triangle of the saw tooth you have a high then a low - high - low so on.
And you get a pulse width modulated waveform in this fashion. So, this is typically used
in many PWM ICs, because saw tooth generation is much easier.
Now, let me add one more op-amp comparator and to that comparator minus I am giving
the triangular carrier or saw tooth carrier. And to the plus I am giving another signal fixed
DC, I am calling that as Vd and it is a fixed DC. Let us see what this does and then the
output of this two. I will erase that and I will put two diodes here and OR them. So, I am
going OR the output of these 2, one is having a control signal Vc another is having fixed
DC signal Vd.
So, let us superimpose on this triangular on this triangular career and see what happens.
So, you see that and here also this we call as Vd this is Vd. So, you see that here there is a
minimum pulse width. So, here whenever the Vc signal control signal is higher than Vd
there is no issue the operation is normal. When Vd become lesser than Vc, then what
dominates is Vd, and Vd will dictate and it will issue a minimum pulse width command.
So, you will be assured of a minimum pulse width and even in the saw tooth case minimum
pulse width is assured. In the saw tooth case you see that always it is starting from one
end. So, one end is determined because of this vertical drop reset one end is determined.
So, modulation is always on the other end. In the case of the triangle, modulation happens
on both ends.
So, both the cases you have a minimum pulse width which will come into the picture. So,
this will declare minimum duty cycle limit for the circuit. So, this way you can achieve
minimum duty cycle even for single ended operations like the forward converter and the
non isolated converters you can use this method.
659
Let us now assume for now that we have a pulse width modulated waveform Va. A single
pulse width modulated waveform and using that single pulse width modulated waveform,
let us develop the mutually exclusive drives that we can obtain or we should obtain for Q1
and Q2. Q1 and Q2 should have mutually exclusive on times and how do we get that. It is
not exactly a 180 degree phase shifted waveform, you should understand that Q1 is on for
some time Q2 is on for some time and then in between there are times when Q1 and Q2
are both off. So, we have to obtain a wave shape which will do that.
So, let us continue the circuit diagram further. Now from Va, let it trigger clock of a D-
flip flop. You have Q and Q-bar and the output of Q bar is connected to the D input. So,
what do we have? We have a toggle flip flop. So, the output of the toggle flip flop, let me
stare it through AND gates and the second inputs of the AND gates I will give it to 1 output
steering control steer steering control pin. This has an important effect on the final output
of the PWM I will discuss this. Just for now assume that this is a pin which can take value
1 or 0.
Now, the output of the AND gates are given to NOR gates like this, and output the NOR
gates are given to the BJT or a transistor which can be further amplified and then driven
driving the power switches. So, these 2 NOR gates we will have the second input coming
from Va in this fashion.
660
So, this will be the steering logic that we will be using for the PWM, let me push this bit
up. So, now, let me mark this is A and this has B and this has AA and this has BB, so, that
we can distinguish the waveforms. Now let me say that Va is a PWM modulated
waveform. Now let me divide the time into segments each of Ts width.
So, within that Ts you are having a portion of the waveform a pulse width modulated
waveform which appears like this which we just develop for Va. So, you have the pulse
width modulated waveform coming like this. Now this is given to the clock of the D-flip
flop.
So, let us say that the D-flip flop will be edge triggered on the rising edge and on the rising
edge let the toggling action happened. So, Q will toggle on every rising edge. So, let us
say it starts from 0 it goes high here and on this rising edge Q will go low. And then again
on this rising edge goes high here again low high low like that.
So, this is toggle flip flop, toggling Q bar is opposite of 180 degrees phase opposite of Q.
So, it starts with a low, then toggles high low high low like that, so, this way you get the
Q bar output also. Now, you see this AND gates if the output steering control is 1 then A
is same as Q, B is same as Q bar.
So, let us say that this are A and B waveforms if output steering control is equal to 1,
meaning that the AND gates are enable. So, this will pass through A and B coming to 1 of
661
the inputs of the NOR gates. So, what is AA and BB? We will see that. So, AA and BB,
now this is waveform A and this is waveform Va this is NOR gate. So, it will be A plus
Va whole bar which is A bar and Va bar is not it.
So, this is Va bar A bar waveform. So, let us, let me look at only those portions where you
will get a pulse. So, I am just marking this portions. So, you see this is A bar, A bar this
will be high and during that time Va bar this portion will be high. So, I am what I am
showing here in the cursor, likewise this portion. So, I am marking those portions then I
am drawing the AA waveform, so, it will be high only on this.
Now BB let me mark it is now B bar into B bar and Va bar. So, you will see these are the
portions where BB will go high and I will write it down like that. So, you see here AA and
BB waveforms are mutually exclusive, you can give Q1 on here Q2 on, Q1 on, Q2 on, Q1
on, so on. and these are pulse width modulated, because they are coming from the primary
pulse width modulated waveform Va.
so, in this way you can generate 2 mutually exclusive dTs period waveforms which can be
given to 2 different switches Q1 and Q2 and where they are never on simultaneously, they
are mutually exclusively on. So, this are the kind of pulse steering that we do and give it
to the switches Q1 and Q2 of the push pull circuit.
Here again you will see that there is no correction for flux walking, typically many of the
PWM Ics do not provide correction flux walking you have to do that externally. However,
you do also find PWM Ics which give flux walking correction.
Another point you have to note is what happens when the output steering control pin is 0.
When output steering control pin is 0, A and B are 0, this latch and this points are out of
action. So, Va comes directly to the NOR gates of these two AA and BB both will be same
except inverted. And you will have a pulse width modulated waveform both same AA and
BB both same simultaneously. So, you can use that one to drive tools drive a single switch
or 2 switches with enhanced drive capability
662
A typical PWM IC of the structure that we just now described is a TL494 pulse width
modulation IC. Very simple around 10 rupees in cost and the block diagram internal block
diagram it is a Texas instruments IC internal block diagram is similar to the similar to the
one that we discussed.
It has a simple block diagram this is a 16 pin IC and easily available in the market.
663
Download its data sheet and you can see there is a test circuit here and you can set the
carrier frequency by setting the RT,CT waveform. And if you monitor the voltage
waveform at pin 5 across CT, you will see that it is a saw tooth carrier.
664
So, this is based on a saw tooth career and then you have 2 output waveforms at C1 and
C2 collectors of internal transistors, collector 1 ammeter 2, collector 2 emitter 1 which you
can externally connect and take out the take out the PWM outputs.
665
The functional block diagram similar to the one that we discussed, you see all an internal
oscillator RT,CT generating the saw tooth. There is a dead time controller comparator,
there are 2 other comparators error amplifier 1 and 2 they are internally ORed. And there
is a PWM this given to the clock, there is a D flip flop with toggle flip flop mode operation.
When you have the steering control logic and 2 AND gates and then the NOR gates given
to the internal transistors. There is an internal reference voltage which you can use.
So, you see the logic is very similar when you give the output steering control 1, then you
will see that you can, you will get mutually exclusive output which you can use for push-
pull circuit or even half bridge circuit and full bridge bits circuits. When you make the
output control pin 0, then the PWM directly comes to both and both are identical, PWM
outputs. You can use the 2 internal BJTs to enhance the drive signal. So, this an interesting
and easily available popular PWM IC available in the market.
Another way of providing pulse width modulated wave forms to Q1 and Q2 in mutually
exclusive intervals of time. Along with providing corrections or flux walking can be done
with this following circuit.
666
So, let us say we have this op-amp comparator and the minus is the triangular carrier and
this triangle career I will invert it and then give it to the minus of another op-amp. So, let
us say Va the pulse width modulate; the pulse with modulation fundamental that is the
modulating signal Va is given in this fashion this is a modulating signal. Now the inverted
wave form will look like this opposite of that one, and let us say the output of the
comparators are A and B they will be the pulse width modulated signals.
So, if I put this triangular carrier like this is the triangle and the inverted triangular carrier
in this fashion. And let us say Va we are giving same Va for both the comparators and this
is Va, let us draw A and B signals. So, the points whenever these are the crucial crossover
points for that black triangle here.
The points wherever Va is more than the triangle you will see it being high otherwise low.
So, this will be the output signal A, and for B you are comparing with the blue where the
inverted triangle, so, you will likewise get a signal like this. So, observe here that they are
low at mutually exclusive intervals of time.
So, if you put an inverter, so, let us say we want to get A bar, B bar by putting an inverter
like this and you can invert these 2 wave forms. And you will see that you will get the A
bar waveform like this and the B bar waveform like this. Now these are mutually exclusive
signal that will switch the transistors at mutually exclusive intervals of time.
667
Now, how do we add correction? Now this is a circuit where you can easily add corrections
independently to the A side or the B side waveforms PWM waveforms. So, let me erase
here and at this point let us have the correction to the control signal analog control signal
itself.
So, the output of the PI the flux walking PI which resulted in voltage correction value. We
can include it at this point here where you are subtracting from the control signal, analog
control signal before comparing, and one side you are adding it. So, in one side you will
add and another side will subtract. So, that you get a difference and we call that as Va1
and Va2, so, if this is Va1, Va2 will be like this.
Now it depends upon the Vcorrection signal sign if it is minus Va1 will be higher Va2 will
be lower. If this is plus Va1 will be lower Va2 will be higher depending upon that the PI
controller anyway will take care of it and correct that part. So, if you see Va1 for A and A
bar, let us say it is we have already drawn. And for if you take Va2 comparing it with the
blue triangle inverted triangle, you will see small changes here and because of the
intersection points. And you will see slight change in the pulse widths.
So, Va1 if it is having d1Ts, Va2 will be having d2Ts where d1Ts and d2Ts are slightly
different. And if Vcorrection is 0, then both will match and both will have the same dTs
pulse widths. So, in this way we can give differential pulse widths to the Q1 switch and
the Q2 switch. So, that you can you bring about correction for flux walking and balancing
out the flux, so, that you have a 0 average flux.
668
Prof. L. Umanand
Lecture – 74
Stimulation of pushpull convertor
Let us, now do the simulation of the push pull converter. I have here in the folder in the
resources, I have a push pull folder; push pull schematic is here, push pull net list, push
pull dot cir, and the edt 01 dot sub which contains the sub circuit files.
669
So, let me double click on this; so, you have the schematic, this is the complete
schematic of the push pull converter. You see the push pull transformer here, the primary
side is center tap, and the secondary side is also center tap. There is this switch I have
used generic switches this is Q1 and this is Q2, switch 2. I have placed here a zero
voltage source to sense the switch current.
I have placed another zero voltage source here to sense the Q2 switch current. Let me
expand that, so, here you see the push pull circuit, and on the secondary side which is
center tap. The center tap is the reference node and to the dot from the dot end you have
a diode D1 ORing at this point with diode D2 which is coming from the non dot end.
And that is passed on the inductor, capacitor and Ro load, they are all connected to the
center tap reference node. I will label the node here the output node o, I have labeled the
gate drive node as g1 and g3 here. I will tell you how I am getting that, and probably we
can label a node here for the pole voltage Vp, so, you have the node Vp here.
Now, edt01 sub is included here; and all the sub circuits x, starting with x, x push the
push pull transformer switches. They are all modeled in this edt01 dot sub, they have the
sub circuit models which we can use. Then down here, I have a PWM bridge which will
which will generate 4 outputs, this can be used even for the full bridge.
670
Therefore, I have made a generic PWM bridge block which will connect to arm 1 top
switch, arm 1 bottom switch, 1T and 1B, arm 2 top switch, arm 2 bottom switch. This is
the label that I have given and here you have the control voltage, I given it as minus 0.5
comparing with a triangular wave form which is going from minus 1 to plus 1.
So, this will correspond to 25 percent switch duty cycle. So, I am using g1 as one of the
PWM signals and for Q2 I am using g3 as the other PWM signal, these two are mutually
exclusive.
So, let me save the schematic.
671
I have now opened the terminal, and let me make the net list file push pull dot net from
the push pull dot schematic file. And then stimulate the push pull dot cir file in ngspice.
So, on simulation you will see the waveform I have plotted the waveform inductor
current waveform il. And the diode D1 waveform, because the diode waveform are very
unique. And therefore; so, it reaches steady state here around somewhere here. Let me
zoom this expand this portion, let me see few cycle here you see.
672
So, you see that this is the inductor current waveform. So, and the diode D1 wave form,
it follows the inductor current during this period when Q1 is on. And during the period
when both are off there is a freewheeling half through this diode and the other half will
freewheel through diode D2.
So, the blue waveform is the diode D1 current waveform. If you look at the schematic,
the output waveform is between o and r and the pole wave form is between p and r. We
can see the pole voltage waveform and the output voltage waveform.
673
Going back to the NGspice environment, I will plot voltage; the pole voltage waveform
with respect to the reference node r, output voltage waveform with respect to the
reference node r. And we will see what we have, let me explain that.
And you see that you have the output voltage waveform and the red one is the pole
voltage waveform. I will show few cycles here when it has reached steady state yes.
674
So, you see the output voltage blue wave form, and the pole voltage waveform. The
negative is due to the diode freewheeling, and this is n*Vi, n we have set it at 2, I have
not shown you that one, let me show you this.
So, this push pull, when I double click I have set Np primary turns as 10 secondary
returns as 20, so, turns ratio of 2. So, you will have 10*2, 20 will be pole voltage ideally.
But there are diode drops which comes in during the simulation, so, it was less than 20.
So, that is what you are seeing here, the pole voltage waveform is hovering at 17.5 volts
or so. So, you have to 2 diode drops coming to the picture, it is actually a diode drop and
VCEsat drop; so, those 2 are the once which are making up for this drop. You see also on
the negative side during freewheeling the drop negative voltage is due to the diode drop.
Let me show you 2 things, one is the edt01 sub and the cir files.
675
So, here if I double click on the push pull dot cir, you have you have this.
So, push pull circuit, I have a dot options relative tolerance and abs tol, I have said. So,
this is the dot trans statement, it is including push pull dot net, this is the control
statements where I am setting the background as white and foreground as black. And
running the plot command to visualize the inductor current and the diode D1 current so,
this is the say our file.
Next I want to also show you the sub circuit file.
676
So, in this sub circuit file; so, what do I have, I have the default diodes of course, diode
micro models.
And I have this switch model which I have used.
677
And there is the electromagnetic transformer, the gyrator; these are called intern by the
push pull transformer. This is the push pull transformer that we are using, sub that sub
circuit which is calling the gyrators.
So, these model we have done using analogue behavioral modeling; of course, it is not
within the scope of this course to cover analogue behavioral modeling for simulation.
But, I am including this file you can study at leisure to understand that.
678
And then, down here I am including another analogue behavioral model for PWM
generation which will generate the 4 PWMs for the 4 bridge switches.
So, this is stimulating push pull circuit, you are free to make changes and explore the
circuit as much as possible, and change the source. You can change the output you can
change the turns ratio, you can also change at the PWM the duty cycle and check how it
will work, how it will behave. And the voltages and the currents at various points in the
push pull circuit is it coming according to what we discussed and according to theory.
And if it is not coming according to theory where is the discrepancy, where is the non-
ideality (Refer Time: 10:50) coming into picture, these aspects you can study.
679
Prof. L. Umanand
Lecture – 75
Half bridge converter
The half bridge converter like the push pull converter is also a buck derived converter,
where the secondary side of the half bridge converter transformer is center tap and behaves
exactly like that of the push pull secondary stage. The only difference comes in the primary
side. So, let us consider where the half bridge converter differs with respect to the push
pull converter. I am drawing here a sine source, ac source, rectifier and a capacitor filter.
This is the ac to dc conversion where you have an ac source, a rectifier and a capacitor
filter. This we have studied, but I want to make use of this I am showing this in gray shade.
I will remove the capacitance and split the capacitor into two capacitors and I will have
balancing resistances there so, that the voltage division is proper. Capacitances being equal
you can expect Vdc/ 2 to be across the top capacitor and another Vdc/2 across the bottom
capacitor. So, if you use this as the dc link for the switched mode converter circuit, the half
bridge configuration looks like this. So, I pulled this out and called this one as Vdc and I
will put this half bridge on like this.
680
So, you have two transistors or MOSFETs or IGBTs in this fashion and across this center
this is Vdc/ 2 point and we will put across this winding the primary of a transformer.
Secondary of the transformer is center tap and you have two diodes in this fashion just like
in the case of push pull and with respect to the center tap you have this inductor and the
capacitor and the output load.
So, behavior of the secondary side of the circuit is exactly same as the push pull. So, I am
just putting in freewheeling diodes or body diodes across the switches for completeness,
you have the dot polarities in this fashion and this is the pole voltage Vp. So, let us say
you have 1:n in this fashion L,C,Vo the output voltage.
So, let us say when Q1 is ON, Vdc is connected to the dot end. So, dot end is positive with
respect to the non-dot end, non-dot end is connected to Vdc/ 2. So, here dot is plus and
non dot end minus and the amplitude across the primary is Vdc/ 2 because this is connected
to Vdc and the non-dot end connected to Vdc/2.
So, totally you have Vdc/2 appearing across the primary. This transfers to the secondary
so, you have n* Vdc/2 on the both halves of the center tapped secondary. Now the dot end
is positive so, this diode here the top diode will be active and the bottom diode is reverse
biased and Vp will be n* Vdc/2.
So, likewise we can also evaluate for Q2 is ON that is Q1 is off and during this time you
see that the dot end is pulled to the ground and the non dot end is connected as usual to
Vdc/ 2.
681
So, let us mark. So, the non dot end is positive, dot end is connected through Q2 to ground
so, non dot end is plus and dot end minus like this. So, if we measure across the primary
you have minus Vdc/ 2 appearing over the primary, secondary likewise n* Vdc/2, here
with the polarity are shown this diode the bottom diode is active and Vp is again seeing n*
Vdc/2 just like in the push pull case.
All other operations on the secondary is same you have periods of time in both Q1 and Q2
both are off under that condition L is freewheeling through the secondary equally dividing
between the top half of the center tap and the bottom of the center tap the mmfs cancel and
𝑑Ф
is 0 and you have voltages across all coils being 0 at which point the center point is at
𝑑𝑡
Vdc/ 2 and the voltage that is across both the halves; across both the switches will be
Vdc/2.
So, therefore, Vp is equal to 0 when Q1 and Q2 are off and what is Vo? Vo is related to
the Vp voltage which is equal to (n* Vdc/ 2 )* 2d here, again there is an oring effect here,
the d is the duty cycle of the switches in a period Ts, but because the oring effect Vp
waveform has a period of Ts/ 2 and therefore, the duty cycle of Vp waveform is 2d.
682
Now, I will put that 2d here, 2 and 2 will cancel of and then you have Vo is equal to
n*Vdc*d, the input-output relationship for the half bridge converter. Here one important
point to remember as in comparison with the push pull is that, in the case of the push pull
we saw the voltage across the device, you see the voltage across the device is ON whenever
the switch is ON and when the switch when the others which is ON, complementary switch
is ON you see you see 2* Vi coming across the switch.
So, whenever both the switches are off you see Vi coming across it. So, the maximum
voltage that each switch has to support is 2* Vi. In the case of the half bridge converter
the voltage across each of the switch max is Vdc or Vi. So, when the bottom switch Q2 is
ON, then you see that this point the dot end point is at ground and therefore, Q1 which is
off will have to withstand Vdc. In the case when Q1 is ON you will see the bottom which
is off. Then this point here dot end point is at Vdc and Q2 will have to withstand maximum
Vdc. So, that is the difference the push pull the devices have to be rated twice Vdc, in the
case the half bridge the device have to be rated at just Vdc. So, you have half the voltage
rating requirement for the half bridge converter that is an advantage.
Another problem in the half bridge converter like in the case of the push pull converter is
that of flux walking. You will see that in the case of the half bridge converter the primary
side is not center tap, but a single winding.
683
So, let us look at the flux walking problem in the half bridge converter. So, let say this
Vprimary voltage across the single winding in the primary and let us look at that primary
voltage when Q1 is ON and Q2 is ON.
So, Vprimary is equal to Vdc/ 2 minus VCEsat1 when Q1 is ON. See normally when Q1
is ON, Vdc comes to the dot end and the non dot end is connected to Vdc/2 here at the
midpoint of the capacitance. So, the voltage across the primary Vprimary will be Vdc/ 2,
but in a practical case because of the drop across Q1, you will see Vdc/ 2 minus VCEsat1
coming into the picture.
Now, the other condition when Q2 is ON Q1 is off. So, when Q2 is ON you will see that
the on state drop of Q2 comes into the picture and therefore, VCEsat2. And we know that
VCEsat1 and VCEsat2 are not same, it will not be equal in a practical case and therefore,
there will be a difference between the primary voltage when Q1 is ON the primary voltage
when Q2 is ON and because you are applying dTs for Q1 and dTs ON time for Q2, the
difference between VCEsat1 and VCEsat2 will reflect in an average primary voltage very
small average primary voltage. And because of the Faraday’s law Vaverage which is equal
to n*dФ/ dt, when you integrate the voltage, Ф will be proportional to time t and it will
gradually increase.
So, as a consequence you will see that the average flux will start rising and then ultimately
reach the saturation flux limit and saturate the core of the half bridge converter. So, how
684
do we solve this? Like in the push pull converter we can use the same method of adjusting
the duty cycle of Q1 and Q2.
In the case of the push pull convertor we had to sense windings, one sensing the Q1 part
of the volt second, in the other sensing the Q2 part of the volt second and these 2 are
subtracted the error the difference e was passed through a PI controller; the output of the
PI controller developed the correction voltage and that was used for correcting the dTs or
the ON times of Q1 and Q2 respectively. In such a way that e becomes 0 and if e become
0 then the volt second of the top part that is when Q1 was ON the volt second of the bottom
part when Q2 was ON would match and thereby achieve the 0 flux average and thereby
avoid flux walking.
Now, this technique can also be used for the half bridge converter and the Q1 and Q2 duty
cycle can be adjusted. In the case of the half bridge converter there is yet another way to
achieve this solution for the flux walking problem that is by placing another component
that is placing a capacitor in series with the primary winding. The advantage here is that
the primary voltage across the center tap here if it develops an average, it can drop across
the capacitance.
There are two components here the coil cannot with stand an average value, but the
capacitor is capable of having an average value and therefore, the average value will drop
685
across that and thereby save the coil from having any average. And this way the flux in the
core will not walk away and the flux will not saturate.
So, this is a nice solution and this capacitor is called the flux walking capacitor Cfw.
Normally, the Cfw is rated for around 10 percent of Vdc, though the voltage that comes
across the flux walking capacitor will not be so high, it will be much much lesser. The one
thing that one should remember you that this Cfw the flux walking capacitor should be
bipolar should be capable of having current flow in both direction and the voltage in both
direction, voltage drop in both direction.
And another aspect is that the entire load current will flow through the flux walking
capacitor and therefore, the arms rating of the flux walking capacitor must be high, and
this will make the flux waling capacitor costlier. However, this is a nice solution especially
if it is for low power circuits.
686
Prof. L. Umanand
Lecture - 76
Simulation of halfbridge converter
Let us do a Simulation for the half bridge converter circuit. I have here in this folder
within resources half bridge folder which is containing the sub circuit files the models,
the half bridge schematic and cir file the net list.
687
So, let us go to the schematic this is a schematic of the half bridge circuit. I have the dc
link set up by these two capacitors equal capacitor 100 µF, 100 µF capacitors, I have
balancing resistor there across this and this forms the dc link with the centre point being
Vdc/ 2. Now on the other side here I have two switches a bridge arm with two switches
Q1 and Q2 or SW1 and SW2 now these 2 are generic switches driven by 2 gate drives g1
and g3.
688
And g1 and g3 are coming from the PWM block which I explained while doing the push
pull circuit similar block which is generating 4 signals. And if it is for full bridge I use all
the 4 otherwise you can use the top 1; 1 top or g1 and you can use the g3 - 2 top. These
two 1 top and 2 top must be mutually exclusive; 1 bottom and 2 bottom must be mutually
exclusive either you can use g1 and g3 or g2 and g4 for these two switches.
So, I am using g1 and g3 and in between you have this full bridge, you have this bridge
transformer, the same type of transformer for both the full bridge and the half bridge. I
am using a sensing voltage source here with 0 volts and this will sense the current
through the primary. And the secondary of this bridge is centre tapped, the dot end is
going through D1 diode, the non dot end through D2 diode they are oring up at the pole
point and then you have the inductor followed by capacitance and the load, r is the
reference node on the secondary side. So, anything that you want to measure on the
secondary side whether it be the output voltage or the pole voltage, you measure it with
respect to the reference node r.
So, now we can simulate this circuit and see how it behaves.
I have now open the terminal I am within that resources half bridge circuit; half bridge
folder, and let me generate the net list first. So, the net list is generated from the
schematics file so, you can generate the net list and then you simulate the circuit Ngspice
689
half bridge dot cir. Before we simulate cir let us have a look at what is within this half
bridge dot cir.
Here half bridge dot cir contains this one you have the simulation commands dot option,
dot tran which is the analysis command, include half bridge dot net. And then in the
control set of commands in the Ngspice environment setting the background to white,
foreground to black, running the plot simulation and plotting iL inductor current their
diode D1 current and the primary current these three proper currents we will see.
690
So, let us go back and do the simulation now. In the terminal I will do the simulation.
Simulation is performed and then you see all this waveforms coming up let me go to the
steady state and you will get an idea of what we are seeing.
So, here you see the red waveform is the inductor current and the blue waveform is the
diode D1 current just like in the push pull case, and the orange waveform is the primary
current. Probably we can zoom in a bit and see you will get a much better view if you
just few cycles yes. Now this gives you a much clearer picture.
691
So, you see the primary current 1 2 3 4, 1 2 3 4 it is an AC current flowing through the
primary that is the AC current flowing through the primary. So, you see that when the
switch S1 is on, you have the current flow in this fashion to into the dot and when the
switch S2 is on you have the current flow into the non dot end coming out of the dot and
into the ground.
So, you have current flow on both directions and it is basically a reflected inductor
current onto the primary.
Here to in the bridge transformer I have said primary turns as 10, secondary turns as 20
you can explore with other terns ratios and the sub circuit for this bridge transformer you
will find it in edt01dot sub.
692
If you open edt01 dot sub, the sub circuit model for the bridge transformer is here. See it
also uses the gyrator sub circuits. In the gyrator is model just above that there is a gyrator
here, as I told you the modelling of this transformers we have used analogue behavioural
modelling and you can study that separately.
And I am using the same PWM Bridge; PWM generator and we can go back to the
schematic and this is what it is. So, I will now leave it to you to change and explore the
693
various waveforms of the half bridge circuit. You can try to see the pole voltage and the
output voltage too, maybe I can show that to you briefly before closing.
Plot v pole voltage with respect to the reference node r, plot output with refers reference
with respect to the reference node r see the outputs.
And when you expanded it you will be able to see; you will be able to see both the pole
voltage and output voltage
694
Here is a pole voltage and the blue one is the output voltage here.
695
Prof. L. Umanand
Lecture - 77
Full bridge converter
One of the most popular converters at high power is the full bridge converter, the topology
is like this; you have 4 switches. In fact, it is comprised of 2 half bridges; see you have
one arm one top switch, one bottom switch. You have another arm, one top switch another
bottom switch connected like this, and both are both arms are connected to Vdc in this
fashion.
Now, if you take the centre points of these two bridge arms, we connect the transformer
across them. So, I am completing the circuit with body diode for the sake of completeness.
They help in during freewheeling operations; so, you have Q1, Q2, Q3 and Q4.
So, across the centre points we will connect the primary of the transformer, the secondary
of the transformer like in the case of the half bridge. You have the 2 diodes and through
the centre tap we have the inductor, capacitor and the load which is operating as a buck
derived converter. It is a buck operating in the buck mode, but with Oring effect as in the
case of push pull and also as in the case of the half bridge converter, operation is exactly
similar.
696
Now, I will put the dot polarities like this. Now, this Vdc; let us consider the case, you are
now operating the switches in pairs, Q1 and Q4 will be turned ON simultaneously. Q2 and
Q3 will be turned ON simultaneously; however, when Q1 is ON Q2, Q3 will be OFF,
when Q2, Q3 are ON Q1, Q4 are OFF.
So, Q1, Q4, when they are ON; Q1 is ON means that the non dot end is connected to Vdc,
the dot end is connected to ground through Q4. So, what appears across the primary plus
minus in this in this fashion. So, the non dot end is minus, dot end is plus, this is the pole
voltage Vp, so, let us look at that. So, when you measure across a primary you get Vdc
measured in this fashion.
And the non dot end being positive, the bottom diode here is active and then it will supply
power to the inductance Vp will be n* Vdc . If you take a turns ratio 1:n. So, Vp will be
n*V dc we will write it down, and during the time when Q2 and Q3 or ON; these 2 are ON
simultaneously. You will see that Q3 is ON means dot end is positive and non-dot end is
negative, Q2 connected to ground.
So, secondary side also you will see the dot end positive non-dot end as negative. So, the
primary voltage is Vdc it will translate into n* Vdc and Vp will be n* Vdc during the time.
And during the time when Vp is 0 when is it, when all the 4 switches Q1, Q2, Q3, Q4 are
OFF.
So, when all these 4 switches are OFF, the inductor current is freewheeling and like as
before in the case of the push pull converter and in the case of the half bridge converter;
the secondary side is centre tap. The inductor current divides splits here exactly by half in
to the top winding and into the bottom winding it goes into the dot end in the bottom
winding, it goes into the non dot end in the top winding.
The mmfs so, generated are opposing and cancel each other in the core and the flux does
not change. So, dФ/ dt is 0. And therefore, the induced voltage across all the windings are
0 and Vp will be 0 and so, also the voltage across the primary.
So therefore, when you take Vo with respect to Vp; this is n* Vdc, Vp is having double
the duty cycle as that of the switches. So, n times Vdc*2*d and re-writing 2* n* Vdc* d.
Now, this is the input output relationship for the full bridge converter. Now, let us say
VCEO rating for the switch, like in the half bridge is maximum Vdc. So, when Q1 and Q2
697
are ON, Q2, Q3 are OFF and you will see that Q2 is connected to V dc, Q3 is connected
to ground through Q4 and therefore, Q3 has to withstand Vdc, Q2 has to withstand Vdc .
Now, when Q2 and Q3 are ON Q1 is connected to ground Q1 has to withstand Vdc, Q4 is

