Evaluating the LMP91050 thermopile amplifier over the last few days using the LMP91050EVBDEVAL eval board. The board worked fine with the input shorted at J2 (CMOUT connected to IN). When I connect the thermopile that we use there is very significant AC power noise at 60Hz. Peak to peak value with GAIN1 set to 250 and GAIN2 set to 32 and using internal signal routing is about 900 codes peak to peak. Using the out of the box external filter the peak to peak drop to 600 codes.
I ruled out the thermopile package as the culprit by installing a 100K resistor between CMOUT and IN. 100K is about the internal resistance of the thermopile so this represents the thermopile without any signal. The noise is still present with the 100K resistor.
The level of AC noise can be changed by changing GAIN2 and you can see the 60 Hz AC waveform on the output of the LMP91050. I can touch a small wire to the IN signal and the AC noise causes the amplifier output to rail. So the noise is definitely on the IN signal with respect to CMOUT.
I did design a bandpass filter using the typical circuit that provides -18dB delta for the target 3Hz frequency and the 60Hz noise signal. Even with the newly designed bandpass filter, the noise is still to high for this application.
I have an existing design that uses an instrumentation amplifier with a differential input and that does not have the 60Hz noise problem because the 60Hz signal is common mode and is cancelled out. It appears the problem is due to the fact that this part is a single ended design and therefore prone to noise.
Is anyone else experiencing this problem? If so, what have you done to mitigate AC power noise. I was contemplating a more aggressive bandpass external filter but wanted to post this first before going down that path.