Trigger jitter on a PXI 5124

[wjm68]

I am using a PXI 5214 digitizer in multi-record acquisition mode.  I am running two timers on a PXI 6608 card to provide the timing for my application.  One counter output (CNTR 0) is a pulse train, software triggered, output to its default PFI pin.  The second counter (CNTR 2) is configured to generate one pulse (1us high, 1 us low routed to PXI Trigger 1) on each falling edge of the pulse train output from CNTR 0. 

This works well and I am using the PXI Trigger 1 as the source for both the START and ADVANCE triggers for the 5124 acquisition.  As I zoom into acquired data, it is apparent I have significant trigger jitter, in the range of 30-40 ns.  I have looked at my 6608 pulses with a high quality storage scope an they look very good and repeatable, so I am confident the source of the trigger is not my problem.

Basically, all of my 30 records of data (20000 points @ 200Ms/sec each) appear to have been triggered at slightly different times (about a 30-40 ns window).  My goal is to average these waveforms after acquisition, but doing so with this kind of jitter is not effective).

I have configured the clock on the PXI 5124 to the PXI 10 MHz source, but noticed no improvement whatsoever.  Any thoughts?

[Jeff_B.]

Have you considered using the Reference Trigger in your application?  You should see better jitter since the resolution on the Reference Trigger is sub sample clock (when using the Internal 200MHz sample clock).  There may be some async delay to calibrate out, but it should be consistent.

If you are only interested in the samples acquired after the trigger occurs on PXI1, set the Reference Position to 0%.  Also, change the Start and Advance triggers to immediate.  If there are some other conditions you need to occur to arm the trigger, perhaps you could use the Start and/or Advance for those.

-jeff

[wjm68]

I already have the reference position set to 0%.  I am pulsing a pulse/receiver unit and want to acquire 1 record for each trigger to the pulser receiver, it may be several hundred.  I am doing a multi-record acquisition so I need to use the START and ADVANCE triggers to get each record, as I trigger the pulse receiver each time.  Right?  I do not see where use of the reference trigger is of use in this application.

The basic steps are:

- Arm the digitizer with the Start vi (it is now waiting for triggers)

- Start my timing DAQmx task. (Pulses to my pulser receiver and PXI Trig 1.

- PXI Trig1 is config'd as the start trigger

- PXI Trig 1 is config'd as the advance trigger also, to aquire the rest of the records after the 1st

- Stop acquisition (Stop vi)

Repeat in loop

[Jeff_B.]

A multi-record acquisition can have the source of the Start and Advance set to immediate, so just because you are acquiring more than one record doesn't necessarily imply you have to have an external Start and/or Advance trigger.  Your current application is using the Start and Advance triggers to determine the timing between records, and these triggers have a resolution of 20ns on the 5124.  I was trying to suggest using the Reference Trigger instead, since it's resolution is better.

The modified flow would be:

1. Configure the digitizer to have Immediate Start and Advance triggers.  Reference Trigger = PXI Trig 1.

2. Arm the digitizer with the Start vi (it is now waiting for Reference triggers)

3. Start my timing DAQmx task. (Pulses to my pulser receiver and PXI Trig 1.

4. Digitizer sees PXI Trig1 edge and triggers 100% post triggered input acquisition

5. Digitizer moves to the next record asap (since Advance is immediate) and begins waiting for another Reference Trigger on PXI Trig1

6. If all records are not done, go to 3

7. Stop acquisition (Stop vi)

[wjm68]

Thanks.  I have implemented the Reference Trigger as you described and notice significant improvement.  If I acquire many records, say 50, all are grouped very tightly, without the 20 ns jitter previously noted.  Now the jitter seems to be about 5 ns or less.

After I fetch all data from the digitizer it is then re-armed via software control.  I seem to lose sync at that point.  The next group of 50 records are all tightly aligned relative to each other, but slightly off from the previously acquisition.   I am confident it is not a function of the pulser / receiver unit.

Now I want to sync the PXI 5124 200 MHz sample clock to the PXI 10 MHz clock (the PXI 6608 card in slot 2 is the source of my trigger pulses).  I am not sure exactly how to use the NiScope Configure Clock.vi.  For the PXI 5124, do I only need to specify the source to syns to (PXI 10)?  What about the 'sync pulse' input parameter?  Does it pertain to the PXI 5124?

In other words: If I sync to the PXI 10 MHz signal, do I need to re-sync just after re-arming the 5124?  On what line does the 5124 receive the PXI 10 MHz signal?  What triggers synchronization?  It seems to go out of sync after acquisition is complete, and another acquisition is initiated.

[Kenn_N]

wjm68,

You can PLL the 10MHz clock to your card using the niScope Configure Clock.vi.  The help explains what sources you can use for a reference input, the default is the chassis CLK10, or 10MHz clock.  Hope this helps.

Regards,

[wjm68]

What about the 'sync pulse' input parameter?  Does it pertain to the PXI 5124?

[Sean_C.]

The Sync Pulse parameter is used when synchronizing older devices such as the NI 5102 and the NI 5112.  It does not apply to the NI 5124.  The following is an excerpt from the High Speed Digitizers Help document:

"For the NI 5102, specifies the line on which the sample clock is sent or received. For the NI 5112/5620/5621/5911, specifies the line on which the one-time sync pulse is sent or received. This line should be the same for all devices to be synchronized."

I hope this helps,

Sean Close
Product Support Engineer: High-Speed Digitizers

[wjm68]

Thank You.

I can clearly see a difference when I sync the sample clock to the PXI_CLK10.

What more can I do beyond synchronizing the 5124 sample clock with the PXI_CLK10?  Jitter is still in the 'nano-second' range.  How do I get closer to the 'pico second' range spec?

This jitter is present when triggered via a 6608 generated pulse routed to the PXI Trigger line, and also when I use an external trigger (an electrical split of the trigger I am presenting to a pulser/receiver unit).  Knowing this, I am assuming it truly is a trigger jitter problem.

Any feedback would be great.

[Kenn_N]

wjm68,

The 6608 is a board that can be placed in slot 2 to drive a high accuracy 10MHz clock.  This is best clock you can get in your setup.  Moving yuor card to slot 2 is the most I can think of and you may have already done that.

Regards,