niRFSG_CheckGenerationStatus fails periodically

[ilyak]

Hello:

I keep observing this strange effect once in a while with my 3 channel PXI-5673 system. Just to make it clear - I have yet to find a way to reproduce this problem.

NI-RFSG 1.6.3.

There are three generators, daisy chained. I set them to generate the same file continuously for 1000 repetitions. Then after some delay I check the completion status with niRFSG_CheckGenerationStatus on all channels. In a normal case I would get 1,1,1 - i.e. all generators finished generation. However, sometimes the instrument returns 1,0,1 - i.e. the second generator didn't finish generating even though the other two did. The status of the generation flag returned by niRFSG_CheckGenerationStatus never changes. Also, niRFSG_CheckGenerationStatus never returns an error.

Any suggestions on what might be causing this error?

Thank you,

-Ilya.

[Dan_K]

Hello Ilya,

Two quick questions.

• You state that this behavior occurs intermitently. How often does it occur? Every other time, 10 times, 100 times?
• Given that the device does not return that generation has completed, is it still actively generating? Basically, is the problem that the generator won't stop or that the driver just doesn't get the update?

This will help narrow down the source of this issue.

[ilyak]

Hi Dan:

1. It happents perhaps every 50 times. Sometimes several times in the row.
2. In my code I have a loop that repeatedly generates the same waveforms on all three generator boards at different power levels. I have the same delay after which the code will start checking the completetion status.
3. As a result, I have a reason to believe that the actual generation is done and that the driver doesn't return the completion status for all cards. The completion status is '1' for two of the three instruments. It is '0' for just one of them.

Let me know if you need more information.

-Ilya.

[Dan_K]

Hello Ilya,

We still need to find out if the generation has actually stopped in this situation. The worry is that the finite generation is actually going continuously. Is there any way you could pause your program and check the output to see if the signal is still generating in this situation. This will give us a better starting place in determining what the problem is.

[ilyak]

Hi Dan:

I'll do this check, but it might take some time (since I can't reliably reproduce the problem).

Are you blocked on this particular test? Is there any other info you may use?

-Ilya.