Frequency from Counter Measurement

[bobbylee]

Hi Mcdan and Jared,

Thanks for the help.  I just wondered which way is more accurate since I wanted to measure a pulse train whose frequency varies from 4000Hz to 72000Hz depending on my engine crank angle encoder speed.  I am still learning DAQmx and the DAQ theory to measure and generate pulses, so there are many confusing parts to me.  Anyways, is it possible to use the "onboard clock" option as the source for Sample Clock istead of using Implicit? I have a USB-6251 and a PCI-6602.  Do they both work?

Thanks,

Bobby 

[Kevin_Price]

No, you're going to need to use "Implicit" timing with either of those boards.  But don't worry, in most ways that'll help you.

With Implicit timing, you can use the "Low Freq - 1 Ctr" method mentioned earlier and still leave your 2nd counter free for other purposes.  Even if you only read a single sample at a time, you'll still get precision of better than 1 part per 1000 (0.1 %)  because your boards have a 80 Mhz internal timebase for doing the freq measurement.  

I would usually recommend doing a buffered read of several "most recent" samples so you can do a little averaging, filtering, outlier-tossing, etc. to make you more immune to glitches that may be caused by spark firing and suchlike.  But first things first.  Get the simpler single-sample method working before worrying about whether you need the advantages & complication of buffering.

-Kevin P

[bobbylee]

Kevin,

Thanks for the adivce.  I should not need the buffer function immediately since I only run my engine for a few seconds.  That is why cycle-to-cycle frequency measurement is critical to me.  I've got the examples work with implicit setup and am still learning other parts.

Thanks again,

Bobby