PCI 5114 Timestamp

[John129]

Hi,

I was wondering if you could help me. The experiment I am working on involves receiving a series of trigger pulses (slow rep rate) and another series of trigger pulses (higher rep rate). The phase between the 2 different triggers are random and there are multiple high rep rate triggers between 2 of the slower ones. I was wanting to use the digitizer card (PCI 5114) to take a trace from one of the channels triggered of the higher rep rate trigger but be able to know the timestamp of each trace relative to the slower rep rate trigger.

Many Thanks,

John

[Ian Colman]

Hi John,

I am currently looking into this for you, but on the face of it, it certainly seems like something that can be done. In order to proceed though I'd like to know what development environment you are/will be using to program this acquisition task (eg. LabVIEW 8.5). If you could let me know I will do my best to suggest a viable approach to solving this problem.

Best Regards,

Ian C
Applications Engineer
National Instruments UK & Ireland

[John129]

Thanks for replying,

I am currently working on Labview 8.2. Ideally I would be looking for something which resets the count number on the timebase to zero upon receiving a digital pulse (on the slower rep signal) although  from reading the forums this appears to not be possible as only a hard reset seems to do that (Although receiving a timestamp for this pulse would enable me to do something similar).

Regards,
John

[Ian Colman]

Hi John,

I think I understand the problem that you are facing. To confirm though, do you wish to know the values t1, t2 and t3 in the attached diagram? Please let me know if that is right or whether I am mistaken with those times. Also could you let me know the aproximate frequency of the 2 pulse trains used (eg. are we talking 2 or 3 Hz or are the trains at more like 2 or 3 Mhz?).

Best Regards,

Ian

[John129]

Yes those are the times im after. As for the time between triggers:

Slow rep rate trigger: ~ 10ms - 1 sec (although for future work smaller times may be desirable but will never be shorter than fast trigger)
Faster rep rate trigger ~ 1ms (currently)
Trace length ~ 10-20 us (triggered of fast trigger)

[John129]

Hi, was just wondering if there was any update on solving this problem.

[John129]

Had another look at the problem and realised that the time between the short pulses is short enough that I can assume the first pulse comes in at time=0 and reference the rest from there. The idea I came up with was to use the long pulse as a start trigger and the short pulses as reference triggers and tell the program to acquire a certain number of them. This works fine but it means I have to initialise the card each time I do so which takes about 0.5 seconds. I was therefore wondering if after aquiring the last reference trigger the card could be told to wait until another start trigger comes in before taking another series of reference trigger.

[Tom O]

Hi John,

I've read over this thread and understand what you're trying to do on a high level. However, I don't really understand what you were proposing in your last post, or how this would help you establish the relative positions of the faster pulses. Could you explain this again?

Also, software analysis be suitable for this? Rather than using the hardware and counting samples between edges etc., you could use software to find the faster pulse edges and find their relative position in time - would this be suitable?

[John129]

Esssentially what I was proposing in the last post was that the exact position of the fast pulse (to within a period or so) may not be exactly necessary. i.e if the first pulse came in at 1ms it could be round down to sayin it came in simultaneously with the slower pulse (as im wanting to see the general evolution of the sampled waveforms over the slower pulses period). Its a bit of a bodge job but it might be possible to work in the short term but the problem is the only way i can get it to give the faster pulses relative to slower pulses is to reinitialise the acquisition each time which takes time (if i don't it seems to ignore the next slow pulse and references all the fast pulses to the initial slow pulse). Personally i would prefer to have an exact measurement relative to the slow pulse but its proving hard to do.

As you say "you could use software to find the faster pulse edges and find their relative position in time", but the problem is relative to what (as I cant find a method to get an absolute time from the slow pulse or a relative time between slow or fast pulses. Chances are I'm missing something rather simple.

[Tom O]

Hi John,

Sorry it's taken a little while for me to get back to you. I understand your problem properly now, and have been playing around with bits of code and various examples on a scope card of my own.

I think the best way to go about finding the relative position of the faster pulses would be:

1. Read repeatedly from the scope on a trigger. This trigger is going to be your slow rep rate signal.
   
• There's an example of this bundled with the driver; simply go to Help » Find Examples, then to Hardware Input and Output » Modular Instruments » NI-SCOPE » Demos » niScope EX Configured Acquisition.vi.
2. On each iteration of the program, read a given number of samples. This is going to be the window of data we analyse.
3. Analyse each of these windows to find the rising edges. You could use waveform analysis VIs or some custom array manipulation. Once you have found an edge, use the sample # to work out the time relative to the trigger.

Bear in mind that this method will require you to specify a number of samples to acquire at a time; if there are fast pulses after the samples you've acquired and before the next pulse, these will be missed.

With that as your general plan of attack, have a try and post back if you have questions on any particular part of that. Good luck!