http://www.ab.com/drives/techpapers/qanda01.

htm

(FAQ) Q AND A ABOUT AJUSTABLE FREQUENCY DRIVES

The purpose of this page is to provide complete answers to questions concerning adjustable
frequency drives and related topics.
Questions and Answers are listed under various topics
Power Factor Correction and AFDs
AFD Harmonic damaging
Reflected Waves and the AC motor

Power Factor Correction and AFDs

Question:
Can power factor correction capacitors be applied to an adjustable frequency drive controlled motor?
Answer:
In general, power factor correction capacitors should not be used with AFDs. Power factor correction
capacitors are used to create a shorter, alternate electrical path for the reactive current component found in
inductive loads, such as motors and transformers. This alternate path reduces the electrical current demand
placed on the utility or supply source. As the load on the motor decreases, the power factor of the motor also
decreases however, since the total current is decreasing, the reactive component of the current also
decreases. In this way, the current through the capacitors remains fairly constant. The capacitors are sized
(selected) as a KVAR device and as such are rated for only a small portion of the total user load.
The capacitors do not control the amount of current flowing through the capacitor. The inductive load controls
the amount of current flowing through the capacitor. The capacitor creates an alternate path for the reactive
motor current. If the capacitors are located close to the motor, the reactive current electrical path is short. If
the capacitors are located at the utility/facility interface, the reactive current path is longer but still reflects a
higher power factor to the utility and allows the distribution and transmission capabilities of the utilities to
handle more real power transfer.
When capacitors exist on a distribution system they are a source of energy. The capacitor voltage will supply
current to any load connected to the capacitor. If a resistor was connected across a capacitor, current will
flow between the capacitor and the resistor load. Real power will exist during this transfer since there is no
reactive current component. The capacitor current is controlled by the capacitor voltage and the value of the
resistor. In a similar manner, when PWM (pulse width modulation) drives are used to control a motor, the
capacitor current is controlled by the capacitor voltage and the input impedance of the PWM drive. Since
PWM drives transfer real power through the drive to the motor, the current demand on the power factor
correction capacitor is real and is a larger portion of the load compared to the reactive portion of the motor
for which the capacitor was originally selected. Thus, the current is forced to carry more current than it was
selected to handle. Some "experts" will state that harmonic currents cause power factor correction capacitors
to overheat, capacitor fuses to open. Although it is true that capacitors can be damaged and fuses will open,
the real reason this occurs is due to the change forced upon the capacitors. Having been selected for KVAR
loads, the capacitor must handle the KVA load and essentially are undersized. Dependent on how effectively
the distribution source (i.e. transformer) works with the power factor correction capacitors will determine how
much extra work the capacitors must do. If the capacitors are located close to the transformer source, the
capacitors are less likely to experience the full change from KVAR to KVA. However, if the capacitors are
located closer to the PWM AFD, they will experience a greater demand for load current and experience more
problems.
The solution to potential problems between PWM AFDs and power factor correction capacitors is to remove
the capacitors from the distribution system, since they are no longer needed to improve the power factor of
the controlled motor. If fixed speed motors exist on the same distribution system with PWM AFD controlled
motors, then locating the capacitors at least 250 electrical feet from the PWM AFD will reduce the stress on
the capacitors however, this adds losses into the system.
It is important that the combination of capacitors and reactive components connected to a distribution system
does not create resonance (i.e. the combination is tuned to resonate at some frequency). The current pulses
associated with PWM AFDs may excite a distribution system that is tuned to harmonic frequencies contained
within the current pulses. If the distribution system is not a tuned circuit prior to adding PWM AFDs, then the
addition of AFDs will not create a tuned circuit. If the distribution system is tuned (i.e. resonant) then
components can be added to either retune or filter the frequencies which can prevent sustained resonance.
The simplest solution may be to retune by relocating the capacitors thus adding or subtracting inductance in
the system circuit.

