After exploring this a bit more, I concur. If I use a virtual o-scope VI monotoring, say channel 1, and connect that channel to the physical o-scope, I can see the signal on both "scopes". If I disconnect the physical 0-scope, the signal on the virtual scope goes away.
The odd part is that if I keep the physical scope connected, and simply exit all of my VI's, the signal on the physical o-scope goes away!
At one point yesterday I measured the signal and saw that the noise was a 20mhz signal. I am wondering if this is some kind of coupling from the internal clock on the card? I would hope not. I do seem to remember that the frequency was lower when I measured it in the past, and since I was using a newer o-scope yesterday than in the past, it may just be a coincidence. That being said, if I enable one of the counters to run as an oscillator, I can see that it disturbs the analog input (as long as the physical scope was connected), so it may related after all.
As I was messing around yesterday, I felt like I was chasing a ghost, but got the strong sense that the issue is a ground-loop of some kind. Sadly, I was not documenting all of the different ways I altered the connections and software settings, making it difficult to "see" a pattern, so I plan to do that next time. We are planning on moving the system to it's "home" soon, however, so I think I will wait until then so that I don't get stuck doing this all again once it has been moved. For now I am just planning on keeping the o-scope away from it and going with the belief that it is clean when left alone 🙂
Thanks for the help guys!
Glenn