How to get rid of interference between signals from two channels in a LabPC1200 card?

[rayms]

I am using a LabPC 1200 card for my data acquisition. I have two input signals, both are in millivolts range; one coming from a thermocouple and the other comes from a photodetector. The signals are amplified and filtered before it is sent to the computer (via the card). Amplification and filtering of the each signal are done in a separate circuit. We think that the signals are both floating even if it is connected to terminal “0” of the power supply of the amplifiers since terminal “0” is also not connected to building ground.
When we use only one channel (that is we input only one of the signals), the measured signal is okay. But the problem is that when we
input both signals (that is when we use two channels), we have observed interference. The signal from the photodectector (that of a sine) interferes with the signal from the thermocouple. We have already tried different configuration of the measurement system (RSE, NRSE, Differential), but none of it work.

[CraigGraham]

You could try connecting the 0V output of the PSU to system ground of the
PC; make sure you're really connecting the 0V output that is going to the
supply of your instrument, since the PSU may have an earth terminal you're
not currently using which is not connected to the output rails. If there's
the possibility of earth loops it may be worth disconnecting the PC from
mains earth, though this isn't something to do unless you can show that you
need to.

Also, try and diagnose where the breakthrough takes place using a scope.

rayms wrote in message
news:506500000008000000BA310000-1005954886000@exchange.ni.com...

> When we use only one channel (that is we input only one of the
> signals), the measured signal is okay. But the problem is that when we

> input both signals (that is when we use two channels), we have