PXI

cancel
Showing results for 
Search instead for 
Did you mean: 

Strange sync problem using PXI 4472B, PXI 6624 and 8196 RT controller

Hi everyone.
I developed an application in LV RT 7.1 to acquire simultaneously  from several PXI 4472B boards synchronized with a PXI 6624 counter/timer board. I use several channels from the 4472 boards to acquire vibrationn signals and 3 counters from the 6624 to acquire the position of 3 encoders, and finally obtain the vibration waveforms, plus the corresponding position arrays (respectively to each vibration sample).
To synchronize the boards I put the 4472B boards on the first slots, so one 4472B resides on slot 2 and is the master. A fourth counter on the 6624 takes the oversaple clock from the master 4472B and generates the appropiate sample clock to be used by the three counters that are performing the encoder positions acquisition. The slave 4472Bs are synchronized as usual.
 
This technique has worked perfect for me using different RT controller models and DAQmx versions, but when I tried to run it on a 8196 RT controller I have weird problems. 4472B boards work together perfecly as usual, but the three counters on the 6624 seem to acquire at a slightly higher rate. The result is that the buffers for counter acquisition overflow and the process fails. It always seems to show the same acquisition rate proportion: about 30% higher for the first counter, 10% for second counter and 20% for the third counter, having 100% as the 4472B rate. It is extremely strange that the three counters are supposed to use the SAME signal as clock (the output from the fourth counter).
 
I have tried with different combinations of DAQmx versions, LV RT 7.1, LV RT 7.1.1, etc. but the problem remains the same.
 
I have only one more thing to try: I have used DAQmx 8.0 and 8.1, but I haven't downloaded version 8.3. I hope this could help, but I won't try it until next week, because now I'm on a trip. Anyway if any of you have some idea about this kind of problem using the 8196, please reply!
 
Any help would be appreciated. Thank you.
 
BTW, another singular behavior is that DAQmx Read vi does not show any error.
 
I attached MAX reports showing the last configurations I tried (of many) using an 8187 controller (named PAB03, which works well) and the 8196 controller (named PAB05, not working properly). Don't care on the 8196 identification shown wrong as 8187 because this was one of my last "experiments" to see if it had any effect.
 
Daniel R.
0 Kudos
Message 1 of 3
(3,567 Views)
Hi. Finally I have installed DAQmx 8.3, but the problem remains the same. Now, I found that the apparent difference between counter rates and DSA rate is proportional to the DSA rate itself. Please look on the attached image.
For example, if the 4472B master is set to acquire at 10000 s/s, counter 0 on the 6624 seems to acquire at 12300 s/s, counter 1 at 11100 s/s and counter 2 at 12550 s/s. Even when the three counters are supposed to use the same counter 3 output as source. Funny, isn't?

Please help! Smiley Sad

Daniel R.
0 Kudos
Message 2 of 3
(3,515 Views)

Hi. Finally the problem has solved. I still don't know the reason for the strange behavior, but I could workaround it. As I told you, I was using three counters on the 6624 to acquire encoders positions, and a fourth counter to generate the appropriate clock signal from the DSA master oversample clock signal. I used counters ctr0, ctr1, ctr2 and ctr3 respectively. To workaround it, I just used any of counters from ctr4 to ctr7 to generate the clock signal, instead of ctr3. Voilá! Works as expected. Seems to be somethig related to the first 4 counters, versus the last 4.

I would appreciate comments from someone at NI about the possible reasons for this phenomena.

 

Thank you,

 

Daniel R.

0 Kudos
Message 3 of 3
(3,280 Views)