I am using a sensor based on AD7746 (2-channel capacitor measurement). The sensor is (of course!) capacitive,
and since I have noise issues, I have replaced the sensor by 2 fixed capacitors. Basically it works, but one of the
channels has quite surprising noise behavior. See attached picture.
The second channel is quite smooth, and the capacitance seems to grow slightly but this might be due to temperature
since the data I used is plain raw data, just read from I2C.
By the way, these graphs have a span of about 10 hours. I use the highest resolution (measurement at 9.1Hz).
On the first graphic, there are smooth parts with a noise corresponding roughly to the noise of the second graphic.
But there are some parts where the noise is 4 ~ 5 times higher.
Has anybody experienced that kind of behavior?
What I do is basically (in pseudo code)
NB: channel 2's noise is about 400 lsb, which means that the last 8 bits are basically noise. It looks higher to the
figures I obtain with the eval board, but on the other hand, the sensor & chip module are hanging at the end of the
wires, which is not an optimal hardware. But I feel this is consistent data. And I also feel there is a serious problem
with the other channel.
Thanks for any hint.