We are currently running into a problem where there is a gradual droop in the measured voltage of our A/D, as seen in the picture below.
The first voltage measured is the correct value (e.g. the correct measurement), however the value measured decreases with each new sample until it reaches a steady point about 40 samples after the first measurement. When we measure the input voltage to this channel via an oscilloscope the value does not change, so we figure whatever is causing the droop must be internal to the A/D. The hardware configuration for the A/D can be seen in the figure below.
This channel (AIN3) is the only one where this phenomenon is noticed, and the only difference in configuration between this and the other channels is the gain of (this channel has a gain of 1 and the other channels are configured to be a gain of 64). Beyond that, we’re suppling a 5Mhz clock to the A/D, collecting data from the A/D at a rate of 300Hz, and our OSR is set to 128.
We’ve tried a number of solutions such as modifying (and even removing) the cap to ground on the input, changing the sample rate of the A/D, etc