Connecting signal to oscilloscope improves performance

[Liquefactionist]

So in a test setup using two USB 6363 DAQ boards, I was getting great results in debugging the board while I would look at signals on a Tektronix TDS5104 oscilloscope.  Removed the oscilloscope from the setup for production test, and problems with testing cropped up.  In the end, not having the oscilloscope on makes the system work better.  Just having a coax attached to a channel (1 Meg/~13 pF input) even if the oscilloscope is turned off, improves the measurements drastically.  I have to assume it is either the lower impedance/higher capacitance of the loading with the scope attached or the grounding of the oscilloscope makes the system work "better". 

It sucks having a $12,000 piece of equipment attached to a system just to make the results "better".  I need to figure the best way to prove what is it about the scope being attached that improves the results, and figure out how to do that without a scope attached.

Any help/insight appreciated.  Thanks,

Chris

[nyc_(is_out_of_here)]

The TDS5104 has a choice of input impedance of 1 MΩ (+/- 1%) or 50 Ω  (+/- 2.5%)

[Liquefactionist]

Correct, and it is in the 1 Meg configuration that I get the better results. It is the problem that to look to see what is wrong the problem goes away.

[crossrulz]

Most likely, it is a grounding issue.  Are you connecting the ground of the DAQ to the ground of your UUT?

[Liquefactionist]

I have tried several ways to connect the grounds.  The signal in question is coming in on a BNC and I have tried the ground switch in the FS/GS settings, AI ground to earth (UPS has a ground connection on it for earth ground), CHS GND, and most other permutations.  I have even tried adding to the BNC going to the channel and putting a BNC shield to earth connection and all these are not as good as the oscilloscope connection.

My only guess is that somehow the lower input impedance of the oscilloscope vs. the DAQ (1 meg vs. 100 meg if I remember correctly) is helping the overall stability of the DUT.

[Henrik_Volkers]

Possible effects:

The ground path added (don't know if the scope shield is hard grounded)

And the bias path added

The capacity of the cable added.

It would be a useful information what type of source and impedance you measure!

And what type of disturbance you see with and wihout scope.

By adding another ground loop you can accidently lower the hum...  where a clean setup would eleminate it 😉