AFD Harmonic damaging

Question:
How damaging are AFD harmonics to a motor?
Answer:
Harmonics are used to describe the voltage and current waveforms that exist in a distribution system. The
harmonics associated with AFDs are not likely to have any destructive impact on the motor. You can answer
your own question by asking how can voltage or current damage the motor. When harmonics or the
waveform shape exceed values which are destructive to motors then problems will occur. With AFDs, there
are two locations where the voltage waveforms are different than the sinusoidal waveform expected from the
utility.
When the AFDs draw current from the utility, current pulses rather than sinewave exist. These current pulses
caused voltage drops in the system, similar to the sinewave voltage drops except there are other frequencies
(i.e. harmonic) which cause the shape of the voltage to become slightly distorted. With a sinewave voltage
drop, a loss in voltage (i.e. voltage regulation) would occur. With current pulses, the voltage regulation turns
into voltage distortion. If the amount of distortion were to cause large voltage changes, motors connected
directly to that changing voltage would be affected. Fortunately, the harmonic content of the current pulses is
small. This results in small voltage distortion at frequencies which would impact motors. Typical harmonic
voltage distortion that exists on the distribution system is too small to impact motors connected direct on line.
Typical harmonic values of 20 volts when compared to the 460 volts of 60 Hz will be ignored by the motor.
Recall that the force or torque experienced in the motor is a function of the square of the applied voltage. At
low harmonic voltage values, the possible damaging forces on the motor results in less than a few degrees
of motion on the shaft of the motor while the motor is operating at its nameplate speed. It would takes large
amounts of harmonic voltages to seriously impact the motor. Which leads us to the next location where
harmonics exist.
The output voltage waveform of the AFD is a series of pulses which change a fixed dc voltage source,
internal in the AFD, to a variable width square voltage to simulate an adjustable sinewave of voltage. The
motor expected to receive a certain volt-seconds area with a symmetrical change in polarity. Whether the
volt-second area is contained within a sinewave or contained within a series of pulses makes no difference to
the motor. The motor performs as long as the correct volt-second area exists. Before the days of PWM, the
voltage waveform sent to the motor by AFDs were single pulses displaced by 120 degrees to create a
simulated 3 phase voltage. The harmonic content of that voltage waveform represented a harmonic voltage
distortion of 33% THD. After many years of installation and application, it was proven that as potentially
damaging as that types of waveform was, the impact on the motor was minor. General practice for selection
and application of motor experiencing that waveform was to select a 1.15 S.F motor and derate that motor to
1.0 S.F. This allows the extra heat that was created to dissipate properly and allowed the motor to operate
without problems. As a side note, this poor waveform did not allow motors to operate at low speeds in
constant torque applications. Today things have changed. With PWM techniques applied to the motor, less
heating, wider operating speed ranges are now possible. The full load motor current resulting from today's
PWM AFD applied voltage waveforms creates less heat in the motor when compared against direct on line
operation. Today's PWM AFD can be fine tuned to the application eliminated all possible losses from the
motor and system.
Note that there are other types of motor controls. Some contribute more harmonics to both the distribution
system and the motor. Since most ac motors are controlled by PWM AFDs, concern about harmonics,
causes by PWM AFDs, impacted either the ac line or the motor should be minimum. Harmonics on the
distribution system, are of concern, however information, not fear should be used to guide in the selection of
equipment which can improve performance and increase operating efficiency.
Another topic associated with PWM AFDs and their impact on motor is the Reflected Wave Issue. This is
sometimes confused with harmonics. It is a significantly different issue and is covering under the topic of
motors and reflected waves.

Reflected Waves and the AC motor

Question:
What are Reflected Waves and will they damage motors?
Answer:
Reflected waves occur due to a mismatch in impedance between the motor cables and the motor. You may
be familiar with the requirement, in audio sound equipment the importance of matching the impedance of you
speakers to the output of the amplifier. Matching impedance insures that the full range of frequencies is
transferred to the speakers. If a mismatch in impedance exists some of the frequencies do not get to the
speaker. They are reflected back to the amplifier.
The PWM pulses at the output of the 460 volt AFD change in value from zero to 650
volts and contain many high frequencies. The leading edge of each pulse represents
the higher frequencies. The faster the voltage changes (i.e. rise time) the greater the
number of high frequencies. These pulses are sent to the motor as the modulated voltage supply. The low
frequencies are accepted by the motor but the higher frequencies are turned away because of the difference
in impedance between the motor cable and the motor. The higher frequencies are turned back because
higher frequencies create a greater difference in impedance than do lower frequencies. The frequencies that
are turned back or reflected combine with the pulses coming from the AFD to form a different shaped
waveform.
Because of the wide variation in impedance between different motor cable and motor
ratings, accurate definitions of the exact shape of the new waveform is difficult to predict.
What is known is that a minimum length of motor cable is required in order to achieve the
maximum possible peak voltage value for the reflected wave. What the means is for short
(10 to 30 ft) cable distances, it is unlikely that the full value of peak voltage will occur.
However, as the distance of the motor cable increases, the value of the peak voltage will increase until a
value of twice normal pulse amplitude is achieved. Increasing the motor cable distance beyond that point will
not cause the peak voltage to increase. Whether the motor is impacted by this reflected wave is dependent
on how good the insulation is in the motor. With smaller HP rated motors, there is less physical space in the
motor so a greater risk for insulation breakdown exists. As the HP rating of the motor become greater than 2
HP, the physical space increases thus decreasing the risk for insulation breakdown.
Of the many motors that have been installed with AFDs, the occurrence of insulation breakdown, due to
reflected wave, is small. Solutions to reduce the peak voltage that can be created include adding a reactor
between the AFD and the motor or add a path circuit at the motor to absorb the high frequencies that are
being reflected. Remember that the high frequencies reflected from the motor will travel in a path that has the
lowest impedance. Adding a circuit at the motor which attracts high frequencies will prevent the formation of
high peak voltages at the motor. Of all the motors that are manufactured, it is possible that some
imperfections may exist in a small percentage of motors.

