NI Modbus TCP Server freezes with ASPENTECH

[xtabi2]

I have a Labview application called CMS built in 8.5.  The CMS program communicates with several instruments using RS485.  It then stores and distributes the instrument data using OPC and Modbus TCP. The program runs on an embedded single core computer using Windows XPe and I am using the NI Modbus TCP Library for the Modbus Server and a 3rd party OPC application.

When I connect the computer to the network and start polling the Modbus server using Aspentech's Modbus Utility the CMS application will at random utilize 100% of the CPU and never recover.  If I restart Windows and unplug it from the network it runs fine.  My only guess is that the NI Modbus Server is crashing, but I've used the exact same Modbus Server code in other applications using other polling utilities (OSI PI) and it works great.  Has anyone else experienced similar issues with the NI Modbus TCP library being inconsistent and have any adivce? It's driving me crazy... 

[Chris V]

Hi xtabi2,

In your other applications, have you been running on Windows XPe?

How does this application compare in scale to your others, i.e. is it much larger?

Is there any sort of repeatability?  Do you notice anything always happening before the CPU spikes?

Have a great day,

Chris V

[xtabi2]

Chris,

Thanks for taking a look at this..

All of our other applications have been running Windows XPe.

It's about the same. We support 5 different instruments with this system thermocouples, pressure/temperature probes, flow meters, differential pressure probes, and our proprietary air in-leakage probes. This is the first system to use all 5, but from a number of instruments perspective it is not the biggest.

There is no repeatibility that we can see so far.  Some days it will reboot 5 times others 20 times, the only slightly dicernable correlation we have found was that it seems to reboot more frequently during business hours (our assumption here being increased network traffic).  We did notice a slight correlation between CPU temperature and a reboot so we installed a fan, but it did not help. We assume the CPU temperature is then an effect of the CPU spiking to 100%.

Some things I forgot to mention that may help is that we have another embedded system connected to the same network with only our air in-leakage probes attached running different labview software, but still using Windows XPe and it runs fine.  It however is being polled by the ASPENTECH OPC Utilities.  I also discovered that the network these are hooked up to is connected to a VPN enabled router.

Thanks!

[Chris V]

Hi xtabi2,

I just want to make sure that we're talking about the same thing.  The library you are using this this one?

So that you know, this library is not officially supported by National Instruments.

Have you tried running your program on XP Pro?  Depending on your specific flavor of Windows XPe, there could be a lot of different things going on.

Are you doing large batch reads/writes?  If so it may be helpful to break these up into smaller batches.

I would highly encourage you, if you have the Standard Service Package, to email into National Instruments support for this issue.

Have a great day,

Chris V

[xtabi2]

Chris,

Yes that is the library I'm using and I know it's not supported. Is there another supported option?

I have not tried a different OS. I will look into whether I can configure a computer with XP pro and have it connected to the ASPENTECH and

instrumentation.

I don't have any large batch read/writes. I write 1 min averaged data to our data files and the most average flaots it keeps in RAM is 8.  The biggest section of memory I maintain is the Modbus registers, but those are read and written 1 instrument at a time (max 8 float values). 

Thanks for your help Chris if I still see issues with XP Pro I'll contact NI support.

[Chris V · Accepted solution]

Hi xtabi2,

National Instruments does offer a DSC Module that might suit your application.

Have a great day,

Chris V