serial port hang by VISA driver when run parallel model in TestStand

[yangyang-rexxar]

Hi Man

I got problem when I run VISA serial write VI (a VI called Serial Write.vi written by another team), when I run the sequence in parallel model, in a step what is called "Enter Test mode" and send command to UUT, the UUT often need several seconds to response. If the two sockets are all in this step, the sequence will hang, if I run only one socket, it will be OK.

I am not so sure if it is a software problem, I know that there are some similar station for a whole series products with similar sequneces run similar "Enter Test Mode" step, but other stations have not raise up similar problem.

And, after the process hang, when I try to close the sequence, VISA error pop up, code is -1073807313, hex BFFF002F. I check related information what are all about VISA EVENT, this made me confused.

[yangyang-rexxar]

Resolved.

Just put long time serial port reading to a loop and reduce the timeout from 20s to 1s, this will let another socket enter the hardware and optimize the sequence.

[dug9000]

You might consider using a teststand lock step type or lock synchronization setting on your steps instead in order to avoid multiple threads accessing the hardware at the same time.

-Doug

[yangyang-rexxar]

Hi Doug

I have 2 serial ports, I think you mean the two ports share only one controller shipset, so I need to lock., if it is DAQ card, I often use two DAQ cards for two sockets, this case I do not need to lock it.