PXI-6255 differential analog input measurement error

[aharris]

Hi all,

I've been having issues with the PXI-6255 DAQ. Attempting to measure a single differential measurement (AI0,AI8 or AI1,AI9, etc) works fine. Attempting to measure multiple channels of differential measurements results in wrong data for the measurements beyond the first channel sampled. Lowering the sampling rate does not seem to change the result. Attached are screenshots from SignalExpress which shows the issue.

For the screenshots, the analog inputs should be reading exactly -1.0V and -1.5V. When AI1 is sampled first, that value is correct (reads -1.0V) but AI7 will read slightly higher (around -1.47V). When AI7 is sampled first, that value is correct (reads -1.5V) but AI1 will read slightly lower (around -1.03V).

Any ideas what may be causing this measurement error and why sample order is making a difference here?

Thanks!

[Dennis_Knutson]

I didn't think you could even select both channels when selecting differential mode. Perhaps Signal Express does things differently. In differential mode, you only select the first channel of the pair. So, you would specify dev1/ai0:1 for acquiring from the first two channels.

[Dennis_Knutson]

Just noticed your screen shot has ai7. My poor eyes read that as ai9, sorry.

[aharris]

*deleted due to previous correction*

[RyanPoPo]

aharris,

What happens if you try to sample an open channel in between those two connected channels? Does it have any effect on the values you read?

Regards,

Ryan

[aharris]

Hi Ryan, 

Yes, I've posted another screenshot which demonstrates the behavior. I'm reading all differential pairs (AI0/AI8 to AI7/AI15) while there are only valid signals connected to AI1 and AI7 pairs). In the screenshot, you can see the order of data acquisition (sequential from AI0 to AI7). AI0 reads properly at 0V. AI1 should read -1.0V but reads approx -0.92V. AI2 should read 0V but instead reads approx -0.07V). AI2 through 6 read correctly at 0V. AI7 should read -1.5V but instead reads approx -1.4V.

The conclusion I've come to understand is that any data acquisition from the previous channel affects the next channel's value. For example, AI1 is pulled up since AI0 is 0V. AI2 is pulled down since AI1 read -0.92V. AI3 reads correctly since AI2 was much closer to the actual value. AI7 is pulled down since the previous AI6 read 0V. Does this help clear up the issue?

Thanks for your help.

[ChrisR.]

http://digital.ni.com/public.nsf/allkb/73CB0FB296814E2286256FFD00028DDF

[RyanPoPo]

Hey aharris,

You are correct in your hunch that the previous channel is affecting the following channel. The phenomenon is known as "ghosting" and it occurs due to the charge from one measurements not haivng enough time to fully discharge from the card's ADC before the next measurement takes place. It looks like Chris already beat me to the punch, but the KB that he provided should solve your issue.

Regards,

Ryan