sporadic problems closing OI

[markus kossmann]

After running our test sequence the standard Labwindows-CVI fullfeatured Operator Interface of Teststand-3.0 doesn't shutdown cleanly. That means click the stop button of the Panel for the input of the serial numbers. The panel closes. Then he clicks the close button of the operator interface. That closing process does not finish. the windows task manager shows the OI as not reacting any more. Finally the operator has to kill the OI process with the taskmanager. That happens in about one of ten times. Any ideas how to debug and solve that problem ?

[PeterAdelhardt]

Hi Markus,

does this happen with any sequence file? Have you tried the example files that are shipped with Teststand? Could you maybe post a sequence file you see the behavior with?

Thanks and best regards

Peter

[SercoSteveB]

Hi Markus

I have had a similar problem using the LabVIEW Operator Interface.  Not sure if it exactly the same as what you are seeing but it's pretty close. 

The first thing the Simple IO - Top Level VI.vi does whan it starts is call Simple IO - Get Engine.vi, this VI uses the Application Manager to get a reference to the TestStand Engine.  The call to the VI works fine and a reference is generated our OI worked fine until you came to shutdown, at which point the OI would hang at the Close Reference call to close the Engine Reference.

The only way I could find the cause in the run-time environment (everything worked fine in the development environment) was trial and error which took ages.  Having identifed it however and realising we didn't need a reference to the TestStand Engine for our application, we took out the call to Get Engine.vi and the corresponding Close Reference and now we shutdown successfully 100% of the time.

More of a fix than a solution, but hey a working system is a working system.

Regards

Steve

Message Edited by Cheesey on 11-25-2009 04:07 AM

[markus kossmann]

It happens with specific sequence files for these testbenches. However when I tried to reproduce the problem, I couldn't do that so far.  The testbench runs a temperature cycle and runs about 6 hours  to 16 hours depending on the number of UUTs (we are using the batch model ).  So the sequence is run only once a day. Everytime I did some debug effort , the problem didn't show up again. So I don't think that running another sequence will show that problem.

The problem is that our sequence also depends on having specific hardware like the temperature chamber. If the sequence is of any use for you without the ability of running it I'am happy to post it.