Circuit Surgery
Regular clinic by Ian Bell
Op amp offsets – Part 1
O
n the EEWeb Forum, user
Deion posted about precision
op amps, first quoting the OP97
datasheet from Analog Devices: ‘The
OP97 is a low-power alternative to the
industry-standard precision op amp,
the OP07. The OP97 can be substituted
directly into OP07, OP77, AD725, and
PM1012 sockets with improved performance and/or less power dissipation and
can be inserted into sockets conforming
to the 741 pinout if nulling circuitry is
not used. Generally, nulling circuitry
used with earlier generation amplifiers
is rendered superfluous by the extremely
low offset voltage of the OP97 and can be
removed without compromising circuit
performance’. Then Deion asks ‘What
confused me is the range of temperature? Should the resistor be balanced?’.
Questions
We are not sure exactly what Deion means
in asking about the temperature range,
as with the reply on the forum, the obvious answer is from the first page of the
datasheet (and repeated in the ‘absolute
maximum ratings’ table) – the device has
an ‘extended industrial temperature range’
of −40°C to +85°C. However, specifications for bias current and various graphs
of device characteristics on the datasheet
cover a larger range of −55°C to +125°C.
The −40°C to +85°C specification is an ‘absolute maximum rating’ which means that
the device may be damaged outside this
range or become unreliable if used close to
the specified limits for extended periods.
However, the op amp may work (at least
for a while) beyond these stress ratings.
The second question concerns the balancing of resistors in precision op amp
circuits – the purpose of this is to reduce
offsets due to bias currents flowing in the
external circuit (although it is not always
appropriate). In this month’s article, after
discussing the basic concept of offsets,
we will focus on op amp internal offsets, in particular looking at how input
offset voltage is defined and modelled.
Next month, we will look at input bias
currents, and address the issues related
to resistor balancing.
60
Offsets
In simple terms, offsets are DC errors in
a circuit’s output due to imperfections in
the circuit or components. This issue is
important in DC and very-low-frequency
circuits, a key example being amplifiers
used for sensor signals (such as temperature) where the quantity being measured
changes slowly – here offsets cause inaccuracies in the measurement.
In other cases, unwanted DC output
may be damaging to a load intended
to only be driven by an AC signal with
zero offset (eg, a loudspeaker). Unwanted DC can be blocked using coupling
capacitors, but this may not be practical if low wanted-signal frequency
leads to a requirement for very large
capacitance values. Even if available,
capacitors with suitable values may
be physically too large, or too expensive or have non-ideal characteristics
of their own which affect the signal in
an unacceptable way.
If offsets were pure DC they would be
fixed for all time and could be removed by
a one-off calibration/cancellation procedure, but real offsets drift due to changes
in temperature, aging and other factors
that influence the circuit. These changing offsets are just like low-frequency
signals and so are amplified along with
the wanted signal – they act as low-frequency noise.
Precision
If offsets are changing in the same frequency range as the wanted signal, then
it is not possible to use coupling capacitors to block the offsets, irrespective of
the capacitors’ characteristics or cost
(they would block the signal too). The
only solution is to design circuits with
inherently low offsets. As with any engineering design, making one aspect better
tends to be at the expense of something
else, for example lower offsets may mean
poorer high-frequency response or higher
power consumption.
Op amps with low offsets as a key
characteristic are often referred to as
‘precision’ op amps, like the OP97 mentioned in Deion’s post. The best offset
performance is obtained by auto-zeroing
amplifiers or chopper-stabilised amplifiers, but standard precision op amps can
also provide good performance and may
be more suitable in some applications,
or where the precision requirements are
not so stringent. We are focusing on standard precision op amps in this article.
In an op amp circuit, offsets at the outputs can occur due to offsets in the op
amp’s internal circuitry, or due to the external circuitry and its interaction with the
op amp. The resistor balancing mentioned
by Deion is an example of the latter situation, which we will discuss next month.
Input offset voltage
For an op amp, ideally, with a differential input of zero, the output should
also be zero, but with real op amps there
will typically be a non-zero output. The
input offset voltage (VIO) is defined as the
DC voltage which must be supplied between the inputs to force the quiescent
(zero input signal) open-loop (no feedback applied) output voltage to zero.
The definition of input offset voltage is
illustrated in Fig.1. Typical values for
standard (not autozeroing/chopping) precision op amps range from about 10µV
to 500µV. Input offset voltage for an individual device can be of either polarity
up to the specified value.
