VISA Alias Corruption

[eric.atherton]

MAX version 4.7.0f4

NI-488 version 2.73

TestStand 4.2.1

LabVIEW 2010

I have set up 3 instruments on a local GPIB network, interfaced to a PXI-8101 controller.  All works fine, except on 3 separate systems now, all the same setup, the devices all of a sudden stopped communicating.

Opening up MAX, all the alias names were correct.  Through MAX, we could querey the devices.  But running out of TestStand/LabVIEW, the devices looked like they weren't there.

Solution:  I went back into MAX, change the name of each of the 3 devices, saved them, then reverted their names back to original and saved again.  This FIXED the issue.

Has anyone else seen this happen?  Is there a solution?  Being in a medical device industry, this is very tough to explain and have a fix for.

[Kyle_M]

Eric,

Just for clarification, are these 3 instruments communicating with a bunch of 8101s, or a series of independent 3 instrument+PXi 8101 setups? Also, is there any commonality in when this occurs? After a certain iteration count, after a certain failed step in MAX, after computer reboot, etc? If so, can you force it to reproduce or is it purely random and intermittent? 

I havent seen any known issues that describe this behavior, so troubleshooting it without some further inquiry into symptoms is a bit tough.  Thank you for any info you can provide.

Regards,

Kyle Mozdzyn

Applications Engineering

National Instruments

[eric.atherton]

Sorry, I mistyped my controller.... it is a PXI-8108.  Anyways, there are 3 devices on the GPIB network, all daisychained to the same GPIB network connected to the GPIB port on the PXI-8108 controller.

There is no common mode that the tests are in when it fails, I have been using one system for over a year, and have not seen it on that system.  Since I work in the medical device field, all 5 systems are setup the same, based on the August 2010 Developer Suite.

I can not reproduce the error.

[Kyle_M]

Eric,

Unfortunately, I do not know of any general known issues which follow this behavior. It is very difficult to troubleshoot if it is not linkable to any specific step and it is as intermittent as we are seeing. It could be anything from a windows update to a virus to an extremely rare race condition/coding error, or even a possible bug. You may want to embed some tracking on the machine to check for what is going on when the failure does occur. 

Im sorry I couldnt be more help.

Regards,

Kyle Mozdzyn

Applications Engineering

National Instruments

[gsinde]

I have many instruments defined in NI MAX and every once in a while when I change an instrument connection method, NI MAX says file is corrupted, I can't talk to any of the instruments, and I have to go back in to NI MAX to set it all up again.

The last time all I did was swap two Ethernet cables between two Instruments and that should have caused ZERO changes to NI MAX. 

Is there a reliable way to save all the NI Max settings to a file so I can read them back in. I need the method to be fairly straight forward if our end customer experiences this corruption. Thanks

[gsinde]

I am building a second system with even more instruments so I will try and find any rhyme or reason to the corrupted config files. 

NiMax could use a better way to recover a corrupted config file. It knows it is corrupted, it could have saved a backup and restored without ever bothering me.