http://www.ab.com/drives/techpapers/pwrqual.htm

POWER QUALITY AND THE AFD OVERVIEW

Overview
In the early 1990's, harmonics became the buzzword in the electrical power quality arena. Electrical
disturbances, causing equipment shutdowns and damage created a growing list of uncomfortable users of
electrical power. Concerns and questions, about overloaded neutral conductors, stray voltages and
overheated transformers, were often met with unacceptable answers and limited solutions.
In an attempt to define these seemingly new power quality issues, the electrical utility industry began to
search for reasons and solutions that would help their customers. Since the problem was not a new problem
but an increasing problem, it was assumed that newly introduced or installed equipment could likely be the
cause. As the utility began to investigate power usage, another equipment problem was identified. As part of
the energy conservation interest, more ac motor controllers (Adjustable Frequency Drives) were being
installed. As the use of AFDs increased so did the complaints about AFDs shutdowns due to low or high line
voltage conditions. This increase in complaints coupled with overloaded neutral conductors, stray voltages
and overheated transformer concerns resulted in overstating the impact of harmonics.
Adjustable Speed Drives (ASDs) became an easy target as those who hoped to make a "buck" and those
who did not have a glimmer of understanding about the wide variation among ASDs began swinging the
"Fear all Harmonics Banner". A hasty renovation of the 1981 version of IEEE-519 was rushed through to
meet the "next 5 year" review time target. Everything but the "kitchen sink" was dumped into the standard
with little thought about how the lack of details would have a significant impact on increasing the cost of
transferring energy and reduce the reliability of motor control systems. Harmonic Voltage Distortion clearly
and correctly defined within the 1981 version of IEEE-519 was quickly overshadowed by a "new" set of
recommendation for limits on Harmonic Current Distortion.
With today’s AFD proliferation, there are more concerns about other factors in power quality. EMI, RFI and
Harmonics have become frequently used terms. The good news is the BPWM (PWM with series inductance)
drives contribute little to the negative side of Power Quality. The bad news is that lack of information and
misleading information results in many potential users missing the many benefits than can be obtained by
applying AFDs. Fear and ignorance prevailed and are supported by “snake oil peddlers”. Customers need
knowledge to assist in decision making.
There are changes within power distribution systems to be reckoned with however, most changes will result
because of poor grounding, poor wire routing practices and momentary loss of electrical power.
This paper will help in understanding the nature of the type of power line disturbances that are encountered
in many installations and will aid in correcting the occasional power problem that occurs. Of all power line
disturbances, Harmonics is the least understood and most feared. This paper will present a broad overview
of harmonics.
Harmonic overview
To begin to understand harmonics, 5 questions need to be answered.
What are Harmonics?
What do Harmonics do?
What are major contributors of Harmonics?
How to determine if a harmonic problem exists?
How to address a harmonic problem.?
What are Harmonics?
Harmonics is a term used to describe the shape or characteristic of a voltage or current waveform with
respect to the fundamental frequency in an electrical distribution system. Harmonics are not a thing but a
way to define current or voltage. With the intent of efficiently transferring power from the electrical system to
loads such as motors, eliminating any reactive component of current is a practical goal. Further ensuring that
the characteristic of any reactive component of current is limited to a single frequency is also important.
Multiple frequencies or harmonic content can increase the heating losses that were previously definable with
a single frequency waveform. If no reactive component of current exists, then the characteristic of the real
component of current can be reduced to a concern for the peak and rms values that exist in the current
waveform .
If no harmonics exist, then the waveform is described as the fundamental sinewave. A pure sinewave of
voltage supplied by the electric utility can be described using a rms value and a frequency. For example: a
480 VAC, 60 Hertz sinewave will have an ideal shape with an inverted image about the 180 degree point and
mirrored images about the 90 degree and 270 degree points. Its peak value is 1.414 times the rms value with
a time base equal to 1/60 Hz or 16.67 milliseconds.
Basically its shape can be described by a single frequency. Because it is a single frequency, its impact on the
electrical system can be determined using simple rules like E x I = Watts. When the waveform can no longer
be described by a single frequency, the waveform must be separated into individual, definable, frequency
components. Then each frequency component or harmonic can be dealt with separately. A harmonic is a
waveform with a multiple of the fundamental frequency. A voltage with a frequency of 120 Hz would be the
2nd harmonic in a 60 Hz electrical system. A 180 Hz waveform would be the 3rd harmonic in a 60 Hz
