Re-arming trigger time with pxi-5122

[Giako]

Hi all!
this is my first discussion in this forum so I'm not sure it's the right place to post it, because i'm using LabView but maybe it is an hardware problem.

So... I have a problem to calculate the re-arm trigger time for the pxi 5122.
From the data sheets I read that it should be about 3 us with the tdc off or 12 us if on.
But I needed an accurate measure of the dead time following each record measured so i decided to find it by myself...
Using an acquisition program I have previous written in LabView, I started single acquisitions of 10000 records and each record is composed of 128 samples; as signal I used square waves with different frequencies, of 10 volts peak-to-peak (my trigger was set on the first channel of 5122 with 1 volt of edge value).

First acquisition: 50 kHz wave. Theoretically I need 0.2 s to capture 10000 records without loosing any signals from the square wave. Choosing an acquisition time for a single record of 15.3 us, I found that the time required was 0.199998 s, very similar to the expected one.
Then choosing an acquisition time for a single record of 15.4 us, I found about 0.4 s.
I can guess that in this second case one every two wave peaks goes lost so I required twice the time to capture 10000 records.

Because a wave period is 20 us I have calculated a dead time of 20 - 15.3 = 4.7 us.
It wasn't as the 3 us described for the 5122 but i wasn't impressed and i went on with my try.

Second wave: 20 kHz. I need 0.5 s to capture 10000 records without loosing signals from the square waves.
What i found was that in this case, choosing an acquisition time for a single record of 39.6 us I required 0,5 s to capture all the peaks, while with 39.7 us I required about 1 second, again twice the time.
As the previous example I have calculated the dead time: 50 - 39.6 = 10.4 us.

Very strange... the dead time I supposed being the rearming trigger (and so fixed) was different for the 2 cases.
Trying with others waves lenghts, values were always different.

This happen also fixing the frequency of the square wave and changing the number of samples per record.
For example, with 128 samples per records as I said before I needed a measure time of about 15.3 us to collect all the consecutive peaks, while with 64 samples I needed 12,8 us and so on.

So it seems to be a dependency between the dead time after a records ( the rearming trigger? now i'm not so sure if I can call so) and the sampling frequency of the pxi 5122.
But I don't know why the DAQ behave in this way.

Is it right? Rearm time should be fixed,shouldn't it?


I know it required a bit of time to read my problem but i tried to be more accurate I could.
Thank you in advance.

Giacomo

[tbob]

Just as with full blown oscope, the cannot select a timing period between the scales shown on the front panel.  For example, the horizontal setting is labled 1us, 2us, 5us, etc.  You cannot select 2.5us.  Here is an excerpt from the PXI-5122 spec manual triggering section:

Divide by

n decimation available where 1<n<65,535

For more information about Sample Clock and decimation, refer to the NI High-Speed Digitizers Help.

What this means is that when you change from 15.3us to 15.4us, the scope changes timing ranges.  That is why it takes twice as long to acquire the samples.

[Giako]

Thank you, tbob!!!

Let me see if i have understood your answer after I read NI High-Speed Digitizers Help...

My Labview software asks the user to insert the measured time and the number of samples.

Then I use these values to calculate the sampling rate (number of samples/measured time) and pass it to the niScope Configure Horizontal Timing.

The decimation method does not allow all the sampling rates to be accepted, but only an integer divisor of the onboard clock.

So changing the measured time just a little, makes the board choose the nearest acceptable sampling rate.

Is this right?

Thank you again...

[tbob · Accepted solution]

Yes that is correct.  However, I don't think that its the nearest acceptable timing rate that is chosen.  I really think it goes to the next lowest acceptable timing rate.  So if one timing rate is 2uS and another is 5uS, and you want a 4.9uS rate, the timing rate will be 2uS, although 5 is closer.  (Or maybe its the other way around)  That is why you see the time period double when changing rates by just a fraction.

[Giako]

Ohhhh finally I can go on with my work...

With some calculations I should be able to find the dead time!

Thank you again!!!

Giacomo