Ground Loop Noise with Aeroflex 8310 Programmable Step Attenuator

[garrettm]

So I got an Aeroflex 8310 programmable step attenuator for use with a project I'm working on and noticed some odd behavior when I connected it to my Agilent 89410A vector signal analyzer for verification of the attenuator ranges.

The attenuator exhibits measureable noise from DC to about 1.5 MHz when the input is terminated into 50ohms or when the attenuator is set to a value greater than 0 dB. Some of this looks like it could be noise from the internal line filter and other parts look to be from the switchmode power supply. The power cables to each instrument have ferrite chokes on them and are connected right next to each other on the same power strip.

Disconnecting the ground wire from the 8310 and using a DMM to measure the voltage between the 8310 chassis and earth ground I found 24.2Vac when off and 55.8Vac when on. Into 1Mohm of the DMM, thats about 24uA of ground current when off and 56uA when on. This ground current looks to be what's causing the observed noise on the VSA.

Connecting a heavy guage aligator cable between the chassis of the ungrounded 8310 to the earth grounded VSA chassis lowered the noise but did not eliminate it.

I did a similar test with my HP 6236B PSU and found 8Vac off and 12Vac when on (between chassis and earth ground). That's significantly lower than the Aeroflex 8310 attenuator.

Is this sign of a bad line filter and would replacing it fix this issue? Would I need to replace the X and Y caps on the SMPS as well? Or is this something that isn't going to go away?