On a PXI-6602 frequency card, can you change the average sampling period

[tmklaves]

I am using a PXI 6602 frequency card to measure the frequency of a rotary flow meter.    The 24V square wave signal coming from the rotary flow meter is knocked down to a 5 V TTL signal and then sent to the PXI card  I am measuring frequencies in the range of 24 - 100 Hz.  My problem is the signal I am recieving from the flow meter is very unstable and looks oscillatory (+/- 0.75 Hz).  I have ruled out the flow meter because similar results were found with another identical flow meter.  As part of the troubleshooting, I measured the frequency at the flow meter using a FLUKE 123 scope-meter.  With the fluke, I got similar results with the time/div set to 10 ms, but when I increased the time/div to 20 ms, the signal smoothed out dramatically and I recorded and stable single with +/- .05 Hz noise.  It seems by increasing teh times/div, I increased the sampling period the FLUKE used to calculate the Hz.  I would like to try something similar with the PXI 6602 card, but don't know who to change the average sampling period. 

Tom Klaves

262-636-8491

[Kenn_N]

Tom,

Thanks for contacting National Instruments.  If I understand your situation correctly you are needing to provide your own timebase to the 6602.  Our 6602 series counter timer boards use the default onboard clock to make it's measurements.  You can provide a different clock source externally and specify the location of this clock in code.  So for period measurement you can specifcy an alternate source to the onboard 20MHz clock.  If you are using LabVIEW then there is the DAQmx Channel Property Node using Property Counter Input >> General Properties >> Counter Timebase >> Source, and you will also have to specify it's Rate in the same property node.  If you are using a different development system there are functions that do this.

You can route this signal from a PFI line, and that PFI line is what you specify as the source in the previously mentioned property node.  Let me know if you have any further questions.

Regards,

[tmklaves]

Kenn

We were able to correct the problem by changing the "Measurement Method" for the PXI-6602 channel within Measurement & Automation Explorer program.  Originally, the method was sent to "1 counter (Low Frequency)" to "2 Counter (Large Range)".  This brings up two questions.  First for the "Large Range" method, what is the math behind calculating the output frequency and what exactly is the "Divisor" specifying.  Secondly there was a second "2 Counter" option which was specified for High Frequency.  This option had an user input for "measurement time".  Again, what is the math behind calculating the frequency and how does changing the measurement time effect the result.  For some background, I am measuring a frequency range between 1 Hz - 75 Hz.

Tom Klaves

[Kenn_N]

Tom,

Glad you've been able to make some progress with your application.  On to your questions.

First Question.   
    2 Counters (Large Range)—Uses one counter to divide the frequency of the input signal by Divisor, creating a lower-frequency signal, more easily measured by the second counter.

Second Question
    2 Counters (High Frequency)—Uses two counters to count pulses of the signal to measure over a predetermined time, specified with Measurement Time.  Increasing this time shouldn't affect the result.  You'll just see more pulses, but you will also divide by a larger time.

For your application, measuring the frequency of 1-75Hz you will only need to use the 1 counter method.  Those are very low frequencies and 2 counters are not needed.  The 2 counter methods are used for frequencies greater than the onboard 20MHz timebase.  If the signal is faster than that there will not be enough time to get a count of how many times the 20MHz clock pulses between the signal's pulses.

Regards,