How to avoid signal drift ?

[raeletro]

I'm using a PCI-6071 reading only one channel in differential mode.
I connected the + on channel 0 and - on channel 8. The signal is good in the beginning but after some time I can notice some drift of the signal.

The system is isolated from the AC Power line ground. I't means that the AC power line ground is not connected to the system. From the AC outlet there is a transformer that provides power to the PC, monitor and an Linear DC power supply. This power supply provides +15, -15 and GND ("virtual"-internal to the system, not the power LINE GND). This "virtual" GND is used as the - signal to be send to the channel 8. My guess is that since the PC is not grounded to AC power line Ground this drif appear over the time.

Any clues ?

[Russell]

What matters most is what the signal reference is compared to the PC reference. You mentioned that you have connected the "virtual" PC ground to the (-) side of your differential measurement. Now, the question is how the signal is referenced. Is it referenced to the power grid? Or is it also floating? I know that it is at least connected to the PC ground, but where else?

Anyway, I think you'll be able to find the configuration you need from chapter 4 of the PCI E-Series User Manual, specifically starting on page 4-14. The important thing to determine is whether your signal ground is isolated from the PC ground or not.

If your signal is floating with re
spect to the PC ground it is possible for the bias currents of the amplifier to build up a charge on the signal lines, causing the input voltages to "drift" upwards. In this case, the manual suggests putting bias resistors between both inputs and AIGND. See page 4-18 for more information.

Hope this helps!

Russell
Applications Engineer
National Instruments
http://wwww.ni.com/support

[raeletro]

Before reading your comment, I inserted the bias resistors and it worked fine.

Thank you anyway.