connected to Vdc and Q4 has to withstand V dc. Therefore, all 4 switches need to withstand
Vdc and they have to be rated for greater than V dc. In the half bridge convertor also the
each of the switches were supposed to be rated for Vdc, only in the push pull converter the
switches need to be rated for 2* Vdc.
Now, in the case of the full bridge converter also we see that there is a requirement for
solving the flux walking problem. Wherever, this kind of a situation occurs where you are
having switches and practical or non-ideal switches and you are having a winding
connected in this fashion, there is a chance of an average voltage occurring across the
Vprimary.
Because when Q1 and Q4 are ON the voltage across the primary will be Vdc minus Vcesat
of Q1 and Vcesat of Q4. And when Q2 and Q3 are ON it is Vdc minus Vcesat of Q2 and
Vcesat of Q3. All the Vcesat Q1, Q2, Q3, Q4 will not match and therefore, there can be a
primary average voltage. And, this will get integrated and a there will be a flux which will
have a proportionally with respect to time, when the flux will be increasing with time and
ultimately saturate this core. Now therefore, you need to provide flux working and just
like in the case of the half bridge and push pull case you can provide flux walking by
adjusting the duty ratios of Q1, Q4 and Q2, Q3 set.
Alternately, like in the case of the half bridge, you have the primary winding which is a
single winding not centre tap. Therefore, you can place a flux walking capacitance which
will prevent an average flux occurring in the winding, and the Cfw or the flux walking
capacitor will absorb the average voltages. So, in this way flux walking can be avoided in
the full bridge also, but note that full rated current has to flow through the capacitance
here.
Now, let me mark i-primary and v-primary here and what is the difference between half
bridge and full bridge. For a given power, given output power; let us say in case of the half
bridge. The primary voltage is Vdc/ 2 recall and let us say the primary current is some
value ip. For the full bridge case, for the same power if you take v-primary voltage is Vdc
698
that is fixed by the topology. And the primary current for the same power has to be half as
that of the half bridge should be ip/ 2.
So, that the product Vdc/ 2*ip and Vdc*ip/ 2 are the same to realise the same power.
Therefore, in the case of the full bridge converter to realise the same output power, you
need to rate the switches half of that of the comparable half bridge converter, and this is
the advantage of full bridge converter. However, you need 4 switches and 4 gate drives for
the full bridge converter, against 2 switches and 2 gate drives for the half bridge converter.
699
Prof. L. Umanand
Lecture – 78
Stimulation of fullbridge convertor
Let us now do a simulation for the full bridge converter circuit. I have the resources folder
within a folder on full bridge and within that you have the full bridge schematic, netlist cir
file and the sub circuit model files.
700
So, when you open the full bridge dot schematic it look something like this. You have a
full bridge circuit diagram here as we have discussed switch Q1 Q2 Q3 Q4, and they are
driven by gate drive signals g1 g2 g3 g4. So, these two are mutually exclusive, these two
are mutually exclusive, these two are switched together. And g2 and g3 are switched
together. So, you can observe the wave forms at a and b and across a-b.
So, I have here like in the half bridge and the push pull, a 0 voltage element put as a current
sensor. On the secondary side of the bridge centre tap bridge you have a 0 voltage current
sensors to sense the id1 and id2. The centre tap is the reference node point, and from the
pole you have the inductance and then the output with the C filter and R naught; so, this is
your full bridge circuit.
701
You see that I have 10 turns in the primary 20 turns on the secondary, turns ratio 2. So,
you will get v pole: 10 volts, at the primary 20 volts for p. But you will have the diode
drops and the on state drops of the switches, there are two on state drops.
So, you will have one on state drop for switch 1 another on state drop for the other switch
which is 4, which will come into picture, then you have the diode drop. So, you probably
will land up with around with 3 to 3.5 volts overall. So, you may have some 16 volts pole
voltage. Then the duty cycle and output voltage comes in here.
702
And if you look go down here you I am using the same PWM bridge, the first arm top
switch drive this is g1, first arm bottom bridge drive it is g2, second arm top switch drive
g3 and g4. And always remember that switch1 and switch4 are switched on
simultaneously. And switch 2 and switch 3 are switched on simultaneously and in a
mutually exclusive manner. Now, let us simulate the circuit and then see we will go to the
terminal and run the simulation in ngspice.
So, I am in the terminal, I am in the folder full bridge, so, let me run the net list. Let us
generate the net list from the schematic. So, you have the net list generated, now run
ngspice from the cir file I have used similar cir file like I used for half bridge and push pull
circuit.
703
So, on running you get the waveforms like this. As before, I have measured the current
through the inductance, output inductance, current through the diode D1 and current
through the primary.
So, if you if I zoom in and look at that, you will see that bit more clearly, the red wave
shape is the current through the inductance going up and down at twice the frequency as
the switching frequency. The blue wave shape is the current through the diode, see that it
704
follows the indictor and then doing the freewheeling time. It takes only half the current the
remaining current flows through the other diode.
And the orange waveform here is the primary current waveform which is reflected on to
the primary of the full bridge circuit. Let me plot v ab, v ab is the voltage across the bridge.
So, this will give you the voltage across the primary, and let us see how it looks like.
705
So, if you plot that voltage and let me take a small portion of that, if you see that, you see
that the voltage when one set of switches are ON, you will see the primary dot end positive.
And it is Vdc minus the saturation drops of 2 switches.
And then when the non dot end is positive, you have ON the other side of the primary
voltage these two should balance each other. And during this time during the time when
all 4 switches are off you will see observe that there is no induced voltage across the coil
the primary voltages are 0.
So, this is typically the type of operation that we expected. And I will leave it to you to
explore this full bridge circuit change the values and try to learn as much as you can from
this simulation.
706
Prof. L. Umanand
Lecture - 79
Area products and mfiles
We discussed 3 converters; push pull converter, the half bridge converter and the full
bridge converter. How do we design the transformers for these converters? We go by the
area convert area product method we have discussed the area product method and then we
used it to design the forward converter and we had developed generic area product
equation. Using the generic area product equation we can find out the area product the
push pull, I will give you the final derived formula for the push pull converter area product
is equal to √2*Po* (1+ 1 / efficiency). Here efficiency is taken to be 0.8 first case.
So, efficiency will always be better than 0.8 and the denominator you have 4*Kw window
factor J current density which is 3* 10^6 ampere/ meter square Bm for ferrites it is 0. 2
Tesla and then fs the switching frequency.
So, this is the area product for the push pull converter. If you use the same generic formula
and then plugin the voltage and the power into the generic equation, that is a half cycle
average voltage you will get this. Likewise for the half bridge converter transformer also
707
the area product I will be providing here, derived along same lines √2*Po* (1+ 1 / 0.8)
divided by 4*Kw*J*Bm*fs.
The full bridge converter and half bridge converter the transformers are very similar and
the voltage wave shape across the primary and the current through them are all similar and
therefore, you will have a similar area product equation. The difference of course, will
come in the actual values of the currents and the amplitude of the voltage. So, this is the
area product for the full bridge converter.
So, you can try to derive these and verify them by using the same generic method, for
deriving the area products. That we discussed for the forward converter as an example and
also the low frequency AC 50 Hertz transformers too. So, all these transformers have the
same generic basis.
Let us now look at some m-files for designing the push pull converter the half bridge
converter and the full bridge converter.
708
So, I have in the resources section a folder called m-files within that like for the forward
converter, we have some common m-files that is the ferrite cores, the metglas core powder
iron core and there is another m-file on wire wires and wire gauges. So, these are common
for all the converter design in the sense that they are data bases of the cores and the wires.
You have 3 design files; one is the push pull design dot m, half bridge design dot m and
the full bridge design dot m and these are the design m-files which you can run and will
make use of this core for right wires and other core and wire details from these other m-
files. So, let me open the push pull.
709
So, here you have the pushpull design dot mfile.
And it has a specification section as specified somethings observe here Vo is having a

