monitoring a voltage signal for peaks

[Peterbon]

hello everyone.

i'm trying to make a "counter" VI that monitors a voltage signal and keeps a running tally of the amount of peaks (for use in counting hits/time to an alpha particle detector).

i've tried using DAQ but the card available to me is pretty old and i can't manage a sampling rate of more than 9Hz, which is too slow for my purposes (the peaks do not last long enough).

my last resort is to use an oscilloscope (Agilent 546000) with GPIB and set up an internal trigger (perhaps a level threshold) where it outputs something through GPIB everytime the trigger is tripped.  i read carefully through the programming manual (http://cp.literature.agilent.com/litweb/pdf/54600-97032.pdf) but i don't understand how triggers work.  is what i'm trying to do even possible?

thanks so much for reading.

Peter

[Collin M]

Peter, I don't know if an oscilloscope will help you solve your problem. Most oscilloscopes can't take continuous acquisitions. Instead, they take "snapshots" in time. The "snapshots" aren't necesarily contiguous in time becuase the oscilloscope needs time to re-arm after each acquisition. This usually isn't a problem when using an oscilloscope because it doesn't matter which event is captured. For example, if I want to capture the rising edge of a digital clock signal I don't care if I capture the 38th clock edge or the 476th clock edge. I just want any clock edge because they are all the same. I can setup the scope to continuously capture clock edges but they won't necesarily be contiguous clock edges.

All that being said, it may be possible to use a scope. If you can assume a maximum frequency of the peaks you might be safe. For example, if you know that peaks can occur at most once per minute the oscilloscope will be able to capture all peaks and notify the host through GPIB every time. Can you assume anything about the frequency or duration of the peaks? Do you know what the maximum frequency and minimum/maximum durations are?

Also, 9Hz seems very low for an NI DAQ board. Which one are you using? You may be able to program it in a different way to get better bandwidth.

You also asked about oscilloscope triggers. A trigger is usually specified as a voltage level and an edge direction. The oscilloscope monitors the voltage of the signal and when it samples the signal at that level it captures some data before and after that, and then displays the data as a waveform. For example, assume I set a trigger for 1.3 volts and a rising edge. As long as my input signal stays below 1.3 volts the scope will not trigger and nothing is displayed. Once the input signal crosses from below 1.3 volts to above 1.3 volts the scope will trigger and display the captured waveform. Note that oscilloscope triggers should not be confused with GPIB triggers. They are definitely not the same thing.

[Peterbon]

thanks so much for the response.

i fear my project is impossible

i've watched the signal on the oscilloscope screen by eye and the peaks can come as often as 3 times per second.

the DAQ board i have access to is very much obsolete.  DAQmx doesn't support it so i've had to use the old DAQ software.  i don't have it in front of me but its called something like PCI 1200A.  i acquired data from it by looping "AI Read."

Thanks for the explanation of oscilloscope triggers.  next time i get a chance, i'm going to see if triggering will keep up with the peaks but it seems like a lost cause.

Peter

[Collin M]

Is the signal you are measuring digital? If so you might be able to use a counter like the PXI-6602.

[Dennis_Knutson]

Even though the PCI-1200 is obsolete, it was rated at 100ksamples/sec and if you are only acquiring at 9Hz, you've got something very wrong in your program. Why don't you post what you've written to see where an improvement could be made.

[f.grasdepot@laposte.net]

Hi,

looking a 3hz event with GPIB is not a lost cause at all,

use a serial poll to detect an event is also quite feasible and not too
hard ...

Could you tell us the level of you event since if of the order of 5V (
amplified perhaps), you can use the parallel port and read it ... This
is free and saves quite an expensive oscilloscope ! ( but has not the
same flexibility ... you are right )

Regards,

FG

[Collin M]

If you are certain that the pulses will come at no more than 3Hz then it may be possible using an oscilloscope and GPIB. The limiting factor will likely be the scope, not GPIB.

In my opinion Dennis's suggestion is the best path to follow. Can you post your vi?