Temperature often has a significant
effect on offsets – the temperature coefficient of input offset voltage specifies how
VIO changes with temperature, typical
values for standard precision op amps are
around 0.1 to 10µV/°C. The datasheet for
an op amp may also have a graph showing offset variation with temperature.
Aging of devices also causes offsets to
vary. Again, this will often be specified
–
VOut = 0 V
+
VIO
Fig.1. Offset voltage defined.
Practical Electronics | February | 2022
Fig.2. The input-referred offset
does not change, but the output
referred offset will depend on
–
–
VOut
VOut
VIn
VIn
the circuit in which the op amp
– +
+
+
is used. In a circuit with feedback the output offset depends
VOS
on the circuit, not on the op
amp’s open-loop gain.
Fig.2. An op amp with offset can be
Use of a voltage source con- Fig.4. Using the LTspice Component Attribute Editor
modelled as an ideal op amp with an offset
nected to the non-inverting to configure the UniversalOpamp2 component.
voltage source at its non-inverting input.
input to model offset can be
investigated in LTspice, but we need to
are needed) via the Component Attribute
on the datasheet, for example, as a longuse an idealised op amp with zero interEditor – see Fig.4. Note the offset paramterm stability value in µV/month.
nal offset to do this. If we select a real op
eter (Vos) for the UniversalOpamp2 is
The circuit in Fig.1, used to define
amp (we often do this for Circuit Surgery
set to zero in our simulation because we
input offset voltage, can also be the basis
simulation examples), it will model that
are using the external sources VOS1 and
for analysing the offset in a circuit. For
device’s offset, which we will not be able
VOS2 to model the offset.
analysis purposes we can replace an op
to control – which is not what we want
UniversalOpamp2 can model op amps
amp with offset with an ideal op amp
here. Instead, we can use the Universawith different amounts of detail. Setting
plus an offset voltage source, as shown
lOpamp2 component from LTspice’s library
the level value for the component
in Fig.2. The offset represented in this
– it is right at the end of the list in the Op
determines which model is used. There
way is called ‘input referred’. Although
amps section of the Component Selection
are four levels (1, 2, 3 and 3a in increasing
we are discussing input offset voltage,
menu. An LTspice schematic with an inlevels of detail). For information on all the
other sources of unwanted output from
verting and a non-inverting standard op
parameters refer to the example schematic
the op amp, such as random noise and
amp amplifier – both using Universainstalled by LTspice (typically at location
the op amp’s response to power supply
lOpamp2 – is shown in Fig.3. This circuit
...\Documents\LTspiceXVII\examples\
voltage variation can also be representincludes voltage sources, as discussed
Educational\UniversalOpamp2.asc on
ed in this way.
above, to model the offset voltage.
Windows). Use of idealised models with
certain specific non-ideal characteristics
Modelling
allows us to explore the effect of those
If an input, VIn is applied to an op amp
UniversalOpamp2
characteristics in isolation from other effects
its output will depend on both the input
You may notice there are two other comto gain insights about circuit behaviour.
signal and the input offset voltage. For
ponents that are not real op amps on the
For this example, we will use the simthe open-loop amplifier circuit in Fig.2,
component menu: opamp (without supply
plest (level 1) UniversalOpamp2 model.
the output will be
connections) and opamp2 (with supplies).
This does not use the power supplies (eg,
These can also be used to include idealised
to limit the output voltage) but power
VOut = AVIn + AVOS
op amps (or real op amps) on a schematic
supplies have been included on the
but require an LTspice subcircuit to define
schematic to allow the model level to
their internal circuit or behaviour. UniverWhere A is the open-loop gain of the op
be switched easily if required. We need
salOpamp2 is more convenient for working
amp. The value AVOS is the ‘output referred’
near-ideal behaviour, so we have set the
with generic op amp behaviour because it
offset. Normally, we do not use an op amp
open-loop voltage gain (Avol model
can be configured by right-clicking the op
open loop, in which case the effect of the
parameter) very high (100G = 1×1011)
amp symbol and setting the various paoffset voltage can be analysed in the context
rameters (if values other than the defaults
of the specific circuit using the approach in
so any errors in output voltage due to
non-ideal (non-infinite) open-loop gain
are not significant. Similarly, the input
resistance (Rin model parameter) is set
very high (10GΩ) so there is little impact
on the currents in the external resistors.
Op amp
with offset
Ideal
op amp
Simulation
Fig.3. LTspice schematic for simulation to investigate modelling of offset voltages.