electrical system. A 300 Hz waveform would be the 5th harmonic. The 5th harmonic is normally associated
with the reactive current in inductive circuits like motors and transformers. 5th harmonic current are
considered negative sequence currents that produce negative power and should be filtered out of the
system. Care should be taken to insure that current that looks like 5th harmonics are truly negative before
using trap filters.
What do harmonics do?
A complex waveform makes the calculation of power and watt loss difficult. One way to make those tasks
easier is to take apart the complex waveform and make individual waveforms, each with a single frequency.
The simple rules can be applied to each frequency to determine how transformers, motors, capacitors, wires,
and other components in the distribution system will behave.
After each individual waveform is analyzed for power and watt loss, the results can be recombined to obtain
the total power and watt loss. Complex waveforms do not create more losses. They will determine how the
simple rules are applied. Operating a 100 watt linear light bulb consumes 100 watts. Operating a non linear
100 watt light bulb consumes 100 watts. The complex current waveform of the non linear light bulb does not
create more losses. It merely transfers the current in a shorter period of time. This means higher peak
current, not more current since the time is shorter.. This may change the rules about rms heating associated
with the current.
One useful document available to qualify waveforms created by power converters and rectifiers is IEEE-519.
This document was intended to establish recommended levels of voltage distortion, where utility power
enters a facility. Recommendations within IEEE-519 are not mandatory but are guides to good electrical
practices. Current distortion recommendations, for injected current, have been presented in IEEE-519 as
benchmarks whose purpose is to limit levels of voltage distortion. Using current distortion percentages, for
the real component of current, as a gauge to prevent voltage distortion only leads to adding waste in
equipment cost and energy usage. When the displacement power factor in near unity, the only features of
current that need to be addressed are the maximum rms value and the maximum peak value. Distributing the
current demand over the full voltage cycle is a useful method to use to minimize current peaks.
Harmonics exist in all electrical systems. Within a 3 phase electrical system, a 3rd harmonic voltage
waveform exists simultaneously with the three phases of voltage. The 3rd harmonic voltage is a natural part
of the 3 phase system. In all 3 phase systems, the sum of the current in the system must be equal to zero. All
current in one phase must find a return path through one of the other phases. If a return path does not exist,
current will not flow. In the event of fault, current will flow from the phase through the fault and return through
the neutral conductor. This is an abnormal condition, so the neutral conductor is not normally sized to handle
the fault current on a continuous basis. The neutral conductor only handles fault current until the fuse or
circuit opens. It is not uncommon to find that the neutral conductor equal or smaller than the conductor size
used for each phase. It is commonly understood and accepted that the neutral conductor does not normally
carry current in a 3 phase balanced linear load system.
Unfortunately, times have changed and more non linear loads are being applied to electrical distribution
systems. Most non linear loads have been computers, TVs, VCRs, and electronic lighting. These loads are
referenced to the neutral so it is not uncommon for currents to flow in the neutral because there is no other
path.
When the system is supplied by a delta source, some current could circulate in the delta. When the system is
supplied by a Wye source, then current will flow in the neutral conductor. With 3 phase equipment like AFDs,
no load current will flow in the neutral because all current flows between phases. Leakage current may flow
in the neutral because of capacitive coupling that exist between conductors and from motor windings to
ground.
One of the concerns about what harmonics do is their impact on the neutral conductor. It has been stated
that the 3rd harmonic causes neutrals to become overloaded. Neutrals overheat because they carry more
current, not harmonics. With balanced linear loads, in a 3 phase circuit, the neutral is not used as the return
conductor. With single phase non linear loads, the neutral is always used as the return conductor. Balancing
non linear loads, in single phase systems, will not prevent the neutral from being used as the return
conductor. This means that the neutral of a distribution system must be sized to handle the square root of the
sum of the squares of the phase current.
Assume that each phase of a 3 phase system is providing non linear loads with 5 amps. The neutral must be
sized to handle the square root of (5 2 + 52 + 52) or the square root of 75 or 8.67 amps. The increased sizing
of the neutral is required because the non linear current pulses are displaced in time and can not find a
return path through the loads on other phases. This results in 6 current pulses in the neutral conductor. It is
real current that is flowing through the neutral, not 3rd harmonic current. It may look like the 3rd harmonic but
it is displaced in time from the real 3rd harmonic.
