I am using the analog front end LMP91000 IC to make a potentiostat circuit. The build-in step resolution of the LMP is not enough for my sensor measurement purpose. I found people having this same issue with the step resolution here:
So I connected the MCP4725 at the Vref of the potentiostat, and just to see that the Vref that is being generated is correct, I shorted the CE, RE pins as shown below, so that Op-amp A1 is used as a voltage follower. (So the variable bias is visible at the output now and I can probe it)
I slowly increase the Vout of the CMP from 0V to 5V, and here is what I see from the oscilloscope. Randomly, the output of the A1 pin generates noise. (I noticed when I re-plug it on power, I do not get any noise for the first minute/seconds. After that, it starts randomly generating noise, even when the voltage is stady):
Noise randomly appears when the Vout of CMP is steady:
Clarifications:
- this noise does not happen when I use the refernce voltage as internal reference (aka the 5V input of the LMP)
- I tried the same layout/ciruit without the 22uF on the output of MCP, same result of noise.
Questions:
- Is the MCP's Vout value too noisy (has relatively high ripple) by default and causes the op-amp to oscillate? If that is the case, why it does not oscillate all the time?
- Why when I plug the PCB on power, (usb power) the output wont oscillate for some seconds, while if I just reset the MCU without re-powering the PCB, it will generate noise faster?
- If I filter the output of the MCP using low-pass filter(s), will the noise issue get improved?