DAQmx PCI 6733 too slow

[mnemo15]

dear all,

I have a NI PCI6733. You can find in attached files the snapshot of the vi and the two-channel oscilloscope :  channel 1 = camera is sending a signal via PFI1 in order to trigger (start and pauyse) an triangle AO=channel 2. the exposure time of the camera is 100 ms but the A0 signal is sent only one per 2 camera signals. I have measured the time to do one loop in my vi and it is 170 ms. But I do not understand why is so slow..

Just for information this problem is gone when I am using Continuous Samples and not Finite Sample for the timing function but I need to use Finite Sample.

Could you help me

thanks

[mnemo15 · Accepted solution]

I found a solution just by removing the Pause trigger

[gaving]

For speed, you should also understand that you can "commit" your task outside of an "inner loop" that just has DAQmx Start and Stop in them.  I know the example given here is an input one, but the idea is the same

https://decibel.ni.com/content/docs/DOC-13564

Effectively, all of the stuff you are doing to create the channel and setup the triggers is all done outside the loop.  This configuration is then started and stopped in a loop.  Once you are finally done, you can clear the task (thus clearing the configuration)

It would be good for you to know that DAQmx has this feature of being in a ready or "committed" state.

-gaving

[TooEagerToLearn]

---

@gaving wrote:

For speed, you should also understand that you can "commit" your task outside of an "inner loop" that just has DAQmx Start and Stop in them.  I know the example given here is an input one, but the idea is the same

https://decibel.ni.com/content/docs/DOC-13564

Effectively, all of the stuff you are doing to create the channel and setup the triggers is all done outside the loop.  This configuration is then started and stopped in a loop.  Once you are finally done, you can clear the task (thus clearing the configuration)

It would be good for you to know that DAQmx has this feature of being in a ready or "committed" state.

-gaving

---

I like this option! But, I don't see a huge time saving as mentioned above. Start and Stop task adds overhead to my data Read. With committing the channel, I am saving about 25ms tops.

Without commit, Start and Stop task is doubling my loop time (200ms for a 1000 sample acquisition at 10kHz)

With commit, it brings it down to 170-175ms.