These same pulses of current flow between phase to phase in a three phase system. Therefore the neutral is
not required to carry any current with a 3 phase load. Only single phase loads that use the neutral as the
return conductor required oversizing of the neutral conductor. Some most electrical distribution systems
provide power for single phase as well as 3 phase load, increasing the size of the neutral conductor is
mandatory.
The pulses of current in a 3 phase system, with a 3 phase non linear load, compresses the time when
current flows to about 2/3 of the normal time. The type of 3 phase load is important regarding the shape of
the pulse. In converters used with DC motor controllers, arc furnaces and Current Sources AC drives, the
control of speed or current is accomplished by reducing the conduction time of the converter by phasing back
the Silicon Controlled Rectifiers (SCRs) used in the converter. By phasing back the SCRs, the shape of the
pulse will become very narrow or extremely compressed so that a high value exists for a short period of time
each cycle. These types of converters also have a displacement power factor that is low. This means that a
high value of reactive current can exist. The existence of reactive current plus the harmonic nature of the
current waveform means that some hot spot heating may exist in transformers built to 60 Hz rules.
Transformers that have a K factor rating greater than 1 are built differently. These transformers are designed
to let the heat caused by load currents exit the transformer more effectively. Typically flat or rectangular wire
is used instead of round wire. An improved winding configuration is used to reduce hot spot heating.
When the converter type of a full wave, 3 phase bridge operating without phasing back the SCRs or power
diodes, then a near unity displacement power factor exists. When the voltage and current in phase, there is
not reactive current injected back onto the distribution system. Since current flows while the voltage is at its
peak value, it takes less current to produce power than would be the case if current was flowing when the
voltage was at a lower value.
In general, a transformer sized to handle an ac motor operating across the line, will not experience any
additional loading when an adjustable frequency PWM drive is used to control the speed of the motor. When
a reactor exists, either externally or internally, the peak of the current pulse is reduced to within a reasonable
value of the rms current required to transfer power to the motor load. Without a minimum of 3% reactance in
series with the flow of power, the peak value of the current pulse can create nuisance problems. The pulses
of current can place more stress on line fuses. The fuse element will respond to peak current and will fatigue
over time. Magnetic circuit breakers may nuisance trip because of these higher peak currents. Although the
rms value of the current has not changed, the 60 Hz rms rules for component behavior will change.
Power factor correction capacitors can be affected by the pulse currents associated with non linear loads.
The capacitor is a friendly component willing to provide all the current demanded by the load. Since its
impedance is low, it does not restrict how much or how rapidly current leaves the capacitor on its way to the
load. The capacitor can not differentiate between real current and reactive current. The characteristic of the
load determines the type of current. Although correctly sized to handle the reactive component of current,
power factor correction capacitors can not take on the additional requirement for the real component of
current.
Power factor capacitors are designed and sized to handle reactive currents. Reactive current is current that
is of a different polarity that the supply voltage. When the voltage polarity changes, the power factor
capacitor begins to supply the load current. With most non linear loads, reactive currents are supplied by the
non linear control device. The use of power factor capacitors and non linear load types like PWM AFDs are
not well matched. The near unity power factor of PWM type AFDs eliminates the need for power factor
capacitor located on the input of the AFD. Power factor capacitors can still be used at the inputs of line
operated motors as long as there is inductance between the capacitor and the input of the AFD. Some
capacitor manufacturers recommend a minimum of 250 electrical feet between the capacitors and the input
terminals of an ASD..
Essentially, power factor capacitors function as kVA devices rather than as kVAR devices, when used in AFD
circuits. This means that fuses protecting capacitors will likely nuisance blow and capacitors may get hot.
What are major contributors of Harmonics?
Harmonics or complex current waveforms are caused by any reactive or inductive load and by any product
that uses a rectifier to convert AC voltage to DC voltage. Single phase equipment like TVs, VCRs,
computers, electronic lighting all convert AC to DC. Three phase equipment like Electric Arc Furnaces,
electric heaters, DC drives, Welders, Uninterruptable Power Supplies, AFDs all convert AC to DC. These
types of products are referred to as Power Electronic equipment. Their characteristics vary with the type of
equipment. The major contributor to harmonics or peak currents would be equipment that using SCRs as
voltage control devices in a high current or high torque applications where the reactive component of current
