BNC-2090 progressive shift in voltage reading

[Carassius]

Hi All,

I have been recording using only CH0 on a BNC-2090 connected to a PCI 6040E. Just two days ago, I realized the voltage was progressively shifting until the card saturates at +10 and I have to restart my computer to get it working again.

I tested with a PCI DAQ card, and still shifts. I tested with a different board and it's solved. So, my BNC-2090 gets charged and at some point delivers +10 V or more. Have you heard about something similar...? I found a post saying that capacitors could be charging inside the board:
http://forums.ni.com/ni/board/message?board.id=170&message.id=61233&requireLogin=False
Any clue about how to fix my board...? Do I have to ground the terminals that are not connected...? It was working well before.
Thanks.

[bushman]

I have a guess. WIthout knowing your complete system, that's all I can venture.

Charging symptoms are typically caused when a voltage source doesn't have a direct return reference.

This can be caused by an open circuit ( you said after resetarting it works fine, so probably not that).

It can becaused by a RSE circuit wired to a differential input without the sig - not connected to the same reference as the signal source.

Remember that current sources are not referenced to system or chassis ground, but to the return of the power supply.

Hope this helps

[Mark_E]

Hi Carassius,

Would you mind clarifying how you tested your setup? When you say you “tested with a PCI DAQ card, and it still shifts,” are you testing with a different DAQ card, and which one? When you “tested with a different board and it’s solved,” do you mean a different BNC 2090?

What happens when you ground the other inputs through bias resistors? Does the drifting voltage go away on CH0? With which mode are you measuring your signal: differential, RSE, or NRSE?

Data Acquisition: Checking Your Board Settings - Analog Input Mode

Field Wiring and Noise Considerations for Analog Signals

[Carassius]

Thanks Bushman and Mark for your reply. First, I characterized the shifting voltage using the Test panels and Igor. It was clear that progresively the voltage was shifting until the card got saturated and it did not recover. Only restarting the computer. Following the DC level with an oscilloscope, it was clear that the signal was correct before entering the board. Before, it was always working well at RSE. To start testing my setup, I used two other DAQ cards (another PCI6040E and a PCI6052). The voltage was still shifting with both of them. Then, when I connected a different BNC-2090 or a homemade board, there was no voltage shifting at all. My conclusion was that something in the original BNC-2090 was charging up to +10V. At the time I did all the tests, the board switches were set at RSE, even for the board that didn't do voltage shifting. Now, I solved everything moving the switch at DIFF for the channel that I need in the original board. BUT, it was working well before when the switch was at RSE. And the board that I used to test didn't need the switch at DIFF to work correctly. In summary, the old board needs to be at DIFF to work correctly. The other one works well even at RSE. What happened lately to the original board to make it adding a DC level when working at RSE...? Sorry for the long message. I am just not able to fall asleep trying to figure out what's wrong with my original board. At least, now I can record again. Thanks so much, Carassius

[Mark_E]

If you have the exact same setup and they function differently, it would indicate that there is something wrong with the old BNC-2090. Please visit www.ni.com/contact to get our Customer Service phone number to discuss sending the card in for repair.