Professional Documents
Culture Documents
Low operating voltage linear systems, especially battery One important factor affecting low voltage applications in
operated systems such as portable instruments or termi- general is the actual input and output operating voltage
nals, tend to have a different set of requirements and ranges available from the operational amplifier. For ex-
constraints placed on them. Usually, for such systems, not ample, an input voltage range of ±11V for a ±15V system
only must the linear circuits operate at low voltages, but may be perfectly acceptable for a given application, even
these voltages tend to vary due to limited power supply though the operational amplifier may need a 4V “overhead
regulation or gradual discharge of a battery pack over margin” from each supply rail. However, for a single 5V
time. Furthermore, power drain of the system is nearly system, such an overhead voltage is obviously not accept-
always a major consideration. In order to lower the able as there would not be any operating voltage left to
operating power dissipation of a circuit, linear active supply an output.
components such as operational amplifiers not only need
to have very low operating currents, but the selection of For a 5V single supply system or a ±2.5V system, a 1V
passive components such as gain setting resistors or overhead from each supply rail of an operational amplifier
network feedback resistors must be at as high values as would mean only 3V linear operating voltage range, which
possible, and still meet the precision and frequency re- is not much room to play with. The same circuit when
quirements of the system. operating at 3V and 2V would have 1V and 0V of actual
useful design operating range respectively. From this it
As an example, a simple DC inverting amplifier connected becomes apparent that for low operating voltages, rail-to-
as a 10X gain amplifier can use a 1kΩ and 10kΩ resistor rail operating voltages may be an imperative rather than a
gain setting network. The maximum current drain through luxury.
the resistors in a ±15V system may be 1.5mA (24.75mW)
and in a 5V system may still be as high as 250µA. Using In order to implement rai-to-rail voltages (or as wide
a low voltage FET input operational amplifier, on the other operating voltage ranges as possible), FET operational
hand, the resistors used may be a 1MegΩ and 10MegΩ amplifiers with CMOS input stages have an advantage.
resistor network without sacrificing the gain precision. The rail-to-rail operating ranges are usually achieved by
However, this requires very high input impedances at the using complementary drive design techniques, best imple-
op amp inputs such that leakage currents at the inputs mented by having complementary N channel and P chan-
would not introduce large errors to the output voltage of nel input stages. Rail-to-rail outputs can be implemented
the FET op amp set by the 10MegΩ resistor. The current by class AB output stages using N channel and P channel
drain due to the resistor network in such a 5V system is MOSFETS for push-pull outputs that span the full supply
drastically reduced to 0.25µA (0.63µW). An added benefit voltage range. MOSFET transistor characteristics tend to
here is that with very low current flow in the resistors, gain assist in accomplishing rail-to-rail outputs because at
errors caused by resistor self-heating effects at different small drain voltages, both N channel and P channel
output voltages are minimized or eliminated. MOSFET have essentially resistive characteristics. With
high impedance loads, such as CMOS data converters,
FET operational amplifiers with high input impedance not CMOS comparators or CMOS logic gates, the output
only help increase the value of an external resistor net- voltages approach within a few hundred microvolts (less
work, but also operate as near ideal voltage buffers. This than a mV) of the supply rail voltages. For ratiometric
is useful for a variety of high impedance sensors and designs that use the power supply rails, rail-to-rail opera-
transducers which are essentially charge accumulating tional amplifiers can provide precision all the way to the
devices that supply a voltage output but cannot sustain rail voltages without introducing significant error.
current flow. Often such devices also produce more
current noise effects than voltage noise effects. In a FET Another design consideration when using low voltage
operational amplifier, where usually the current noise precision operational amplifiers comes from analyzing
characteristics are far superior to its voltage noise perfor- some of the specifications themselves. The PSRR and the
mance, the overall system noise may be reduced while the CMRR (the power supply rejection ratio and the common
voltage noise spec alone might have indicated otherwise. mode rejection ratio) of a given operational amplifier are
NOTICE: Advanced Linear Devices (ALD) reserves the right to make changes and to discontinue any product and or services as identified in this publication without notice. Current specifications for any product and or services
should be verified by customer before placing any orders. ALD warrants its products to current specifications in effect at time of manufacture in accordance to its standard warranty. Unless mandated by government
requirements, ALD performs certain, but not necessarily all, specific testing and procedures as ALD deems necessary to support this warranty.