is high.
When comparing loads, do not assume that single phase loads are too small to be of concern. With AC to
DC converters, the demand current occurs around the peak or high point of the voltage sinewave. A
thousand 100 watt lights consumes 100 kW of power. If the lights are non linear type loads, then the peak of
the currents will add directly. This can cause a dip in the voltage waveform. If the dip in the voltage waveform
drops below the threshold voltage for electronic lights, flickering will occur.
How to determine if a harmonic problem exists?
In most applications with PWM AFDs, there will be no harmonic problems. With other converters loads such
as arc furnaces, dc drives, current source drives and other high reactive current loads, harmonic problems
may exit Some of the following problems may indicate a harmonic condition but may also indicate line
voltage unbalance or overloaded conditions.
Nuisance input fuse blowing or circuit breaker tripping
Power Factor Capacitor Overheating, or fuse failure
Overheating of supply transformers
The problems listed above will be more common on single phase systems.
Problems that are not harmonic problems are:
Overheating of motors
Overcurrent tripping of AFDs
Interference with AM radio reception
Motor failures or insulation breakdown
Wire failure in conduits
Harmonic problems are rare. There have been no documented harmonic problems with PWM AFDs that use
a series reactor in the dc bus or in the input ac line. There have been many problems blamed on harmonics
but investigation resulted in identifying some other cause. There have been overheated neutrals but this has
been with single phase, non linear loads where the neutral conductor was sized to the 60 Hz, linear load
rules.
How to address a harmonic problem?
If a true harmonic problem occurs, the solution is to change the characteristics of the distribution system.
Changing inductance, resistance and capacitance will change the way the distribution system behaves.
Adding filters is sometimes recommended. This method should be done with great care since filters will act
as a preferred path for all currents in the system. The filter could easily be damaged by outside sources that
are added later or were not identified while attempting to solve the problem. It should be noted that tuned
filters are normally used to trap or divert negative sequence (negative power producing) currents.
Magnetizing current for transformers and motors can create 5th and 11th harmonic currents that are negative
sequence currents.
In severe cases, it would be good practice to insert tuned filters to create alternate paths for these currents.
The converter or 6 pulse rectifier circuits of DC drives, Current Source Drives, and Arc Furnaces can also
contribute to 5th and 11th harmonics that, under some cases contain negative sequence currents. This type
of equipment, not PWM type AFDs, may cause problems if a problem is going to occur. With PWM type
AFDs, all currents on the input are positive power producing currents. It would not be good practice to insert
tuned filters to divert positive power producing currents from their intended load, the motor.
A safe assumption is that harmonic problems, with PWM type AFDs, do not exist. Peak current problems
may exist. The only solution to peak current problems is to reduce the current peak with inductance or to
time shift individual peak currents so that they do not add directly. Using multiple voltage sources with
different phase shift will reduce the accumulation of peak currents from non linear loads.
Derating the supply source as the non linear load content increases will prevent overheating and overloading
problems. With DC drives, the input transformer should be derated by at least 30%. Using PWM type AFDs
without line reactor or a dc bus reactor would require a 20% derating of the input transformer. A 10 %
derating would be recommended when using a PWM type AFD that uses either ac line reactors or contains a
dc bus reactor.
There are many more details covering the topic of harmonics. For this overview the important thing to
remember is that injected current harmonics do not exist with PWM AFDs and should not be the issue, rms
and peak current are the issue. Using Harmonics as the only method to describe the shape of the current or
voltage waveform only leads to confusion. Real current can cause instantaneous voltage changes, however
the reactive component of current increases the demand on the distribution system. This greater demand will
result in more voltage distortion. How much the voltage distortion impacts the other components on the
distribution system will depend on the design of those components. Derating components may be required
on some systems. Relocation of other components may be required. For the majority of installations, there
will be no harmonic related problems. The greatest problem will continue to be the problem of overcoming
the fear and lack of understanding that consultants have of harmonics. Specification will continue to carry the
requirement for meeting IEEE-519. Fortunately, in most installations, the recommended voltage distortion
limits defined in IEEE-519 can be meet. Using PWM adjustable frequency drives results in no injected or
reactive currents reducing the losses in the distribution system and increasing the efficient use of electrical
power for the control of ac motors.

