Absolute Accuracy PXI 6123

[Shehzaada]

Please see pg. 4 of the specification for the device. The signal I'm trying to measure goes through a 30A:300mV shunt, but in most cases, the maximum value is about 10 A/100mV. The readings we get are not clean and are quite noisy, when the expectation is that they should be quite clean.

It seems the smallest voltage range for the card is +/- 1.25 V. According to the manual, the absolute uncertainty according to the table is 740 uV (or 0.74 mV). I looked at this and thought it was quite high. I mean at 2 A (20 mV) an absolute uncertainty of 0.75mV  is 3.75% and this is quite high. I realize that we are only using approx. 4% of the voltage range and that is a problem but I still thought it might be better than that - am I interpreting this right?

If I am interpreting this correctly, then is card about par with industry standards?

[FinchTrain]

Hi,

I am not surprised that your reading is not showing up cleanly when using only 4% of the available range. Do you have access to another module with a smaller range capability? It could be the case that you are also encountering excess noise from an external source.

Field Wiring and Noise Considerations for Analog Signals

http://www.ni.com/white-paper/3344/en/

Regards,

Finch Train

[Shehzaada]

Thanks for the reply FinchTrain!

I forgot to mention in the earlier post, I had previously measured the signal with another device that had -0.5 to 0.5 V (NI cFP AI-118). The signal didn't jump around nearly as much as it did with the PXI. I guess my question is, am I understanding Absolute Accuracy correctly? It's defined as the absolute worst case accuracy of the DAQ card at a specified accuracy range. It seems that at 20 mV, the absolute accuracy will be 0.75 mV - and at 3.75% that seems high to me for such a low signal range. Is this on par with other DAQ devices at this range or any competitiors that NI may have?

At its core, a major problem is definitely that the signal requirements do not take advantage of this card properly (by not taking advantage of the device input range). I'm just trying to understand the impact of absolute accuracy in this context too.

[FinchTrain · Accepted solution]

Hi,

Thanks for clarifying, as I was more in the troubleshooting mode to try and get the best representation of your signal! Take a look at this article by NI detailing Absolute Accuracy. You can calculate this using the values from the table on page four of the manual as you mentioned.

How Do I Calculate Absolute Accuracy Or System Accuracy?

http://digital.ni.com/public.nsf/allkb/8BA2242D4BCC41B286256D1D00815B90

There are three ways to calculate the absolute accuracy pending your system design, though I will add the specification on page four of the manual I believe is the number provided and tested by NI. I apologize as I am a bit busy at the moment or I would run the numbers myself.

The absolute accuracy specification is for a signal that suitably uses the +-1.25V range. A 0.74mV absolute accuracy in the scheme of a 2.5Vpp signal is less than two tenths of a percent of error (0.185%). There are modules (as you mentioned) designed for smaller input ranges, which is why you saw a much better representation with the 0.5 to -0.5 range. You could consider amplifying the signal to the +-1.25 range to get the best representation with the PXI 6123. While the distortion seems significant in your application, using only 4% of the available range is simply a case of not using the best suited hardware for your application.

Regards,

Finch Train

[Shehzaada]

Thanks a lot Finch Train! Looks like we were understanding it correctly. Next time we'll pick a better DAQ card!