PXIe-5124 measuring time delay between analog triggered signal and external TTL sync

[alexianna]

Hi there,

I am using PXIe-5124 card to acquire signal peaks (2,5 microsecond width) coming from a sensor. For this I am using an analog trigger on Ch0 of PXIe-5124. I need to know the phase (or delay) of this peaks with respect to the TTL Sync-OUT signal coming from an external waveform generator. The TTL signal is periodic, and within each period a high number of randomly time-distributed peaks can occur and must be measured, and thei phase(delay) must be known.

For the moment, the easiest working solution I found, is to use another scope card present on the PXI chassis, the PXIe-5105. I feed PXIe-5105 Ch0 with the Sync-OUT signal. Therefore I compare the timestamps of each peak with the one measured from each rising edge of the TTL signal (named t0, time zero reference) and I have the delay. To do so, I thought to run in parallel the two acquisitions, with two parallel while loops.

My solution works, but I think is highly inefficient and inelegant, with a lot of physical resource needed.

Furthermore, it can work nicely only when the Sync-OUT frequency is relatively low (<10 Hz). At higher frequencies,  I start loosing peaks or missing syncs signals.

Is there any better way to measure the delay between the peaks and raising edge of the periodic TTL signal?

Maybe using less resources (only one scope card for example) and that be possibly more robust with the varying frequency of TTL signal (my target is 100Hz max).

I was thinking to feed the PFI1 line of the PXIe-5124 with the Sync-OUT, so it would trigger at digital edge, reset the timestamp to zero and then acquire on Ch0 of the same scope-card with a different analog trigger. And retrigger everything on the next rising edge of the TTL signal. Do you think is this solution viable?

Looking forward for replies, my best regards,

Alessandro

Acquisition timing sequence

[alexianna]

Hi there,

I tried to use the PFI0 of the PXIe-5124 as Start and Advance trigger. I connected to the PFI0 the Sync-OUT signal coming from the waveform generator. 

The analog signal coming from the sensor is then configured as a Window Reference Trigger on Ch0.

It "works" but I only have a single acquisition on the Analog channel Ch0  after each Start/Advance trigger. This acquisition is correctly triggered by the Analog Window Trigger on Ch0, but all the successive pulses after the first peak are not triggered and therefore missed.

Any help would be appreciated!!

Alessandro

[Henrik_Volkers]

If the sync pulse frequency is known and stable, you can use a multi record aquisition.

Trigger on sync with a PFI digital input, record length is one periode minus trigger/record rearm delay. See manual how much ns that might take.

with every record you get a timestamp, that can be used to adapt to the periode length.... (unless you missed a pulse... 😕 )

Use a producer consumer architekture, so the producer capture the records and pipe it into a queue .. the consumer does the phase measurements..

Depending on the needed timing resolution (and if the pulse amplitude isn't needed).. a counter would be my solution. On every sync readout the counter values and reset the counter .. on every pulse record the countervalue.... (A timeintervall-analyser would be the tool :).. are they still produced? ) 

However with an analog readout, you can fit the peaks and maybe get subsample timing resolution...

[alexianna]

Dear Henrik_Volkers, thank you very much for your reply!

The sync pulse frequency in theory is very stable, indeed. This is true at least in theory... 

But I think I did't get very well your solution: you suggested to trigger on the digital edge of the incoming sync pulse with the PFI line, and then acquire the full period on an analog channel (Including all the peaks)? This will result in a post-processing of the acquired data. Did I understood correctly?

I was looking more for an on-line solution, this is why I wish I can use a kind of known counter exactly like the solution you proposed.

Isn't there any simple way to start/reset a counter on every Sync pulse, and read the counter value (instead of timestamp) at every analog reference trigger on the analog Ch0? Could someone help me with this way?

I am little surprised that a PXI system does not offer this capability :(! I hope to be utterly wrong!

By the way, the pulses from the sensor I am trying to record looks like the one in the picture attached: they have a maximum horizontal amplitude of 2.5 microseconds and vertical one of 50 milliVolt. And it is important to me to have the whole waveform recorded (not just the peak).

[Henrik_Volkers]

I also work with the 5124... but I don't know everything ... so it's a good idea to send NI helpdesk a mail and point them to this thread 🙂

With multiple record recording you should be able to capture each pulse with some pre and post samples ... but this is also a question of the minimum time between two pulses.. so some more information on the timing and would be nice 😉

And how precise you want the timing resolution. TDS can give you subsample (<4ns) timing resolution, but slows down trigger rearming ...

The problem I see is the use of two trigger sources..  in your first diagram there is a timedelay from the sync pulse to the first sensor pulse.. maybe one can switch /reconfigure the trigger source inbetween, but that's a question to NI.

How many sensor pulses do you want to aquire ? How many sync pulses  (sync frequency?)

Number of sensor pulses per periode known?

Simple solution: producer/consumer architekture. Sync to Trig input , capture one periode as one record, start fetching and postprocessing while recording the next periode. Depending on your on board memory you can capture  some fractions of a second.  (I have up to 256MSample on some cards, good for ~ 500ms)

Or couple both signals into the Trig-input (50pF? decoupling amps?) use it to multi record both channels (sensor and sync) and decide by the data what each record is from. This might enable longer aquisition time.

Or sync the generator that sources the sync pulse with the PXI system (10MHz or 100MHz;) )..   or ...

More information about your system and your goal would be helpful....

And keep in mind: A meter 50 Ohm cable is already about one 200MHz sample delay 😉

Most counter offer the functionality you need, but without capturing the analog waveform..