http://www.ab.com/drives/techpapers/pqpwm.htm

WHY PWM AFDs DON’T CAUSE HARMONIC PROBLEMS

The Chances of Harmonic Problems are less than a Big Win in the Lottery

Howard G. Murphy P.E.

If the chances of having a Harmonic Problem with Adjustable Frequency Drives that use Pulse Width
Modulation and contain internal reactors or use input line reactors is small, then why do many electrical
specifications contain the requirement to meet the guidelines of IEEE-519.
There are two reasons why electrical specifications include the IEEE-519 requirement. The first reason is
fear of an unknown phenomenon called Harmonics. The second reason is peer pressure. If your peers
(including utility “Experts”) tout the importance of Harmonics and refer to “qualified” references, then it would
look bad if your specifications did not include the IEEE-519 requirement.
So what if the IEEE-519 requirement is included in your specification. Are there any negatives? The answer
is YES when the following conditions exist.
AFDs with a rating between 5HP and 1000HP
AFDs are specified as PWM with internal or external reactors (3% impedance min.)
AFDs connected to a distribution system that use less than 90% of the source capacity.
The first negative for the project is that it will cost more than required. The second, and most important, is
that specifying IEEE-519 does not guarantee that other electrical problems will not exist. The last negative is
that equipment, like harmonic filters, added to make the current waveform “pretty” can create some of the
problems that are explained in IEEE-519.
Harmonic Filter traps can act as “sinks” for harmonic currents created by the normal operation of a system.
Turning on power, starting across the line motors and load changes on across the line motors can create
voltage disturbances which will force currents into filters which were not designed to handle these
disturbances.
Perhaps a review of IEEE-519 would be in order. First the document explains many of the characteristics of
electrical distribution systems. It creates a “bag of electrical bad guys” and dumps all Adjustable Speed
Drives Types (AC and DC) into that bag. It sets some numerical guidelines for all types of drives without
consideration for the major differences between those drives types.
In 1981, when the original version of IEEE-519 was released, the importance of that document was to set
guideline limits for total percent harmonic voltage distortion (%THD) covering the use of Silicon Controlled
Rectifiers in variable voltage controllers like DC drives and arc furnaces. The phase control and importance
of balanced firing could and does place huge demands on electrical systems. The possibility of harmonics
problems with those types of controller (converters) demanded that input transformers and line reactor were
used.
During the 1980s, many imported Adjustable Frequency, PWM, AC drives were introduced. These AFDs did
not contain internal reactors and created overloading and fuse failure problems. Where problems existed, the
addition of line reactors eliminated the problem. In response to these earlier AFDs, the 1992 version of IEEE-
519 was released. This document carried through the importance of limits for total percent harmonic voltage
distortion, but added guidelines which defined limits on percent injected current distortion. The document did
not clarify that injected current is reactive current or current that is sent back into the distirbution system.
Without that simple clarification, percent distortion of the forward current or power producing current was
subjected to a demand that its shape be made to look like a sinewave.
It is the IEEE-519 limits for percent current distortion that has done a disservice to the industry. There is no
quantitative data which shows that reducing the current distortion (making the waveform pretty) makes the
system operate better or more efficiently. The problem with IEEE-519 has become a numbers game. The
equipment used to measure percent current distortion compares an ideal waveform with an actual waveform.
There is no indication of how the actual waveform impacts the electrical system. It is merely a comparison of
a pretty waveform versus a not so pretty waveform. Since IEEE-519 specifies that "INJECTED" (reactive)
current back into the distribution system should be limited, trying to force the forward, power producing
current to be "pretty" is outside the recommendations of IEEE-519.
The present interpretation of the IEEE-519 guidelines presents the following picture.
Given a garden hose capable of 100GPM, the guidelines state that using the hose at 50GPM but wildly
varying the flow from 30 to 80GPM will create problems.
Following the guidelines would require that a flow regulator be added so that the flow could not change
rapidly.
What would happen if the same limits were applied to the city water supply. Restricting flow within some
arbitrary limits would create a lot of unhappy customers in the early morning when lots of flushing and
showering occurs.
In the preceding case, the solution would be to regulated pressure. In the same manner, it is important in
electrical systems to regulate voltage. The guidelines of IEEE-519 regarding % voltage distortion limits are
important and practical. However, to ensure voltage regulation, sufficient system capacity must exist. When
the system is too small, there is no way to regulate the voltage or pressure when current or flow are rapidly
changing. To set limits of how rapidly flow or current can change does not yield the kind of performance that
customers require and demand.
Including IEEE-519 as part of the specification should still be considered. However, some rational thinking
should be employed when certain types of AFDs are specified. To raise customers costs by blinding
specifying 12 step AFDs or harmonic filters where neither is required to obtain reliable and trouble free
operation is not only bad judgment, but bad engineering. The solution is to dig into the IEEE-519 “bag of
electrical bad guys” and deal with the ones that may cause problems. Discard those which will not create
problems. To continue to specify IEEE-519, without clarification, would be like specifying earthquake proof
and watertight equipment because those possibilities exist.
Don’t let the measured percent harmonic number become “the problem” you need to solve, use the percent
harmonic measured number as part of a data base to show how your electrical system is changing. You are
likely to reach the limits of the electrical system long before you encounter a problem with harmonics.
The correct method for sizing any motor system is to pick an input source (i.e. transformer) suitable for the
ac motor operated across the line (suitable for inrush). Using an AFD will reduce the total rms current and
never allow typical inrush currents. Do not use PF correction capacitors within 250 electrical feet of the AFD
input. If capacitors are required because of other fixed speed motors on the system, place a line reactor
between the capacitors and AFD input.

http://www.ab.com/drives/techpapers/meet519.htm

WHAT IS MEANT BY MEETING THE REQUIREMENTS OF IEEE-519-1992?

