OPA182: Question on PSRR and CMRR characterization

Part Number: OPA182
Other Parts Discussed in Thread: OPA387,

Tool/software:

I am currently evaluating the OPA387 and OPA182 precision amplifiers for a high-accuracy application. While reviewing their datasheets, I found some details on PSRR and CMRR characterization that I would like to clarify.

From the OPA387 datasheet, the typical CMRR is listed as 140–150 dB at VS = 5.5 V, while the PSRR is given around 115–138 dB depending on conditions. For the OPA182 family, the electrical characteristics table specifies extremely high values: a typical CMRR of up to 168 dB and an open-loop gain of 170 dB.

Looking at the corresponding figures for noise and frequency response, the CMRR and PSRR curves show degradation with frequency as expected, but the very high DC values (140 dB and above) raise a question regarding measurement methodology. In practice, when attempting to measure rejection ratios in excess of 140 dB, the instrumentation and environmental noise often limit the achievable test floor.

Could you please provide more details on:

  1. How the PSRR and CMRR values reported in the datasheets were characterized for these devices.

  2. What specific measurement techniques or setups were used to achieve reliable results above 140 dB.

  3. How these very high rejection values relate to the noise plots presented in the typical characteristics section.

I would like to understand whether these specifications are based on direct measurement, extrapolation, or modeling, and how to properly interpret them for practical design considerations.

  • Hi Alejandro, 

    I was the lead validation engineer who collected the data for the OPA182 and improved our CMRR measurements over the years. We have an article out that describes the test circuit here:

    Op_amp_Test_Basics_from_TI

    The shunt resistor between the inputs needs to be small to reduce the noise. You can use 10 ohm shunt and 10k feedback in the two amp loop described in the article. Then you use a PGA in a gain of 100. The ADC CMRR needs to also be measured and removed from the measurement. Bench correlation is performed between two completely different systems. 

    Best Regards, 

    Chris Featherstone