Upgrading NI-VISA impact on Old Version of LabVIEW

[N8D11]

We are still developing in LabVIEW 8.2 and have had some troubles with our GPIB system. I have upgraded NI-VISA to the latest release thinking this would help if there was a known issue with the driver that fixes the problem. After looking at the compatibilty chart for NI-VISA and LabVIEW, I see that the last compatible version is NI-VISA 4.6.

My question is should I revert my machines back to VISA 4.6?

Would intermittent communications issues in an executable be caused by an incompatability within the LabVIEW application I am developing?

Any insight would be greatly appreciated.

Thanks,

Nate

[N8D11]

Also most of the target machines are using an Agilent GPIB card, with the Agilent VISA installed in side by side mode.

[smercurio_fc]

Intermittent communication issues are likely a hardware problem, such as a loose connection or a bad cable. I would suggest checking your cables and connections. You might also have a bad instrument. We've seen our fair share of these, as when an Agilent signal generator starts going bad and creates "intermittent communications issues".

[Jesse_D]

Downgrading VISA would be a good troubleshooting step also, just following the compatability chart. What kind of issues were you having before you upgraded? 

[N8D11]

We are talking to a Delta Design Oven and all goes fine for many interactions. Seemingly randomly the communication bus goes down(no intruments seen by the agilent connection expert nor MAX). This bad state is resolved by power cycling the oven in question. Then all instrument communication resumes as normal. I have attempted to resolve the bus issue through sending an interface clear but that does not fix the oven issue.

Is it possible that Performing a read immediately after the write could be causing this? Perhaps a VISA Clear would help...

[Jesse_D]

Hi Nate, 

A read should just monitor the com port for incoming data. A clear just empties the computer input and output buffers. It is an easy step to try a clear but I doubt that it is problem.

I tend to agree smercurio that this is a hardware issue. Is there a way you can eliminate the bus and test the instrument directly?