Practical Electronics | February | 2022
The op amp amplifiers in Fig.3 have the
same resistor values. Using the wellknown formulae for the gain of these
circuits we get the following. For the
inverting circuit, the gain is: −RF1/RI1 =
−50kΩ/1kΩ = −50. For the non-inverting circuit, the gain is 1 + RF2/RI2 = 1
+ 50kΩ/1kΩ = 51.
With 10mV input provided by source
V1 we would expect outputs (with no
offset) of −50 × 0.01 = −0.5V for the
inverting circuit and 51 × 0.01 = 0.51V
for the non-inverting circuit. However,
the circuit includes offsets, introduced
by VOS1 and VOS2.
We just need to find the DC output, so
we can use an operating point simulation
(.op SPICE directive). The results are:
61
--- Operating Point --V(in):
0.01
V(n003):
0.0105
V(out_noninv): 0.5355
V(vp):
5
V(vn):
-5
V(n001):
-0.0005
V(out_inv):
-0.5255
V(n002):
-0.0005
V(n004):
0.0105
voltage
voltage
voltage
voltage
voltage
voltage
voltage
voltage
voltage
In both cases the output voltages V(out_
inv) and V(out_noninv) are shifted
by 22.5mV from the values obtained in
the calculation above. For the inverting
circuit we have −0.5 – 0.0255 = –0.5225
and for the non-inverting 0.51 + 0.0225
= 0.5355. The magnitude of the offset at
the output (with the same input offset)
is the same for both circuits despite the
different circuit gains. This is because
the offset model is applied directly to
the non-inverting input in both cases.
As far as the offset voltage sources are
concerned, both circuits behave as noninverting amplifiers. We can see this
using some circuit theory called the
‘superposition theorem’ – if we have a
linear circuit with multiple voltage and
current sources the output is equal to the
sum of effects of the individual sources,
with the others set to zero.
For the circuit in Fig.3, if we set the
offset sources to zero (as for an ideal op
amp) we get the standard amplifier circuits for which we have just calculated
the output. We can also consider just the
effect of the offset sources by setting the
input voltage to zero (equivalent to shorting the input to ground). Doing this results
in the circuits shown in Fig.5. We have
the same circuit in each case – a non-inverting amplifier with the offset source as
the input. Thus, in both cases the output
due to the offset will be the non-inverting
gain times the input offset voltage, in this
case: 51 × ±500µV = ±22.5mV.
The sign of the offset was set in the circuit of Fig.3 to increase the magnitude of
the output, to make the result more obvious. Changing the polarity of the offset
sources would reduce the magnitude of
the total output in both cases (offset opposite in polarity to wanted output), but
the offset magnitude would be the same.
noise gain, rather than
signal gain, is used when
calculating output offset
(or noise) for an op amp
circuit. The term is commonly used in technical
documents by op amp
manufacturers discussing
offsets and noise.
Although the offset
(or noise) voltage source
is commonly depicted
connected to the noninverting input, moving
it to the inverting input
does not change the fact
that offset (or noise) is
amplified by the noise
gain (non-inverting gain).
This can be verified by
the LTspice circuit in
Fig.6, which produces
the same output voltages as the circuit in Fig.3.
This may not be obvious
at first sight, but consider Fig.5. If we consider just the contribution of the offset for
the fact that in a feed- two amplifiers in Fig.3, then we find that in both cases
back circuit the op amp is offset is amplified by the non-inverting gain.
maintaining (in the ideal
case) zero volts between its inputs, so
of the inverting circuit is smaller than
moving the source past the inputs does
the non-inverting gain by this factor
not change the effect it has on the output.
(50 = 51×(50/51)).
Another way to understand that the
offset is not amplified by the circuit gain
Offset nulling
in the inverting amplifier is to consider
Even with a low offset, op amp users
the fact that in the inverting circuit (in
may want a circuit to have the facility
Fig.6) the offset source is connected dito manually adjust the offset to minirectly to the op amp input, so it is not
mise (or null) it. This feature is provided
going to be affected by the resistors in
by some op amps, including the OP97,
the same way as the input signal. The
despite as quoted above, the datasheet
input signal is attenuated by the potenindicating that it is ‘rendered supertial divider formed by RI1 and RF1 before
fluous’. Typically, if offset adjustment
is available, the op amp will have two
arriving at the op amp (by a factor of
pins labelled ‘Null’ (see Fig.7 for the
50kΩ/(50kΩ+1kΩ) = 50/51 = 0.9804 in
OP97 pinout).
our example). The magnitude of the gain
Noise gain
The fact that offsets, and random voltage
noise (which is modelled as input referred
to the op amp’s input in the same way),
are amplified by the circuit’s non-inverting gain, irrespective of the amplifier
circuit configuration, leads to the idea of
‘noise gain’. The noise gain of the circuit
is the gain which applies to a voltage applied directly to the op amp’s inputs. It
is equal to the non-inverting gain. The
62
Fig.6. Moving the offset voltage source to the other input produces the same results.