vector of 2 elements which means it is an isolated multi output. There are two ORed
secondarys. You have the Δv ripple and 1 amp and 0.5 amps for the 2 different outputs.
And you have the set of designer variables and switching at 40 kilo Hertz, flux density
here 0.2 Tesla for transformer, maximum 0.25 Tesla for inductors, Dmax of 0.45, J I have
taken 3 amp/ mm-square or 3*10^6 amp/meter-square for the current density and window
factors I have chosen like this.
710
So, there is the power calculation and after you do the power calculation the turns ratio
calculation and this is where it will be different from the forward transformer design which
is a push pull transformer. The area product of the push pull transformer I have taken it
according to the area product of the push pull that I have just listed down. And the part of
the design is similar to that in the forward converter where we call the cores.
And the cores selects the cores and then wires for selecting the wire gauges and cross check
the design.
711
Then you have the inductors; inductors for every isolated output, we have 2 isolated output.
So, 2 times it has to be done. So, inductor value calculation followed by energy calculation.
Then the selection of the core from the cores function file and then the wires from the
wires function file.
712
Then you have the design of the capacitors the output diodes power switch and then finally,
to display the results.
713
So, this is the push pull design, I have given the specification here you can go in to octave
or even matlab for that matter go into the respective folder resource folder I already have
and I have here full bridge half bridge and push pull. So, you type in push pull design and
it will ask for the type of the core material I am going to choose 1, which is Ferrite and
then it will ask for the shape of the core material I am going to choose EE.
And I am going to choose 2 and you see here the area product calculated and then the
window area cross check is not fitting. So, again you will have to choose the next bigger
714
size core you can change the material if you want, but I will stick to ferrite, go ahead with
ferrite and the EE core.
Let it choose the next higher one, it has chosen then it has gone for inductor a winding 1 I
will choose Ferrite and EE CORE to keep the same inventory list, then inductor for
winding 2 again you choose ferrite and I am choosing EE core.
And then that is it the calculations are done and the results are displayed, you have the
push pull converter specifications then the designer variables.
715
Then followed by the POWER SWITCH RATINGS diode ratings, inductor design.
Inductor design with its core, air gap all those things capacitor filter rating and everything
that is needed for rigging up the converter.
So, the push pull converter design is designed in this fashion, I will allow you to play
around with this specification and look into the equations and try to get more insight into
the multi output push pull converter m-file.
716
Likewise, I have the half bridge design m-file, you can look into that also. I have kept the
same specifications and so that you can compare. Slightly I have changed the output specs,
but otherwise I have essentially kept the same simplification.
Here also everything is same except for the half bridge transformer design, the area product
which this equation will change just like as I have listed shortly a while back.
717
So, you can execute this design like I did for the push pull and try to find out the design
values.
I will just run it through half bridge converter design, it goes in the same fashion I will
choose Ferrite EE CORES again Ferrite EE CORE for the inductor. So, you will get the
design values, you can read through the values later on and here in the case of the half
bridge I am also calculating the flux walking capacitor ratings for so, that if you want to
718
use the flux walking prepare prevention using capacitors you can use it, for small powers
this is really fine. In a similar way I have also with me here the full bridge design m-file.
So, very similar except for the area product. The area product in the case of the half bridge
and the full bridge are same. So, even those equations remain the same. You just have to
run full bridge and do the same process of selecting the core, selecting the shape and go
through the process of selecting the core for the inductor shape, inductor 2 shape and then
you get the results.
So, just walk through the design values, you will probably get a hang of the numbers for
the typical design of this specification. So, I will allow you to play around with the specs
and get try to get more insight into the design of the converters using this m-file as a
template.
719
Prof. L. Umanand
Lecture - 80
Intro for drive circuits
In this session we shall discuss on the very important topic of drive circuits. You see that
we have used semiconductor switches at many places in DC-DC converter. In the
beginning while we were discussing and studying the buck converter we have used a
BJT at this position. Now instead of this BJT it could as well be a MOSFET or it could
as well be an IGBT.
Observe here that there is this control signal we have just always left it floating in all the
circuits saying that if we give the proper signal as per the drive requirements, then this
will be on and during the dTs period. Q will be on during 1- dTs period, Q will be off
these are just only information signals. How do you convert the information signals of
this dTs period high pulse and 1-dTs period low pulse to ultimately drive this switch on
and off now that is what we need to study.
Because all these semiconductor switches are have their own characteristics. For
example, the BJT is a current driven switch you need to give the proper base current to
720
switch this ON and you have to remove a certain amount of base charges to turn this
OFF.
How do we provide for all those things in the base or the drive electronic circuits? In the
case the MOSFETs you need to provide a certain amount of gate charge for it to turn ON
and remove a certain amount of gate charge for it to turn OFF. So, like that each of the
semiconductor switch will have their own characteristic and we will have to design the
drive circuit appropriately. So, that you can turn ON the switch and turn OFF the switch
at will.
We have seen also in the forward converter there is this BJT, now we have shifted the
BJT from the rail based to the ground based saying that the drive circuit is simpler if it is
driving a ground based switch, we will see all those issues why it becomes simpler to
drive a switch which is ground based.
721
The push pull converter has two switches which are mutually exclusive and they need to
be switched in a particular pattern, it is easy for us to get the ON OFF signal information.
But how do you convert this ON OFF signal into a proper base current or a charge drive
for BJT or a MOSFET is the matter that will be concerned, now that we will be
discussing in this particular session.
In the case of the Half bridge converter and the Full bridge converter you have top
semiconductor switch and a bottom semiconductor switch. The bottom semiconductor
722
switch is ground base drive, the top semiconductor switch is a floating drive. So, how do
you make such Complimentary drives for the half bridge and the full bridge converter
switch arms.
In case of full bridge like the half bridge you have two switches semiconductor switches
which are floating drive and two semiconductor switches which are ground based drive.
So, these are the aspects that will come into the picture.
Apart from that we need to provide protection or stress relief for the switches because
there will be di/ dt so Ldi/ dt voltage surges across the switches when they are turning off
and when they are turning on there will be currents surges through the switches.
And we need to definitely protect them or limit the stress on these switches by using
some circuits which we will call as snubber circuits. And these snubber circuits are
important component of the switch base drives.
723
In this session we shall discuss the drive circuits of two very important most popular
devices one is the BJT, the Bipolar Junction Transistor. The BJT drive circuit they are
very important because they are ubiquitous there they come in the drive circuits of
almost all the power semiconductor switches whether it be the BJT power switch or
whether it be a MOSFET power switch or whether it be IGBT power switch or the now
more recent silicon carbide and GaN switches.
The BJT comes as one of the circuit elements in the drive circuits. So this is very popular
and very useful we need to study that if you have to understand the drive circuit. The
second one that we will be studying is the MOSFETs, the MOSFET switch are very
popular in all DC-DC converter circuits and we need to study that.
Of course, in the high power and high voltage circuit’s IGBTs are the most popular, but
have drives similar to MOSFETs and if you know how to do the drive circuits for
MOSFETs you could as well do it for IGBTs to. So, this is basically the focus of our
discussion in this coming session and that is what we will be talking about circuits drive
circuits for BJTs and MOSFETs.
724
Prof. L. Umanand
Indian Institute of Science, Bengalure
Lecture - 81
BJT base drive
Before we try to design drive circuits for practical BJTs and MOSFETs let us try to
understand what an ideal switch is. It will give us more insight while the designing the
drive circuits for practical switches. Now, consider a switch like this, it has a pole and it
has a throw and there is a control pin which will control the single pole single throw
switch.
Now, consider it is an ideal switch. Now, what is the i-v static characteristic of an ideal
switch? So, let us draw the i-v characteristic x-axis is the voltage axis, voltage across the
switch which is vTP and current through throw and pole that is iTP.
Now, for an ideal switch we would like that when the switch is ON the current to be able
to flow both in the positive and also in the negative direction meaning T to P and P to T
also and during the time on the switch is OFF, we would like the voltage across the
switch that is to be supported by the switch. That it should be capable of supporting both
the forward voltage and also the reverse voltage, positive voltage and also the negative
voltage.
725
Now, this would be the characteristics of an ideal switch, a static characteristic where
you have the i-axis covering both the positive and the negative current, v-axis covering
both the positive and the negative voltages. Let us now list down the characteristics of
the ideal switch. Now, let us say in the ON state when the switch is on, as we saw from
the static characteristic we should say that it should allow current flow from T to P and
allow current flow from P to T also, that is support both positive and negative switch
currents.
Next the zero conduction loss meaning when the switch is ON there is a current flow
there should not be any i^2*R loss within the switch or there should not be any drop
across the switch only under such conditions will zero conduction loss happen. So, let us
say one of the character of an ideal switch is that when the switch is ON there is no
power loss in it. As a corollary feature you have zero conduction resistance. So, when the
switch is ON there is no resistance the switch resistance is zero and therefore, the drop
across the switch is zero and therefore, the conduction loss is also zero.
In the OFF state when the switch is OFF again looking at the characteristic it has to
support voltage. So, it should be capable of supporting both positive and negative voltage
across the switch, this is one character of the ideal switch and also like during ON state,
during OFF state also there should be 0 off state loss no leakage current should be
flowing through the switch. And the resistance when the switch is OFF should be infinite
should be infinite meaning 0 leakage currents flowing through the switch and also
meaning 0 losses in the switch.
Another important character of an ideal switch is that, switching the switch from ON
state to OFF state, OFF state to ON state should be instantaneous it should be in 0 time
which is not so in a practical switch, but this is a desirable feature in an ideal switch.
This also implies that if it is instantaneous if this switching ON and switching OFF
happens in zero time it also means that there is no power loss during switching.
So, zero power loss during switching. So, switching transitions happen in a lossless
manner. So, therefore, switching losses are zero, then this also implies that the drive
circuits to drive this control pin of the single pole, single throw switch these drive
circuits will have zero drive power requirement. So, therefore, just only the information
signal is sufficient to be given to turn the switch ON and OFF.
726
So, these are some of the classic features of an ideal switch; however, you should know
that none of these features are fully met in a practical switch, as we have seen even in the
static characteristic of BJTs Diodes and MOSFETs they do not reflect the static
characteristic of the ideal switch.
So, we have to consider the shortcomings in each of the switches and try to design drive
circuit which will try to make up for the shortcomings and take it to as good
semiconductor switch electronics switch as possible.
We see here the static characteristic of an ideal switch, during the ON time we see that
the ideal switch is capable of supporting both positive currents and negative currents to
flow through it, throw to pole and pole to throw. And while the switch is OFF it can
support a positive voltage across the device and also negative voltage across the device.
Now, as again this how does the static characteristic of a practical semiconductor switch
appear? Let us take BJT for an example and see how its static characteristic looks like.
So, let us say this is the symbol of the BJT we have the collector, this an NPN, BJT,
base, emitter the voltage across the collector and the emitter we will indicate it like this
and the current through the collector iC.
Now, you can have on the x-axis vCE voltage drop across collector and emitter of the BJT
and collector current iC as the current through the switch. So, we know that the collector
727
current is unidirectional, it can only flow the positive direction from C to E, and there
cannot be a current flow from E to C. So, this can support only positive direction of
current flow when the switch is ON, and when the switch is OFF the BJT can withstand
only positive voltages across collector and emitter. And therefore, only this portion of the
static characteristic or potential operating points. So, this is how a practical switch will
look like.
Now, let us say that there is an operating point here, what does it mean? It is on the x-
axis it is supporting a voltage’ current is zero which means it is in the OFF state. So, the
switch is in the OFF state zero current is flowing and it is supporting some positive
voltage across the collector and emitter.
Now, let us call this as the OFF state. Now, let us take another point on this axis like this
are shown, here there is some finite current for this operating point and the voltage
across the switch is zero, what it basically means is that this is the ON state and it is
supporting a positive current. Now, if I have to transition from OFF state to the ON state
one way is to go along this line with zero current, go along this line at zero current then
at zero voltage go along this line and reach this operating point.
Now, if you take this route at every point on this route the power loss is 0, in all this
portion to the route current is zero and then once 0 voltage is reached in all this portion
of the route voltage is 0. So, therefore, in transiting from the OFF state to the ON state
you do not lose any power. Likewise while coming back you also do not lose power if
you are using this route.
So, in this route you have zero switching losses so which basically means that if I am
using a switching with zero voltage and zero current then you have zero switching losses.
On the other hand the practical switches try to reach the other state directly crossing
through the linear region. So, when you cross through the linear region there is finite
voltages and current in between and there will be loss in the device. So, these are called
the switching losses and the device should dissipate them through heat sinks and remain
cool; however, the efficiency will come down.
So, this route is a lossy route and the lossy transition, we should actually try to avoid that
such a route, but all practical semiconductor devices do have this. They directly go
through the linear region to reach other state and back again through the linear region to
728
reach the OFF state. We should try to make this as quick as possible to reduce the
switching losses or you we try to put additional circuit such that it does not go to the
linear region, but goes through this route.
So, let us design the drive circuits keeping this kind of a background in mind. In order to
design drive circuits just the static characteristic is not sufficient we need to understand
the dynamic behaviour of the switch.
So, let me now, write down an NPN BJT switch like this. So, an NPN BJT switch
voltage across collector and emitter let me indicate like this, this is collector, the base
and the emitter; the collector current flows in this fashion and there is a base current
here. So, the base current here initiates the triggering of the switch to turn it ON or OFF.
So, base current is the control signal or the trigger signal.
So now, let us divide the time into four parts. So, let me make the division. So, let us say
here first the transistor was OFF, and then there is a portion of time here OFF to ON.
This is switch ON of the transistor, then there is the ON state of the BJT and then there is
an ON to OFF or switch OFF of the BJT back to the OFF state. So, the cycle repeats. So,
this time period is the turn ON time period and this time period is called the turn OFF
time period.
So, in order to understand the switching behaviour of the switch let us try to look a bit
729
more in depth into this these parts turn ON part and this part the turn OFF part in a bit
more detail. So, let me first take the turn ON part and let me zoom it so, I am going to
plot four waveforms on this.
So, first this is called the turn ON time. Now, here it was OFF state so, the switch was
OFF and there was a positive voltage vCE which was being which should supported by
the OFF switch. And there is a zero base current here, at this instant I am initiating a base
current to flow through this device here. So, let us say the base current has a finite rise
time and then moves in this fashion and that is called iB
Now, because of this iB the vC which was here tends to drop like this in this fashion.
Now, this is called vCE; and what happens to the collector? As the vCE is dropping the
collector current is increasing. So, the collector current increases and during the turn ON
time it increases more than what is required and then comes down to whatever is dictated
by the load.
Now, if I multiply instantaneously vCE* iC, I will get here the power instantaneous power
curve during the turn ON period. So, this is multiplication of iC* vCE. The instantaneous
value of iC and vCE keep on multiplying and then plot this curve. So, this looks like this.
Now, with relative to this iB let me make few changes so that you try to get some more
insight.
So, next I will place another similar segment turn ON segment and let us view the same
waveforms. Now, iB I slightly increase, a higher value of iB. If I give a higher value of iB
what happens to the vCE waveform. So, that vCE and how does it evolve? So, you will see
that it goes much more smoothly and quicker towards the ON state and you will see a
slight rise in the collector during the turn ON period and then again it comes back and
settles at the same value which is then later controlled by the external load.
Now, what is iC* vCE looking like during this period here. So, you see that vCE drops
quickly here and then the height is reduced and this is the iC in to vCE the instantaneous
power curve, you will see that it has reduced significantly.
Now, I will make one more change to the turn ON base current, I am going to give a base
current in this fashion I give a kind of a small surge base current during the turn ON and
then bring it back down to the normal operating base current. Now, this turn ON surge
730
base current we will make the vCE to go down much faster, it will go into the ON state
much quicker and then there will be an iC which goes high and then comes back to the
normal operating iC.
What happens to the power curve? So, you see that during the time when iC is large vCE
has already reached the ON state. So, most of the power dissipation will be in this zone.
So, we will see that there is a peak of the power dissipation iC * vCE curve and then goes
down quickly. So, you see that the power dissipation is much reduced in this case.
So, typically we would like to give here the base current with a turn ON surge current
which is higher than the normal operating base current. So, you give a turn ON surge
current and then let it settle down to the normal operating base current you will get the
best switching speeds and not only that the power dissipation is lower.
Now, what happens during turn OFF? So, during turn OFF I am going to focus here on
this portion, you are going from ON state to the OFF state. Now, let me see the same
waveforms how it looks in the turn OFF condition.
Now, I am going to split the turn OFF into two zones here I will explain what that zones
is it will become evident the moment you look at the waveforms. So now, look at the
base and the base current was at this level and then I am going to initiate turn OFF of the
base current. So, let the base current be turned OFF so, base current is turned OFF
actually you are removing the base drive, but the base current here then starts going
negative means there is some base charge which is the removed from the base the trans
base of the transistor.
Remember that the BJT is a semiconductor device and there is some base charge which
is stored in the base capacitor of the transistor and that has to be released and that has to
flow out in the negative direction. So, that is negative base current and then as the base
charge gets removed you will see that the base current finally, dies down to zero.
So, even though you have initiated removal of the base current here you have stopped the
base drive, here it takes some time for the base charge to recombine and this is the area
under this curve is the base charge that has been removed and that takes a finite time.
Now, this base current and the collector current actually stays almost unchanged till this
731
vertical line that I have drawn and then it starts to decrease and then go to 0 into the OFF
state. So, this period where nothing much happens in the collector current or even in the
collector or emitter voltage this period is called the storage period.
So, this calls the storage period because most of the base charge is in the process of
getting removed. So, majority of the base charge will get removed up to this period and
then only later when there is a only a small amount of base charge removed this actually
shows a trend towards switch OFF.
Likewise if I look at the vCE waveform nothing much happens up to the storage period
looks as though the transistor is still continuing to be in the ON state then when majority
of the base charge is removed here you will see that the voltage starts to increase and
then starts to support the forward voltage.
What about the power curve? Even in the power curve here during this period vCE is 0,
not much of power dissipation there. So, if I take iC * vCE here there is only a slight
increase, a slight power dissipation and majority of the power dissipation happens here
when it is passing through the active region. So, this moves in this fashion. So, this is iC
*vCE curve and this is actually the turn OFF loss, instantaneous value of the turn OFF
loss
So, our job is to keep this portion as small as possible by providing turn-ON surge base
current and then allowing it to settle to the nominal base current and during turn OFF we
need to provide path for the base current to reverse so that the base charges of the BJT
are removed. And during first portion of the turn OFF period which we call as ts or the
storage period you will see you will not see much change in iC * vCE because the base
charges are in the process of getting removed and majority will get removed at the end of
storage time.
Further on the remaining time is called the tf or fault time and that is when the current is
actually falling down to 0 and vCE is rising to block the power or voltage and that is when
actually majority of the switching will turn OFF switching loss occurs and we need to
keep that one to a minimum.
So, the nature of the base drive that we have to give look at this is the turn OFF portion
you give a surge base drive for ON and then you will have to have a ON state base
732
current and then while turning OFF we will have to have facility for negative base
current so that there is a possibility of current flowing in the negative direction and
sinking somewhere only then the base charges can be removed. So, if we give a base
drive for the BJT where the currents are of this form of this shape then you will be
actually generating a proper base drive for the BJT. So, keeping this background in mind
let us design some base drives for the BJT.
Now, we can summarise the BJT base drive requirement. So, base drive for the BJT
means base current and how should be a typical base current be for switching ON and
switching OFF the BJT device. So, let us say the base current during the OFF state, the
base current is 0 and during the turn ON portion you give a base current surge and then it
will settle to a lower value which will be the normal ON base drive.
So, this we saw will reduce the power dissipation turn ON power dissipation and then
you keep it continuously on such that the BJT is in the ON state you have to give a value
of iB greater than iC/ hFE-sat or iC/β. And then during the time when you want to turn
OFF you need to make it go negative the base drive circuit should have sinking
capability and then it will settle and go back to 0 in the OFF state. So, this will be the
typical base drive current wave shape that will be required for BJT switching.
Now, let us mark this peak value we will call it as iB+ and this normally ON value will
call it as iB-ON and this negative base current peak value will call it as iB- we will use this
733
symbols labels so that consistently we know what we mean and we can visualize when
we use when we use these iB+, iB-ON and iB- symbols. So, iB-ON is the normal ON state of
the BJT switch which means you will be providing at least greater than iC/ hFE-sat. So,
normally iB-ON we will give 2* iC/ hFE-sat.
So, when you give 2* iC/ hFE-sat it is going to drive the transistor into saturation. So,
you are sure that the transistor the BJT is going to be in saturation. So, this is a thumb
rule that normally designers follow. Then iB+ the turn ON surge that magnitude value
will be 1.5 to 2* iB-ON and iB- the turn OFF negative magnitude is 1.5*iB-ON in magnitude.
Now, these are typical design guidelines there is no formula for this, these are thumb
rules empirical by nature. So, these have worked for many designers all over the world.
So, it is a good starting value to start with these design and then find unit from there on.
So, let us now draw our first base circuit base drive circuit. So, this base drive circuit that
we are going to draw now will be the basic building block of most base drive circuit very
important and very popular. So, it is very simple it consist of the transistor, this is the
transistor BJT that needs to be switched ON and OFF. It consist of just two components
for driving the drive circuit, it has resist one resistor and another resistor like this
connected in this fashion.
So, this BJT is the BJT to be operated and these two are the drive circuits. So, you give a
734
base signal here and the currents are limited by this R1 and R2. So, there is this current iC
which is decided by the external load and here you have to give a iB-ON to turn ON the
BJT to full ON condition.
We will call this one as R1 and this one as R2 and the character of this BJT after your
selected you will know from the datasheet what is hFEsat value or it is also sometimes
called the β, this is V BEsat. So, when the BJT is ON as the saturated condition the base
emitter voltage is called V BEsat it will be around 0.7 volts.
So now, let us design the value of R2. We know the voltage across R2 is V BEsat. So, V
BEsat divided by how much current will one allow through R2, normally a current
equivalent of iB-ON will be allowed through R2. So, this will be V BEsat/ iB-ON. And what
is iB-ON? iB-ON is 2* iC/ hFEsat. And therefore, we can say V BEsat* hFEsat/ (2* iC), will
be the equation that one can use for choosing R2.
You can also choose the power rating of R2 because you know the current that is going
to flow through that iB-ON * root of the duty cycle will give the rms current and thereby
you can find the irms square* r, power dissipation of the resistance.
Now, how to find the value of R1? So, there is a drive signal Vb, let us say coming here
and will name it as Vb. Now, what is the voltage drop across R1? Vb minus V BE-sat.
So, Vb minus V BE-sat is the voltage drop across R divided by R1 will be the current
throw it. What is the current throw it? There is iB-ON flowing through R2 into the base of
the BJT. So, 2*iB-ON will be flowing in R1. So, 2*iB-ON is 2*iC/ hFE-sat. So, (Vb -VBE-
sat)* hFE-sat/(4*iC), this will be the equation to select R1.
735
For this circuit let us see how the dynamic characteristic look like. So, let us plot against
time t, this is the OFF state, ON state you have the turn ON you have the turn OFF and
this is the OFF state. So, during OFF state iB is 0. So, that is the current here is 0 and
during the time when you want to turn it ON the ON state i B-ON; however, there is no
special arrangement in the circuit to provide a turn ON surge, it will just go and stick to
iB-ON right even during turn ON. And during turn OFF because of R2 there is a path for
reverse current to flow here so that the base charges recombine.
So, we have a negative base current flowing in this fashion. So, this value is V BE-sat / R2
current will be flowing through that and here it is i B-ON. So, this is how the drive currents
will look like for a circuit of this type. There is no speed up, but otherwise you have
other aspects of the drive current requirements being met. So, this is a very simple base
drive circuit and it is a building block for many other complex base drive circuits, which
we will see later.
736
Prof. L. Umanand
Lecture - 82
BJT base drive example
We saw the base drive circuit-1 where we had used an NPN BJT. Along similar lines we
can also use a PNP BJT. It is the dual of the NPN circuit. The PNP BJT also is quite
popular in base drive circuits as you will see when we develop other base drive circuits.
Now, here you have this emitter, base and collector and the PNP is a rail driven that is
positive power supply driven in the case of the NPN it was driven with respect to the
ground because emitter was connected to the ground. Here the emitter is connected to the
positive of the power supply. So, the drive is rail driven, but the circuit are exactly similar
between emitter and the base have this resistor that is the positive supply and has its
resistant between the emitter in the base even in the NPN circuit between base and emitter,
you had a resistance.
But only now that resistance is connected to the positive rail and between the base and the
signal source, you will have one more resistor and you have the collector flowing into the
load as before in the case of the NPN of the collector was connected to the load.
737
So, you have this ic, the collector current which is the load current and let the transistor
has the hFEsat some value, ic/ hFEsat will be the just saturation value. You have to provide
iBON in the base drive. So, that it goes deep into saturation, iBON is 2* ic/ hFEsat, so that value
should be given here. R1 and R2 are the two resistances that you need to design and iBON
is what you should give.
So, how do you choose the value of R1 and R2? This is V BEsat like in the NPN circuit. R2
is given by V BEsat by unequal amount of i BON flows through like this. You have iBON
flowing through in this fashion in this direction and Vcc minus this potential divided by
R2 will be the current flowing through this and that current should be 2* iBON will flow
through R1.
Now, let us calculate R2. R2 equals Vcc minus this potential is nothing, but V BEsat. So,
VBEsat divided by i BON will be V BEsat * hFEsat / 2*ic, same like in the case of an NPN drive
circuit.
Now, if you have Vb the signal voltage, so this will go low. When this goes low, then there
will be a current iBON flowing from through R2 and iBON coming in from the base: 2*iBON
flows through R1 and sinks into the signal source. When Vb is high, this potential will
match the potential here and there will not be any flow of current in this direction. So,
when V b is high, the PNB BJT is off and when V b is low, the BJT is on.
So, let us say you have an iBON coming from here and iBON coming from here that will flow
through R1 and that current will be 2*iBON. Now, what is the value of R1? R1 is Vcc minus
V besat minus V b divided by 2*iBON.
So, iBON is nothing, but 2*ic by hFEsat. So, I will say hFEsat divided by 2* 2*ic: 4ic. So, this
will be the equation for finding out R1 in the case of the PNP drive. So, this is exactly the
dual of the NPN drive and many a cases many complex drives are built from these two
drives, NPN PNP complements because they complement each other and many stages can
be achieved by this NPN PNP combination. We will of course have a look at this
combination complimentary drives later.
The simple base drive circuit just two resistance in the base NPN and PNP base drive
circuit are very popular and used in many base drive circuits. Let us see if we can apply
them to our DC-DC converters. Now let us take the simple non isolated buck convertor. It
738
is something like this.
Now, here on the rail side on the positive side let me put a PNP transistor like this. If you
are doing it on the ground side, you can use a NPN transistor. So, let me put this PNP
transistor and let me complete this buck converter circuit by putting this diode, inductor,
capacitance, resistance and then you are taking the output across these two terminal. This
is Vi and this Vo measured in these two fashions.
Now, for the PNP based drive circuit from the emitter to the base let us put a resistor. A
resistor from the emitter to the base and from the base to Vb that is the source from where
the signal is coming we put another resistor. So, this is our simple two resistor base drive
for PNP which is driven from the rail. This is the positive, this is R1 this is R2 and this is
V b.
Now, what to give here for Vb? Let us see how we can give Vb. In fact, to make this a
practical circuit I can recommend that you try implementing this. It is possible you just
take a 15 volts laboratory power supply. This will be your 15 volts input, you can design
for 5 volts output, you can use 2N4033 BJT transistor. This a PNP transistor easily
available, then a diode you can use FR304 diode easily available in common electronic
shops, you can use a 1.5 mH inductor wound on a core you have to design and wind it on
a core, put 100 µfF capacitor and then that is it. You are having almost all the components
and you can design Io again 5 milli-amps to vary between 5 milli amps and 500 milli amps,
739
max is 500 milli amps. So, the resistance here is known.
Now, what do we connect here and then at what frequency do you switch this BJT? So, let
us put that down switching frequency of the BJT 20 kilo Hertz. So, almost everything is
ready except that we need to give the pulse width modulated waveform here at V b.
So we had looked at how to generate pulse width modulation. Pulse width modulated
waveform sometime back and I had also told you that there are commercial IC is available.
One such IC’s TL494 Texas instruments. TL494 is very inexpensive around 10 rupees or
so. You should be able to buy that a new set.
So, let us make the connection for TL494. Let us say pin 12 is Vcc pin and 7 is ground
pin. Let me connect the ground and for the pin 12 ground. Ground means this, these two
are joined together and this is Vcc. So, let us say this is Vcc and let me connect Vcc to Vi.
So, Vi itself will be the power supply for the TL494, 15 volts.
Next, I will connect two important components resistance timing resistance and a timing
capacitance to pin 5 and 6. So, a timing resistance RT and a timing capacitance CT these
are precision components use a metal film resistance here and use a metallised polyester
capacitor for timing and if you want 20 kilo Hertz, you use one typical component value
for resistance is 5.6 kΩ and the CT is 10K pF. There are various other possibilities you can
use from the datasheet. You can look at nomograph and pick up the values of RT and CT.
Next let me give something to vary the duty cycle. So, I will use a potentiometer and two
resistances in this fashion resistance fixed resistance connected to Vc, fixed resistance
connected to ground and in between I am having a potentiometer.
So, this is the variable resistance. It will this centre tap will vary from a particular minimum
resistance to a maximum resistance here and I will have a voltage here in between such
that it is between the sawtooth. If you measure at pin 5, you will get a sawtooth waveforms,
so, that the voltage here should be in between the sawtooth waveform.
So, connected in that fashion this we can say 1K, 5K and 1K this connected to pin 1, pin
2 and 3. Now, 1, 2 and 3 is pins of an error amplifier, pin 1 is the positive input and 2 is
the minus input and 3 is output. So, you can short 2 and 3 to make it in the voltage follower
mode, there is also another amplifier which we can disable.
740
So, let us say 4, this is the duty DTC or the Dead Time Control pin. There is a staring pin
and the other error amplifiers positive, all these let us ground it. We do not need it now,
ground it and the plus of the other amplifier, the minus of other amplifier and the output
of that one we short it just like here in the emitter follower and the voltage follower mode.
Now, that is 15 and 14. Now internally there are two transistors that will get enabled or
disabled depending upon the pulse width modulated waveform that is coming to the base
internally of these internal transistors. So, this is pin 11, 10, 8 and 9. You can use any one
of the transistors.
Now, let us say I use this 8 and 9, where 8 is the collector, 9 is the emitter. Ground the
emitter and the collector let me connect it here. You have PWM generator IC where the
duty cycle can be varied and it is controlled and at this collector you will get pulse width
modulated waveform which is given as get base drive signal to this PNP BJT which will
switch this buck converter and you will get the required output. Now this is the open loop
control of buck converter and is and it can operate as designed. R1 and R2 you know how
to design it. Just we just saw while designing this PNP base drive.
So, how does it operate internally? When this internal transistor turns on, you will see that
this will be shorting the collector will be connected to the emitter and the Vb will be at
ground potential ideally. So, there is a path for iBON to flow here and iBON to flow here, so
iBON flows through R1 into Vb and into the ground and complete the circuit.
Now, let us say the internal enabling pulse is removed for the internal transistor. This is
open. When this is open, pin 8 is floating. So, there is no chance of current flow, current
iBON cannot flow, BJT is off. So, in this way BJT can be turned on and off by controlling
the internal enabling pulse to the internal transistor of the TL494.
But our focus here was just demonstrate to you how the base drive works. This is the
simple base drive to resistance base drive and a PNP. If you put if you use the ground side
switching you can use a NPN transistor and to resistor base drive. However, as you start
going higher in power you will need more complex base drives and that is what we will
look at further.
741
Prof. L. Umanand
Lecture - 83
Multi-stage base drive
The two resistor NPN and two resistor PNP drive I am going to rewrite in this fashion
you have the NPN and BJT. And I am going to put the resistors vertical like this as a
potential divider or an attenuator and take it in this fashion, I am calling this one as Vb
this is R 1 this is R2, Q having an hFEsat and there is an ic load current flowing through
it and this external load.
Now, this is a 2 resistor NPN drive which we discussed, only instead of putting R1
horizontal, I have just put them all vertical. You will understand the reason why I am
putting it vertical, because it is much easier to form more complex base drives in this
fashion.
The other dual is the PNP. So, the PNP where it is rail driven just like you are having
ground as the common point, in the case that PNP you have the rail or the V cc as the
common point. So, you have the base emitter resistor and the drive resistor here. So, this
is the load Vcc, Vb, R1, and R2. And you have the Q with some hFEsat and you have
collector current or the load current flowing in this fashion.
742
So, this is also two resistor, but PNP base drive. So, in both these drive circuits it is only
single stage you just have two resistors. So, if you have ic the current that should flow
here, would be iBon which is 2* ic / hFEsat; and the current that is flowing from R1 is
2*iBon, which is 4* ic / hFEsat is the current that is flowing
So, likewise here also the current that is flowing or sinking into the Vb port will be 4*
ic/hFEsat. Now if hFE is low, let us say as the power level increases the hFEsat value
decreases that is you may find the hFE at 10 amp level to be around 3 or 4. So, ic 10
amps divided by 3 would be around 3.3 amps on the base side. So, 2* ic/hFEsat would
be 6.6 amps and 4* ic/hFEsat would be 12 amps flowing through that.
So, one stage will not give you much of a reduction in the current. So, you may have to
put multiple stages because, 12 amps you cannot source from a PWM IC or from a
microcontroller. So, you will have to reduce the current to a level, up to let us say micro
amps whatever the sourcing or the sinking capability of the signalling source is here,
based on that only you can directly connect it to the signalling source only if the current
is made compatible.
So, therefore, if I want to make it still lower now let us say I have an 10 amps flowing
through that, this stage is giving you hFEsat of let us say 5. So, here it will be 2 amps,
2*2 amps, which is 4 amps and what will flow here, 4 and 4+ 4, 8 amps. 8 amps has to
be sink into Vb and in this case 8 amps should be source from V b.
So, let us say this becomes the load for the next stage. So, further I will put another stage
which has let us say hFE of 10 and then you will get 0.8, 2*0.8 is1.6, 1.6+1.6 = 3.2
amps. Then further another stage from 3.2 amps will keep decreasing till you go to
milliamps and micro amps, where the PWM ICs or microcontrollers will be able to
source or sink. So, multiple stages you may have to put to achieve particular level of
current capability so that it becomes compatible with the signalling source.
So, for this we need to concatenate the BJT drives and this would be the basic drive
alternately we will complement NPN - PNP, NPN - PNP and make multi stage drives.
So, that is what I would like to show multi stage complimentary drives.
743
Let us now design a complementary multi stage drive or it basically means is that we
will use NPN and PNP stages, alternately and make a multi stage drive. So, let us say we
have this basic drive, you have an NPN BJT and this is the power BJT, the final stage
which is supposed to handle the load external load and ic current flows through this, and
to drive this we have this two register drive like this.
Now, let us say that the current drawn here is a large and it cannot be supplied by this Vb
source. So, we need to reduce the current further by putting multiple stages. So, let us do
that. So, let me make some space and at this point instead of Vb we will put one more
stage, now I am taking the PNP stage here. So, I am taking the PNP stage and attaching it
here, so this PNP stage the collector current for the PNP stage, will become the drive
current for the next stage. So, this collector current will be much lesser than this because
of the hFEsat division.
Now, if you say that the current here is still further is still not low enough for the Vb
source to handle and you can add one more stage and reduce it still further, let us do that.
Now let me make this connection, then at this point I will remove and now at this point
you attach an NPN stage with a collector connected here, so that the collector current
here will become the drive current for this.
So, let me place it there in that fashion. So, this is an NPN stage and this stage here you
can give the drive signal input. So like that you can keep building the stages. Now, this is
744
a three stage drive so complimentary drive because alternately you have an NPN stage
you have the PNP stage and you have the NPN stage like that.
So, you can number the components in this fashion say we call this Vb. So, this is one
drive circuit if you want you can take an example to see how those look like, let us say
that we would like to drive it with a pulse like this. Now whenever this pulse is low,
whenever Vb is low what happens, when Vb is low there is no drive current through R5
and therefore, there is no iBon to this transistor. So, if there is no iBon to this transistor this
transistor is off; if this transistor is off R3 current cannot flow, which means there is no
current iBon flow through this branch. And therefore, this transistor is not having iBon
therefore, this transistor is off, if this transistor is off, then there is no drive current here
and therefore, this transistor is off.
So, whenever Vb is low this transistor is off, whenever Vb is high there is a drive current
through R5, there is a current through the base of this transistor so this transistor is on
and there is a drive current flowing through R3 right down to ground. And if R3 drive
current is in this direction, PNP will conduct and it is on. If PNP conduct there is a drive
current flowing in this direction through this base. And therefore, Q is on and a collector
current and therefore, the load current will flow.
So, in this way it will operate; so if you make the Vb signal low and high you will be
able to switch the final transistor here on and off. So, if Vb is low this is off, Vb is high
the final transistor is on.
Now, let us say for example, the hFEsat of this final power transistor is 10 and the
hFEsat of this transistor here PNP transistor is 100; and the signal NPN transistor is
another 100. So, let us just go around calculating some numbers to get a feel. Now, let us
say the max load current that needs to be handled is 10 amps.
So, if it is 10 amps what should be the iBon base current here, base current here should be
iBon value which is 2*ic/ hFEsat, ic/ hFEsat is 1, 2*ic so this is 2 amps. And iBon current
should also flow through R2. So, here R1 should handle 4 amps. So, 4 amps now
becomes the collector current for this PNP transistor. So, if 4 amps supposed to be the
collector current for the PNP transistor 4 by 100, 40 milli amps is the base current at
critical saturation turn on condition.
745
But if you allow ib arm value which is 2* ic/ hFEsat, it will be 80 milliamps. 80
milliamps flows in this direction, 80 milliamps will flow through R4 and 160 milliamps
flows through R3 and that becomes the collector current for this transistor.
So, 160 milliamps by 100 is 1.6, 2 times that will be 3.2 milliamps which is iBon for this
transistor, 3.2 milliamps supposed to flow here also. Then the drive current here in R5
will be 6.4 milliamps. So, here in this circuit for this particular set of numbers if you
have a source Vb having this kind of a pulsed nature and is capable of sourcing 6.4 milli-
amp. Then at least greater than 6.4 milli-amp. Let us say the sourcing capability of Vb is
10 milli-amp then this circuit will work nicely it will be able to source its 6.4 milli-amp
and then operate this final power transistor at 10 amp.
So, this is how the drive circuit behaves. In fact, you should always design from the load
side and start working back towards the low signal side. So, you will see the current
values decreasing gradually as it goes stage by stage to the lower power side then you
know that you are on the right track and not making mistakes.
Let us discuss another variant of the multistage drive, now with an inductive turn off
improvement enhancement. So, essentially the multistage drive that we discussed similar
circuit we will use, but with a modification in just one component. So, I am going to
introduce an inductor here, apart from a resistance here an inductor is also added to the
main power transistor base. So, this is the PNP so which means the PNP drive stage
746
followed by an NPN drive stage. So, as before we have 2 resistor drive circuits like this.
So, each transistor is having a two resistor drive, this is another two resistor drive apart
from the two resistor you also have a inductor.
So, now let us say Vb which is having a pulsed waveform like this, this is actually duty
cycle control. And the final stage where the power is actually drawn to load so that a
load current flows through the external load. So, this is the external load and you have
the load current flowing as collector current here ic, this is Qp, Q1, Q2 and that is Vcc.
Now, you see the operation exactly similar to the multistage complementary drive
circuit, when Vb is low 0, there is no drive current flowing here Q2 is off and there is no
drive current flowing here. Q1 is off and there is no drive current flowing here, no
current through this RL branch, no base current Qp is off so ic is 0.
Now, when Vb goes high, during this period, you will see that there is a drive current iBon
flowing through the base of Q2, it is on. Once Q 2 is on there is 2*iBon flowing through
this, iBon through the the base of Q1 and Q1 is on, there is a drive current flowing through
in this direction. So, there is a current flowing through this resistance and inductance
branch is a base drive, Qp is on and collector current ic will flow. So, this is the basic
operation of this circuit, let me mark this current iL to indicate it is a current which is
flowing through L will call that one as R1 this is R2.
So, now let us say what should be the value of L, we know how to find the value of R1
and R2 and in the steady state the voltage across L is 0. And therefore, the entire V bsat
is coming across R2 and V bsat by iBon will be the current that flows through R2.
Now, from the faradays law L*dib/dt is equal to the voltage in this loop. So, there is a
Vbsat drop across the base emitter of Qp positive here, negative here and therefore, vL
will be R2* ib-. So, when you are turning off at that instant inductor is not going to allow
a instantaneous drop in the current therefore, there will be a change in the polarity of the
voltage across the inductor and it will drive a current in this direction as I am showing in
the cursor. So, that it will enhance, it will speed up the base charge recombination and
turn off will be faster.
So, during the time when you are turning it off here Q1 will go off and this drive current
will be cut off, but this drive current will not be immediately cut off the inductor current
747
will see to it that, there will be a freewheeling that needs to happen here. And this will
recombine the base charges and then quickly go to switch off condition. So, during the
time of turn off the current here will call that one ib- and there is this drop and then there
is this drop opposing minus V bsat. Ib- we know is nothing, but 1.5* iBon and we know
during turn on of Qp, iBon is nothing, but 2*ic/ hFEsat. So, you have 1.5*ic/ hFEsat
minus Vbesat.
So, therefore, L is 3* 2* ic/ hFEsat minus Vbesat and this whole thing divided by diL/ dt
will give you the value of L. Now what should be diL/ dt, it is a designer choice and we
normally have some empirical rules for good design 0.15, 15 percent of max ic amps/µs
will be this value. And this is for high voltage transistors 700 volts or greater than 700
volts because, it takes longer for the voltage to transit from a larger voltage to 0 and 0 to
the V cc or whatever the rating of this transistor voltage rating of this transistor is in that
particular circuit.
For low voltage circuits applications we can use 0.5 ic amps/µs for less than 200 volts
transition. Because it is just lower voltage transition, you can afford to going for much
quicker recombination and turn off. So, you should also remember that higher the faster
the switch off higher will be the dv by dt of the power semiconductor switch and that
particular power semiconductor switch should be capable of withstanding the dv by dt
which means that in the datasheet you will have to look at the dv by dt rating also. This
0.15 ic amps/µs and 0.5 ic amps/µs are reasonable starting values and then after that you
can fine tune the design to get better values of L and turnoff values.
748
In literature you may find slight variation to the circuit in this fashion, let me remove the
inductor and then R2 is connected to a minus power supply. So, this multistage drive
circuit, you will see in some drive circuits this R2 immediately after the first stage the R2
that is connected from base to emitter, sometimes the other end of the R2 resistor is
connected to minus power supply.
So, what changes in the calculation of R2 is this? Now there is a V bsat and there is a
minus Vee divided by ib- for turn off. So, that is the current that will flow through this
and because of the existence of this minus power supply it is able to pump the necessary
reverse base drive to turn off this BJT quickly, so therefore, the turn off is faster.
So, if you continue with this ib- which is 1.5* iBon you will see V bsat plus Vee divided
by 3 * ic /hFEsat. So, I put hFEsat on the top here numerator. So, this is the value of R2
so, this is the modification that will happen here so that the turnoff of Qp is speeded up.
749
Prof. L. Umanand
Lecture - 84
Base drive with speed-up circuit
Let me now discuss another interesting Base drive circuit with speed-up; speed-up of
both turn on and turn off. Let me start with a PNP transistor, the PNP transistor the
collector is connected to two resistance divider like this. And from the centre point of the
resistance division I will pick of a line and then connected to the base of an NPN
transistor like this.
The emitter of the NPN transistor on the collector of a PNP transistor are connected in
this fashion, collector is grounded and the same base is given to this PNP transistor also
and the collector of the NPN is connected to the Vcc. So, you will see that across this
Vcc you have an NPN, PNP transistor combination connected in this fashion is called the
Totem pole combination.
Now, when the NPN transistor is on because the base is more positive with respect to the
emitter the base of the PNP is also connected to the same point. So, the base of the PNP
is also more positive with respect to the emitter and therefore, the PNP transistor is off.
So, when the NPN transistor is off and the PNP transistor is on, the emitter is more
750
positive as compared to the base and the PNP transistor is on and because the emitter is
more positive than the base, NPN transistor is off. So, the on and off of this NPN and
PNP complements are always mutually exclusive. So, there will never be a direct short
from Vcc to ground.
Going further let me mark that a Vcc and the base drive for this PNP we have already
discussed, it is 2 resistance combination and at this point here you can give Vb, the base
drive from a base drive source. Now, from this centre point of this totem pole I will pass
it through one network R and C combination like this in series. And another in parallel
resistor like this and connect it to the base of the final transistor or the power transistor.
Final power transistor is connected to external load and ic of that is what is supposed to
be switched on and off that we need to control, this is Qp. Now let this; let me call this as
1 this is R2, C this is Q1, this is Q2, this is Q 3, R 3, R 4, R 5, R 6 and Vcc. This portion
of the circuit R1, R2, C is the portion of the circuit that we will be calling as the speed up
circuit or the speed up network.
So, in this speed up network this branch R2, C branch is the first dynamic branch R2*C
is having a time constant of the order of turn on times or the turn off times of the Qp
BJT. So, during transition, during turn on and turn off R2 C branch is very active and
during the steady on time R1 is the active branch.
So, let us see how this operates. Consider now this point here is the pole voltage Vp.
Now, consider the state when Q1 is off. So, there is no current flowing through in this
direction, So Q p is in off state.
Now, at this stage let us switch on Q1; how do we switch on Q1? Q3 is switched on.
How do you switch on Q3? Make this low, so when you make this low there is a
possibility for current flow like this. So, you have iBON flowing like this then you have
Q3 on. When Q3 is on there is a drive current flowing through 3 into this branch here
and it is flowing in this direction Q1 is on; Q1 is on Q2 has to be off. So, when Q1 is on
there is a drive current that can flow through into the speed up network.
So, how does the drive current flow? Let us say a drive current flows like this and
through this R2, C network in this fashion because initially when the Q1 is turned on the
capacitor C acts like a short R2 is much smaller compared to R1, major portion of the
751
current flows through like that in this fashion through the R2, C branch.
There is also a parallel division current flows through R1 branch also, but the current
through R1 branch is lesser than the current that is flowing through R2 branch. Then as
time progresses this exponentially decays; exponentially decay in the R2, C branch
because C gets charged up plus minus and when C gets charged up ±2Vcc the current
through this branch C is only the current through this branch that flows, and this should
be the steady on state current.
So, let us say plus minus, the C charges in this fashion and let us draw the time
waveform and see how it looks. Let us say this is iB current flowing in through Qp, how
does this waveform look like, will plot iB. So, iB was 0 during the time Q1 off and Q1
was turned on there is a surge here. So, that surge is iBON which is flowing through R1
plus there is as surge due to R2, C branch. So, let us call that one as i bplus.
So, from i bplus there is a steady decrease because C is getting charged plus and minus
and gradually this branch will not conduct current and only this branch will conduct
current. So, whatever the steady on current is flowing through R1 will be iBon and then
again during turn off, now Q1 is turned off, this Vb is high here. Q3 will turn off and Q1
will turn off and because of that Q2 here will now turn on because C here is having
positive charge, this is positive and the base of Q2 the PNP is now connected through R4
to ground which is at the lower potential.
So, there is a possibility of current flow through like this and current flow through like
this. So, you will see that Q2 will turn on and there will be a current flow through Q2.
So, we will call this I bplus and this is iBon and then during the time Q1 is being switched
off.
752
We will see that charts current starts flowing in this fashion down through like this surge
flowing like this and through in this direction. So, that the base charges start recombining
and comes back and completes the circuit here. So, this is the negative current, that needs
to flow through Qp. So, that Qp turns off as fast as possible and there is base charge
recombination.
So, this point here we will call that an as i bminus. So, this current which will flow
through starts with i bminus and gradually exponentially decay the charge on this
capacitor. So, as the capacitor charge start decreasing the current here starts decreasing in
an exponential manner. So, the current will decrease in an exponential manner to 0.
So, once this decreases to 0 all the base charge has been recombined and Qp has regained
its off state. So, here we know how to design R3, R4, R5, R6 and the two resistance
drive, the new part is this network. So, let us calculate R1; R1 is supposed to handle
steady iBon current. So, let us say during the steady iBon portion the potential on this side
of the R1, Vp is Vcc minus VCEsatQ1 that is this point. So, we will write that down Vcc
minus V cEsatQ1 and on that side you are having V BEsatQp, the whole divided by iBon
current this will give you the value of R1, R2 needs to be much smaller than R1. So, you
use a constraint let it be one-fifth of 1/5 th of R1.
In order to calculate C we know the time we should start with the time constant R2*C: τ
which is the time constant, now how do we choose that τ. So, the turn off time let us say
753
is comprising of storage time plus fault time, turn on time is comprising our ic time, look
into the datasheet of the Qp BJT and choose τ comparable to the turn on or turn off times
of the BJT. So, let us say you have ts storage time and you have tf fall time, ts plus toff,
tf fall can be fixed as the time constant and R2 is known, C you can calculate in this
manner.
So, by making τ (the time constant) smaller and smaller the speed up can be made much
faster. So, this is the base drive which has both turn on speed up and turn off speed up
and it is an interesting base drive circuit and this effects only that portion of that network
base drive network just immediately before the main power transistor. Then following
that you have this totem pole following that all the sections and if you have still further
more multistage sections all will be just like as we discussed before 2 resistor NPN PNP
combinations. You have a 2 resistor NPN drive and 2 resistor PNP drive alternately
coming into the picture to form a multistage drive with speed up.
754
Prof. L. Umanand
Lecture – 85
Base drive with isolation
Let us now discuss a base drive circuit with transformer isolation. Let me begin by drawing
an NPN transistor like this and I will drive this by a 2 resistor drive circuit. We are familiar
with this kind of a 2 resistor drive circuit. We know this we know how it operates. In the
collector of this NPN transistor I am going to place the primary of a coil transformer coil
in this fashion call that as Vcc.
On the same core the secondary coil also placed and look at the dot polarity whenever this
is positive the dot on the secondary side is also positive, looks similar to the forward
converter configuration. Now on the secondary side I will place a diode like this so, that I
take only the positive signal and use that for driving the main power transistor.
So, let me draw the main power transistor which is supposed to be switched ON and OFF
and this is the main NPN power transistor, the collector is connected to an external load
and the collector current there is what is supposed to be switched ON and OFF. So, we
connect it like this and the drive for this is a 2 resistor drive, I will put a 2 resistor drive in
this fashion.
755
So, this is the 2 resistor drive driving this one. We can now connect on the primary side a
freewheeling circuit for freewheeling the magnetising current of the primary. Here you
will give Vb and Vb will be the pulse wave form which will go high and low. So, when
Vb is low what happens, when Vb is low you will see that Q1 will be turned OFF.
I will call this one as Qp and this is ic which is supposed to be controlled, name the parts
R1, R2, D1 this is Rf and Df, R3 and R4. So, when Vb is low there is no drive current
flowing through R3 therefore, no base current for Q1 and Q1 is OFF.
When Q1 is OFF the voltage across the primary the dot is negative; the dot is negative the
diode is going to be reversed biased OFF there is no current flowing through R1, no base
current to Qp; Qp is OFF. Now when Vb goes positive or high you have a drive current
flowing through R3 base current for Q1; Q1 is ON. Now when Q1 is ON there is a current
flow through the primary, dot end is positive; dot end is positive, diode D1 conducts.
So, there is a drive current through R1 flows into base and turns ON Qp then the ON state
current flows. So, this way the Qp can be switched ON and OFF based on this signal Vb
through an isolation barrier. You see that the power side of the circuit does not have any
physical connection to the control side of the circuit.
So, the control side of the circuit is isolated and the power side circuit is isolated. So, there
is some degree of protection for any spikes and surges coming on the power side where
we will not get, it will not get transferred on to the control side and protect the control side
circuits. So, this is a transformer isolated base drive circuit. The operation of this
transformer portion using Q1 as a switch is very similar to that that of a lossy forward
converter.
So, by that same argument you can convert this lossy forward converter into a loss less
converter, loss less based drive circuit by putting one more coil 1:1 with dot polarity
inverter like in a lossless forward converter and put a diode here and then put back the
magnetising energy into the supply can do that one however, that I will leave it to you for
doing that and understanding how that works.
756
Let me now discuss another based drive isolated based drive, but instead of transformer
isolation I am going to discuss optocoupler isolation or the light isolation. So, let me start
by drawing now here this is the power transistor Qp and that is connected to the external
load the collector current of which we need to switch ON and OFF.
The drive for this Qp is coming from a 2 resistor combination we know how this works
and in order to reduce the drive current from ultimately the Vb source, let us put in multiple
stages. So, there is a PNP stage here and the PNP stage we know how to drive that by 2
resistor combination in this fashion. This is connected to another NPN stage, the NPN
stage I am going to drive by another 2 resistor combination.
And this is being driven by a PNP stage and the PNP stage is driven by another 2 resistor
combination, like this you can keep going. Now let us say at this point normally we would
collect connect Vb that is the base drive signal, but we need isolation. So, in order to
provide isolation I am going to connect another device which is actually a BJT, but driven
not by current, but by light generated from a photo diode.
So, this whole photo diode combined with this transistor this whole device is called an
optocoupler. So, that transistor which is the optocoupler transistor I am going to connect
here and ground this. So, this will be the drive transistor for this base drive. However, the
drive for this is coming from light and the light source being obtained whenever there is a
current flow through the diode.
757
So, the diode is now connected to another ground another circuit ground totally physically
different from the power side ground. So, there is a resistance here and Vb there is a pulse
in this fashion. When Vb is low there is no current flowing through this resistance and the
diode and as there is no current flowing through the diode there is no light source and
therefore, the transistor of the optocoupler is OFF. When Vb is high there is a current flow
through this diode and the transistor of the optocoupler is on. So, that is how this
optocoupler operates.
Let us first name the parts. So, here is ic, this is the current that you need to control switch
ON and OFF, this is Vcc let us call this Qp power transistor to be switched ON and OFF.
Q1 this PNP transistor, Q2 NPN transistor, Q3 another PNP transistor, Q4 optocoupler
NPN transistor and this is the optocoupler.
So, let us see the operation, let us say Vb is low, there is no current flowing through the
diode; the diode does not emit any photons the Q4 is OFF so there is no drive here, there
is no base current for Q3; Q3 is OFF and there is no drive here, there is no base for Q2.
Therefore, Q2 is off; Q2 is OFF there is no drive here, no base current for Q1; Q1 is OFF
so there is no base drive for Qp; Qp is OFF, ic is 0. Now when Vb is high, here is a current
flow in this circuit limited by this resistor.
So, the diode once there is a current flow emits photons and the Q4 is turned ON; once Q4
is ON there is a current flow through this resistor. So, there will be iBON flow, Q3 will be
ON when Q3 is ON there is a drive current flowing through this path into the base of Q2;
Q2 is ON.
When Q2 is ON there is a drive current flowing from the base of Q1 into this and then
there is current flowing through in this fashion Q1 is ON. When Q1 is ON you have a drive
current flowing into the base of Qp and Qp is ON and ic flows. So, in this way Qp can be
turned ON and OFF by this base drive signal.
See the number of stages. So, the current here will be iBON : 2*ic/ hFEsat and then you
have here 4*ic/ hFEsat then divided by the hFEof Q1 will come in here and divided by the
hFE of Q2 will come in here divided by hFE of Q3. So, this will be a very small current
and optocouplers normally available or generally handling currents in the order of
milliamps.
758
So, you reduce the drive current to low value by this multiple stages interpose this
optocoupler; the optocoupler gives at least 3000 volts isolation between this part of the
circuit and this is a very good isolating base drive. Now, note this different grounds here
how do you give the power supply for the power portion how do you give the power supply
for the base portion?
Recall that we studied during the rectifier, how to generate multiple isolated power
supplies? So, just let me show that you have 230 volts ac and let me say you have a step
down transformer and you have 2 secondaries. So, in one of the secondaries let us say have
this rectifier diode we know how this circuit operates put the capacitor filter. To the output
of the capacitor filter connect a linear regulator and the output of the linear regulator you
will get a regulated supply.
Now, this let us mark this is a Vcc and let us say this is the power ground. So, this portion
of the winding you can use it for the power side here across Vcc and this ground. So, that
will be the power side supply. Now the other winding, so, you can pass it through a similar
rectifier just like I have shown here bridge rectifier, like a block.
The output of that bridge rectifier you can have a capacitor filter like this and the output
of the capacitor filter can be passed through a 3 terminal regulator and that regulator output
can be used for the control side that is with this power supply in ground, look at this ground
759
symbol. So, I can use that ground symbol another power supply which can be given to the
control side.
So, this portion I will give it to the power side and this supply I will give it to the control
side. So, you see that this power supplies themselves are isolated there is no physical
contact with 230, there is no physical contact between the power side and the control side
in terms of the power supply because of the optocoupler.
So, there is a at least 3000 volt isolation between the power side and the control side here
and there is also around 3000 volt plus isolation between the 2 secondaries. So, the power
side and the control side are totally isolated.
As compared to the transformer isolated base drive, transformer isolated base drive very
simple there is no need for secondary side power supply, it is just like a forward converter
you have 1 power supply here and then the secondary side the power transformer is of a
power transfer is occurring through the transformer. So, the base drive power is coming
right through that transformer and turning the switch ON and OFF.
Whereas, in the case of the opto isolated base drive, you need a power supply on the
secondary side of the optocoupler and therefore, you need to have an arrangement like this
where you are having dual power supply, one power supply isolated given to the secondary
side of the opto coupler and another power supply given to the control portion of the
optocoupler. Only then the galvanic isolation is achieved and isolated base drive will work
with sufficient voltage isolation between the power and the control portion.
760
Prof. L. Umanand
Lecture – 86
MOSFET gate drive
MOSFET is a very popular semiconductor switch, especially for DC-DC converter we

