Kepserver Modbus TCP IP Dropping and Timeouts

[AC_10026208]

Kepserver scanning modbus devices, each channel has three devices preset, which all three have the same IP and ID. 
Port 502 is used. Setting 1 socket per each device on channel is set.

I am getting big number of connection timeouts and random data collection disabled. All timeout/retires settings are already maxed out.

Attached are event log from offshore live system, some screenshots to present config from demo/test system and a .csv export example of one device from three on one channel.

[cmorehead]

@AC_10026208,

The provided Event Log shows two separate issues going on. There are a lot of "Attempt to Add Item" failures which indicates the Client application is making requests for an address that does not exist in either the KEPServerEX space or the device logic. Here is a link to the Kepware knowledge base article that covers this error:

Article - CS272971 - Error Attempt to add item '<Channel.Device.Tag>' Failed in PTC Kepware Products

There are also a lot of "Device not Responding" errors in the log. This means the KEPServerEX application is making a request, but the device cannot handle the request. It's possible the device is having a difficult time handling 3 simultaneous connections.  In order to optimize communications, it is recommended to have a single device per channel. Are you able to test this by reducing the 3 devices you currently have configured down to 1 device and see if the communications improve?   

You may want to go into the device properties and increase the Request Timeout setting which will allow more time for the device to respond. 

If neither of these recommendations work, I recommend opening a support ticket with the Kepware team so a more in-depth troubleshooting session could be possibly had. Here is a link to the login page where a support ticket can be opened:

Log In | My Kepware

Thanks,

*Chris

[AC_10026208]

@cmorehead Thank you for your response.

As advised I will try to create new Kepserver configs with one device per one channel. I should have some window to test it on live system in a week or two.

Request timeouts/retires are already maxed out in the device settings.
In the meantime our problem is that the Data Colllection is getting disabled probably by the number of timeouts. Is there any option to stop that from happening despite the number of timeouts? Is there any mechanism to force the Data Collection after it is dropped for the device?

[cmorehead · Accepted solution]

@AC_10026208 

There is no way to override the timeout setting. The timeout setting is hard-coded and designed to keep the server running in an optimized fashion in the event that a device truly becomes unresponsive. 

Thanks,

*Chris