Ground loop issues with National Instruments PCI DAC and zero-drift opamp measurements

Hi,

I'm having some problems reliably zeroing a signal using my oscilloscope, and wondering if it could be do to a ground loop.

The signal comes from a national instruments digital-to-analog PCI card on my PC, then goes into a linear zero drift operational amplifier that should have a maximum offset of 4 uV. There is an analog out ground pin on the PCI card that is connected on my PCB to the common power supply ground. The bipolar power supply for the opamp is made from 2 lab power supplies connected in parallel, so the power supply "ground" is at the negative terminal of one, and the positive terminal of the other. The power supplies also have a dedicated ground terminal, but that is not connected to anything.

So far, I don't think there should be a ground loop as the only path to ground is through the DAC card and ultimately through the PC's power cable.

I have been connecting the output of my opamp directly to the positive on the scope probe, and the negative on the scope probe to ground to zero the signal, but I have reason to believe it is not actually zeroed. I assume this is because the ground on the scope is an alternative path for the current, or ground loop.

If I use two scope probes and an "A minus B" method of looking at the signal, I see that there is actually about 200 mV offset, though I can't zoom in quite as precisely as I'd like.

Does this ground loop hypothesis sound plausible to you? If so, are there any common solutions to how to avoid the ground loop on the scope in a case like this?

Thank you.

Best,
Michael

You are correct that the ground lead on the scope probe forms a ground loop. What do you see when you touch the tip of your scope probe to the same point the probe ground is connected to?

When using two scope probes and "AminusB" differential mode, connect the two probe ground leads to each other and not to the PCB. When you touch both probe tips to the same point, you should see zero signal on the scope. You could try measuring the DC offset using a battery-powered digital voltmeter.

You can eliminate any ground loop in your lab power supplies by powering your opamp from two 9-volt batteries, if your opamp is rated to run from +-9volts.

Try waving your hand close to, but not touching, the opamp. If the offset changes, that indicates the opamp is oscillating at a frequency too high for your scope to see.

Try listening to the opamp output through an audio system. If you hear music or speech, you have radio-station pickup.