First and foremost, IEEE-519-1992 describes the RECOMMENDED Practices and Requirements for
Harmonic Control in Electrical Power Systems. The scope of IEEE-519-1992 is clearly stated as the intention
of establishing goals for the design of electrical systems that include both linear and non-linear loads. The
document describes the voltage and current waveforms that may exist throughout the system and establish
waveform distortion goals. It defines the interface between sources and loads as the point of common
coupling with observances of the design goals to minimize interference between electrical equipment.
Secondly, it is the responsibility of any reputable equipment supplier to provide their customers with
equipment, at the best possible cost per performance ratio, that will meet the known operating requirements
of the customer. Included with the purchase of that equipment are the less tangible but equally important
application experience that the supplier can share with the customer.
Lastly, the equipment supplier should be able to supply any necessary service and application support
directly associated with the performance of that equipment and its impact on other electrical equipment
utilizing the same point of common coupling.
Specifically addressing the objective of meeting the requirements of IEEE-519-1992, it can be stated that all
variable speed drives (both AC and DC) meet the requirements in most installations. With some types of
variable speed drives (DC and Current Source Inverters), it may be necessary to add an isolation
transformer or ac line reactors between the source and the drive to prevent interference with other electrical
equipment. The requirement for isolation transformations or ac line reactors stems from the fact that those
types of power converter produce voltage distortion in the form of line notching. Those types of converters
also may contain higher levels of reactive currents which may place additional stress on the components of
the distribution system when they are injected back into the system.
Typically, if the system capacity is greater than twice the demand, voltage distortion is minimal and no
interference with other electrical equipment is experienced. Often the recommended goals established within
IEEE-519 are exceeded without creating any interference problems. Because of the diversity of the individual
loads, the recommended goals defined within IEEE-519 can not identify the demarcation line between an
electrical system that is problem free or a system that is subject to problems.
The principal goals established within IEEE-519-1992 are for voltage distortion.. The document clearly
dedicates its volume to describing power converter types where phase controlled rectifiers are used. The
main applications for that type of converter is in power control for arc furnaces, dc motor controllers and
current source drives. The converter used in PWM ac drives is mentioned briefly in a few paragraphs of the
100+ page document. That fact should not go unnoticed.
It is generally accepted that PWM ac drives which contain internal bus reactors or input ac line reactors do
not create electrical interference with other electrical equipment. Although some rare occurrences may exist,
more than 500,000 installations where PWM ac drives are installed have not experienced any interference
problems. In the majority of these installations, the voltage distortion goals defined within IEEE-519-1992
have not been exceeded.
In many or even most of the installations, the current distortion goals have been exceeded (if current is
measured at the input of the drive and not at the utility). These facts clearly show that the description for
current distortion defined within IEEE-519-1992 is correctly stated as an aid in determining voltage distortion.
In section 10.3 (Development of Current Distortion Limits), "..... The objectives of the current limits are to
limit the maximum individual frequency voltage harmonic to 3% of the fundamental and the voltage THD to
5% for systems without a major parallel resonance at one of the injected frequencies." It further goes on to
state that; "..If individual customers meet the current distortion limits, and there is not sufficient diversity
between individual customer harmonic injections, then it may be necessary to implement some form of
filtering on the utility system to limit voltage distortion levels." It can be clearly noted that achieving low
voltage distortion is the principle goal. Meaurements for current distortion are to help identify potential
voltage distortion problems.
Meeting the requirements defined within IEEE-519 does not mean crossing every t and dotting every i
without applying common sense. It is a system evaluation which requires going beyond the easy approach of
putting in bigger and more expense equipment that is not required. It is the responsibility of the equipment
supplier to take their best technical experience and share that experience with their customers to provide the
lowest cost, most efficient solution to the application. If we are to efficiently use the limited energy that is
available to us at this time we must not implement the bigger, less efficient, and more expense solutions
when it is obvious that they are not required for the proper performance and most efficient use of electrical
energy.
Finally, we must understand that we are dealing with electrical systems which are dynamic. It may be
required, in the future, that different solutions are necessary. As the demands on the electrical system
change, it is likely that other solutions, that are more expensive today, may be required in the future. There is
every reason to believe that, when those solutions are needed, the cost and reliability will be more favorable.
IEEE-519 is also a dynamic document. As technology changes, that document will change. As suppliers offer
new products, those products will help the customer meet their electrical needs when those needs are
required and they will likely be achieved at favorable costs without an atmosphere of fear. If no interference
between electrical equipment exist in an installation, then the scope of IEEE-519 has been met. The
harmonic distortion percentage limits are guides to achieving that result. Those who utilize IEEE-519 with
that in mind clearly have met their responsibility of designing and specifying on the basis of intelligence not
fear.
As a final note, It is recommended practice to consider the effect on power factor correction capacitors that
were applied to the system to improve or counteract the lagging power factor of any motor operating across
the line. When applying drives, any capacitors on the system may have to be relocated or electrically
buffered to prevent an increased demand caused by forcing the capacitors to provide kVA power instead of
the kVAR power that they were designed to handle. Discussions with the equipment supplier can obtain a
more appropriate solution to a potential system problem.