Instrument handle through a shift register generates error code "-1073807346"

[NNS-CAO]

I get an error when I pass an instrument handle through a shift register but not if I pass it straight to the next subvi. I am using LabVIEW 2011, NI-Sync 3.3 and a PXI-6652 timing module. This vi worked fine previously in LabVIEW 7.1. Did something change that would cause the shift register to no longer work for instrument handles?

[Ben]

I certainly hope not!

Please post images of your code for us to review.

The VI with the SR is not going idle is it?

Curious,

Ben

[NNS-CAO]

Ben,

I have attached an image of the problem.

[Ben]

If the two calles to the AE are made from the same running top level VI (the AE never goes idle) then there is something very strange happening.

Q:

If you replace the For loop with a While loop (with constant to stop while loop) does that change things?

The only time I have seen trouble with your code is when the VI went idle between the two calls.

As a sanity check you can put an indicator on the ref to make sure it does not change from the first call to the second.

Ben

[NNS-CAO]

Both calls are made from the same top level VI. I have tried replacing the For loop with a while loop and have the same result. I added an indicator and did not see a change in the value between calls.

[Ben]

You have both stumped and troubled (thank you!).

Last Q:

Is the action a type def and the constant used to select the action copies of that type def?

Ben

[JÞB]

There's got to be something fishy.  Two places to look:

Check and make sure you don't have tunnels overlapping in any of the cases (I've had it happen with a unwired tunnel set to use default if unwired with a wired tunnel to nowhere on top of it) a CTL+U usually exposes these.

Check and make sure you aren't accidentally calling a close case (probe and breakpoint on the mode control) 

A conditional probe on the output SR's input (after the case) set to suspend if invalid may help too

[NNS-CAO]

Ben,

Yes, it is a typedef and the constant in the top-level vi is tied to it.

[NNS-CAO]

Jeff,

I checked for overlapping tunnels and have none. There are no accidental calls to the close case.

This code worked previously in LV 7.1. I have made no changes to the code other than converting it to LV 2011.

[JÞB]

OK, something really odd then (possibly a bug, possibly a mix-n-match on NI-SYNC API )  We'll have to look at the code itself.. first thing Check the VI hierarchy show vi.llb and full paths - is the API mixed from two different LabVIEW versions?  If you can't find it that way send the working AE <save as- to new location-preserve Hierarchy-include VI.llb as a zip and same-same for the bad AE. 

But it almost has to be a mixed NI-Sync API at this point or the 2010 NI-sync got corrupted- we can compare the .zips to clean installs