V 570 and kepserver

[Seversean]

Yes, those were the ones, Modicon 984. 

I did actually have one of the sockets mistakenly set to Master so that has been corrected.  I'll have to power cycle and see if there's any change.  We have 12 high pressure coolant pumps and filtration systems here running with V570's.  The objective is to get them on Wondeware for higher visibility on their status and alarms.  I'll get back to you after reboot and if still having issues upload the program.

Does anyone have any experience from the Kepware side of it?

[Cam]

I have done project with KepServer before mostly with AB but some with UniStream,  Couple questions

1)Are you running a licensed version of KepServer or is it in demo mode?

If your in Demo Mode it's limited to a 2 hour runtime and just closing the window doesn't reset that timer(Re-intilizing it doesn't do it either), you actually have to stop the process and restart it to get another 2 hour window.  Scroll through your the KepServer event log(default location is a docked window at the bottom of the KepServer window) and look for Licensing under the source header, and then check and see if the event is Time Limited usage period on feature MODBUS TCP/IP has expired.      If it's present either your demo time has run out or if you have a licensed version something is wrong and you'll have to contact PTC support about it.

2)If it is a licensed version (or non-time expired demo) have you looked at the diagonstics in KepServer to see that it's actually on the channel you set up for this device?

•  Right click on your channel select properties and verify that diagnostics capture is enable
• Do the same for your device

Right click on either of them and select diagnostics it will cause a new window to open  that will let you see data KepSever is Transmitting and Receiving,

As Joe said can you post the code so we can take a peek inside the FB's

 

 

[Seversean]

Yes it is definitely a licensed version of Kepware, we have many other devices connected for a couple years.  Allen Bradley, Koyo, and now trying Unitronics.  Current situation is I am back to IO server 1 collecting data while IO server 2 is unable.  Screenshot below is IO server 2, not getting any data at all.  IO server 1 showed good reads.  Strangely, at one point IO server 2 was getting data while IO server 1 wasn't.  They seem to be fighting over the same port still even though I've pointed the to two different network ID's and ports. 

 

 

 

[Seversean]

In the past I had one other device, a wireless sensor gateway, that would only accept one connection.  My work around was messaging the data to an AB Micrologix and reading it out from there.  That was only one device, since this is 12 devices I'd like to figure out how to get dual connections working in this case. 

 

I've attached the project file. 

 

2-4-2021.vlp

[Seversean]

 

I thought maybe this screen would be helpful.  Sockets 0-3 initialized.  Sockets 1,2 utilized.  1 being my PC and 2 being IO Server 1.  No other connects. 

[Joe Tauser]

Your PLC code looks correct.   You've put together quite a program here so I think you've adapted 😄.  If you ever do a Unitronics program again use a V700.  Way better graphics and more Ethernet power than a V570 for the same price.  PM me if you want to pontificate in private.

What version of Visilogic are you using?

I'm not experienced on Kepware, so I would default to basic troubleshooting procedures.  What happens when you turn off IO Server 1 in Kepware with IO Server 2 still enabled?

BTW - you can call Socket Init and Modbus IP Config while the program is running if you want to change parameters and test them.   If you change the port to Address of MI you can go online and enter any number you want into that MI and re-kick the socket by setting your kick bit.   This is upper-level stuff, but it sounds like this isn't your first rodeo and I think you can handle it.

 

Also, I'm not entirely sure what's under the hood in the way Unitronics handles being two Modbus IP slaves at the same time.  I'd give myself a way to turn on and off the slave functionality for each one:

 

 

You can rename the Modbus config blocks so everything's easier to read-

 

 

 

Joe T.

[Seversean]

I got it! 

Turns out the last issue was the fact I missed a port setting in Kepware on IO server 2 and it was still set to 502 instead of 503.  There's port settings in Kepware under both the channel and the device so just missed one of them.  Pulling data on both IO servers now, I'll keep an eye on it to see it continues to work.

 

Joe, wish I could take credit (or maybe not with some of the bugs) on putting together this entire program but I didn't.  I just inherited these machines running and was tasked with installing LAN adapters on all of them and getting remote connectivity capability as well as data monitoring.  Sounds like the V700 would have been the way to go but that was the machine builders decision for whatever reason. 

 

Thanks for all the help.  I'm sure I'll have other questions eventually. 

[Flex727]

14 hours ago, Joe Tauser said:

Also, I'm not entirely sure what's under the hood in the way Unitronics handles being two Modbus IP slaves at the same time.  I'd give myself a way to turn on and off the slave functionality for each one:

I've done this multiple times and there is no issue. The Vision PLC handles this without the need to add logic. I've also done multiple MODBUS R.H.R commands in the same PLC cycle with no problems.

[Seversean]

One issue I've come across is losing communication randomly.  Usually both IO servers will stop communicating Modbus, reinitializing the connection fixes it.  But yesterday I lost both Modbus and PC connection and couldn't even connect to the PLC at all to troubleshoot.  In the mean time I've put an initialize comms button on the touch screen and have the comms reinitializing automatically each day at 6am.  When I've lost comms the SB bits all show connections still. 

Not sure if anyone has any idea or fixes for this?

[Flex727]

Do you have some entries in SI 103 - SI 106?

[Joe Tauser]

Don't forget SI 101 packet timeout.  It defaults to 2 (200 ms).  Set it to 10.

Joe T.

[Seversean]

I've had issues with this still off and on.  Sometimes one Kepware server loses connection, or both.  Sometimes I lose PC connection all together.  For the time being I've put a comms reset button on the HMI to reinitialize the ports.  Even this was a pain because I'd find I'd reinitialize and only one Kepware server would connect.  I then found having a delay between the two different Modbus connections initializing fixed that and they will both return to connectivity.  I'm not sure why I lose connection, perhaps it was SI101?  It was set to 2 and I changed it to 10.  Even when Kepware wasn't communicating the bits showed there was a connection so not sure how to detect no connection automatically and reinitialize with that. 

[Seversean]

On 2/18/2021 at 4:20 PM, Joe Tauser said:

Don't forget SI 101 packet timeout.  It defaults to 2 (200 ms).  Set it to 10.

Joe T.

Looks like maybe this fixed the disconnections. 

[Seversean]

Actually I'm still getting disconnected from Kepware here and there.  Is there a bit or a way to detect this disconnection?

[Ausman]

Are you sure the issues are NOT something to do with any devices between the PLC and PC?  Or some device far away, but near a comms cable, that might do a big inductive now and then which could possibly upset things.

cheers, Aus

[Joe Tauser]

Aus brings up a good point.  Look at the physical connection path between the device running Kepware and the PLCs themselves. 

The PLCs are running as Slaves, so there's not much that can go wrong on that end.  I have a customer that is harvesting data from some V130s using PCOM and they kept losing the connection.  Of course this was my fault.  I finally went out with a maintenance guy and a manlift and  we went up to the ceiling where the Ethernet switch was installed by a collaboration between their IT people and an electrical contractor.  I found an office-grade switch secured to the roof trusses with zip ties and home-brewed CAT 5 cables plugged into it.  I measured the distance between the office and the switch and it was over the 100m length limit on CAT 5.  That is a real thing, unlike the 50 foot limit on RS232.   Also, the temperature in the ceiling of a manufacturing facility can get a lot higher than the 50 C rating on those switches.    

All the contacts on the unprotected switch had turned green and when I touched it everything went offline, including their workstation computers.  We put another switch in to get them up and later built a couple of small enclosures with industrial switches and cable glands to do a better job of sealing the environment out.

Joe T.