Electronic – CMRR test methodology and effect of open loop gain

operational-amplifiertest

I am trying to test the CMRR of opamp and so measuring offset variation across Common mode range in unity gain mode. I have gone through many articles on CMRR measurement. All those are suggesting a differential configuration of Opamp for testing CMRR.

Why shouldnt I test for offset across Vcm in Unity gain mode for CMRR ? In unity gain mode, varying the common mode voltage and measuring the offset as the difference of output and input, CMRR can be measured.

If I have to measure in differential mode, there will be 4 resistors besides opamp. The accuracy of these resistors will effect my measurement across Vcm, Vdd and Temperature.

Someone stated that the infinite open loop gain will effect the offset in unity gain mode across common mode range and give pessimistic values of offset as compared to differential mode. I tried with math to understand the effect of openloop gain on closed loop gain and found that the effect is the same whether it is unity gain or differential mode.

Please help me understand right methodology.

Best Answer

Just short-circuiting the inputs will not be possible. Usually, the small input offset voltage will be amplified by the opamp and will in most cases lead to saturation of the output.

Using a amplifier in buffer configuration reduces gain, so saturation will not happen at the output. However, it will change the output voltage as you sweep the input. Measuring the difference with the input will therefore give you the sum of the input offset voltage, a contribution due to finite gain and your figure of interest: the CMRR. So you have one more term in there. If the differential gain is much much larger than the CMRR then the approximation will be OK. But I don't think this is necessarily the case. The input offset voltage is more or less constant w.r.t. the common mode, so you can compensate for that.

When working with the differential amplifier configuration, you can add some gain to the CMRR, making the measurement more sensitive (if the resistors are matched well enough). The gain can be limited so output saturation due to an input offset voltage will likely not happen. And because the output stays close to constant, the extra term in the previous paragraph (caused by the finite gain) will also be much more limited.