Can a NI-9775 housed in a cDAQ-9181 be synchronised with a USB-6356?

[RodosC]

I am working on a project that requires synchronisation between a generated signal and an acquired signal. Presently, I am using a USB-6356 with LabVIEW NXG to generate and acquire signals, as well as sycnhronise generation and acqusition. 

I would like to use the NI-9775 module my new analog input because of its higer sample rate. My question is, is it possible to house the 9775 (need only two analog inputs for signal acquisition) in a cDAQ-9181 and synchronise it with the USB-6356 do i need a different cDAQ?

[Kevin_Price]

Well, you're gonna have your work cut out for you.

Normal methods of synchronization depend on sharing digital timing signals between the two devices.  Your chassis and module provide no I/O connection points for any such timing signals, so you're kinda dead in the water on that front.

   Even with them, you'd have some further work to do since the 9775 uses a Delta-Sigma converter which induces an unavoidable delay in the signal path.  You'd need to compensate for that as well.

For "pretty good" sync, you can capture a trigger-like signal with each device as a means of aligning their data streams after the fact.  You'd still need to make sure this extra task on the 6356 was fully sync'ed with the AO generation task.  (On that device, a simple shared sample clock is sufficient.)   

    And this would technically only align the *start* of the acquisition.  After that, the two devices are governed by their own internal timebases which will make their timekeeping skew a bit over time.  I know a lot of MIO boards are rated at 50 ppm accuracy, which works out to 3 msec per minute of acquisition.   Another way of looking at it is 50 samples worth of (potential) skew per million taken.

-Kevin P

[RodosC]

Hi Kevin, 

Can I instead use a cDAQ-9185 which features a single PFI0 line? Or do I need two digital lines to share the clock as well as the trigger? Can one of the digital lines be substituted with one of the analog inputs?

[Kevin_Price · Accepted solution]

A cDAQ-9185 4-slot chassis gives you a good chance.  I don't have a ton of experience and knowledge about signal routing limitations on cDAQ, but I'd be trying something like this:

Option 1:
    Consider the 9775 as the master task and export its sample clock out to the chassis' PFI0.  Wire that over to (probably) any PFI terminal of the 6356 and configure DAQmx Timing for your AO task to use that PFI terminal as the sample clock 'source' input.  Then be sure to start your AO task first (before any sample clock signals get generated) and your 9775 task last.

    I *think* the sample clock signal will start showing up *after* the inherent Delta Sigma delay of the 9775.  So at the moment the first clock pulse occurs and you acquire your first 9775 sample, it's telling you about the state of the analog input terminal a short time in the past.   That's also the moment you (finally) generate your first analog output sample.  So you'll be misaligned until you figure out how to realign.  The spec sheet will show that there will be an integer # of samples that'll get you very nearly aligned though not quite perfect.

Option 2:

   Do your AO with a 2nd module in a multi-slot chassis.  You'll need to do most other config things similarly, but can route the sample clock internally within the chassis without the need for a PFI terminal or physical wiring.

-Kevin P