will discuss its drive. Let me draw the MOSFET symbol. So, this is called the gate and
this is called the drain, the substrate and the source are connected. So, you have the drain
source gate, there is a natural internal body diode, the arrow here pointing to the n
channel inversion layer that will get formed. So, this is an n channel MOSFET.
The MOSFET is a high input impedance device meaning it takes a very very low gate
current unlike the case of the BJT. So, the gate power requirement for the MOSFET is
very low, it is more a voltage control device as compared to the BJT which is a current
control device. So, it is important that we give an amount of charge for the MOSFET as
required by the datasheet of the MOSFET and if you give that amount of charge to the
gate source then that will turn ON the device.
So, you can think of the current now the iD which is flowing through the drain and then
there is internally an equivalent capacitance gate-source let us say and then you need to
761
put a charge Qg into the gate source so that the device turns ON. So, imagine it in that
fashion. So, let us look at the drive requirement.
So, let us say this is the time I will split it in 2 time zones, this is the OFF state of the
MOSFET, this is the ON state of the MOSFET again the OFF state of the MOSFET
there is an OFF to ON transition this called the turn ON period, there is a ON to OFF
transition this is the turn OFF or the switch OFF period. So, during this time let us look
at how the gate current ig looks like.
So, as I said this is a charge control device or a voltage control device it is actually this
charge that is important. So, therefore, we will see how the gate current looks like based
on that argument of giving a specific amount of charge to turn ON the MOSFET.
So, during the OFF state no gate current is needed, so, that is 0. During the ON state also
no gate current is needed in an ideal sense you need 0 gate current to maintain the switch
in the ON state and therefore, also finally, in the OFF state 0 gate current. So, you see the
gate current requirement how it looks like almost all 0, 0 and 0. Now only difference
comes in the turn ON period and the turn OFF period.
So, during the turn ON period you want to switch ON the MOSFET, you supply a
specific amount of charge called Qg as defined in the data sheet for that specific
MOSFET. After supply that charge the MOSFET will turn ON. So, let us say I will
supply some a current with some profile like this, the area under that curve is the charge
Qg.
Now if I supply this Qg amount of charge to this MOSFET then the MOSFET will turn
ON the voltage across the MOSFET will be 0 and there is no charge or no current needed
to maintain the MOSFET in the ON state. So, once turned ON it will remain ON that is
why we say it is a high input impedance device.
Till you want to turn it OFF. When you want to turn it OFF all you have to do is remove
the charge that you have given to it while turning ON. So, whatever charge you are given
remove that charge and you would have switched OFF the MOSFET. So, this is how the
requirement for the gate current will look like. Now as it is a charge control device or a
voltage control device, you are actually not so much interested in the shape of the current
as long as I am able to give the required gate charge.
762
So, you can also give the gate currents in this fashion, a pulse gate current like this the
area under that should be Qg. So, as long as I am giving Qg amount of charge the
MOSFET will turn ON or instead of that I could give an RC an exponential decay like
this as long as the area under this is Qg during on time. Then remove that amount of
charge that you had supplied to turn it ON.
So, this is also fine as long as you are giving Qg amount of charge to turn it ON and
remove Qg amount of charge to switch it OFF this is also valid. In fact, this is the type of
current wave shape that we will see mostly because it is much easier to put an RC drive
circuit rather than trying to achieve constant current in a constant current pulse drive
circuit. So, this is typically the behaviour of the MOSFET, the dynamic behaviour of the
MOSFET.
You need to supply the MOSFET at the gate to source some fixed amount of charge Qg
which will turn ON the device. If I am supplying this Qg in shorter time interval then it
will turn ON faster, if I am supplying Qg over a longer time interval then the turn ON
will be slower, likewise for turn OFF also.
Let us now see how the basic or the fundamental gate drive for the MOSFET looks like.
It is very simple it consists of only one component that is one resistor. So, let us start
from the requirements the gate drive requirements, I have this timeline I have split the
timeline in to the turn ON period the ON state, turn OFF period OFF state.
763
So, we have just seen how the gate current looks like gate current is 0 during the OFF
state and let us say it raises and there is an exponential decay where it pumps in a
positive gate charge Qg in to the MOSFET gate, turns ON the device and then after that
you do not need any current to maintain the device in the ON state. Then when you want
to turn it OFF you have to remove the gate charge and then bring the device to the OFF
state.
So, let us say this is ig+ and this is ig- and the average of this in the turn ON period we
can call that one as ig average and if I take a straight line approximation it is
approximately equal to ig+ / 2. Let me call this period as t-ON, let me call this period as
t-off. So, now, let us draw the gate drive circuit. So, you have the Vg the gate pulse being
given in this point then pass it through a resistor and connect it to the gate of the device.
The drain of the device is connected to the external load and it is the drain current that
you want to switch ON and OFF and the source you connect it to the ground. There is an
internal equivalent capacitance which is going to absorb the charge that you are going to
pump in, we will call this Vg and this is R1. So, how to calculate R1? Very simple
during the time when you want to turn it ON, at the time of turn ON dynamics capacitor
is a short.
So, the current that is going to flow through will be the full Vcc/R1 or Vg/ ig+. So,
whatever this ig+ is decided then Vg/ ig+ is the value of R1 and Vg normally is the pulse
amplitude which is Vcc of the control gate drive. Now what is ig+?
Now look at this drive requirement, you take this average the charge in this period to turn
ON the device is Qg obtained from the datasheet and that is equal to ig average* t-ON. t-
ON is the designer requirement how sooner how fast you want it to turn ON and during
that period I should have given the complete Qg to the MOSFET.
So, ig average * t-ON will be Qg and ig average is approximately ig+/ 2 and therefore,
ig+ is 2*Qg/ t-ON. So, this can be the value of ig+, you plug it in here and then you can
get the value of R1, Vg you substitute Vcc control power supply voltage. So, if you
actually give a pulse like this, you can expect the current here gate current here to flow
something like this, like a differentiated value of this square pulse.
764
So, whenever you have a discontinuity here going from low to high you will get a
positive current spike due to this RC. And this will decay as the capacitance here charges
up to plus minus and then when you want to remove it bring this to ground, here you are
making it 0 and the Vg should have sinking capability there should be current being
capable of flowing in a negative direction. So, whatever charge is there in the
capacitance, internal capacitance will discharge through R1 and the external Vg circuit.
So, therefore, you get this negative pulse and charge removal.
So, this way the MOSFET will be turned ON and OFF. Remember you need some
charge to turn ON the MOSFET, you do not need charge to maintain the MOSFET in the
ON state, you do not need charge to maintain the MOSFET in the OFF state. You need
only current for turning ON and you need current to flow in the opposite direction for
turning OFF.
And it is a charge control device, it is very very important to see that the area under this
curve is equal to Qg as declared by the datasheet for that MOSFET; so, that MOSFET
can turn ON. So, this is simple basic building block of gate drive for the MOSFET and
all complex MOSFET gate drives are built using just this. So, this is a 1 resistor gate
drive. Just like in the case of the BJT you had a 2 resistor gate drive as the basic gate
drive in the case of the MOSFET a 1 resistor gate drive is the basic drive for the
MOSFET.
765
Let us discuss another MOSFET drive gate drive with buffers CMOS buffers, this can
improve the turn ON speeds and turn OFF speeds. Let us see the circuit I have the
MOSFET here the drain is connected to the external circuit and the drain current is what
you need to control switch ON and OFF. The source as usual is connected to the ground
is a ground based gate drive is an internal body diode and the gate is connected to the
resistor the 1 resistor gate drive.
And now here you apply the gate drive pulse and the normal one drive, 1 resistor gate
drive. So, when you apply the gate drive pulse you have the charge being pumped into
the gate and that MOSFET turns ON. But the turn ON time is decided by the amount of
ig+ that can be drawn at that given instant of turn ON and that is dependent on this
sourcing capability of the source.
If the sourcing capability of the source is low then even if you reduce this resistance the
charge that will be pumped in to the gate will be over a period of time and therefore, turn
ON time may be longer the turn ON may be slower.
In order to speed this up we can use a buffer, I could use a non inverting buffer like this.
Typical non inverting buffers like the CMOS buffers, CD4050 this a typical non
inverting buffer which can sink 5 milliamps and source slightly more than 1 milliamp or
you could use an inverting buffer like this CD4049 and apply Vg at the input.
So, when you apply an inverting Vg pulse, you will see that when the pulse goes low the
buffer output goes high and then pumps in the gate charge and turns ON the MOSFET.
Now in order to improve the speed of turn ON you can put more buffers in parallel, if I
erase there and then make some space. Let us say I put one more buffer here and then let
us say one more buffer here connect them in parallel in this fashion.
In fact, I can put many more buffers in parallel. So, normally the CD4050 and 4049
series come as hex buffers. So, there will be 6 buffers in a pack and then you can connect
all 6 of them in parallel and then you can effectively get close to 10 milliamp sourcing
and then of you can get around 30 milliamp sinking capability.
So, the idea here is that now we enhance the sourcing capability of the input. So,
therefore, by decreasing R1 you will be increasing the ig+ value and therefore, more gate
charge can be given in a given time or could say the same amount of gate charge can be
766
given in a lesser amount of time and therefore, t-ON can be reduced. This is evident by
the equation that we saw here ig+ is 2*Qg/ t-ON. So, as ig+ increases, t-ON can reduce.
So, in this way you can improve this speed of turn ON and turn OFF. So, this is one
interesting speed up gate drive for the MOSFETs.
All BJT drive circuits can be used for MOSFET gate drive. Let us see an example I will
draw the MOSFET, drain connected to the external load, source connected to the ground
there is a body diode and we need to monitor iD in the switch the drain current the gate
drive is the 1 resistor R1.
Now, the source for this gate is coming from the BJT drive circuit and you can adjust the
sourcing and the sinking capability to whatever value of ig+ you would like to give to the
gate drive of the MOSFET. So, therefore, with BJT drive circuits stage by stage you can
enhance the ig+ value to faster turn ON and turn OFF speeds.
So, at this point let me put an NPN PNP combination totem pole directly across Vcc and
let me connect both the bases together and let me drive these bases with a 2 resistor base
drive circuit which we are familiar with and the drive for this is coming from let us say a
PNP stage the PNP stage gets its base drive from another 2 resistor drive stage which is
driven by an NPN stage, which is driven by another 2 resistor drive stage.
767
So, you see like this stage by stage the current starts decreasing, the current here the ig+
whatever requirement here would be lesser here on the base by base side of this. On the
base side of the PNP it will be still further lesser here till at the source Vb it can be of the
order of micro amps to milliamps.
So, let us say we need to give a pulse in this fashion here. Let us say this is Q1, R1, Q2,
Q3 and Q4, we have 4 transistors. So, whenever Vb is low there is no drive current
flowing through this resistor path no base drive for Q4; Q4 is OFF, there is no current in
this path, therefore no base drive for Q3 base therefore, Q3 is OFF. Q3 is OFF no base
drive current flowing through this path and no base current for this and Q1 is OFF.
When Q1 is OFF there is no drive current to charge up the gate of the MOSFET, so, the
MOSFET is OFF. So, when Vb is low MOSFET is OFF, id is 0. Now when Vb goes high
there is a drive current flowing through this 2 times iBON flows through this and iBON into
Q4; Q4 is ON. When Q4 is ON there is an ib. 2*iBON is flowing through this resistor
corresponding to Q3 base drive there is an iBON here Q3 is ON, there is a drive current
flowing through this path into Q 1 and Q 1 is ON.
When Q1 is ON Vcc minus V cesat appears here and it will drive a current through R1 in
to the gate source capacitance of the MOSFET and charge up the capacitance. So, Qg
amount of charge will be pumped in and the MOSFET will turn ON. So, in this way the
MOSFET will be switched ON and OFF. The value of ig+ can be controlled by having
the number of stages. So, the more number of stages means I can have more
amplification. So, let us say this particular at this particular point you have a 10 milliamp
sourcing capability, it could become 100s of milliamps sourcing capability at this point it
can increase by another factor of hFEsat here. So, you can get in terms of amps another
factor by hFEsat you can get in terms of 10s of amps.
So, you can get really very high sourcing capability at this point when you are having
multistage BJTs to amplify whatever the current here to a large value of ig+. So, once
ig+ is very large you can charge up you can give the gate charge for the gate capacitance
of the MOSFET in a very short time.
So, the turn ON times can be really small likewise the turn OFF times can also be really
small because you will be discharging through these high current devices. So, this is a
high speed MOSFET gate drive circuit using BJTs.
768
Prof. L. Umanand
Lecture – 87
MOSFET drive with isolation
Like you had isolated base drive for the BJT, where the power side and the control side
were physically isolated galvanically isolated. Even in the case of the MOSFET we can
have galvanically isolated gate drive circuit. Now we saw the BJT based gate drive circuit
where the MOSFET is driven by this Q1, Q2 totem pole BJTs which is driven by Q3 and
which was driven by an NPN stage.
Now, this NPN stage I will replace it with the NPN BJT of an optocoupler, the optocoupler
where this transistor is being driven by the light emitted from a photo diode. So, let me
ground this and then connect the pulse source to this optocoupler diode in this fashion and
we have Vb the pulse source here. So, observe here when Vb is 0 there is no drive current
to the optocoupler diode current is not there diode will not emit photons and the
optocoupler transistor is OFF. So, when this is OFF there is no drive current here Q3 is
OFF.
When Q3 is OFF there is no drive in this path and therefore, Q1 is OFF when Q1 is OFF
there is no drive to charge up the capacitance of the MOSFET and therefore, the MOSFET
769
is off. So, when Vb is low MOSFET is OFF. When Vb is high there is a drive current
flowing through the diode, the diode emits light and the optocoupler transistor is ON there
is a drive current here Q3 is on.
Q3 allows drive current to flow through in this path to Q1; Q1 is ON and there is a drive
current to charge up the capacitance of the MOSFET and the MOSFET turns ON. When
you want to switch OFF the MOSFET, you are switching OFF the pulse here goes low,
this is OFF Q3 is OFF, Q1 is OFF but there is a charge ON the capacitor here plus minus.
So, that is coming to this point emitter of Q2 the base of Q2 is low and because emitter of
Q2 is positive Q2 will turn ON and because Q2 turns ON there is a reverse current flow in
this fashion and discharges the gate charge capacitance. So, this is the discharge path and
turns OFF the MOSFET.
So, this is an isolated gate drive circuit for the MOSFET, remember the power side are
having different grounds the control side is having a different ground and just like in the
case of the BJT you need to have a power supply with isolated secondary windings and
one winding is used for generating the power supply for the power side and another
winding is used for generating the power supply for the control side. So, in that way total
isolation galvanic isolation is maintained between the power side and the control side and
the optocoupler gives an isolation of 3000 volts between the 2 sides.
770
Like in the case of the BJT base drive we discussed with transformer isolation, the
MOSFET base drive can also be done with transformer isolation similar circuit let me
draw that. So, the primary of the transformer is connected to an NPN transistor and the
NPN transistor is driven by a 2 resistor drive which we are familiar with this fashion.
Observe this is NPN transistor which I am using here can also be done with replaced with
the MOSFET.
Now, the secondary of the transformer is passed through a diode like in the case of the
BJT drive and it is connected to the power MOSFET. The power MOSFET is connected
in this fashion drain connect to the external load source is connected to some other point
in the power circuit. The transformer non dot end is connected to the source and the dot
end to the diode and the diode is connected in series with the resistor in this fashion.
So, this will be a 1 resistor drive for the MOSFET. Let us name the parts is R1 this is D1
Vcc, Q, R3, R4 and this is Vb the base drive pulse given at this point. So, when the base
drive pulse Vb is high. So, you will have a current flow through R3 there will be a base
drive current for Q; Q will be on.
When Q is ON the transformer primary is active with dot end positive; dot end positive
means ON the secondary side dot end positive and there will be a drive current flowing
through D1, R1 to charge up the MOSFET and the MOSFET will turn ON. So, as long as
dot end is positive the MOSFET will be ON.
Now, when Vb goes low there will not be drive current through R3, Q is OFF and here the
dot end becomes negative non dot end becomes positive the diode will be reversed biased
and it will click the wave form. However, the MOSFET has MOSFET capacitance that
there is no discharge path for the MOSFET capacitance. So, it will not turn OFF. So, you
have to provide a discharge path for the MOSFET so, therefore, we will put a resistance
here R2.
Only then the MOSFET can turn OFF, there is a discharge path. Also note on the primary
side we will put a diode resistor combination like this the freewheeling diode, in the
freewheeling resistor this is necessary to demagnetise the core. So, that it is ready to
operate in the next cycle. So, this is the transformer isolated drive for the MOSFET.
771
In the transformer isolated gate drive shown here there is not much control in speeding up
for the turn OFF of the device because the presence of R2. You cannot make R2 too low
because the R2 is made too low while turning ON R1, R2 becomes an attenuator. So, if
you are applying 15 volts here, you can have a very low voltage applied at the GATE; gate
source and the MOSFET may not go into full enhancement it may be in the linear region.
Therefore, it is better to replace R2 and have an active pull down to speed up the turn OFF
process. So, let us do that modification. So, I will remove this make some space and
connect it here and here I will put an active pull down circuit. So, let me use a diode and a
PNP transistor in this fashion, I will put a resistor here and connect the base to this point
and let us see how this operates.
Now, here there is an internal capacitance remember. So, we will have the gate source
capacitance indicated by this greyed line here. Now, when let us look at the wave form at
the secondary when Vb is having a pulse wave shape and the in this fashion whenever Vb
is low Q is OFF and whenever Vb is high Q is ON. So, let us say Vb is high Q is ON, dot
is positive, here will be positive. So, you will have a wave shape something like that and
then when here Vb goes low Q turns off.
So, when Q turns OFF there is a freewheeling action happening here. So, the voltage
exponentially will start coming down, so, it will go negative and start exponentially
coming down. Remember that this area positive area negative area should always balance
out volt second balance is always valid. So, you will have a wave shape like this going
negative, that is why we have put this diode D1 and on this side the diode will clip it and
will give you a nice neat positive only the positive pulse.
So, during the time when it is positive there is a current flow like this and this second diode
that we have put will have a polarity plus minus forward drop of 0.6 in this fashion. So,
there is a current flow and it will flow through R1 and charge up the CGS capacitance. So,
a charge of QGS is deposited in the MOSFET and the MOSFET will turn ON. So, during
that time when you are driving the MOSFET on diode is plus minus in this fashion look at
the PNP transistor emitter is minus base is plus.
So, this transistor is off, so, it is out of the picture. Now after this is ON and then you
decide to switch OFF the MOSFET you bring Vb low. So, when Vb is low you are in this
portion of the region where is in this portion of the region of the pulse. So, let us say that
772
at this point capacitance is fully charged to whatever this peak Vcc value and diode is now
not conducting. This capacitance potential appears here emitter is positive this is at 0
potential with respect to this there is no current flow.
So, therefore, there is a possibility of base drive to flow in this fashion. So, emitter is
positive because of the capacitor. So, you will have a base drive flowing in this fashion
here, it cannot go in this direction because the diode. So, it has to come in this direction
and complete the circuit. So, if the base drive flows then the transistor Q can be ON and
you can have a larger collector current to flow in this fashion and discharge the capacitor.
So, the capacitor in fact charge the gate source capacitor, in fact it makes this active pull
down work. So, this is a much better circuit, much faster turn OFF can be achieved with
this.
773
Prof. L. Umanand
Lecture – 88
Over-current protection
Let us now discuss about the current protection for the MOSFET and the BJTs. Though I
will be describing this current protection scheme with respect to the MOSFET power
switch, it is applicable even to the BJT power switch. So, let me first draw the MOSFET
power switch like this you have the gate, you have the drain connected to an external
load and the source and there is an internal body diode.
So, this is the MOSFET it has a one resistor drive single resistor drive as shown and
there is an active PNP pull down, as we discussed there is a diode here. And I am going
to put a resistance here and the base connected to this point. So, this is the active pull
down that we discussed and we have a diode here. So, this part can be isolated or non
isolated, but this is essentially the drive part which is there just before the power
MOSFET.
Now, how do we go about doing current protection, we have here the drain current and
we need to do some action if the drain current exceeds the rating as particular set limit.
774
So, if there is an over current then this drive must be switched off the MOSFET must be
switched off even if the drive current the gate drive is present.
So, how do we do that? The principle that we are going to use is that as the current
increases the voltage Vds across the MOSFET or VCE across the BJT is going to
increase. The voltage across the device is going to increase. Now we have to detect this
increase set a set point and once the increase goes beyond a particular value, then we say
that over current has happened and we will switch off the MOSFET. So, that is the
principle that we are going to take.
So, let us do some modifications to the circuit here, I am going to push the diode back
here and I am going to put a series pass a series switch I am going to put a series PNP
switch here and this PNP switch is having a two resistor drive which we know how it
operates. So, if this point here is high if the voltage is high then there is no current
flowing in this resistor.
And therefore, no base current flowing up and therefore, this series pass switch is off and
the drive is cut off the capacitance will discharge and the MOSFET will switch off. Now
if this point here potential is low, then there is possibility of current to flow through two
resistances and there will be a base drive and this switch will be on. .
So, now this is a series pass switch we will operate. Next let me put an high slew rate
comparator Op-amp, plus minus it has its power supply, it is a unipolar power supply
Op-amp and to the op-amp minus I will give a reference how do I generate a reference I
will have the resistor, Zener diode and then potentiometer across this, this is a typical
shunt regulator which we have discussed some time ago. So, this output I will give it to
the minus, I will adjust the pot value here such that I have some Vref value this is the
reference value. If the positive value goes above, the reference value this will be high.
Now, let this plus terminal be connected to the over current detection circuit. So, how do
I detect over current?
Now, I will put a resistor a diode and then connect it in this fashion. So, this point now
this is the ground here for this circuit, the power supply here is obtained from. The
secondary of an isolated transformer and the power supply is specifically for this gate
drive circuit. So, it is isolated now this is the ground, now from this point I am going to
tap off and connect it to v-plus.
775
What is this potential? When the MOSFET is on there is a drop across that I will say
Vds-ON. Vds-ON is coming across that and this diode is conducting in this fashion to
the ground from the power supply there is a current flowing through this diode through
the on device and to the ground. So, therefore, this potential will be diode drop plus Vds-
ON drop we will just write that down. So, we have Vd0 forward drop plus Vds-ON. So,
this will be the potential here
So, Vd0 is around 0.6 or 0.7 volts and this is the Vds drop and as the current increases
the MOSFET has an on resistance Rds-on. So, or Rds-on* id will be the drop across this
and as id increases the Rds-on * id drop across this increases Vds-ON increases and this
will increase. Beyond a particular reference value set point value this will go high the
output will go high and that if I am connecting it here when the output goes high.
This point is high and this PNP transistor is switched off and once this is switched off
there is no drive to the MOSFET, the MOSFET will discharge through this PNP
transistor in this fashion and this will turn off. So, whenever there is a over current Vds-
ON is acting as the sensor for the over current and that will shoot up and that will go
beyond the reference value, and it will make the output of the op-amp comparator high
and switch off this series switch element and cut off the drive. So, this is the principle on
which the over current protection comes into the picture. However, there is a small
problem in this.
When the MOSFET is off the voltage across D and S is a very large voltage as dictated
by the external circuit because, when the MOSFET in the off state it is going to
withstand the entire Vdc link voltage, which may be a very high voltage value. Now,
once that is a large value Vds during the off state plus Vd0 will be coming at this point
and therefore, this would always be high and this will be disabled. So, you will not be
able to turn even if I give a turn on pulse you will not be able to pass the drive on to the
gate of the MOSFET because this is always off.
So, you have to provide a small period of time when you want to turn on a small period
of time when this is disabled means this is on, it should allow the MOSFET to switch on
turn on Vds should come low to the on state; and it is the Vds-ON that we are interested
for the fluctuation and the variation in Vds-ON during the on time and the id can vary
776
depending on the external load. So, allow it turn on and then enable this part of the
circuit. So, that is what we will be doing slight modification in the circuit.
So, let me use an AND gate and the input of AND gate I will pick it off from this point,
this is the gate pulse itself. So, the gate pulse comes in here now the gate pulse from the
output of the AND gate is passed to an RC filter. So, this is an RC filter and as this an
analog signal I will pass it through a Schmitt and so this is going to rise slowly, the other
input of the Schmitt I will give it directly.
So, this is going to rise immediately, it is going to fall over the gate signal, this is going
to rise slowly. So, which means a delayed gate signal and the output is going to be
delayed by as by sometime as dictated by this time constant. Now the output of this I will
call this A and I will call this B, the output of this and this should be ANDed and given
here. So, how does this operate? So, whenever there is a gate pulse this circuit will delay
the enabling of this means this will be low for some time where a small amount of time
turn on time during that time this is low.
So, even if this is high the output of the and gate is low therefore, the PNP transistor is
enabled this is going to cause the drive to flow through to the MOSFET a charge is put
and the MOSFET turns on after the MOSFET has turned on you need to enable this
portion of the circuit. So, this will go high, this will be enabled and this will go low
because when Vds has turned on and the current is within the limits this will be low and
this PNP will be on as long as there is no over current and this have this is low.
Once there is an over current this v-plus will go high output of the Op-amp will go high
and this is already enabled from here therefore, this AND gate output will go high and
disable the transistor. So, this operation is very important let me say that this is vg let me
just indicate how this delay action comes in, let us say vg is like this, this is the vg signal
gate signal that you are going to give, look at vb here the point B how the signal comes.
Now, see vg is coming directly to this Schmitt AND; and the other input the Schmitt and
is coming through an RC. So, this is delayed signal. So, therefore, output here will be
delayed by a small amount of time as we decide by the RC time constant. Now because it
is delayed here the AND gate output is low for a small period of time here.
777
And this is the enabling that this is called the blanking time and it is during this blanking
time why we say blanking time? The current protection unit is blanked out at that time;
and PNP transistor is on. So, during that time the MOSFET turns on and Vds becomes
low and goes into Vds-ON mode, at which time if the current is within limits this will be
low. So, if this is low then even if this is high the output of this AND gate is lower and
the PNP is on.
So, it goes high after a small amount of time which means this will become the enabling
pulse for this after a small amount of time and then after that any fault that occurs here
will disable this PNP and switch off the drive and turn off the MOSFET. So, this is how
over current protection operates and this is one of the most popular over current
protection circuit which is used in almost all the gate drive ICs. The current protection
gate drive ICs uses this principle.
778
Prof. L. Umanand
Lecture – 89
Snubber circuits
Let us discuss about couple of Snubber circuits. Snubber circuits are used to relieve the
stress on the power semiconductor switches. There are essentially two types of Snubber
circuits we will discuss couple of them. Consider a switch like that, and the switch is on,
it is conducting the current and switch is off it is supporting the voltage and it is a control
switch. Now, let us say there is track inductance and let us say it is having a value L and a
current iL flowing through the switch.
When it is flowing the polarity is in this fashion plus or minus and the voltage across the
switch is given as VSW, there is a current iL flowing through that, now you TURN OFF
meaning you are cutting open the switch which means iL is suddenly dropping to 0. So,
𝑑𝑖 𝑑𝑖𝐿
there is a large − 𝑑𝑡 so the polarity will reverse and VSW will be L 𝑑𝑡 plus the external
voltage I will put it as Vcc, but this is a very large quantity. So, you will see a large spike
across the switch and it will be so large that it can blow up the device or it will deteriorate
it to a large extent.
779
So, this is a large problem and we need to address the term of stress. The other case is
during turn on. So, you have a switch connected to the external circuit and through the
external circuit there can be capacitances, stray capacitances connected to the ground and
let us say there is a switch current when you TURN ON the switch at the time of turning
ON the switch, all the charge that is stored in these stray capacitances can flow into the
switch.
𝑑𝑣 𝑑𝑣𝑐
And that can be a very large current because C 𝑑𝑡 . Let us say on TURN ON isw is C 𝑑𝑡
𝑑𝑣𝑐
plus all the external currents which is supposed to flow normally. The C 𝑑𝑡 can be really
large and this TURN ON current can be so large that it can blow up this switch. So, you
need to protect against such a stress on the device.
So, during turn off the protection circuit snubber circuit is called turn off snubber or even
called a shunt snubber in many cases because we try to put a capacitance across this device
𝑑𝑖
to address this L𝑑𝑡 problem so it is called a shunt snubber. That dual is this the turn on
𝑑𝑣𝑐
snubber or we put an inductance in series to address the problem of a large C 𝑑𝑡 current is
called the series snubber. So, let us see how we design the basic turn off snubber and the
basic turn on Snubber.
Let us now discuss the shunt Snubber or the turn off snubber and see how it works? Let
me draw the switch or semiconductor switch it is a BJT switch that I am drawing here, but
780
it could as well be a MOSFET switch or any other switch and this has a problem due to
the winding inductances or the track inductances, the stray inductances and there is a
current flowing through that and during the time of turn off there is going to be a large
𝑑𝑖/𝑑𝑡 spike due to these inductances.
So, let us say there is a current iL flowing through that through Q and the voltage across
Q is Vce and let me plot Vce with time. So, during the time when Q is on, so Vce is 0, let
us say and during the time when Q goes from on state to the off state turn off because of
this iL flowing through the stray inductances there is very large di / dt.
But we want to prevent that stress appearing on the device and then we would like to see
that the voltage across the transistor rises in a controlled manner to Vcc and settles down
at Vcc. How can we do this? So, this is done by putting a shunt element a capacitance
across the device across the switch.
𝑑𝑖
So, we pass this current iL to freewheel. So, instead of having a large L𝑑𝑡 spike when you
turn off Q, we will allow the current iL to freewheel through this diode into the
capacitance. So, the capacitance will get charged and during the time when the transistor
turns on again, we would like to discharge the capacitance and for that you need to have a
discharge path. So, I will put a resistance path here because you cannot connect directly
the capacitor across this it will be shorting and then there will be huge current.
So, therefore, you pass it through the resistance in a controlled way there will be a
discharge through it when Q turns on again and then it will be brought back to its original
state. So, that it can receive again the stray energy from the stray inductances. Now that
this is this portion of the circuit this D, R and C is called the snubber circuit or the shunt
snubber. So, I will call this as C this is D, R and the current through C as icap and the
current through the collector of the transistor as ic.
Then let me also plot ic; ic would be high when the transistor was ON and at the time of
turn off the collector current linearly falls down to 0 in this fashion. Now, let us look at
this wave shape and try to find out what should be the value of C. So, this iL is equal to ic
+ icap from K.C.L you get that. And then if I take only tf fall time period during that
period, ic that is current through the switch is of this shape it is linearly following I am
approximating it as a linear fall. So, it is iL* (1 – t/ t f) linear fall plus icap.
781
Now, icap see iL is on this side. So, icap I can take it to one side and then rearranging you
𝑑𝑣𝑐
can find iL is t/tf. So, this is the current through the capacitance. So, C 𝑑𝑡 is equal to icap
or the voltage across the capacitance because the diode will be the same as Vce.
1
So, you can say Vce is 𝐶 ∫ 𝑖𝑐𝑎𝑝 ∗ 𝑑𝑡, 0 to tf, let us take only that time period then integrate
it, because icap is known you substitute here. And at tf apply Vce should be Vcc at tf
boundary condition and then C will work out iL*tf/ 2*Vcc and this is the value of the
capacitance that you need to use. So, that in tf time period in a controlled way you can give
a smooth rise to the voltage across Q to reach Vcc, there will not be any spikes. So, this is
the effect of the turn off snubber or the shunts snubber. How to find the value of R let us
see that.
So, the capacitor would have charged up to ±Vcc and now when the transistor is turned on
next, there are two currents that flow one is iL itself the other one is this capacitor will
discharge through this R into the Q. So, that is one path this discharge path of the
capacitance and the other one is iL itself.
So, these two currents will be the TURN ON instantaneous current Vcc/R plus at that
instant was iL. Now this should be less than Icm rating of the device of the Q is very
