08-25-2020 01:33 AM
Evaluating NI 6365, on p. 3 of https://www.ni.com/pdf/manuals/374654e.pdf I observe 1.66mV accuracy in the example that assumes 10,000 measurements to reduce noise impact if I only change the number of measurements from 10,000 to one, I get ~2.6 mV, my accuracy is ~13 bits (20V range +/- 10V and the 2.6 mV absolute accuracy), does that stack-up?
We would like to measure +/-140V time-variant (3MHz peak frequency content) signal with 15-bit accuracy, so expect to precision voltage divider and a good trigger and then our source resistance & settle time becomes a concern!
Appreciate your confirmation and suggestions regarding our application.
09-02-2020 10:57 AM
Try to find a calibration lab to qualify that 😄
15bit is about 30ppm, I'm not aware of any gear that can do that, but also interested in solutions 😉
My guess: be happy if you reach 0.5%
Here you can find the spec of a Fluke 5790 AC voltage transfer standard (40k+ gear):
http://download.flukecal.com/pub/literature/5790B___omeng0000_0.pdf
09-03-2020 02:47 AM
Additional remarks:
The 6365 is a multiplexed card .. with 2MSPS max for one channel, 60ppm INL , >166ppm accuracy (I guess for DC one channel 😉 ) not quite the rigth gear for that task.
So, do you need use that card (and have the 'accademic' wish)?
or do you need these specs (and have resources) ?