0
\$\begingroup\$

I'm trying to implement an instrumentation amplifier to measure pressure from an MPX2050 transducer. After deriving resistor values to give me a suitable gain and adding an adjustable reference voltage I moved on to simulation and everything was looking good.

After building the circuit on the breadboard, it was amplifying differential signals as expected. However, I found the circuit was not rejecting common mode signals. I've tried rebuilding the circuit, testing the breadboard for shorts, replacing the op amps and at this stage I'm out of ideas. Do you guys have any ideas on what might be the cause or how to fix it?

Schematic:

schematic

Breadboard: enter image description here

Op amps: OP27G

Pressure transducer: MPX2050

EDIT: Thanks for the help guys. I've reuploaded the breadboard layout with labels. I completely agree with you guys saying I should use an off the shelf instrumentation amp, however, it's a college project and we're limited to using this configuration.

I can understand tolerance issues being a cause of reduced CMRR, but I'm getting pretty much zero common mode rejection. Connecting the 68k resistor directly to ground doesn't affect the output.

\$\endgroup\$
4
  • \$\begingroup\$ The only variation from a 'standard' circuit is the variable ground voltage control (270k-10k-270k) at the end of the 68k. Depending on the voltage output of the lower op amp this voltage will shift. Try the circuit by connecting the end of the 68k to a real 0V. Also, please label components on your circuit (R1,R2 etc) Its a pain in the rs trying to describe which component you mean. \$\endgroup\$ Mar 25 '17 at 12:40
  • \$\begingroup\$ You don't mention the size of the common mode signals. At any rate, careful PCB layout is needed to come close to the potential of precision parts like that. Using one on a breadboard is asking for trouble. They don't appear to be decoupled either. \$\endgroup\$
    – user133493
    Mar 25 '17 at 12:40
  • \$\begingroup\$ Why aren't you using a monolithic in-amp instead? \$\endgroup\$ Mar 25 '17 at 13:04
  • \$\begingroup\$ I reuploaded the image with labels there. After connecting the 68K directly to ground, there was still no change. The common mode signals are ~7.5 V, and the differential signals are 0-40 mV. Unfortunately I'm limited to using the breadboard, OP27 amps and pressure transducer as it's a college project. \$\endgroup\$
    – Pudar
    Mar 25 '17 at 18:33
1
\$\begingroup\$

There is a VERY high chance you are facing component tolerance issues with a discrete-base instrumentation amplifier enter image description here

For ideal components the overall gain is 10.119 (gain POT == 0)

This complication comes from how well matched the two 47k, 18k and 68k resistors are. EQUALLY the additional bias network tolerance & balanced rail.

If the upper resistors are +1% and the lowers are -1% the issue starts to become clearer. A slight bias.

Even if you had a the best resistors

your bias network is a POT + your supplies will cause a bigger error.

Perfectly matched resistor network BUT +-1% on the POt and +-10V supply.

enter image description here

So depending on how big of an offset you are seeing, it will come down to component mismatch & the bias pot will be the biggest source of offset

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the comprehensive response. Would tolerance errors be large enough to result in nearly zero common mode rejection? \$\endgroup\$
    – Pudar
    Mar 25 '17 at 18:39
0
\$\begingroup\$

Two things. First, you buy instrumentaton amps for good performance, as you will get great CMRR that way. When you build them, tolerance issues on resistors cause problems. Next, that's not how you provide offset. You need low impedance on the second stage's voltage divider. Buffer with a voltage follower.

\$\endgroup\$

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.