important. So, continuous max the rating should be greater than this value.
So, from this condition R is one variable which is not known you can say R should be
greater than Vcc/(Icm – iL). So, this is one condition, the other constraint is when this Q
is on during that time the hole of Q charge on the Q should get discharged. So, there should
be a minimum time for which Q should be on, so that it allows enough time for Q to
discharge its charge and be ready for the next cycle.
So, the minimum on time Ton-minimum should be greater than 5* RC time constant, there
is a RC time constant, this discharge is exponential in nature having a time constant RC,
5*RC time constant means the steady state is reached. So, therefore, this constraint leads
to R as Ton-minimum/ 5*C, C of course, you know from this relationship.
782
From these two constraints we get the range of R that which we should select from. So,
Vcc/(Icm – iL) should be less than R which would be less than Ton-minimum/ 5*C. So,
we get the range of R that we should choose and the power dissipated in R: PR,we know
1
that the energy that is dissipated is 2 C*Vcc^2; and this is the watt seconds divided by time
1
Ts the period switching period, this will give you the power and therefore, 2 C*Vcc^2*fs
will give you the power dissipated in R whatever be the value of R within this range.
So, in this way we know how to find the value of C and find the value of R and this is how
you design turn off snubber or a shunt snubber as shown here like this.
783
The dual of the turn off snubber or the shunt snubber is the TURN ON snubber or the
series snubber, it has a dual effect. So, let us draw the switch I am using a transistor switch
it could as well be a MOSFET switch.
Now, let us say there is a stray capacitance charged capacitance coming across the switch
due to the external circuits; now this is charged to some Vcc. So, when the switch is turned
on, you can have a very large current flowing through the switch and it could also be fatal.
So, let us draw the time waveform and I will take the current ic through the switch.
Now, ic is when the switch was off the current is 0 and now we are turning it on. When
you turn it on there can be a huge discharge, huge surge current, because of the discharge,
because the on resistance of the power semiconductor switch is almost 0, you can have
this Vcc voltage divided by almost 0, a very large current flowing through that and this
large current can blow the transistor. So, instead we require that the current through the
transistor ic rises in a controlled manner, and let us say this is the turn ON time it rises in
a controlled manner and reaches iL gradually.
So, for that you need to interpose something a series snubber circuit and that is composed
of L and if you have L you need to have a freewheeling path the demagnetising path diode
and resistor as shown here. So, let me say this is L, D and R and this voltage is Vcc, V ce
and this voltage is vL. So, I will mark it Vce, vL and the voltage together will be Vcc now
let us see what Vce would look like with time.
784
So, when the switch was off during this time Vce would have been high at Vcc and then
gradually comes down to 0 when it turns on. Using these waveforms and these Kirchhoff’s
relationship let us develop the equation for finding the value of the L in the series snubber.
So, Vcc equals vL+Vce and if I am talking about the time of interest tr, rise time.
Now, during this time you see that Vce is falling in a linear fashion I am making a linear
approximation. So, it is Vcc*( 1 – t/tr) + v L. Now if I take vL into one side, vL can be
found to be Vcc*t/tr. Now ic that is the current through the current through the collector is
1
also flowing through the inductor. So, L*dic/dt will be the voltage vL. ic is ∫ 𝑣𝐿 ∗ 𝑑𝑡.
𝐿
Let us consider the period of interest 0 to tr. So when you do the integration you substitute
this Vcc*t/tr here and integrate, at t = t r you can take ic = iL. So, L is equal to Vcc*tr/2*iL.
So, this will be the relationship for L
Now, if you look at the time when the switch is off the inductor current is freewheeling in
this fashion. It is freewheeling in this fashion; and there is a positive drop from positive in
this side a drop across R and drop across D in this fashion. So, what are the voltage across
the transistor: Vce, Vcc+iL*R+VD this is coming across. The transistor and it should be
less than VCEO rating of the transistor. So, therefore, this condition gives R to be less than
(VCEO- Vcc-V D)/iL this is one condition.
Another condition is the energy in the inductance should get released by the time this turns
on again. So, there should be a minimum off time for this transistor, Toff-min should be
785
greater than 5*L/R time constant so that inductor releases all its stored energy, this means
R should be greater than 5*L/ Toff-min.
So, by this you get 5*L/ Toff-min is less than R which is less than (VCEO - Vcc – Vd) / iL,
this is the relationship selecting the range of R. And what is the power dissipation in R?
1
*L*i^2 amount of energy is removed every cycle divided by Ts or into fs this will be
2
participated in the R for whatever be the value of R that you choose in this range.
So, this is the series snubber or the turn ON snubber. So, every switch whether it be a BJT
or a MOSFET or an IGBT should have a series snubber and a shunt snubber together. So,
the series snubber, shunt snubber and over current protection along with the gate drive
forms the complete power semiconductor switch.
786
Prof. L. Umanand
Lecture - 90
Intro for close loop control
In this week’s session, I will discuss about one of the more important aspects of DC-DC
power converters and that is close looping. You see we have studied the non isolated
DC-DC converters, the buck the buck boost, the boost converters they are called as the
primary DC-DC converters. We have studied the open loop operation, the steady state
operation and we have seen how to design that. We have also seen isolated converters
like the forward converters, the various forms of the forward converters and then the
derived forward converters like the push pull, the half bridge and the full bridge. And,
also the buck boost derived converters like the fly-back these were some of the isolated
converters that we discussed and also studied how to design.
And we also looked at the magnetics, how to design the magnetics L and the
transformers. So, having done all that we know how to operate the DC-DC converters in
the open loop. Now, the important part is that we need to regulate the output voltage or
the inductor current or some state parameter within the converter. So, let us say that we
want to regulate the output voltage then you have to feedback the output voltage and then
787
take some corrective action. So, a controller has to be built. A PI controller is the one of
the most robust and industry standard converters we will discuss them.
So, we will discuss two aspects in controls, one is we will look at the traditional close
looping and then I will introduce a slightly a modified form of the traditional control
loop that is introducing feed forward control. It is a slightly faster form of the traditional
control, here also we will use PI control, but the stress on the PI controllers will be much
lower. And then after that I will follow it up with another very important popular control
mechanism especially for DC-DC convertors and that is current control or current
programmed control.
So, current control is an important aspect of most DC-DC converter operation, used in
many applications like battery charging and we will discuss a battery charging example
with the current control. Of course, we will also look at simulation, NG spice simulation
of these close looping aspects with a normal close looping, close looping with feed
forward control and then close looping with current control. Then after that I would like
to close the session with an introduction to single phase inverters. Single phase inverters
which are driven by sinusoidal PWM control patterns strategy.
So, the switches I will take a full bridge converter as a typical topology, example you
have four switches q1, q2, q3 and q4, the drive signals for q1, q2, q3 and q4 will be so,
generated such that the fundamental component in the PWM will be a sinusoidal
waveform. And how do we generate such a PWM, how do we give the signals to the four
switches, and then in what way the fundamental PWM can be extracted from the center
arm of the bridge are aspects that I will be discussing.
788
Prof. L. Umanand
Lecture - 91
Close looping dc-dc converters
In this session let us look at Close loop operation of dc-dc converter. Let us take an
example of a buck converter, this is a buck converter we are familiar with this topology,
but note that you can work with any converter the principle of closed loop operation is the
same. Call this Vi, L, C, Ro, Vo and we would like to regulate Vo and the control input is
here in the form of duty cycle dTs.
So, let us say that is the control input d. Now, let us build the control portion of the
schematic. So, let us have a reference a feedback quantity plus and minus we are
comparing it and let us pass it through a control block and how about the control block
will give you the control signal which will be compared with the triangular carrier to
generate pulse width modulation and steering that we have discussed.
And then from the pulse width modulation it will go to the gate drive circuit, we have
discussed the gate drive circuit on protection and then goes to a drive the switch. Now, this
feedback here will be the output feedback. So, we are sensing the output, if we are
controlling output we have to sense the output and pass it through a sense circuitry which
789
means, it could be made up of op amps for stepping down, filtering and then scaling and
the output of the sense circuit is given here to this compare for to this compare circuit in
order to do the control.
So, you have Vo reference and Vo feedback. So, how it operates is that now let us say Vo
here increases this will make Vo feedback to increase. So, Vo feedback increases Vo ref
minus Vo feedback will decrease so error will decrease. And the output of this will
decrease the pulse width modulation signal will change in such a direction to decrease the
drive duty cycle and therefore, once that duty cycle decreases output will decrease and
come back to its original position operating point.
And this is the close loop action for one direction where we starts with some Vo increase
and the Vo increase could have occurred due to any disturbance, disturbance in the load,
in input voltage or even temperature.
If Vo decreases also the action will be in similar way, Vo feedback decreases, error will
increase this will increase pulse width modulation d will increase and then therefore, Vo
will increase and it will pull it to normal operating point.
So, this is the control action that you will see. It is now required that we design the value
of this controller gain K. So, if we properly choose this controller then we can ensure that
the error here will be 0 and the Vo feedback will match Vo reference and your output will
be regulated as per the reference desired reference. So, let me introduce some variables,
this is error e and output of the controller will call it as a control voltage Vc.
This control voltage in fact, will be comparing with the carrier triangular to generate the
pulse width modulated waveform. Now, let us focus on this part irrespective of all the
other parts of the system. Now, if we see here the relationship is: Vc is nothing but e* K,
error into controller gain. So, let us say e: error is equal to Vc / K, when can we get the
error as 0? There are only two possibilities, either Vc should be 0 or K should be infinite.
Now, if we take the case where Vc is 0 which means you are grounding this portion. So,
if you are grounding this portion and making Vc always permanently 0 then there is no
close loop, becomes an open loop system.
790
So, Vc has to be allowed to vary and make K infinite. If you make K infinite whatever
may be the variations in Vc, e is always 0, the error is always 0 which means that the
output is regulated the output Vo feedback and Vo reference are the same and which means
that the output here Vo is as per the desired reference Vo ref. So, in order to achieve desired
control we have to make e going to 0. So, if e tends to 0 which means that you have made
K tending to infinity. Now, this is the important rule or important control principal that
you have to remember whichever be the system you have to make the controller gain as
high as possible to achieve error tending to 0.
However there are many problems and trying to make controller gain infinite. So, we need
to apply some constraints and under some operating conditions we will try to make K
tending to infinity. So, let us have a look at how we go about doing that.
So, now, let us look at the controller gain K, now I am going to draw a graph of frequency
on the x axis you have frequency, on the y axis you have the gain the magnitude or the
gain of the controller and I am going to express as 20 𝑙𝑜𝑔10K in the dB scale.
So, now let us use this graph to understand K a bit better. The system is going to be in
various states: it will be in the steady state, it will be in a transient state and it will be
transiting on transition state to the steady state. So, when you say steady state it means
omega is 0, that is it is this area here near omega 0 will be the steady state zone. And as
omega is moving more towards the right part of the x axis or the omega axis frequency is
791
high which means that you are in the transient zone here. So, let me mark this portion, this
is roughly this steady state zone closer to the dc and therefore, low frequency region.
Therefore, we can call this as the steady state zone or the low frequency zone. All this
portion the higher frequency region can call the transient zone, all the dynamics are
occurring in this region of the omega scale. We can call this as L.F. Low Frequency region
or this as high frequency region. I have drawn here a curve a green curve here and this is
the noise curve, see from here that the noise is low in the low frequency region and the
noise is more in the high frequency region the systems are more sensitive to noise the high
frequency region. So, therefore, if we go by this rule that K should be infinite the controller
gain should be infinite it will amplified noise throughout all the frequencies.
So, if I place the controller gain K infinite here means this is very high gain it is going to
amplify not only the signal. It is going to amplify even the noise everywhere and the noise
which is pretty significant in the high frequency region especially the measurement noise,
sensor noise you will see that they will get also amplified to very high value large value
and Vc will be mostly noise, it will be swamped by noise and then you will not get any
meaningful control.
So, therefore, we cannot just put K equal to infinite throughout spanning all the frequency
regions, you need to have K is equal to infinite only during steady state let us say at more
closer to the dc here and then K should become lower, smaller at higher frequency regions.
So, now that is the technique that we will adopt to shape the K curve as frequency changes
as a function of frequency.
792
Let us begin with the gain of infinity near DC in the near the steady state part. So, if there
is a gain of infinity then the error is 0 so you will get 0 steady state error. To start from
here, let me draw a gain curve K like this which means it is going down at minus 20 dB
per decade. So, here it is infinite and this type of shape of the gain curve is given by our
familiar integrator which is written as 1 / S in the as a transfer function. So, as omega tends
to 0, omega is tending towards dc the gain K will tend to infinite value.
So, around this region it is infinite, gain is very high error is 0, almost close to 0 and you
have steady state accuracy of 100 percent. Now, as the system moves towards higher
frequency zones during different points on the operating; different points in the operating
region, you would like to have the gain K lesser and lesser as the frequency goes higher.
The reason being as I told before the noise is larger as you start going towards high
frequency on therefore, you have to reduce the gain attenuate the gain if you want to get
meaningful control signals.
Now, this is just plain I control, you just have 1 by S: integral and that is I control; I could
have chosen instead of this -20 dB per decade line, I could have chosen this line. This also
is a -20 dB per decade line starting from another infinite value, it is a family of infinity or
I could have chosen another -20 dB per decade line like this or I could have chosen a -20
dB per decade line above the one that I have chosen.
793
So which one should I choose? So, let us introduce one more degree of freedom K: is Ki
by S. So, I have put one more variable instead of 1 by S, I put Ki by S; Ki is equal to 1 is
this take blue line which I have chosen. For any other value of Ki you have this parallels.
Now, if I have to include Ki into this figure the difference between this and this would be
20 𝑙𝑜𝑔10Ki. So, I know which parallel to pick if I start picking parallels as it starts to
intersect the omega axes towards the right the bandwidth is higher. So, this is K is equal
to K i is equal to 1; now you see the points of intersection this is the unity gain bandwidth
product. So, when you compare this parallel intersection with respect to the blue line that
we have selected you see that this parallel is having a higher gain bandwidth, the one here
as a lower gain bandwidth.
The one here are still lower gain bandwidth. So, the speed of response is the one that is
going to be affected by choosing the parallels. So, let us now write down our controller
has a Ki here then passed through an integrator and you have the control signal, this whole
put together is Ki by S. Now, let me draw the time graph. So, this is time t and then let me
have a step response what is the step response to a step in Vo ref. So, what happens to the
output, the output can be like this depending upon what value of K let us say I have chosen
a Ki which is corresponding to this line.
So, you have a very slow response and let us say I choose this line which is having slightly
higher gain bandwidth product. So, it will be slightly faster. So, you see that as Ki you
choose the parallels, with Ki increasing you will have faster and faster response so here
also Ki increases in this fashion. So, let us say the one that we have chosen we could
probably have a Ki coming like this and you can see it reaching steady state quicker.
Like that if we choose a higher one it would probably go much quicker and then you choose
one that satisfies you on the oscilloscope. Then let me now introduce one more component,
a component like this in parallel I will call that one as Kp just a proportional gain and this
proportional gain is added up to the proportion the integral one also.
So, you add it up in this fashion. So, this is error and this is we see the control signal. Now
together you see this is Ki /S + Kp will be that transfer function between Vc and e. So, this
can be written as S in the denominator Kp into S plus Ki by Kp. So, you see that there is
as 0 coming into the picture. So, this is an integrator that is the integrator that continues to
stay, there is 0 at some frequency point on this x axis Ki by Kp; now let us say at some
794
point I design and place this as the Ki by Kp. So, at this Ki by Kp value of frequency I
place the 0.
So, what should happen? The pole is giving a -20 dB/decade zero will give +20 dB/ decade,
the +20 dB/ decade and the -20 dB/ decade here will cancel and flattened out in this
fashion. So, what have I achieved with this. So, you see if I had not put this Kp the gain
would have fallen down like that, but now I am gaining extra gain as frequency changes.
So, I am improving the high frequency gain in this fashion, the result is that in the time
axis this is the high frequency part the transient that is the high frequency part transient
part this is the steady state part.
So, you will be pulling up the transient part because the gain has increased you will pull
up the transient part and then after it goes starts going to the steady state part you will see
it is only I control that comes into the picture the baton is handed over to the I control. So,
if you see the whole response may look like this with a PI initially the P comes into the
picture, that P is actually improving the transients and then after the transient has pass
going to the steady state the baton is handed over to the I control then you will see a
response something like that.
So, this way you try to include a PI controller. How do I choose the values of Ki and Kp?
The step is very simple you do not include Kp first, you only start with Ki very low value
start with very low parallel.
So, keep on increasing Ki gradually keeping Kp: 0, Ki gradually you increase till you get
to a response with pure I control that is best and then you stop at that value and then Kp is
only going to improve the dynamics. Steady state Kp is not going to touch because Kp
comes only at the higher frequency zone as the low frequency comes into the picture that
is only I control. So, you have taken care of I control first you have to take care of the I
control and fix your study state response. So, that is these responses.
So, once the steady state response is fixed then you try to pull up the dynamics by
introducing Kp. So, Kp also you try to increase from 0 onwards gradually increase Kp that
keeps adding up here the improvement in the dynamics and then it will try to pull it up and
then at some value you will see that it is best beyond that it will start to probably have
damped responses.
795
So, do not go over to that and then you probably may have over shoots. So, you stop at the
best response you think you can achieve by looking at the scope and at that value you stop
at the value of Ki and Kp and those values of Ki and Kp you plug in and you have your PI
controller.
We saw that the controller K can be a PI controller and how we can get parameters of the
PI controller Ki and Kp we just now discussed. I want to introduce one more concept call
the feed forward control which will help the PI controller to be more robust and faster.
So, let me how this x y axis on the x I am going to have the duty cycle d, on the y axis it
is Vo. So, Vo and d are related; Vo is the output of the buck converter and that is the
controlled variable d is the input to the buck converter it is the control input. So, the
relationship is d is control input this is the controlled output and actually when you talk of
transfer functions it is Vo by dS in the Laplace domain.
So, these two have a relationship so let me draw that the slope is Vi. So, Vi into d so let us
say for example, d varies from 0 to 1 and let me take an operating point where this x
intercept is d, I will put it as uppercase D and the y intercept is uppercase Vo. So, the
relationship between Vo and D is Vo is equal to D*Vi. This slope of this transfer curve is
Vi which is fixed constant.
796
So, Vo isD* Vi, D is actually the control input. So now, let me say that due to many reasons
V i itself can vary. So, therefore, the slope of the blue line can vary or there could be a
disturbance in the output or there could be some disturbance which is reaching the output
and Vo can vary. So, therefore, this operating point can vary in the neighborhood of this
steady state operating point that we have shown here.
So, let me say that D is actually not D, but D ±d hat indicating the small variation about
the neighborhood. And likewise on the Vo side also on the y axis I will say that this is Vo
+ v naught hat where vo hat is nothing but the variations in the neighborhood of the
operating point. So, in the normal PI control that we discussed till now without the feed
forward control, the controller this I will just write down D is Vo / V i and this is coming
from the model where Vo is a specification and Vi is measured from the circuit.
Now, why this is important is that all these variations all the variations in control input
variation D is supposed to handle the complete variation of Vo. So, the PI controller that
we discussed till now is supposed to handle all these variations which means that the
operating point can swing between 0 to the actual operating point. And the PI controller is
supposed to handle this large signal variations as opposed to this with feed forward control
what is it that we are going to do we are going to use this. We said that this is the model
of the system we know that Vo is equal to D*Vi we have derived develop this model earlier
while starting the buck converter.
So, I know Vo is coming from the user specs, Vi is coming from the instant by instant
measurement. So, Vo by V i is known. So, this is the nominal operating point or the steady
state operating point that the system should be for that given for that measured Vi and a
specified Vo. So, this information is available to me, let me use that information which the
PI controller was not using before. So, you use this information and set the operating point
at this, then due to any disturbance if there is any variation in the neighborhood the
operating point the PI controller should only handle this small deviations.
So, d hat and vo hat only the deviations about the neighborhood the operating point that
the PI controller has to handle. So, these are small deviations and therefore, the PI
controller is much faster to handle the small deviation and also because of the feedforward
term this is the feedforward term which is giving a guidance or direction to the controller
that any major deviation it is supposed to come and settle there it will be more robust. So,
797
this way it is more robust and faster. So, let us see how we include this feedforward term
into our control.
Let us now incorporate the feedforward part of the controller into this system. So, I have
the controller as it is existing now. Now you see that K is nothing but the PI controller and
this PI controller is handling the entire large signal swing of Vo. Now, I will remove that
because I need to include feedforward portion, also apart from PI let me extend that make
the connection and here I will include the PI controller.
Now, this PI controller has two parameters Ki and Kp which we need to select just like as
I told you first make Kp= 0, Ki you gradually increase from a very low value to take the
parallel with low bandwidth and keep on increasing till you get a very good response time
response and then fix that and after that increase Kp from 0 gradually to pull up the
transient response performance.
So, this is the error and the output of this is connected to another sub merge and that is
where we attach the feedforward term and that feedforward term is the steady state duty
cycle which will decide the nominal operating point or the steady state operating point.
Now, what is being fed back is Vo the nominal or the steady state Vo plus the disturbance
or the variations. Deviation in that Vo reference has to contain only the nominal steady
state operating point Vo, it does not contain the deviation because this is the reference.
798
So, this minus this would give you the error which is basically the deviations about the
nominal operating point about the normal steady state operating point and the controller is
supposed to take care of this and make this deviations 0. So, you see that PI is now only
handling vo hat deviations, small deviations in the neighborhood of the operating point.
And what is the output of the PI controller? The output of the PI controller is vc hat. Now,
Vc which is a voltage which is proportional to the nominal duty cycle is given at this point
and this vc hat is proportional to d hat which is deviations about the nominal because they
are all proportional to time.
Now, on adding these two you have Vc plus vc hat which is the voltage proportional to
the deviation in duty cycle, is together proportional to D plus d hat. So, that is the voltage
signal which is actually going to get modulated with this particular triangular carrier and
that PWM is driving this gate and switching on and off this buck converter transistor.
So, you see here the PI controller is only handling the hat terms; it is only handling the
deviation terms which is small signal terms and the nominal steady state operating point
is given from the model was D is equal to Vo specified by Vi measured.
So, D is a feedforward term here this handles the steady state part and the deviation any
deviations any disturbance PI takes care and it is handling only small signal deviations.
And therefore, this PI will be very robust and very fast because the deviations are small.
So, in this way you can introduce the feedforward term and make the controller fast and
reliable.
799
Prof. L. Umanand
Lecture – 92
Simulation of close loop control
We shall now look at the Simulation of buck converter in closed loop. So, let us open
this buck close dot sch, which is the closed loop schematic for the buck converter.
You can use the same approach for any other converter too, let me zoom in, so you must
be able to recognize this portion, you see that this is a dc source 12 volts, this is the
power switch which we will be controlling through some duty cycle. And you have the
buck converter portion L and the C which we used in the open loop simulation and the
inductor will freewheel when the switch is off and when the switch is on the inductor
will charge.
And I have put some load of five ohms that will give me around one amp of current.
Now, the changes that we have made is here in the control portion, which ultimately is
going to control the gate drive of the switch. So, what is it that we have done? We had
actually included the PWM in the open loop itself and we had given a constant voltage
here. Now, instead of the constant voltage, we are now having a controller, a comparator
and the feedback.
800
Now, the feedback is given through this label and you see that when we discussed in
theory we had connected this directly to the output, but then instead of cluttering up the
schematic, you recall that we had given a label here called o and same label if I copy it
and put it here, it is equivalent to saying that I have made a physical connection between
these two points. So, this output voltage is actually now fed back to this point. So, here
you are having a difference amplifier plus and minus to the plus I am giving a reference
voltage.
The reference voltage as you see here is not a constant I have given a kind of a pulsed
disturbance. So, that we will actually see that the output is tracking this pulse disturbance
but; however, in actual practice you will see that this will be a pure constant set point
value. Now, the difference between the set point and the feed back voltage, the error is
fed to the PID controller and the PID controller has some parameters Ki parameter for
the integrator, Kp parameter for the proportional, Kd for the differential. I am not
utilizing it, I am setting that to 0, there is lsat that is the lower saturation limit set it to
minus 10, usat the upper saturation limit setting it to plus 10. You have control on setting
it to still lower values too.
Now, this output is finally fed to the PWM which will generate the appropriate duty
cycle pulse and give it to the power semiconductor switch here. Now, this is the total
closed loop circuit of the buck converter and this is the closed loop portion and this in
fact, is the controller with these parameters. Now, let us see what happens when we
simulate, all these things now you know how to import the file and this is the inclusion
of the edt01 dot sub.
Now, before we go to the simulation what is it that we have added? We have this is one
block; this is a new block that you would not have seen till now, and this is another block
that you would not have seen till now. This is nothing but a summer, this is one of this is
this has been built using the analog behavioral model of Ngspice, likewise PID is also
built using the analog behavioral model of Ngspice. I will just show you how that block;
the model for that block has been incorporated in edt01 dot sub.
801
Now, these two blocks are available in the ablock library. So in the ablock library you
see add. So, this is the symbol and that is what we have pulled and kept it there this
symbol you recognize that.
And then the other file is the PID this is the PID and this is what we had; we have used
here. So, this is available in that ablock folder of the library. So, now, let us save all this
and this schematic let us use our well known process which is first we will go into that
directory and generate the net list.
802
So, this is the net list or command and I think you recognize this, I am using the buck
close dot sch and generating the buck close dot net. So, let us generate the net list and
you would see here that the net list has been generated. And now, let us go onto the
Ngspice environment and call buckclose dot cir.
Plot; now, I want to plot the reference and the output, just to see whether the output is
tracking the reference. So, let me plot r and the output o.
803
So, you see the red waveform is the reference waveform and the blue waveform is the
output waveform it is trying to track the reference waveform. Of course, though the
tracking is not too good we may have to tune the PID parameters, there is a measure of
control and you see that it is getting controlled between 5 and 6 volts. Now, what I would
like to do is quit from this environment, I will go and adjust these parameters. Now Ki is
the integral constant, Kp is the proportional constant.
804
And let me change the proportional constant to a larger value, so that the dynamics is
improved. Now, we will have to do all these things in step by step, but I am trying to do
it in the one big step just so that we save some time, but you can play around with these
parameters here. Now, let me save this and then let us go back and generate the net list.
And I will go into the Ngspice environment, run that.
And let me do these settings and we can now see the same waveform vr and vnaught.
805
So, you see it is much better you see that the waveform is more closer in tracking the
dynamics, basically because you have improved the dynamic response by the choice of
Kp.
Of course, I will leave it to you to play around with all the different parameters and also
the parameters of the buck converter, L and C’s and just try out through that. This pulsed
kind of a reference is just for this purpose of learning so, that you see that there is some
effect the output is trying to track the input and output is coming to a desired value. In
practice it will be constant let us say you want to give an output of 5 volts your set point
will be 5 volts, it will reach the set point 5 volts and continue to stay here without all this
jiggly-wiggly wave shapes. So, this is just to show that the buck converter output can be
regulated. So, now here you see that.
806
This of course, is the same buck converter circuit you have the PWM block here, which
is giving you the appropriate duty cycle and on this side you have that reference; and
then you have the feedback coming from this o label here; o label which is connected at
this point, this is the error is passed on through the PID controller. And the output of the
PID controller corresponds to d hat; corresponds to the deviation in the duty cycle a
voltage corresponding to that.
Now, here on the other terminal of the hat adder, we are having a dc voltage and here I
have given minus 0.05 to correspond to 45 percent duty ratio. So, this is the nominal
value or the operating point value which I am giving it as a feed forward term and this
plus this will be this control voltage and this is D plus d hat, with even the variation that
is coming due to unwanted reasons and it is coming from the controller. This goes into
the PWM and controls the switch. So, this is how we have this feed forward time here.
So, I would also like to simulate this, I will put back the those numbers which I used last
time during the simulation of the other buck closed circuit.
807
So, let me save it here now recognize that x add is another new block which I have
included, PID and xSUM you saw. And x add is also available in the ablock this is the
add and you can use that and this is the SUM which is used here ok. So, this is saved and
we can go to the folder here.
Now, here the buckclose and dot cir and the edt01 dot sub, I did not show you what is
there inside edt 0 dot sub you should understand that I have added some things here.
808
I have added the PID sub circuit, this is again composed of all analog behavioral
modeling elements; and these are well defined in Ngspice manual. And then I have made
a summer and that is also by the analog behavioral model.
I have made an adder here and that is also Ngspice analog behavioral model. I included
here a multiplier also, but of course, we are not using it anywhere just kept it here for
your reference. So, you will close this.
809
Now, let us go to the terminal and let me go to the directory; cd to the directory and let
me generate the net list and you know how the net list generated here, you see and
Ngspice buckclose dot cir into the ngspice environment, you ran the program.
And now let me set the background color to white and set the foreground color to black.
Now plot, I want to plot as usual the reference and the output vr and v naught. So, you
would see that it gives you also similar kind of performance measures. You could also
810
see what happens here and what is the gate pulses which are given to the switch you
could probably see plot vg which will give you the gate signals.
So, this will be the gate pulses which are varying continuously as you would see here and
you could also see the inductor current plot iL along with v naught.
So, you see that because v naught is no longer constant you see that the induct current is
also having a lot of variations.
811
But you could see here probably, in this time span where it is constant it would it will be
like what we expect if I increase the amplitude.
You see this is what we would expect when it is constant. So, you can play around with
the waveforms and the values and you should be trying to get more inside into this whole
close loop system.
812
Prof. L. Umanand
Lecture – 93
Current control for battery charger application
A very common and popular application of DC-DC converters is in battery charging,