ALD assumes no liability for any circuitry described herein. Applications for any circuits contained herein is for illustrative purposes only. No representation of continued operation of said circuits under any operating conditions
are implied. Any use of such circuits are the responsibility of the user. No circuit licenses, copyrights or patents of any kind is implied or granted. ALD does not authorize or warrant any of its products or designs for use in life
support applications, and hereby expressly prohibit any such use. All rights reserved.
© 1998 Advanced Linear Devices, Inc. 415 Tasman Drive, Sunnyvale, California 94089 -1706 Tel: (408) 747-1155 Fax: (408) 747-1286 http://www.aldinc.com
usually specified in dB. However, the effect of these minimum input impedance under all circumstances is
parameters when applied to a low signal voltage applica- acceptable. This reduces the need for current compensa-
tion is of interest. Use of large supply voltages, with a tion with bias balance resistors and noise bypass resis-
large common mode voltage supply, will introduce a larger tors. The high impedances inside a circuit network also
error signal voltage into an already small signal to be tend to reduce current spikes in the power supply line,
amplified. minimizing any signal and load induced variation of the
power supply. This is especially true for high impedance
As an illustration of this point, first let us take the case of power supply sources such as batteries. Lowered current
PSRR. A circuit operating at ±15V and using an opera- spikes in a circuit reduce the need for local supply bypass
tional amplifier with PSRR of 80dB would have introduced capacitors, and frequently, even when a bypass capacitor
an error voltage of 100µV for each 1V change of the supply is deemed necessary, a single one can be used to serve
voltage. A 10% ±15 V power supply would have total many operational amplifier circuits.
supply voltage variation of 3V. The equivalent input offset
error voltage introduced by PSRR would therefore be 300 IDEAL FOR HIGH PERFORMANCE SIGNALS
µV. Now consider an operational amplifier having the
same PSRR spec but operating at 2V. A 10% variation of Low voltage FET operational amplifier circuits excel when
a 2V supply now introduces 0.2V change to the supply high source impedance device applications are consid-
voltage for the part. The operational amplifier with the ered. These applications include a class of devices such
same 80dB PSRR spec would now only introduce a 20µV as high impedance bridge networks, capacitive sensors,
error. pH probes, humidity sensors, diode detector arrays and
pressure sensors. A single supply FET operational ampli-
REDUCED POWER SUPPLY AND COMMON MODE fier is often a suitable choice for interface circuits that
ERRORS perform the tasks of buffering, amplification, signal condi-
tioning and linearization, and temperature compensation.
Now consider CMRR which is also usually specified in dB. A low voltage power source may, in this case, bring
An operational amplifier with 80 dB CMRR at ±15V with a significant additional cost savings to the system due to
voltage signal equal to 50% of the supply range would see reduced power rating requirements for the system. High
±7.5V voltage excursion and a 1.5mV equivalent offset in-circuit element impedances can also mean easier single
voltage error introduced into the signal. In comparison, supply to dual supply circuit conversion. For example, in
the error voltage for a low voltage operational amplifier at single supply linear systems there is often a need to
2V supply, with 80 dB CMRR and 50% signal range would generate a virtual ground which acts as a reference point
be 100 µV instead. In this case the percentage error for some of the positive and negative transition analog
introduced into the signal may be the same, but a 1V peak signals. This reference can be generated simply and very
signal compared to a ±7.5V peak signal, both having 50% inexpensively by using two large resistors in a voltage
signal voltage range, would in actual fact have quite divider with a bypass capacitor, providing that only a
different absolute error voltages introduced. voltage with no current supplying capability is required. In
a FET operational amplifier circuit, conditions are fre-
OUTPUT CURRENT DRIVE quently right for just such a reference.
Another anomaly to consider in adjusting the intuitive High in-circuit impedances do make the system more
process to low voltage circuit design, especially for an vulnerable to noises, both generated internally within the
experienced analog designer accustomed to using ±15V system components and coupled from outside. Internal to
operational amplifiers, is the output current drive spec. the system higher impedance means less di/dt caused
For a 2kΩ load in a ±15V (30V) system, to drive the output coupling noise. However, higher internal circuit imped-
to a voltage close to full scale would require 7.5mA output ances may require careful grounding of the circuit board
drive current. The same 2kΩ load in a ±2.5V system would and perhaps some degree of shielding. The amount of
only require 1.25mA from the output stage. Of course high grounding and shielding is a function of a given application
power output drive is not quite compatible with low voltage and its level of susceptibility to the environment in which
operational amplifier applications, but the point is that one it is expected to operate.
