Hello,
We see the DAQmx -200279 error during extended runs of continuous acquisition. We can go for hours without seeing the issue. I have attached the main blocks of our data acquisiton module (daq_code.c).
Some observations / comments:
- Debug build
- We saw much improvement setting sleep policy to VAL_SLEEP_MORE
- The nSamples value in callback when error occurred appeared to be a value of 8 (at least in CVI debugger). We had setup the callback to be called every 800 samples.
- We do use remote desktop connections to view extended runs. Although, it does appear that the issue occurred at least once when RDP was not being used.
- Source module for DAQmx acquisition is built into a DLL. The application and DLL share a thread safe queue.
Questions:
- Is it advisable to override the default DAQmx buffer size?
- Is it OK to assume that both DAQ1 and DAQ2 will have the same number of samples when Callback function is called? (since they are synchronized in initialization)
- Is there a way to ignore the -200279 error and not lose the proper order of samples based on channels
- Could the counter task have any impact on AI tasks?
Thanks!