11
\$\begingroup\$

When we have a two-stage instrumentation amplifier, such as the following. enter image description here

Why do we need the first stage of the two Operational Amplifiers? Couldn't we just input V1 and V2 into the differential amplifier?

\$\endgroup\$
  • 2
    \$\begingroup\$ Look at the input impedances \$\endgroup\$ – PlasmaHH Sep 9 '18 at 20:19
  • \$\begingroup\$ Infinite for the first stage, and R2 for second stage, right? \$\endgroup\$ – Bee Sep 9 '18 at 20:20
  • \$\begingroup\$ Input resistance is one point which is sometimes important but increasing CMRR electronics.stackexchange.com/questions/343096/… is another key benefit \$\endgroup\$ – carloc Sep 9 '18 at 20:34
  • \$\begingroup\$ No, different input impedences for the two inputs of stage 2 \$\endgroup\$ – Scott Seidman Sep 9 '18 at 20:42
  • \$\begingroup\$ If you have any imbalance in capacitance on Vin+ versus on Vin-, the common-mode rejection is in peril. \$\endgroup\$ – analogsystemsrf Sep 9 '18 at 22:23
13
\$\begingroup\$

The 3 op-amp design has three main advantages over a single op-amp differential amplifier.

  1. The input impedance is much higher, since the inputs drive directly into an op-amp input rather than into a resistive divider.
  2. The gain can be set by changing a single resistor, so the critical parts can be easily integrated on to one chip (maximizing symmetry) with a single external resistor for setting the gain.
  3. In high gain configurations the common mode rejection is much better because the gain of the first stage effectively multiplies the common mode rejection of the second stage.

Note that in general it is better to use a specific instrumentation amplifier chip than to try to build it yourself out of separate parts. Having everything on one chip improves symmetry and hence common mode rejection.

\$\endgroup\$
  • \$\begingroup\$ Excellent point about using a single IC; even using 1% discrete resistors lowers the theoretical CMRR to no more than 34dB \$\endgroup\$ – Peter Smith Sep 10 '18 at 13:00
14
\$\begingroup\$

One of the biggest benefits of the 3 op amp INA is the equal and high input impedance. The op amp's non-inverting pins' input impedance can be up in the \$T\Omega\$ range. I'll leave it as an exercise for you, but if you look at the difference amplifier circuit, the input impedance of the negative input varies with the positive input.

\$\endgroup\$
  • \$\begingroup\$ Why do we want a high input impedance? \$\endgroup\$ – Bee Sep 9 '18 at 20:29
  • 6
    \$\begingroup\$ @Bee so that we can measure from sources that have a high output impedance \$\endgroup\$ – BeB00 Sep 9 '18 at 20:33
3
\$\begingroup\$

In addition to input impedance concerns, gain in two stages offers better frequency response.

\$\endgroup\$
  • 1
    \$\begingroup\$ While this is true in principle. most instrumentation amplifiers have all the gain in the first stage with the second stage having unity gain. \$\endgroup\$ – Peter Green Sep 10 '18 at 22:59
1
\$\begingroup\$

There is a single stage, high input impedance, differential amplifier.

schematic

simulate this circuit – Schematic created using CircuitLab

Without Rg, this has a gain of (f+1). Rg can be used to increase the gain.

However it has performance compromises over the 3 amplifier version.

a) It has less open loop gain, so very high gains are not as stable
b) The two signal paths have different phase shift, so common mode rejection only works to low frequencies. You can improve this a bit with a strategically placed capacitor though
c) At low gains, common mode range is limited by the power supply headroom
d) It's more complicated to draw correctly than the 3 amplifier version. It's only recently that I've created a mnemonic to get the resistors in the right places from memory.

But, if you have only 2 amplifiers left, then it does work. It does allow you control of the gain with a single variable resistor, just like the 3 amplifier version, but unfortunately, just the 3 amp version, this resistor is floating.

\$\endgroup\$
-2
\$\begingroup\$

The above answers are reliable, but i want to add something. Consider the differential amplifier: enter image description here

When a person wants to vary the amplifier gain (for instance to exploit the maximum resolution of ADC) the 2 resistors with value K⋅R must be adjusted perfectly synchronously, via electromechanical adjustable resistors so an slight runout or wearing of these resistors results in imbalance between the value of this two resistors then it results in not to be neglected common mode factor. Usually the common mode in the previous stage (Wheatstone bridge for example) is way bigger than differential mode then it results in false measurements.

\$\endgroup\$
  • 5
    \$\begingroup\$ I wouldn't call this an instrumentation amplifier, I would call it a differential amplifier. You need to provide a citation for the image you included. Also, please don't assume that all of the people using this site are male. \$\endgroup\$ – Elliot Alderson Sep 9 '18 at 21:30
  • 2
    \$\begingroup\$ @ElliotAlderson maybe the implication is that women are too smart to use this circuit design anyway ;) \$\endgroup\$ – alephzero Sep 9 '18 at 23:53
  • \$\begingroup\$ @ElliotAlderson Yes this is a differential amplifier, the OP asked what if we use only differential amplifier and i tried what happen when we use only differential amplifier. And by men i meant 'man' , so all the human races regardless of their gender. It is perfectly possible to make such a mistakes, that's why there is 'EDIT' button. \$\endgroup\$ – Sam Farjamirad Sep 10 '18 at 6:17
  • \$\begingroup\$ If you know it is a differential amplifier, you shouldn't say "consider the instrumentation amplifier:". It's true that others can correct your answers, but in the meantime some new users may get more confused. \$\endgroup\$ – Elliot Alderson Sep 10 '18 at 12:35
  • \$\begingroup\$ You need to add a citation for the graphic in your answer so the original creator may receive proper credit. \$\endgroup\$ – Elliot Alderson Sep 10 '18 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.