can be fooled with the intuition that low voltage operational
amplifier applications require the same high output cur- HIGH RELIABILITY
rents as their higher voltage cousins when driving a similar
impedance load. Often, low voltage FET operational amplifier designs not
only imply low power operation but also provide improved
Using FET operational amplifiers, circuit designs are often reliability. For example, a circuit with 5 V and 1 mA power
simplified due to design considerations such as input bias drain uses 5 mW whereas a circuit with 30 V (+/- 15 V) and
and offset current effects, where these effects are consid- 1 mA burns 30 mW of power. The difference in power
ered to have negligible effect upon the performance of the dissipation means less self-heating for a low voltage
design task at hand. Variation of a very high input circuit and therefore lowered operating chip junction tem-
impedance would not impact a design if the worst case perature.
R
R 33 R
R22
250K
250K 100K
100K
+5 V
+5V
R
R11
1K +5V
1K
-
VOUT
+
V IN
VIN
1/2 ALD2701
RR22
10MΩ
100K
VOUT = R2
- VIN
R1 R1
1MΩ +2.5V
1K
VIN -
VOUT
+
-2.5V
Note: Gain of 10 amplifier
Input impedance is limited to R1.
ALD1706
Total typical current drain of 20µA
250K 100K
25K 10K +5V
+5V or GND
or GND 100Ω
1K +5V -
- VOUT
+
+ 1/2 ALD2701
VIN
1/2 ALD2701 ~ 10,000
Gain =
VOUT = 1 ∫
VIN dt
RC
i2 C
i1 = VIN
R
R
VIN - dVc dVOUT = i
i2 = C =C 1
dt dt
VOUT
i1 +
Note: This circuit forces a current i 1 on the summing junction and causes the output to ramp at a rate of
∆Vout/∆t equal to i1. It is set by input resistor R and input voltage Vin. It can be seen that any
leakage current on the summing junction directly subtracts from i1 and affect the ramp rate. For
long integration time, dt is large and imply small i2 or large C since voltage swing ∆Vout is usually
limited. For precision applications, amplifier input leakage directly affects the value of i2 and may
be a significant contributor to overall error of integrator.
i1 dQ
R i1 = = C dV = Vdc
dt dt dt
-
VOUT
+
Charge Generator
Q= CV
R2
900K
R1
100K +2.5V
-
VOUT
VIN + VOUT = { 1+ R2 } VIN
Gain of R1
-2.5V
10 amplifier
900K
= { 1+ } VIN
100K
=10VIN
+5V
- VOUT = VIN
VOUT
VIN +
This is extremely effective in providing isolation between input and output. Essentially the
source impedance can be open circuit (i.e. capacitive source) while providing drive to
resistive load as low as 1kΩ amd capacitive load of 400pF.
VOUT = IINR1
R1 IIN range from 1nA to 1mA
+5V
-
IIN VOUT
+
R3
VOUT = IINR1 { 1+R3 }
R2
R2
-
VOUT
+
IIN
R1
- R1 R2
VIN +
C1 -
VOUT
+
C2 100pF
R2
= 99 C1= C2 { 1+ R2 } = 100 X C2
R1
R1
VC
VC VC VC
VC
0.1µF
-
VC VOUT
VIN +
VC 0.1µF
V+ = +5V V+ = +5V
-
VOUT
1MΩ +
ALD1701
APP 11 EPS WK
VIN can be set to within 1mV of either power supply rail. The total power dissipation of the
circuit is essentially that of the operational amplifier power dissipation of approximately 0.5mW.
RL
R1 iL
VIN -
VOUT
+
iL = VOUT = -VIN
RL R1
RS RL
R1 V+ iL
-
VOUT
VIN +
R2 R3
iL = R2 { 1+R3 } VIN
10K 9K
R1R5 R4
R4 1K
R1 = VIN
VIN 10
- RL
10K
+ VOUT R5
iL 100Ω
0.1µF
IN914
R1 2N2222A
VIN -
VOUT
+
R1 ALD1701
0.1µF R1
IN914
VIN -
VOUT
+
2N2222A
ALD1701
LOG
LOG B
R R
A LOG
LOG
AMP
AMP
- A/B
ANTI-LOG
LOG LOG A R LOG A/B AMP
B +
AMP