Minimum time between conversion - NI 9205

[Sina_RM]

Hi all;

I use one NI 9205 for acquiring data from 20 different channels (NI compact RIO 9114, LabVIEW FPGA).

I need sampling frequency of 12.5 kHz or higher.

As far as I understood, with default settings (minimum time between conversion = 8 us), the maximum sampling frequency for my case is:

fs = 1 / (20*8e-6) =  6.25 kHz.

To increase the sampling frequency, I am going to set minimum time between conversion to 4 us, so I can have:

fs,max = 1 / (20*4e-6) =  12.5 kHz

In the link below, it has been stated:

https://zone.ni.com/reference/en-XX/help/373197L-01/criodevicehelp/conversion_timing/

"The default minimum time between conversions for the NI 9205/9206 is 8 µs. The accuracy specifications in the NI 9205 and NI 9206 hardware documentation on ni.com/manuals are based on this default value. If you set the minimum time between conversions to at least 8 µs, the accuracy of the module is not affected. If you set the minimum time between conversions to less than 8 µs, the accuracy of the module degrades if you sample data from multiple channels."

My question is: how much accuracy will I lose?

In my case, my readings on the channels should have maximum error of 16.5 mV.

Does setting minimum time between conversion to 4 us produce more error than 16.5 mV?

I understand that the error and accuracy depends on the temperature, calibration, noise, etc.

I want to have a rough idea about the amount of error.

Please let me know if you need more information.

Thanks a lot,

Sina

[santo_13 · Accepted solution]

Your question is answered in the instrument datasheet

The accuracy will be thrice worst than the 8us covert interval.

[Sina_RM]

Hi Santhosh

Thank you very much for your answer.

I didn't notice that the datasheet has mentioned the accuracy.

Now I have two more questions:

1- How should I calculate LSB?

For example, if I work with +-10 V, range, is it true to say that:

LSB = Full Scale/(2^n-1) where n = number of bits (n =16 for NI 9205) --> LSB = 10/(2^16-1) = 0.153 mV?

Then the accuracy for 8 us interval would be +-2*0.153 = +-0.306 mV

and accuracy for 4 us interval would be +-8*0.153 = 1.224 mV ?

2- In the datasheet, what is the difference between the accuracy mentioned here:

and the absolute accuracy?

Which one should be considered when we want to estimate the maximum error?

Thanks again,

Sina

[santo_13]

I would add the conversion error on top of the absolute accuracy if you are measuring just 1 sample +/-(6.23mV + 1.224mV)

As per specifications, it uses a 100 sample average to arrive at the specification numbers.

[Sina_RM]

Thanks a lot Santhosh!

Best regards,

Sina