06-15-2017 04:35 AM
Hello community,
I would like to ask you about your practical advice on calibration:
My NI Devices (PCI-4472 and PCIe-6323) are due for calibration.
Working with these cards I don't need absolute voltage accuracy!
But I don't have a full overview of what's done during such a calibration and what can be caused if I systematically ignore it.
In this case, do I really need to calibrate my DAQ cards?
I appreciate your advises!
Best
06-15-2017 09:23 AM
Usually calibration is needed because you need a traceable uncertainty and/or you need the 'stamp' for the auditor 😄
You didn't tell us your usecase. So why and what do you measure?
How about the timing, linearity, input impedances .... ?
Next would be to ask your quality manager (let sleeping dogs lie 😄 ).
So if you identified the properties that are important for your task, you can ask NI or other cal-labs for a quote.
Example: We use PXI 5124 to capture heterodyne signals where the information is in the frequency/phase of the signal and the PXI clock is tied to an atomic clock... I don't care about the vertical accuracy, but I still have to verify the phase 🙂
06-20-2017 07:25 AM - edited 06-20-2017 07:27 AM
We read sensor voltages from the PCI-4472 and are interested mostly in what the effect is if we don't recalibrate:
* Can we expect drifts in the absolute voltage detected or also in the relative voltage (i.e. the voltage corresponding to +- 10V input range of the card).
* How large are those effects Similar for the timing of the samples: will the mean clock frequency change over years? By how much?
06-20-2017 11:13 AM
Without calibration you don't know 😉
With a lot of regular calibrated units out in the field and some years experience .. one could make an educated guess, but you never know for your unit.
(My Fluke handheld DMM is on spot even after >10 years ... ok, prooven by calibration 😄 )
Your internal reference can drift, your XO can drift....
You can do a check .. read a stable (~battery??) voltage with all voltmeters and DAQs you have (cycle more than once) , compare the results ... capture a stable frequency ... (ask the volt nuts and the time nuts 😉 )
BUT you don't know unless you do a calibration with traceable sources.