I am getting data from an AD7961 ADC with an ADA4941-1 in front as a single-ended-to-differential stage.
I am getting a surprising noise of +/-2mV when I short the ADA input to ground. This noise goes down to +/-0.2mV if I skip the ADA4941-1 stage and instead short both the differential inputs of the AD7961 to ground (the inputs still have their resitor-capacitor network as indicated in the part's datasheet).
I am working on home made prototyping PCB with the ICs mounted onto fine, commercial SMD-to-thru-hole adaptors. I know that this is not actually an ideal situation to get pure data in, but I can't figure out how the mV noise range is possible. I have used all the precautions, such as capacitors near the chips and precise voltage reference by ADR430. I am using this temporary solution to evaluate the circuit and move to a more professional PCB later on.
Any suggestion is appreciated.
Many thanks in advance,