battery chargers are made of switched mode DC-DC converters. So, you see that here is
the battery which needs to be charged, that battery of course will be used to drive some
load. Now this battery is interfaced through a DC-DC converter with some DC source.
Now that DC source could be a fuel cell or it could be a mains grid followed by a
rectifier or it could be photovoltaic cells. Very popular photovoltaic cells charging the
battery or wind hydro system connected to an alternator or a generator followed by a
rectifier which is used for charging the battery.
So, there are many many sources which can be used as the input source which can be
used to charge the battery. And as far as the DC-DC converter is concerned, you could
use any DC-DC converter. It could be the buck, the boost, the buck boost converter and
we have discussed the isolated converters like the forward converter, the dual switch
forward converter, push pull converter, half bridge converter, full bridge converter, fly
813
back converter all these converters are possible candidates to be used for the DC-DC
converter to charge the battery.
We have seen that in a DC-DC converter, the output voltage which is across the output
of the DC-DC converter can be regulated, and we also saw how to do the closed loop
operation. In this battery charging application we need to control the charge current that
you are pumping into the battery. So, it is a slightly different nature of control where you
are controlling the current the output current io of course, in the case the DC-DC
converter if it is a buck converter which is used, the output current and the inductor
current of the same.
When many cases the inductor current that is the state is controlled and which will
indirectly control the output current. If you are using a boost converter, the inductor
current which is the input current will be controlled. In the case the buck boost, the
inductor current is controlled. Generally for battery charging, the charging current which
is being pumped into the battery needs to be profiled. So, it is properly profiled then the
chemistry in the battery will work properly to extract as much life out of the battery as
possible.
However, what we will be discussing is not just output current control. We will be
discussing the inductor current control that is controlling in the inductor state, using any
one of the converters the buck, boost, buck boost converters.
The concept of current control is similar whichever be the converters, I will initially
explain with the buck converter using let us say photovoltaic cell input as the source
here, being the one of the most popular and growing application areas. You can use it
with either the boost or the buck boost or any other converter also. So, let us discuss the
DC-DC converter operation with current control, controlling the inductor current. So, in
the case of the buck converter that I will be discussing the inductor current control and
the output current control will become same.
Some of the advantage of current control is that there is inherent over current protection.
So, cycle by cycle instantaneous over current protection can be achieved with current
control. Another nice advantage spinoff benefit that you would get from current control
is that no flux walking will happen, you do not need flux walking capacitor. And that is a
814
significant benefit because as you start going for higher and higher powers, the capacitor
can become significantly costly in order to prevent flux walking.
So, let me consider a photovoltaic cell as the source because it is very popular source.
Now rooftop solar PV cells becoming more and more popular and many applications
related to photovoltaic cells are involved in charging the batteries. So, I will replace the
generic source with a photovoltaic cell. You may use any of the voltage sources, but I am
just using a photovoltaic source. It may as well be mains grid rectifier which we have
dealt with much earlier in the course.
So, this photovoltaic cell is used to charge the battery through this DC-DC converter, and
the DC-DC converter will be a buck converter that I will be using. You can use any
primary converter and we will be sensing the output current which will be the inductor
current and you that to control the current of the inductor. So, let me mark this is the
terminal voltage of the source, iPV is the current input current to the DC-DC converter.
So, now let us see how this battery charger is built. Let us try to understand the principle
of operation and then let me go into the circuit and give you the schematic for controlling
the DC-DC converter for current control and later I will be explaining some problems
that we will also later address by providing slope compensation.
815
Let us now discuss the battery charge a circuit where MPPT is integrated inbuilt into the
charger scenario. So, let us consider this PV module or PV array is connected to a DC-
DC converter and then the output the DC-DC converter is connected to a battery that
needs to be charged. So, you will say this is the DC-DC converter. It has a control input
d, the duty cycle input. The voltage across the terminals of the DC-DC converter output
is connected to the battery and therefore, the output of the DC-DC converter the voltage
is fixed by the battery because the battery is a source. Across the battery you may have a
load connected like this. So, you have a load connected like this.
Now, how do we charge the battery? The terminal voltage is vB and this voltage is
defined then and what would be the charge current. So, let us say we want to put the
peak power from the panel into the battery. Now let us say pm is the peak power, Pm by
vB will give you the iB the peak power that is going out of the DC-DC converter.
So, iB can be controlled by controlling the duty cycle because vB is fixed and therefore,
power can be controlled by controlling the duty cycle as vB is fixed. So, that is the
strategy that we want to adopt. Sense this current io which is flowing out of the terminals
of a DC-DC converter and use that as the feedback you also provide the PV current ipv
and also the terminal voltage of the PV vT which will be used for sensing the power that
is being drawn from the PV module.
Now, what is the control algorithm that you need to put in here? Let us consider this
boundary here the terminals of the DC-DC converter to which the battery is connected.
Now, normally if you consider these terminals and the battery is connected here, all
those portion PV source DC-DC converter can be considered as a controlled source.
Now let us say if it is a controlled voltage source plus and minus like this. If you have a
controlled voltage source, you cannot parallel to voltage sources because there is no
impedance in between and there can be huge circulating currents which can blow up
something in the control voltage source.
So, whenever you are having a control voltage source, you need to have a series
impedance which will limit the current. So, if you are having a controlled voltage source
have a series impedance, this is one possibility. But you do not want to control the
voltage because battery voltage is fixed you can directly control the current. In fact, a
controlled voltage source with the series impedance and if the impedance is inductive
816
non-dissipative that will form a controlled current source itself. So what can be also done
is across this boundary here you have the battery, you have a controlled source and that
controlled source you make it as a controlled current source. It has the inbuilt impedance,
so that there will always be a current limit, current source and the voltage sources can be
connected in parallel. So, we will follow this model and that is what we will introduce
into this control algorithm which is a current controller.
Let me draw and discuss the schematic for a buck converter based current controlled
battery charger. Now, what it basically means that the PV module or the PV array is
interfaced with a buck converter which will charge the battery. So, let us draw the circuit
schematic. So, you have the PV module, PV panel connected to a buffer capacitor CT at
the terminals and this was followed by a buck converter. So, I have a BJT switch. Now
this BJT switch can be replaced with a MOSFET switch or an IGBT power
semiconductor switch followed by a diode and I have an inductor and following the
inductor I am not going to place a capacitor. Instead of the capacitor I will connect a
battery in this fashion.
And across the battery you may have the load Ro like this. So, because the battery here
fixes the potential across the output terminals of the DC-DC converter, the current row
Ro is decoupled from the rest of the circuit operation. So, the current io through Ro is
Pbattery by Ro. Now here you have the battery voltage vB, we will call that as vB. The
817
current through the inductor is as shown here, this will be the current through the battery
iB and this current will be io which is flowing through the external load.
Now, our point of interest is this current, the current that is going to the battery load
combine should be such that maximum power is drawn from the PV array or the PV
module. So, let us draw the control schema. I have a comparator here plus minus so, it
compares the feedback signal with reference signal I will call the feedback signal as iL.
So, what I am going to feedback is the current? Recall our discussion voltage vB being
constant because of the voltage source connected across the terminal, it is sufficient to
feedback just the current and you start one to control the power and here you will have a
iL reference for this.
Now, this error goes through a comparator. So, I will connect this to the minus terminal
to the plus terminal I connect to ground 0. Of course, there is power supplies for the
comparator. The output of the comparator goes to the SR latch. So, I am going to connect
the out of the comparator to the R reset pin of the SR latch. There is set pin also and
there is a cube the output pin of the SR latch. The output of the pin of the SR latch is
going to get connected to this get drive circuit or base drive circuit in this fashion.
So, output of the SR latch Q gets connected to the base drive circuit which will drive this
on or off. So, how does this operate? Now, before that there is something we need to
connect to the set pin also. So, let me have clock block. So, the clock output is connected
to the set pin of the SR latch. How does this clock look like? So, let me have the time.
So, I will mark these sticks. Each of the stick space is this switching time period Ts.
So, the clock is in this fashion at the beginning of every switching time period, you will
have a very narrow, very small duty cycle pulse coming like that at the same period the
period being defined by this clock. So, this moment is small duty cycle pulse comes to
the set pin; it comes and sets the output of the latch on the rising edge. So, you have the
rising edge. So, at the rising edge we will see that Q gets set and at every start of the
period Q is set and then switches on the transistor.
Now, let us look at the various waveform to understand how this operates. So, I will
draw the various axis time axis basically and I will divide this into two portions. This is
the first cycle Ts and this is the later cycle. So, at the start of the cycle you have a small
duty cycle pulse and that is this clock pulse and it happens at every starting of the period.
818
Now, the moment this clock pulse is given, it sets this SR latch Q is high. Moment Q is
high there is base drive for the BJT or gear drive for the MOSFET and this is on and VT
gets connected and the inductor is charging up the inductor current is growing.
Now, let me set some reference here. Let me for the moment set a reference which is a
constant, clock then I will set a reference which is constant iL reference. So, when Q is
high, it is providing the base drive or gate drive to this for semiconductor switch and
there by the inductor current is integrating and having a positive slope and that positive
slope is vin minus vB; vbattery divided by L. So, it keeps increasing in the fashion.
This is iLref which is higher than i L. So, when iLref is higher than iL, the output of this
algebraic summer will be positive, but that is connected to the negative of the comparator
then the output will be low, the reset pin will be low and no change on Q. So, it keeps
increasing till the point when iL reaches and crosses over iLref. So, the output of the
algebraic summer becomes negative and this becomes positive, reset becomes positive
and therefore, Q will reset and go low. Moment Q goes low, there is no drive for this
power semiconductor switch. It will switch off and the inductor will start freewheeling
and start falling; current will start falling.
So, if you take this portion of the operation you will see that, if you look at the output of
the Q of the flip flop, it would have been high till this point and then it resets. So, during
this period when this is low inductor is freewheeling, you know the buck converter
operation it is freewheeling. This is connected to 0 because it is freewheeling like this
and what is connected to the industry is minus vo by L or minus vo in this case is vB,
minus vB by L.
And then again at the new cycle there is a clock coming in. This will trigger the SR latch
the Q goes high the cycle repeats. So, you will have the inductor current rising and then
at the point when iL reaches iLref, there is a change in the slope so, it keeps happening in
this fashion. So, observe that whatever current value you are setting iLref, the inductor
current the current that is going to the battery load combined never exceeds the iLref. So,
it is always below the iLref and gets controlled in this fashion and this is called current
control. It is as though every cycle there is a check for the current and the moment there
is an over current within a cycle, it is switched off. And then again the power
semiconductor switch is switched on, the next cycle when it is on till the point when the
819
current exceeds the reference current and again it switches off. So, it keeps on doing and
cycle by cycle there is current limiting and therefore, this is a very safe and fast method
of doing battery charging through current controlled mechanism.
820
Prof. L. Umanand
Lecture – 94
Instability in current control and slope compensation
In current control there is a problem of instability that will occur for duty cycle operations
greater than 50 percent. For duty cycle operation less than 50 percent the current control
is stable inherently. But for greater than 50 percent any error or any perturbation can get
amplified and this is a problem that you will see in most of the current control type of
topologies.
We will have a look at that, we will study this and then also see how we will solve this
problem and there is a standard solution for this called slope compensation. So, let me
draw the time line and I will split the time line into various periods dTs, 1- dTs and dTs so
on. So, let us take the inductor current, now this the current that we need to study. So, let
me draw this average inductor current which is Io and let me superimpose on that the actual
inductor current which will be in this form of a linearly rising and linearly falling wave
shape.
So, this slope rate you know the voltage across the inductor during dTs period divided by
L and the voltage across the inductor during the 1 - dTs period by L, but I am going to use
821
generic slope terms. So, let me say that this is the iL reference. So, whenever the inductor
current rises up and hits the iL reference then the switch will be turned off and the inductor
current starts to fall and then at this point the new period starts and the clock will reset,
then we will restart the switch and the switch will turn on and the inductor starts to charge
up.
So, this operation we just now had a look at how current control operates. Now, let me
indicate the slope, the rising slope as m1 this is a generic slope term it can be used for
representing the slope on any converter whether be buck converter or the boost converter
or any other isolated converters as well. So, this m1 is nothing but Δi by Δt during that
time which is the voltage during that time divided by the value of L and then here on the
down slope also you have m2. So, let me indicate that peak to peak is the ripple current Δ
iL.
So, now with these let us just formulate the problem. So, this slope m1: the rising slope is
nothing but ΔiL divided by this time period dTs and m2: the falling slope which is -ΔiL
by 1 - dT s. So, this is pretty evident from this wave shape. So, let me write m1 by m2, I
will take the modulus, I am going to take out the minus sign out of it which is d divided
by 1 – d, we will call this as equation 1 this is the ratio of the slope which is d by 1 - d.
Now, let us look at the m2 by m1 in from a different perspective, let me draw the inductor
current waveform once again inductor current iL and this is iL reference. Let me draw the
m1 slope part of the current waveform and then this is the m2 slope part the current
waveform, I am just going to draw for one cycle. Now, I am going to perturb let us say
due to some reason whatever the reason there has been a perturbation and that has come
at this point, it may have occurred previously it may be accumulated.
There is a perturbation it should have started from this, but it is going to the operating point
are shifted and then there is going to be the inductor current going in this fashion because
of this perturbation this error. And then once it goes in this point, at this point it will not
turn back because it has not yet hit the current reference.
So, it will go further hit the current reference here and then the switch turns off and then
you will see the inductor current going in this fashion this is the end of the period. So, if I
say this is the perturbation δ0, now this perturbation goes on to here and then appears on
822
the side as δ1. Now, the question that we need to ask is δ1 higher than δ0 or lower than
δ0? if δ1 is less than δ0 then I know that it is decreasing and then the next cycle it will
further decrease the next cycle further decrease and actually it will converge.
So, if the perturbation error converges, so if mod δ1 is less than mod δ0 then the
perturbation error will converge and ultimately it will decay down to 0. But the problem
is if δ1 is greater than δ0 then you are going to have an instability situation. So, let us check
that out. So, now, here I am going to draw this line and I am going to shade this triangle
look at this triangle and this is having a slope of m1 and then I will mark this horizontal
error as ex. Now, you can express this from basic trigonometry m1 is nothing but δ0 by
ex.
Now, let me draw the triangle this distance is the same it just rises up, this is the same
distance horizontal distance still remains the same ex. So, now, from this let me shade this
triangle. So, you can now say m2 is δ1 by ex, I have taken mod absolute value and now
with these two you can formulate m2 by m1 which is δ1 by δ0. And from here we can
substitute m2 by m1 with d by 1 - d so, you can say δ1 is δ0 * d by 1 - d. Now, this is the
important relationship that we have here.
So, what does this mean? Means that for d < 0.5 and 1 - d > 0.5 and therefore, this term is
less than 1 and δ1 is less than δ0. So, therefore, δ1 is less than δ0, δ2 is less than δ1. So,
on δn is less than δn - 1. So, therefore, gradually from δ0 you will see the perturbation
823
decreasing to δn and it is converging to 0 therefore, it is decay. So, this implies that the
perturbation progressively reduces to 0.
Now, for d > 0.5 : So, 0.5 divided by something less than 0.5 this is greater than 1 and δ1
> δ0, δ2 > δ1. So, on δn > δn - 1 which means the perturbation is progressively increasing
and it does not decay. So, therein comes the instability issue. So, this leads to instability
for d greater than 0.5. So, this is what we mean by saying instability in current control for
duty cycles greater than 0.5. So, there is a solution for this; there is a standard solution for
this which we call slope compensation. Let me let us also discuss about that.
The inductor current I will draw the time versus inductor current and I will keep the
markings in such a way that the duty cycle is less than 0.5. I want to show few cycles.
Now, let us say this is one cycle, second cycle, third cycle, fourth cycle now, these are the
duty cycle gaps. So, in the first cycle duty cycle less than 0.5 second cycle same duty cycle
less than 0.5 third cycle like that.
So, let me mark the upper tip and the bottom tip of the inductor ripple. So, let me mark it
like that and the inductor current will stay within this ripple. So, during the time of this
duty cycle, dTs inductor current is going to rise like this and during the time of the 1 minus
dTs it is going to fall; rise fall rise fall rise fall so on. Now this would be the normal
expected steady state value of the inductor current. Now, let us see what happens if there
is a disturbance.
824
Now, I will introduce a disturbance here, it could have occurred due to any number of
reason. I will say if this disturbance occurs, the inductor current will start on from here, it
is an integral, start on from here with the same slope Vin minus Vb divided by L.
So, it go parallel to this line. So, let me draw it parallel to that line and then after it hits this
top limit it will go parallel to the down slope of -Vb by L then up slope, down slope so on.
You see that, it progresses in this fashion. And take away from this is that whatever there
is the disturbance, finally decays to 0 and then if the disturbance is removed and the normal
inductance steady state value current flows.
Now, take the example of another case t versus iL. Now, I am going to increase the duty
cycle to beyond 0.5, same time periods duty cycle is higher, same time period duty cycle
is higher and so on. So, now, when the duty cycle is greater than 0.5 what happens? Let
me again mark the top and bottom of the inductor ripple, I will keep it same and what is
the steady state normal expected value the inductor current let me first draw that.
So, this will be Vin minus Vb by L, this will be -Vb by L negative slope, positive slope,
negative slope, positive slope, negative slope so on. So, this is the normal expected steady
state value the inductor current under duty cycle greater than 0.5 also.
Now, here also we will introduce a disturbance, the disturbance has occurred due to
whatever the reason. So, when you introduce the disturbance, now from this point onwards
it will start going parallel to this line, it will have the same inductor slope. So, let us say it
takes this line it hits the top value here that is the iL reference value, then goes down resets
again. Where this is the time duration when the Q of the flip-flop is low, hits the time the
end of Ts when the clock again gives a small pulse and sets the flip-flop, again it will rise
then goes down continues going down here it will get reset so on. And you see that the
disturbance starts growing, cycle by cycle it starts growing becomes unstable.
So, it never converges to the expected steady state value of the inductor current. So, when
the duty cycle is greater than 0.5 the inductor current will start diverging and it will not be
stable when you do current controlled operation in this fashion where this is the iL
reference, the top of the inductor current ripple is the iL reference. How do we solve this
problem let us seek a solution and try to find out how to do that and implement that.
825
Let me consider the problem case waveform. Here the duty cycle is greater than 0.5 and
we see that if we give a disturbance, the disturbance keeps growing and leads to an unstable
inductor current waveform.
Now, this line here, this top line here where the inductor current is comparing with that
line is the iL reference line. So, let me highlight that with a different color and let me
indicate that. So, that is iL ref. So, whenever the inductor current reaches iL ref and tries
to cross that then the error changes direction becomes negative and then resets the flip-
flop, but that is not happening in a proper way when the duty cycle is greater than 0.5.
Now, let us try to change the shape of this iL reference line instead of being a straight flat
line let us introduce some slope.
So, what I will do is try to extend this axis and then let me draw a line which is having the
same slope as the falling slope of the inductor. Now the falling slope of the inductor here
is -Vo by L or -Vb by L for the case of the buck converter. So, we will trying to use that
same slope continue it upwards, now this is the slope line that we will use for this time
period Ts we will repeat the same sloped line for every time period. So, using that as a
marker let me repeat the same slope line for the next time period then the next time period
too and so on. Now, this saw tooth type of waveform will become the new iL ref. So, when
you use the iL ref of this nature you see something nice happening.
826
Now, let us take the same disturbance and let me allow the inductor current to go parallel
to what is supposed to be the steady state value. So, it goes parallel, now hits the boundary
here it hits the iL reference here and at this point it tries to cross over. And the comparator
goes; the error goes negative the comparator output will see to it that the reset of the S - R
latch resets is asserted and resets the Q value. So, thereby it will then start having a down
slope and the down slope is going to be along the same parallel because the falling slope
for the inductor current is -Vo by L or -Vb by L in this case.
So, you see that in just one switching cycle the error has been reduced. So, the error is
from there on continues to follow the steady state value. So, you see that any error quickly
converges and the error is removed within a switching cycle, this is the beauty of having
this type of a sloped saw tooth shaped iL ref. So, you should give the slope here for the iL
ref same as the falling slope of the inductor current then you are safe and the error will be
removed within a cycle. So, how to give a slope now this was the old iL ref to this old iL
ref you have to compensate and see that you have this kind of a slope so this is called slope
compensation.
So, current control with slope compensation will give you the best results where even if
there are errors, the errors will be removed within a cycle and the system will be stable.
So, this can also be proved mathematically what I have just shown by graphs, you can refer
to literature on dc-dc converters or any of the NPTEL courses on dc-dc converters.
We shall now see how we can generate an iL reference which is of this form. It should be
in this kind of a saw tooth form and the slope being same as the falling slope of the inductor
current. And once we generate that then our current controlled converter is ready for
performing battery charging operation with MPPT.
827
Prof. L. Umanand
Lecture – 95
Slope compensated current control
Let us now try to understand what slope compensation is and how to obtain the slope
compensated reference waveform. Let us understand it with the help of some waveforms
with respect to time, I will plot iLref which is nothing but a constant dc value. So, the
inductor current waveform is compared with this constant dc and then that is how the
current controlled output is generated. But this has a problem for greater than 50 percent
duty cycle where it will become unstable.
So, therefore, we need to generate a slope compensated waveform. Let me now draw this
type of wave shape and this slope should be VB / L. So, this will have a magnitude of VB /
L and this is called i_slope let us say and the new reference will be iLref minus i_slope
and that will look like this. So, this will be the new current reference that we need to use
for doing the current control.
So, the inductor current should get compared with this slope compensated current
reference. So, this will be the slope compensated reference that we will use. Now, how do
we get this wave shape where the slope is same as VB / L that is the down slope of the
828
inductor current waveform. So, the question, how to generate this waveform? So, if we
look at the inductor current.
The inductor current here what we have drawn during 1 - dTs period during the time when
the switch is off, the slope of the inductor current is -Vo / L or -VB / L, the magnitude of
that slope is VB / L. So, this inductor by the Faraday equation the VL voltage across the
𝑑𝑖 𝑑𝑖
inductor is given by L𝑑𝑡. Let us write that down VL is equal to L 𝑑𝑡𝐿 and during the 1 - dTs
𝑉𝐵
period iL is - ∗ ∫ 𝑑𝑡 and for i_slope because here we are talking up of positive slope I
𝐿
remove the negative sign. Taking on considering only the magnitude of the slope will be
𝑉𝐵
∗ ∫ 𝑑𝑡.
𝐿
So, how do we obtain this current wave shape? How do we get this wave shape by
implementation? So, what we can have is we can have a controlled current source like this
and the controlled current source is connected to a capacitor like this. And let us say CS is
the capacitance and then you can monitor the voltage VCS across here and then let us say
there is a current I that you have decided for the current source. So, what is the voltage
1 1
across the capacitance? So, the voltage of the capacitance is 𝐶 *∫ 𝐼 ∗ 𝑑𝑡 which is 𝐶 because
𝑠 𝑠
I is a constant.
829
Now observe this and this, the coefficients of integral dt, VB / L and I / CS. Now comparing
𝑉𝐵 𝑉𝐵 ∗𝐶𝑠
these two; comparing these two we can write I as *CS. Now let us make I here ∗
𝐿 𝐿
1
𝐶𝑠
∫ 𝑑𝑡 would be the voltage across this capacitance. So, the voltage across this capacitance
will follow in this wave shape.
So, voltage equivalent value of i slope will follow this wave shape with this exact slope
VB / L because CS / CS will cancel. Now this was an integrator and as long as a current
here is constant this will keep on integrating, it will not come down to 0 so, there has to be
a reset mechanism. So, let us connect a switch across this. So, the current source I will set
𝑉𝐵 ∗𝐶𝑠
it at .
𝐿
Now, I will connect a switch there across the capacitance and drive that switch on or off
from the clock. So, whenever there is a clock so that will switch this on it will discharge
this capacitance bring it to 0. So, during that period of clock duty cycle this will come
down to 0, then again it starts integrating and then at this point clock will come into the
picture this will come down to 0. The clock waveform looks like this, we have seen this
earlier it is a set of narrow pulses occurring at every Ts period; at the beginning of every
Ts period is the very very narrow duty cycle pulse. It will turn on this switch short circuit
the capacitance the capacitance discharge through this one.
You can put a small resistance in series to probably limit the current. Now, this if you place
it along with these set of waveforms it will look in this fashion so let me draw so, these
are the positions when the clock will appear. So, you will see the clock signal. So, a very
narrow duty cycle pulse occurring at the start of every Ts period, it will reset the
capacitance and then bring it to 0.
Then the capacitor will keep integrating it will build up and at the occurrence of the next
clock it will reset to 0. Again it will start integrating and the occurrence of the next clock
starting on the T s period it will reset to 0 so, this will keep happening. So, in this way we
can generate this i_slope waveform and use it for obtaining the slope compensated
reference.
830
Now, let us take our current controlled converter charger circuit earlier we had iLref directly
here, iLref here was a constant dc value. So, we will replace it with this slope compensated
reference. So, what we will do is that we will remove iL there and I will say that this point
we are going to give a slope compensated reference. Now, how do we generate that? Let
us have iLref here shifted back a bit and I will compare it with this I slope. So, let me have
a controlled current source.
This controlled current source will have a current value i which is dependent on the sensed
𝑉𝐵
value of *CS. So, put a capacitance there and across the capacitance I will put a switch
𝐿
and the switch I will turn on based on this clock signal. So, whenever the clock gives a
high this will also turn this on reset this capacitor and then when this goes off this will start
integrating.
This resistance is put here just to limit the current and it will also put in a time constant
CS*R time constant into picture, but it will be a small time constant which will reach five
times time constant by the time the clock duty cycle finishes a duty cycle pulse.
So, this value what you would be giving there is i_slope and you subtract it from iLreference
and give this as the compensated slope compensated reference. Now if you look at this
wave shape, this wave shape is just the dc and what about this wave shape? So, let us first
𝑉𝐵
set the current reference here the current value is *CS, observe that VB is the sensed
𝐿
831
value. So, the current reference here will change according to the output voltage here and
if you do that then here you will get a positive slope sawtooth waveform like this; iLreference
minus this will give you a sawtooth waveform like this. And this will form the slope
compensated reference that you will provide to this controller.
So, now, iL whenever it is greater than the slope compensated value then the error here
will go negative and this will go positive and reset the S R latch. So, this is how the slope
compensated current controlled converter will work and charge the battery.
832
Prof. L. Umanand
Lecture – 96
Simulation of current control
Let us now see how we go about simulating the battery charger circuit with current
control with slope compensated current control. So, let me open this charger schematic in
gschem.
So, this is the schematic that I have for you; pv dot sub I think you are now familiar with
how to include the sub circuit file now this is the schematic. I discussed while discussing
in theory, I discussed about the buck converter, but here in the implementation I thought
we will use the buck boost converter.
This is the buck boost converter circuit, so that you will get familiarized with the one
more convertor in doing current mode control. So, we have this photovoltaic source it is
the same source that we had modeled earlier a short circuit current of 2 amps, voltage
scale factor of 20. And then I am having a buffer capacitor Ct this V iPV with value 0, V
iq with value 0 or the current sensing voltage sources have a switch here; this is the
inductor current inductor current measurement sensor a diode D at the battery. Observe
there is a battery is put in a position where here it is one is plus, two is minus.
833
Because the buck boost circuit you will have a negative voltage coming across the output
and therefore, you have to appropriately connect the battery terminals and of course, the
load. Now, how are we getting the PWM signals for duty cycle control? This is the SR
latch we have discussed about that, now there is this summing junction where I am
having iL measured. So, I am using the sense there is this source here and I am sensing it
from that source.
So, you see that this is a B source which is giving a voltage equivalent of the current
which is flowing in the inductor. So, voltage is equal to I current flowing through the V
iL voltage source. So, here basically I have a voltage equivalent of the inductor current
and I have the i Lreference coming in here which is i Lref minus the slope compensated
sawtooth waveform.
Now, I have interchange the plus and minus here if I zoom in, you will see that I have
made this plus I have made this minus this is used to make a simulation block simpler.
We have put a comparator here with a minus and plus with an inversion here and this
comparator block I have removed and then did a sign change here itself minus and plus.
The inversion caused by the comparator is brought forward and I have inverted the sign
here.
Because, there is a Schmitt comparator at the input to the SR latch itself. So, that is the
only change then if you go to the slope compensation thing now here I am having a B
834
source. And, B source I am getting a voltage equivalent of vo by L; L is 10 milli Henry
this one micro farad is coming because of this VB * CS by L. So, that is the formula that
I have used that we have discussed.
So, inductor value is 10 milli Henry that is what I have used. So, this is the formula.
Now, this voltage will drive this voltage controlled current source. So, there will be a
current that is coming out through this and then it will be charging up this capacitor. And
if we monitor the potential here you will get the voltage equivalent of I slope.
And, the every time there is a clock giving a pulse, it will also reset the switch short
circuiting the capacitor and discharging the capacitor. So, you will get a sawtooth
waveform, this sawtooth waveform subtracts from the reference waveform and that is
what is given as reference slope compensated reference.
So, this is exactly the block schematic that we are simulating and here this VC in the
case of MPPT case it will come from the output of the MPPT controller and we will
define the i ref. So now, let us simulate this. Here a point that you need to note is that this
SR latch is introduced now the sub circuit of this SR latch is included in pv dot sub.
And the symbol file for this is also included in the set of files and folders which had
provided as resource in the last weeks session. This is a Voltage Controlled Current
Source: VCCS and it is a spice simulation block itself. So, the rest all you are familiar
with and let us simulate this.
835
So, let me open the terminal window let me go to PV sim the folder, which contains all
these files. I will generate the netlist first charger dot net by executing the g netlist and
then now I will run ngspice charger dot cir. So, it will start simulating.
Let it simulate and after simulating we will look at some important waveforms.
836
Let us plot the i Lref and the inductor current and we will see. So, the red one is i Lref
which is at 3 amps, and let me expand it. So, and the blue one is the inductor current.
Now, you see that the inductor current is not reaching up to i Lref the reason is that we
have a slope compensated modified reference which is falling down from i Lref value.
And, when it is falling down from i Lref value the inductor current will cross it over
somewhere lesser than i Lref value, and the change state and then start in to the down
slope mode.
837
Let us observe the clock waveform and the voltage across the Cs capacitance. The
voltage across the Cs capacitance is supposed to be like a sawtooth waveform with a
positive slope. Let us compare them and see if we are getting waveform like we studied
in theory. So, let us plot voltage of the clock and the voltage across the Cs node.
So, let me expand that, let me select a very small narrow region and you will see that this
red pulse is the clock pulse very small duty cycle. So, whenever red pulse comes, there is
a reset of the voltage across the capacitance Cs and then it rises and again and then
starting of the next cycle clock pulse comes reset and then rises and so on.
You have control on deciding the duty cycle of the clock. So, this is as per what we are
expecting. Let us now see this modified i Lreference that is the slope compensated
reference here which is the i Lref minus i slope and compare it with the inductance
current iL so, plot.
838
So, let me plot the difference between these two and the iL. So, it will appear something
like that let us take a very narrow region so that we will be able to zoom it like that.
Now, here you see the blue line is iL and the red line is the modified or the slope
compensated reference. So, you see with the falling whenever the clock comes in here
then it starts falling subtracting again rises back, the i Lref was at 3 amps and from 3
amps it starts falling down, with the same slope as the inductor down slope. And so, the
839
inductor current, the moment it hits this or the down slope the SR latch changes state
switches of that device and then the inductor is falling so, on it keeps going.
So, this is also behaving exactly like what we anticipated and you see that the inductor
current is controlled to whatever value that you set. Now, let us see the voltage across the
PV panel and the current through the PV panel and just check how much amount of
power is being transferred to the load to charge the battery and the load.
So, if we check the power drawn from the PV panel see that it is hovering at around 24
close to the p power operating point. So therefore, you see that adjusting this i Lref value
plays a role in actually deciding the amount of power that is being drawn from the PV
panel.
Because the battery potential is more or less fixed more or less constant at around 12
volts. So, in this way this buck boost regulator that we have used here can be used as a
slope compensated current controlled charger for the battery, drawing power at
maximum power from the PV source. So, I will let you to explore this simulation on
yourself and try to gain more insight from this simulation schematic.
840
Prof. L. Umanand
Lecture – 97
Single phase inverter with sinusoidal PWM
Let us discuss now a single phase inverter with sinusoidal PWM. You have already studied
the full bridge circuit while studying the DC-DC converter the full bridge DC-DC
converter. So, let me draw the full bridge circuit. So, this is one arm of the full bridge
circuit you have a top switch semiconductor switch, bottom semiconductor switch it could
be either BJT or MOSFET, if there is an internal body diode for the MOSFET and for the
BJT if there is no internal body diode you have to put externally this freewheeling diode.
So, you also put another arm just like this arm. So, you have another set of BJT, top BJT
and a bottom switch in this fashion this is Vcc, this is ground. So, let us call this Q1, we
will call this is Q2, this is Q3, this is Q4 and in between here we will have the load. So, we
will have an R-L load let us say L and R. So, this is our full bridge converter. This full
bridge converter depending upon the way you modulate it, it could provide DC-DC
converter. If you put a transformer at center with center tap secondary or in this case you
are putting L-R the voltage across the centers or center points of the bridge will be AC
voltage.
841
Now, we want to get a special quality of the AC voltage which is having a fundamental of
50 Hertz sinusoidal. So, that when you filter it out you get back your fundamental
sinusoidal PWM. So, we need to modulate the switching pattern of the switches in such a
way that when you filter out the voltage waveform which is occurring here, you will get
the fundamental sinusoidal wave shape. So, let us see how we go about doing this.
Now the switches Q1 and Q2 should be mutually exclusive when Q1 is on Q2 is off so

