Modbus implementation issue
TMacMoney
So please explain how this makes sense.
I have a PLCs that monitors my data centers and takes actions based on events.
I picked up a gateway, sensor hub, and 8 temperature and humidity sensors for testing.
Based on the logic of the way modbus is devices are assigned to "nodes" to the hub.
if I code my PLC to talk to sensor 4 which is in cold row 3 using the correct node / register addresses I get my temperature reading as expected.
But if that sensor 4 fails, cable fails or gets cut, whatever reason the hub can no longer communicate with it the whole Node table shifts.....
Now my temperature reading the PLC is getting for that same register is now assigned to a different temperature sensor for a different part of the data center because the original sensor is missing to the hub......
If on boot the gateway made a node table for all plugged in sensors and statically held it that would solve the problem. If it required a reboot to rebuild that table that would be fine. But for the whole table to shift while active because some issue comes up with one sensor is a bad design. A device failure can cause me to get incorrect data for a specific location which could cause the incorrect action to be take by our PLCs.
I have a PLCs that monitors my data centers and takes actions based on events.
I picked up a gateway, sensor hub, and 8 temperature and humidity sensors for testing.
Based on the logic of the way modbus is devices are assigned to "nodes" to the hub.
if I code my PLC to talk to sensor 4 which is in cold row 3 using the correct node / register addresses I get my temperature reading as expected.
But if that sensor 4 fails, cable fails or gets cut, whatever reason the hub can no longer communicate with it the whole Node table shifts.....
Now my temperature reading the PLC is getting for that same register is now assigned to a different temperature sensor for a different part of the data center because the original sensor is missing to the hub......
If on boot the gateway made a node table for all plugged in sensors and statically held it that would solve the problem. If it required a reboot to rebuild that table that would be fine. But for the whole table to shift while active because some issue comes up with one sensor is a bad design. A device failure can cause me to get incorrect data for a specific location which could cause the incorrect action to be take by our PLCs.
Comments