Practical Electronics | February | 2022
conforming to the 741 pinout if nulling
circuitry is not used’. If an op amp is
swapped with one using a nulling cir–
2
7
–IN
V +
cuit connected to the opposite supply,
it is likely to be destroyed. The device
+
3
6
+IN
OU T
datasheet must always be checked for
OV E R
5
V – 4
the exact details of using nulling pins
OP 9 7
C OM P
where they are available.
For the OP97 the trimmer can be in
Fig.7. OP97 op amp pinout.
the range 5kΩ to 100kΩ and can adjust
the offset between 300µV and 850µV
Typically, to use the null pins they
depending on the individual device.
are connected to a trimmer potentiomUse of nulling pins is often not a good
eter, with the wiper going to one of the
solution for excessive offset. The pins
supplies, as shown in Fig.8. For the
can introduce noise into the circuit if
OP97 it is the positive supply. For the
used – particularly with poor layout,
741 it is the negative supply – hence
such as long leads to the trimmer. Offset
the comment in the quote above that
drift with temperature and age may mean
the OP97 ‘can be inserted into sockets
that the offset has to be nulled more
often than just at initial setup, which
R 2
may be inconvenient. Adding nulling
R 1
may make temperature drift worse. For
VIn
–
VOut
these reasons it is often better to find a
device with an offset which is inher+
ently small enough for the application.
Of course, higher-performance devices
A dj ust to
may cost more, which may limit options
r emov e offset
V S upply
in some cases.
Offset nulling can be achieved withFig.8. Example offset nulling circuit for an
out the use of null pins using circuits
inverting amplifier – check device datasheet such as the one in Fig.9. This creates
for details (eg, which supply the trimmer
a small, adjustable voltage at the nonwiper connects to).
inverting input of the op amp in an
N U L L
1
8
R 2
N U L L
R 1
VIn
–
VOut
+
R
R 3
R
– VR
A
R
B
P
+ VR
Fig.9. Example of offset nulling circuit
using external circuitry rather than op
amp null pins.
inverting amplifier. The VR voltages can
be the supply and RP can be equal to the
parallel value of R1 and R2 – which relates to the resistor balancing which we
will discuss next month. More details
on offset adjustment circuits like this
can be found in the MT-037 Tutorial
document from Analog Devices, see:
https://bit.ly/pe-feb22-ad
Simulation files
Most, but not every month, LTSpice
is used to support descriptions and
analysis in Circuit Surgery.
The examples and files are available
for download from the PE website.
