I am currently experiencing a strange behaviour from my AD8304. I am using the AD8304 evalboard (the evalboard is described on pages 18 and 19 of the datasheet). To test it, I had connected a Keithley current source to the input and a voltmeter to the output. With this setup, I get the result shown in this image:
This is exactly how I want the circuit to function. However, when I put the evalboard in its housing, the current source is connected via a BNC to banana-plug adapter, the characteristic curve gets distorted, as shown in the next image:
For clarification, the insulation resistance of the adapter cable was tested with a electrometer and is above 2 GOhms, so the cable should not be the issue. The housing is made of varnished cast aluminum, a short circuit or other connections between parts of the evalboard PCB are not possible. My guess is that the missing shielding for the current source could be the issue.
How can this behaviour be explained (source of the problem) and what would be an appropriate solution to fix it? Adding a BNC or SMA port to the housing is not possible. I already posted this question in the AD Engineer Zone, however I got no answers there.
