TLV8812: the input signal and power supply is not stable because of switching input channel

Hello,

Is one of the mux MOSFETS "on" at all times? If neither MOSFET is activated, then the input to U6A is floating. Because of the low input bias currents and capacitance of the MOSFET, the output may "drift" as the MOSFET capacitance charges or discharges. U6A does not know what to output when the switches are off - so it is wandering.

The U6A input needs a defined DC bias path when the switches are off. Try adding a 10M resistor between pin 3 and GND. This should force the U6A output to go low when the switches are off.

You do not show the output loads of U6. Remember that these are nano-power devices and they cannot drive much capacitance (<20pF). Do you have the recommended isolation resistor (~50k) in series betweenthe output and ADC input cap??

It looks like you are supplying the ADC power from U6B. If there is a large bypass capacitor on the VCC line, it will cause instability. This is probably the "instability" you are seeing.

Remember that the TLV8812 is a nanopower device with very little sourcing ability. The current limit is just a few mA, and the linear range is less than 1mA.

I doubt the TLV8812 even has enough output current to power the ADC properly. You may need to use a different amplifier, or a ADC with a reference voltage input separate from the power supply.

Even the 20K feedback divider (R16 + R17) is a "heavy" load to this nanopower devce (and you are wasting more power in the feedback resistors than the amplifier quiescent current!). Consider changing R16/R17 to 100k, or even 1M. The TLV2333 is a higher power device and slightly more tolerant of capacitive loads - but not much! It can have the same issues.