STEWART OF READING
17A King Street, Mortimer, near Reading, RG7 3RS
Telephone: 0118 933 1111 Fax: 0118 933 2375
USED ELECTRONIC TEST EQUIPMENT
Check website www.stewart-of-reading.co.uk
Fluke/Philips PM3092 Oscilloscope
2+2 Channel 200MHz Delay TB,
Autoset etc – £250
LAMBDA GENESYS
LAMBDA GENESYS
IFR 2025
IFR 2948B
IFR 6843
R&S APN62
Agilent 8712ET
HP8903A/B
HP8757D
HP3325A
HP3561A
HP6032A
HP6622A
HP6624A
HP6632B
HP6644A
HP6654A
HP8341A
HP83630A
HP83624A
HP8484A
HP8560E
HP8563A
HP8566B
HP8662A
Marconi 2022E
Marconi 2024
Marconi 2030
Marconi 2023A
PSU GEN100-15 100V 15A Boxed As New
£400
PSU GEN50-30 50V 30A
£400
Signal Generator 9kHz – 2.51GHz Opt 04/11
£900
Communication Service Monitor Opts 03/25 Avionics
POA
Microwave Systems Analyser 10MHz – 20GHz
POA
Syn Function Generator 1Hz – 260kHz
£295
RF Network Analyser 300kHz – 1300MHz
POA
Audio Analyser
£750 – £950
Scaler Network Analyser
POA
Synthesised Function Generator
£195
Dynamic Signal Analyser
£650
PSU 0-60V 0-50A 1000W
£750
PSU 0-20V 4A Twice or 0-50V 2A Twice
£350
PSU 4 Outputs
£400
PSU 0-20V 0-5A
£195
PSU 0-60V 3.5A
£400
PSU 0-60V 0-9A
£500
Synthesised Sweep Generator 10MHz – 20GHz
£2,000
Synthesised Sweeper 10MHz – 26.5 GHz
POA
Synthesised Sweeper 2 – 20GHz
POA
Power Sensor 0.01-18GHz 3nW-10µ W
£75
Spectrum Analyser Synthesised 30Hz – 2.9GHz
£1,750
Spectrum Analyser Synthesised 9kHz – 22GHz
£2,250
Spectrum Analsyer 100Hz – 22GHz
£1,200
RF Generator 10kHz – 1280MHz
£750
Synthesised AM/FM Signal Generator 10kHz – 1.01GHz
£325
Synthesised Signal Generator 9kHz – 2.4GHz
£800
Synthesised Signal Generator 10kHz – 1.35GHz
£750
Signal Generator 9kHz – 1.2GHz
£700
HP33120A
HP53131A
HP53131A
Audio Precision
Datron 4708
Druck DPI 515
Datron 1081
HP/Agilent HP 34401A Digital
Multimeter 6½ Digit £325 – £375
HP 54600B Oscilloscope
Analogue/Digital Dual Trace 100MHz
Only £75, with accessories £125
(ALL PRICES PLUS CARRIAGE & VAT)
Please check availability before ordering or calling in
Keithley 228
Time 9818
Practical Electronics | February | 2022
Marconi 2305
Marconi 2440
Marconi 2945/A/B
Marconi 2955
Marconi 2955A
Marconi 2955B
Marconi 6200
Marconi 6200A
Marconi 6200B
Marconi 6960B
Tektronix TDS3052B
Tektronix TDS3032
Tektronix TDS3012
Tektronix 2430A
Tektronix 2465B
Farnell AP60/50
Farnell XA35/2T
Farnell AP100-90
Farnell LF1
Racal 1991
Racal 2101
Racal 9300
Racal 9300B
Solartron 7150/PLUS
Solatron 1253
Solartron SI 1255
Tasakago TM035-2
Thurlby PL320QMD
Thurlby TG210
Function Generator 100 microHz – 15MHz
Universal Counter 3GHz Boxed unused
Universal Counter 225MHz
SYS2712 Audio Analyser – in original box
Autocal Multifunction Standard
Pressure Calibrator/Controller
Autocal Standards Multimeter
o er
lifier
Voltage/Current Source
DC Current & Voltage Calibrator
Modulation Meter
£250
Counter 20GHz
£295
Communications Test Set Various Options
POA
Radio Communications Test Set
£595
Radio Communications Test Set
£725
Radio Communications Test Set
£800
Microwave Test Set
£1,500
Microwave Test Set 10MHz – 20GHz
£1,950
Microwave Test Set
£2,300
Power Meter with 6910 sensor
£295
Oscilloscope 500MHz 2.5GS/s
£1,250
Oscilloscope 300MHz 2.5GS/s
£995
Oscilloscope 2 Channel 100MHz 1.25GS/s
£450
Oscilloscope Dual Trace 150MHz 100MS/s
£350
Oscilloscope 4 Channel 400MHz
£600
PSU 0-60V 0-50A 1kW Switch Mode
£300
PSU 0-35V 0-2A Twice Digital
£75
Power Supply 100V 90A
£900
Sine/Sq Oscillator 10Hz – 1MHz
£45
Counter/Timer 160MHz 9 Digit
£150
Counter 20GHz LED
£295
True RMS Millivoltmeter 5Hz – 20MHz etc
£45
As 9300
£75
6½ Digit DMM True RMS IEEE
£65/£75
Gain Phase Analyser 1mHz – 20kHz
£600
HF Frequency Response Analyser
POA
PSU 0-35V 0-2A 2 Meters
£30
PSU 0-30V 0-2A Twice
£160 – £200
Function Generator 0.002-2MHz TTL etc Kenwood Badged
£65
d
£350
£600
£350
POA
POA
£400
POA
POA
POA
Marconi 2955B Radio
Communications Test Set – £800
63