that there is no direct shoot through current. And when Q2 is on Q1 should be off, likewise
Q3 and Q4 should be mutually exclusive, Q3 is on Q4 is off, Q4 is on Q3 should be off.
Now, therefore, we can say that if this is A and B and if you are giving a PWM here what
you give here is through a NOT gate.
So, it will be the inverted signal likewise. So, if we call that once PWMa, likewise on this
side also you have to give to Q4 the inverted signal. So, you can say this is PWMb. So,
now, how do we get this PWMa and PWMb and how do we get the sinusoidal PWM
pattern across the center arms of the bridge voltage VAB. So, let us have a look at that.
So, now let us draw some waveforms and try to understand how sinusoidal PWM can be
established. Now, let us take a time waveform and let me draw the modulating signal this
is supposed to be the sine wave modulating signal and it could be a 50 Hertz sine wave if
it is for a 50 hertz application. Now, across A-B after filtering we would like to have a sine
wave shape like that. So, let us compare this sine wave with a carrier. So, the carrier is a
triangular carrier which is between these two limits. So, let us draw that carrier wave form.
So, the triangular carrier is in this fashion, it is between these two limits and let us say
these limits this is the triangular carrier and it is between the limits of minus 1 to plus 1
this is 0.
So, in general you choose a carrier wave waveform frequency, which is greater than 20
times the frequency of this fundamental modulating signal. So, keep that in mind when
you are choosing the carrier frequency. So, the carrier frequency if you choose at least 1
kilohertz for a 50 hertz waveform, then it is exactly 20 times. So, anything greater than 1
kilohertz is a good choice so, but normally the frequency of switching of these switches
are much higher. So, it will be carrier frequency will be of the order of 20 kilohertz or 50
kilo hertz and the fundamental frequency is the order of 50 hertz. So, there is a sufficient
ratio much greater than 20.
842
So, next let us plot the waveform that you get PWM a. So, we will plot PWM a. So, these
are the intersection points, critical points where the modulating signal is cutting
intersecting with the carrier signal. So, let us mark that points and pull that down in this
fashion. So, the PWMa is generated in this fashion. So, you have a comparator plus minus.
So, you are having the modulating signal sine wave given to the plus carrier given to the
minus and the output of that is the PWM signal. So, which means that whenever the
modulating signal is higher than the carrier, you will get a high and whenever the
modulating signal is lower than the carrier you will get a low.
So, let us draw that. So, we get a low here then it goes high because modulating signal is
higher than the carrier, then in this time period you see that the modulating signal is lower
than the carrier let us mark that here the modulating signal is higher than the carrier. So on
you can plot draw the waveform PWMa in this fashion using this logic here.
Now, here this at the bottom Q2 switch, we have let us say PWMa bar. So, we can plot
PWMa bar which is let me drag down these critical time lines and invert PWMa to obtain
the PWMa bar in this fashion. So, this is PWMa bar. So, we have the switching pattern for
the arm-1 the A arm Q1 and Q2 given in these two fashion likewise we should generate
Q3 and Q4 also.
Now, for that do not use the same modulating signal, see if you are giving a modulating
signal to the A like this then for B you give the modulating signal which is 180 degrees
phase shifted so that when this is positive peak high this will be negative peak high so,
there will be voltage difference across A and B.
So, you do the following to generate PWMb and PWMb bar. So, let us give a 180 degrees
modulating signal, the same modulating signal pass it through a 180 degree inverter and
use that compared with the same triangle do not change the triangle with the same triangle,
and plot draw the PWM waveforms. So, let me mark the critical time points there is a
critical time point here with the red waveform. So, 180 degrees phase shifted waveform
there another critical point here; critical point here so, on you mark the points of
intersections.
Now, after having mark the points of intersections use the same logic as here whenever
the modulating signal is higher than the carrier you have a high at PWMb and whenever
843
the a modulating signal is lower than the carrier, you have a low. So, we have a high here
a low here, high and low in this fashion you see PWMb.
Compared PWMa and PWMb, PWMa you have more pulses more wider pulses during the
time when this is positive likewise PWMb has more wider pulses when the 180 degrees
phase shifted is more positive. So, PWMb in like manner can be obtained by basic
inversion. So, let us invert PWMb to obtain PWMb bar. So, this is a simple. So, now, you
have PWMb and PWMb bar. So, all the switches Q1, Q2, Q3, Q4 have the pulses in this
fashion.
Now, let us plot VAB and across that should be the pulse width modulated waveform it
should be the sinusoidal pulse width modulated waveform because it should contain within
it the fundamental sinusoid that is this fundamental sinusoid which we should be able to
extract by filtering. So, how does that wave shape look like? So, if you look at the
waveform if you look at the potential at A it is, if Q1 is on this is Vcc connected here if
Q2 is on A is connected to ground. So, this potential VA with respect to ground swings
Vcc to 0 so which will be something like the PWM waveform itself, but only the amplitude
will be Vcc.
Likewise potential at B with respect to ground will swing from Vcc to 0, it will be like
PWMb only the amplitude will be Vcc. So, the difference between PWMa scaled with Vcc
and PWMb scaled with Vcc, the difference between those two waveform should give you
the voltage waveform VAB. So, let us draw that VAB and let me draw down all the critical
points so that we it will be easy for us to draw the waveform.
So, now take this first part PWMa, PWMb. So, PWMa the A part and the potential at b
they are similar this is also at Vcc; at Vcc they will cancel and you will have 0 here. And
then here PWMa is having a pulse PWMb is not having a pulse. So, you will have a
positive pulse in this fashion where this is Vcc going after amplitude of Vcc. Then PWMb
is 0, PWMa is 0, VAB is also 0, PWMa is having a positive pulse, PWMb is 0. So,
therefore, AB it will be having a positive.
Now, here you have PWMa is positive pulse, PWMb is also having a positive pulse. So, it
will be 0 at the output VAB. And here PWMa is having a positive pulse PWMb is 0 you
have a positive VAB and both are 0 PWMa and b are 0 at that point. So, here it will be 0
again PWMa goes high PWMb being 0, here PWMa and PWMb are same cancelled and
844
become 0 and here you have PWMa, both are 0 here and at this point you have PWMb
which is positive PWMa is 0. So, it goes negative here and you have PWMa and PWMb
coming in together. So PWMb is positive PWMa is 0. So, you have negative VAB is
negative and 0 here again it is negative 0 because these two cancel negative 0 so on.
So, in this fashion you get VAB which is having pulse width modulation and this is the
pulse width modulated waveform, the widths are varying according to the sign pattern, and
this will have within it the fundamental sinusoid which was which was used for modulating
the carrier. So, when you filter it you will get the sinusoidal wave shape out of that filtering
out the high frequency carrier.
In the generation of the PWM signals for the various semiconductor switches Q1, Q2, Q3
and Q4, there is a practical problem and to solve that problem there is dead time that we
need to generate called the dead time generation. If you look at these waveforms here you
take PWMa and PWMa bar or PWMb and PWMb bar they are mutually exclusive 180
degrees out of phase. So, when Q1 is on Q2 is off, Q2 is on Q1 is off look at these two
waveforms.
They definitely are inverted 180 degrees out of phase; however, the problem occurs at the
transition look at this transition which is indicated as an instantaneous transition, and at
the instant when Q1 is going off Q2 is going on at the same instant. However, in practice
845
in reality things do not happen the switching on and off does not happen instantaneously
there is a finite time.
So, in the finite time you will see that Q1 is going from the on state to the off state and in
a finite time Q2 will be going from the off state to the on state and in between they will be
in the active region for some period of time when both will be conducting. During that
time there will be a shoot through directly from Vcc, Q1, Q2 to the ground. So, that is
called the shoot through current. This is not desirable because that will cause Q1 and Q2
to deteriorate or you may have to overrate the devices which is not actually a good thing
to do because it can be very costly.
So, in order not to have these shoot through current stresses, we provide dead time that is
we turn on this a bit earlier and we turn off this a bit earlier and we turn on Q2 a bit later
so that you can fix a period for period of time call the red time period when both Q1, Q2
are off that you are insured Q1 is off and Q2 is still not started on. So, this is a problem
that we need to solve. So, we are solving this by this dead time generation and let me
explain to you that let me take just only PWMa and PWMb, but it is valid for PWMb and
PWMb bar.
So, the PWMa and PWMa bar how do they look like? Let me just draw for one period,
one switching period PWMa bar is inverted in this fashion like this. So, this is the problem
this point of switchover transition PWMa is falling low PWMa bar is going high and also
during these times and what is it that we expect let me make some space here. So, what is
the new PWMa and PWMa bar that it should look like.
So, let me create a dead band zone and during this zone, both PWMa and PWMa bar should
be low, which means both the associated switches Q1 and Q2 should be off. So, which
means that PWMa should go low earlier and come up later in this fashion, going down
should be earlier say likewise PWMa bar should go up later and come down earlier in this
fashion. So, in this fashion you have a dead band here and that is called the dead time and
this can be a fix at time and it is normally very very small percentage of the on times and
this is the order of 1 to 2 microseconds.
So, how do we implement that? So, normally if you take the PWM generation scheme,
you have v m that the modulating signal which is the low frequency sine wave vc which
is the carrier signals the high frequency triangular carrier compared in this fashion plus
846
and minus and you get the PWM output. And that PWMa that is inverted and you will get
PWMa bar. So, this is the normal way of generating as we discussed, but let us now
generate with which will give you this dead time generation inherently.
So, let us first have this carrier, now this carrier this is now translated. So, let me have this
vepsilon a very small DC voltage which is added and subtracted to this carrier. So, this
carrier plus and minus here. So, if you see this is vc plus which is small epsilon DC bias
is added to the carrier here, and to this vm, vm is now compared with vc plus and that will
generate PWMa.
And in another case you have vepsilon subtracted from the carrier which means it is
brought down a bit translate at level translated towards the negative side average, and that
is called vc minus and that is actually added that is plus sign and vm is minus subtracted
and you will get the PWMa bar. So, this is the scheme that we will use for generating
PWMa and PWMa bar and if it is for the other phase PWMb and PWMb bar similar such
thing. Let us just have a look at the wave shapes to understand this a bit.
So, let me have between these two limits, what is supposed to be the actual carrier? But
this is not the carrier that we will use, we will add a small epsilon to this carrier and get vc
plus and vc minus. So, when you add a small epsilon vc plus will look like this, it is shifted
up by a small epsilon. So, it is going to give me the epsilon is converted as a time shift
847
here and if I subtract it vc, vepsilon is subtracted from vc you will get vc minus. So, it is
in this fashion it is now slightly lower.
So, you use vc plus and vc minus for comparison, vc minus and vc plus is used for
comparison. So, with vc plus we are doing the regular comparison of the modulating signal
to the plus terminal of the comparator vc plus is given to the minus terminal PWM a is
obtained. So, let us get that. So, this is the modulating signal now let me draw the crucial
points here these are the crucial points where the modulating signal is intersecting with vc
plus waveform the green modulating signal intersecting with vc plus modulating signal
intersecting with vc plus so on.
Now, let us draw the PWMa waveform PWMa; what is PWMa? Whenever the modulating
signal is greater than vc plus it will be a high. So, here it is greater than vc plus high, here
vc plus is higher you get a low at PWMa then high here at this point. So, in this fashion
you get the PWMa waveform, what is PWMa bar waveform? vc plus is given to the plus
terminals and modulate signal the minus terminals. So, whenever vc minus is greater than
vm you get a high, whenever vc minus the now let me draw the waveform for v it PWM a
bar and let me mark the critical points of intersection of vm with vc minus that is a red
triangle.
So, whenever vc minus is higher than vm it is having positive like this. So, you see that
you have this dead band a fixed set dead band as determined by vepsilon. So, vepsilon will
determine what is this fixed dead band irrespective of the pulse width this fixed dead band
will be given between PWMa and PWMa bar. So, this is the dead band that will be used
to protect or avoid any shoot through currents in the arms of the full bridge circuit.
So, in this way you can implement a sinusoidal PWM with dead band generation or a dead
time generation to obtain single phase sinusoidal pulse width modulated inverter.
848
Prof. L. Umanand
Lecture – 98
Simulation of sinusoidal PWM
Let us now look at the Simulation of the sinusoidal PWM circuit. So, this is the
schematic of the sinusoidal PWM circuit. There are some things which have written
outside the page marker. So, here is the power supply, I am using power supply here and
this is the reference sign sources, modulating sources. And on this side I am having a
filter just to see whether you will be able to get back the fundamental modulating signal
and what is within the page border is the actual PWM circuit.
So, let us examine this. So, let me expand. So, this portion here this is a simple triangle
generator which I have made using op-amps, so you see here this op-amp across the
minus in the output, I am having a capacitor and then a resistor connected to the minus
point coming from some signal. So, this portion here, this portion which I am indicating
through the cursor is nothing, but an integrator. So, this is integrating, the output of the
integrator I am giving it to a comparator is a comparator with hysteresis. So, giving it to
the comparator, so, once the value of the integrator crosses this threshold, the trip points
you will see the output here going high or low.
849
So, if it is going high, then this integrator is integrating down. If this goes low negative,
then it is integrating up because the inverting integrator. So, the output here will be a
triangular waveform going up and down both positive and negative. So, this is the
triangle portion. So, to the triangle what do we do? So, to the triangle we will add a small
epsilon voltage to get the vc plus and that is the upper triangle and to this we will
subtract a small voltage or the epsilon voltage to get the lower triangle. So, how do we
get the epsilon voltage? I am having this Vcc to Vee potential division. This potential
division is not of equal resistance. It is slightly offset on the upper side I have 10.1 k and
it is 10 k here on a similar attenuator.
On the lower side I have 10.1 k and it is 10 k here. So, when you take the centre points
here, this will be slightly negative and this will be slightly positive. So, I am calling this
one as n d lower, n d upper. So, use that one to add to the triangle. So, to the triangle you
are subtracting get using a different stage and you will get the triangle upper and then use
the other one to subtract different stage, I get the triangle lower. So, this will be vc plus
and this will be vc minus carrier plus and carrier minus. Now, use these two upper and
lower carriers to obtain the PWM with dead band or the dead time.
So, now if you go to this portion, the comparator section, so you have the triangle upper
connected to the minus and to the plus you are having the sign reference. So, the sine
reference is this is again a hysteresis comparator and the output of that will be the pwm.
Now, the sine reference given to the minus and a triangle lower or Vc minus given to the
plus, this will give you the PWM bar PWM a bar. If this is the for the PWM a this will
represent PWM a bar with the dead time. Likewise for the other arm, you use the
reference B compared with triangle with the triangle upper career and you will get n B
and then use the reference to minus and the triangle lower triangle lower to the plus, you
will get n B bar.
So, these 4 signals you used to drive the switches top bottom of one arm, top bottom of
another arm. So, if these are given to the full bridge switches, they will generate the
desired PWM waveform at the centre of the bridge. For the modulating signal reference I
am using sine just giving a amplitude of 3 and 50 Hertz and for the other arm B arm I am
giving the same sine signal, but minus 10 milli seconds 180 degrees phase shifted. Now I
am having this library. Within the library is the sub circuit for the op amp. Also I have
used a generic op amp. Have a look at that one.
850
And this here filter across the bridge arm A to B, I just put an RC filter to see if you will
be able to extract the modulating signal fundamental modulating signal. Of course if you
put a higher value of capacitance, you make it a much better purer sign signal. So, now
let us simulate this circuit and to simulate this let us go to the terminal window.
So, I open the terminal window. So, here I will run the simulation. So, before I run the
simulation, I would like to show to you the dot cir file. Let us see the dot cir file. So,
within in the dot cir file, so we have the trans statement I am going to see up to 40
milliseconds which is 2 cycles.
I am going to use this PWM net and the usual making the background colour white
background colour, black and running the simulation. What am I going to see? I am
going to see nA and nA bar and the voltage across the centre of nA minus nB that is
across the centre arm of the bridge of the full bridge. Let us now go back to the terminal
and run the simulation. I will use runsim pwm. So, it will generate the netlist and then
run execute the simulation.
851
So, this is basically the PWM waveform. The red one here is nA, the blue one is nA bar.
You are not able to see the blue one here because it is hidden behind the red one.
Probably I will expand that and show, you will be able to see the blue here.
So, the nA bar nA, nA bar nA they are mutually exclusive that are being generated; the
angle rise time fall time is because of the print steps. So, it is calculating at much higher
resolution, but it is printing at a lower resolution of 10 micro seconds. Then coming back
852
here the yellow portion is the PWM signal which you are seeing across the full bridge
center arm the center portion.
Just to cross check I can see the voltage across the filter. Observe that I have given nA;
nA to here nB to here and I have put an RCR filter and across the C, I am observing there
is there are two labels nlA and nlB and across nlA and nlB you will see the filtered
portion of the PWM signal. So, let us do that you should see a sine. So, let me plot
voltage.
nlA comma nlB, this case in sensitive.
853
So, you will see that here that the filtered portion I have put a very low value of C, but
you can as well put a large value of C and get a much smoother wave shape. Also note
that there will be some harmonics due to the dead time that we are giving and the dead
time will introduce the harmonics and it will not be as pure as sine wave has expected.
The smaller the dead time, the better will be the sinusoidal extraction of filtering. So,
now I will leave the circuit for you to explore.
854
THIS BOOK IS
NOT FOR SALE
NOR COMMERCIAL USE
(044) 2257 5905/08 nptel.ac.in swayam.gov.in

