Period measurement timebase apparently lower

[Daniel_Chile]

I'm performing buffered low frequency period measurement using 1 counter, with an NI 6624. I am using the maximum timebase, which is 20MHz. The gate signal has a frequency of about 25kHz.

The expected period "resolution" due to quantization should be 1/timebase, in this case 1/20MHz=5E-8s.
The problem is that I am getting an effective "resolution" of 5E-7s, or an apparent timebase of 2MHz. What could be wrong?
I've checked digital filtering (DAQmx Channel Properties) on the source and gate signals, and are both enabled with the timebase of 20MHz.
Any comments apreciated.

[Bill B]

Hello Daniel,

I agree that you should have a period measurement resolution of 50ns. This means that if you have a perfect 25kHz input signal, you should count 800 pulses of the 20MHz timebase within the 40us window. If you count 799 or 801 pulses, you will get 1/(39.95us) = 25.031289kHz or 1/(40.05us) = 24.968789kHz.

However, the thing to keep in mind is that you must have very accurate 25kHz source in order to determine if this is the case. Using the calculations from above, you would need a signal source with an accuracy of at least +/-0.1249% at 25kHz.

You may want to verify that your signal is at least this accurate before determining if something is wrong with your board or configuration.

Regards,
Bill B
Applications Engineer
National Instrume
nts

[Daniel_Chile]

Hi Bill. Thanks for your answer.
Returning back to my problem, actually it doesn't matter to me if I have slight frequency variations on my 25kHz signal. In fact, I can vary the frequency of this signal, since it comes from a signal generator.
Now, either when the frequency of the generator is set to (about) 25kHz or if I sweep it, the succesive (variable length) periods are given on multiples of exactly 500ns, which is the apparent behavior that I could expect if I used a 2MHz timebase, but this is not supposed to be the case.

I still don't know what's happening.

Thank you.

[Daniel_Chile]

Hi. Another detail I've noticed from a DAQmx channel property node is that the digital filtering minimum pulse width property is set to 1us.

[Bill B]

Try setting the filter setting to 100ns. Let me know if you see any difference.

Bill