PCI-6221 digital start trigger is triggering even when input is on a different pfi pin than what is configured in my task

[Dan_Alias]

I have a PCI-6221 37-pin version.  I'm passing in a digital start trigger to PFI0 for one task, into PFI1 for another task, and into PFI2 for a third task.  If I'm running task1, it should only monitor PFI0's input as a start trigger, however pulsing on either PFI1 or PFI2 causes PFI0 to trigger.  I've monitored the inputs to the board separately with a scope and they're behaving correctly, independently.

I noticed in the NI 622x Specifications pdf page 6, it says “DO or DI Sample Clock source3..... Any PFI,” but surely they don’t mean literally ANY signal on ANY PFI pin (regardless of the one configured in the task) will trigger the start trigger?  Surely it means that ALL PFI pins can be used as digital start triggers, right?

The specifics of this application are that we’re using the digital start trigger on a buffered acquisition of about 3000 samples on 3 analog channels configured in differential mode.  One of the channels.  The goal is to plot the stimulus (the digital signal being used as a start trigger) against the responses (the two analog signals).  So we’ve wired the digital channel both to PFI0 and AI0 so that we can sample it with the AI.  There may also be a common mode issue or something going on.  I gather that you’re not supposed to tie the digital ground to the analog ground, but I don’t think that should be a problem as we’re connecting the digital ground to AI0- (pin 20 aka AI8 in single-ended mode) and the digital signal routed to PFI0 is also routed to AI0+ (aka pin1).  Since all the digital start triggers (on PFI0, 1, 2) share a digital ground, that ground is tied to three AIx-'s... but again, we're in differential mode so it should reject common mode on the AI... does it screw the PFI pin's reference though?

Another caveat that will have you rolling your eyes is that we’re using DAQmx 8.0.2 for Linux… come on you say you support it!

[Joe_S]

Hi Dan_Alias,

Let me get some more information from you about your setup.  You said that you were using three different triggers for three different tasks.  What tasks are you using and how are you configuring them (sampling rate, sources)?  Also, could you provide a connection diagram for your setup (pins, signals)?

You are correct, all PFI lines can be used as digital start triggers.  And yes, we do say we support it 🙂

Regards,

Joe S.

[Kevin_Price]

Nothing concrete, but it sounds like a hardware & signal issue to me, not likely a linux & driver issue.

I don't speak fluent EE, but I'd be putting my attention into the way you've tied together digital grounds with analog negative references.  Do an experiment where you keep those suckers separate and see if your ghost trigger problem goes away.

Even if you don't wire the analog negative reference at all and produce garbage ai data, you can still troubleshoot whether

the triggering problem is tied to your common grounding.

-Kevin P

[Dan_Alias]

We have 3 tasks that run separately in time.  Each is nearly identical, but for different inputs.  The application is a flight simulator.  We're doing a total system latency test, so we have a start trigger that comes from the control inputs.  There are 3 stick motions that are handled by a 3rd party controls system.  They turn any acceleration on one of the axis of the stick into a digital start trigger.  We've verified that they work.  For whatever reason, instead of making all 3 stick motions produce a start trigger on a single digital output, they put them on 3 separate outputs.  We don't really care that much b/c we have plenty of spare PFI pins. 

For each of the 3 tasks, the 'stop' trigger of the acquision is a photo sensor (photo sensitive diode that produces an analog output between 0-5V). For each task we measure the 200ms of data.  The sample rate is 13kHz.  For each task we have a digital start trigger configured on different PFI pins.  Then we have a reference trigger with 10 pre-trigger samples.  Finally, we measure 3 analog inputs; 1) the photo sensor, 2) another AI that you don't care about, and 3) the digital start trigger.  The reason we measure #3 with an AI is because it is required that we plot the start and stop signals on a single chart. 

All this is working.  The one strange part (the real issue here) is that if we run the "elevation" test which montiors one of the 3 axis of stick input (so one of the 3 digital start triggers), but then move the stick in azimuth axis, we can see the correct digital output from the 3rd party system going high with a scope, and the others stay low... but the 6221 triggers anyway.  This may be a ghost signal issue as the other poster mentioned.  Maybe a ground loop, etc. but our EE's are pretty sure we're okay there.  However, I wanted to rule out the strange wording on the 6221's specs and make sure that one of the low-cost limitations of this board isn't that a signal to any pfi triggers an input... since the spec's wording is strange. 

Thanks,

-Dan

[Dan_Alias]

The digital start trigger comes from left side SCT Power Tray box.  The NI 6221 is on the right side.  The middle is just a patch panel changing connector types and extending the signals over a distance.  The pin numbers on J02 are correct per NI's 37-pin i/o pinout, but the signal names may say "AO" on an "AI" channel.

I can disconnect all the AI for Elevation, Azimuth, Extension, Photosensor, etc. and just keep the digital start triggers going to the PFI pins and also to their AI channels.  This shouldn't cause a problem b/c the differential inputs are supposed to reject common mode voltage (up to the input range of the AI at least, and we're only doing 0-5V inputs, on a +/-10V range, and the common mode is not measureable with our isolated DMMs). 

I had a similar concern as you at first, so I removed the digital ground coming from the SCT box start triggers to the NI's digital ground (pin 14).  I'm working on really old knowlege, but when I was an NI AE 11 years ago, I feel like I remembered something about not tying digital ground to analog ground, but electrically that makes more sense on the RSE/NRSE configuration than differential AI.  I'd have to exceed the common mode voltage specs on the AI for that to cause an issue, and I'm pretty sure I'm not.  At any rate leaving the dground disconnect from the NI's dground didn't solve the triggering issue.  Still triggered on any PFI input not the one configured in the task.  Any thoughts would be appreciated.

[Kevin_Price]

Again, I'm not an EE, but I'd be trying to isolate the trigger stuff from the ai stuff at least long enough to troubleshoot.

I'd leave the connection from the SCT box to DGND, but would remove the connections from DGND to AI12,13,14

and would also remove the connections from PFI0,1,2 to AI4,5,6.   Basically, let the trigger signals into the digital side,

and let the analog channels float empty long enough to troubleshoot.  Then see whether you get the same bleedover

effect into multiple PFI's when the connection to AI can be ruled out as a contributor.

-Kevin P