Ref Nptel1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ref Nptel1

Uploaded by

Copyright:

Available Formats

INDEX

I am Vivek Agarwal. I am a faculty at IIT Bombay in Mumbai, and welcome to this

(Refer Slide Time: 06:09)

(Refer Slide Time: 36:47)

(Refer Slide Time: 42:35)

Thank you very much.

(Refer Slide Time: 04:51)

(Refer Slide Time: 14:24)

(Refer Slide Time: 25:38)

(Refer Slide Time: 32:23)

I thank you for your attention.

(Refer Slide Time: 00:57)

(Refer Slide Time: 03:06)

(Refer Slide Time: 04:23)

Here is another example of a renewable energy system in which a solar PV panel is

(Refer Slide Time: 08:17)

(Refer Slide Time: 09:37)

(Refer Slide Time: 11:13)

(Refer Slide Time: 20:25)

(Refer Slide Time: 22:09)

(Refer Slide Time: 25:20)

(Refer Slide Time: 27:17)

(Refer Slide Time: 34:17)

(Refer Slide Time: 37:56)

(Refer Slide Time: 39:12)

Similarly, there is a possibility of a switch which is an inductive switch which actually

(Refer Slide Time: 40:40)

Pac = d1( 1-d1 ) (V2-V1) (I2-I1)

(Refer Slide Time: 42:52)

(Refer Slide Time: 44:12)

When we are talking about DC to AC conversion, in fact, even in DC to DC conversion

Thank you very much for your attention.

(Refer Slide Time: 00:33)

Welcome back to Fundamentals of Power Electronics. Any power electronic system

(Refer Slide Time: 05:28)

(Refer Slide Time: 07:10)

(Refer Slide Time: 08:11)

(Refer Slide Time: 11:03)

(Refer Slide Time: 11:34)

(Refer Slide Time: 12:33)

(Refer Slide Time: 14:22)

(Refer Slide Time: 18:08)

(Refer Slide Time: 28:21)

(Refer Slide Time: 29:44)

(Refer Slide Time: 30:28)

where L is the proportionality constant between the flux linkage and i.

where, dA represents an elemental surface area.

(Refer Slide Time: 36:43)

Where we used the fact that

F = φ / (π d2/4) . (2 πa/ μo)

R = 2 πa/ μo(π d2/4) A/wb

(Refer Slide Time: 41:22)

Eturn = N dΦ/dt = d λ /dt

The value of inductance L can be obtained as

(Refer Slide Time: 44:12)

(Refer Slide Time: 46:52)

(Refer Slide Time: 47:42)

(Refer Slide Time: 04:13)

(Refer Slide Time: 06:29)

(Refer Slide Time: 09:17)

(Refer Slide Time: 10:05)

(Refer Slide Time: 13:55)

And if I define this to be equal to d/dt then in my equations in my differential equation

(Refer Slide Time: 16:07)