Visa basics question

[wd8ivl]

Here's the short version. (long version on request)

I've had a test application, written in LV8.2 and built into an EXE, running on the shop floor just fine for the past several years, in a PC that has two network cards.  One NIC is on the company network, set for DHCP, and gets an address in the range of 10.16.xxx.xxx.  The second NIC is not connected to the company network, and has a fixed I.P. of 169.254.130.1.  Connected to that NIC is a Keithly 2701 DMM with an I.P. address of 169.254.130.3.  I wrote my own VI to communicate with the meter using Visa, and it's been working great, without errors.

Just recently we had an "event" (arc?) on the power system at work.  The PC and Keithly were not powered (isolated by an open contactor), but the company network hardware (specifically, the managed switch that the PC is connected to) was on.  

Now, once in a while, (since the "event") the PC will hang when the application will attempt to query the DMM. (the application takes 100% of the CPU and everything else stops) 

I have found that if I disconnect the PC from the company network, it doesn't hang and reads the meter just fine.

So my question is this.  When Labview does a Visa write to the DMM, is it sending that write out of both network cards in the PC?  My guess is that something got screwed up in the company network switch that is allowing this Visa write to pass to somewhere else in the network and my application is waiting for a response.

I do know that other network switches were damaged by the "event"...but getting IT to look into it has been like pulling teeth (it works fine with Lotus Notes & SAP, so it must be O.K.!)

Any help would be appreciated.

Thanks,

Mike

[JÞB]

Really short version

It worked reliably

An event happend that damaged simillar equipment

it does not work reliably now

Sounds like you've done a great job on the tech side-- but, (comma - pause for effect)

Let's back that THEORY with fact to "load the gun" for getting IT to buy into your theory.  Exactly what errors are being reported?  your VI hangs - we can't analize why without the code !! post it. please and we'll have a better idea WHAT and WHERE the vi "hangs"

VISA has timeouts- so the VISA API can protect the calling application from hardware issues.  So it is not VISA (most likely) since it is uncommon for files to be corrupted by AC mains surges.  Files are 1's and 0's on media- pretty safe.

Your app MUST be what is hanging.  It might not hang when the hardware is OK but, it hangs now.....Hmmmm.....  I smell something here.....

[Dennis_Knutson]

It is the os that actually handles the low level TCP/IP communication. It decides which nic to use and I can imagine there are several scenarios where windows could get confused with problems at the switch side. Perhaps some network monitor program could see what is going on.

[wd8ivl]

Jeff & Dennis,

Thanks for the replies.  Also, I'm sorry for not posting any code, but it is at work, and I am at home.

The longer version...

There are 11 identical test systems.  The two that are acting up (that have been working fine for over a year) are connected to the same managed switch on the company network.  The others are connected to different switches and are testing along all fat, dumb & happy.

The problem original showed up on one of the PC's....then several days later, the other (the first one gets about 7 hours of use per day, the 2nd, only an hour).  The problem is intermittent.  Durring the test application that the test operator manually steps through, the DMM makes 20 to 50 different measurements...AC Volts & Amps & DC Volts at several different ranges from different parts of the product.  It has worked for as long as 10 days, and as little as 10 minutes.  It never freezes when the test application is sitting idle, waiting for the operator to advance to the next measurement, only when trying to take a measurment.  And it has frozen the application on every type of measurment. 

Thinking it was a hardware problem with the first PC, we substituted an identical (brand new) PC - it was loaded with a known good working image made of the original PC about a month ago (before the "event" on the power line)...we even disabled Windows updates & Symantec thinking it was something the I.T. guys pushed down...and it started to act up the same day.  We swapped the configurations on the two NICs in the PC, thinking it was a NIC problem.  We've replaced everything connected to the fixed I.P. NIC....Fieldpoint rack & I/O cards, DMM, O'scope, network switch & router...and even the network cables...even though everything is in one enclosure.  (It's almost like there is something on the company network with the same I.P. address as my DMM...and my VI is requesting a measurement and waiting for a responce.)  Anyway, I think I've done the due diligence on my end...

The I.T. issue is this...we're a big company...the I.T. guys that do trouble calls on the PCs (that help me out) are separate from the I.T. guys that take care of the network hardware who are separate from the I.T. guys that do I.T. security who are separate from the I.T. guys that do I.T. desktop applications who are separate from the I.T. guys that do I.T. network applications...etc, etc, etc.  So it's up to me and the PC guy to convince the I.T. network hardware guys that there's a problem.  (Right now they have my I.T. guy installing the image on a Dell, thinking that there is something odd with my Allen Bradley industrial P.C.s)  So it would be real helpful if I could tell the network I.T. guys what to look for.

Not being a network guru, should the IT guys be able to monitor the traffic on the managed switch that's connected to the DHCP NIC and see if it's getting any requests that should be going through my fixed I.P. NIC to my DMM, O'scope, Fieldpoint, etc?

Any ideas would be appreciated.

Thanks,

Mike