Network Question

Hello All,

Not sure if a network question is appropriate here...sorry if not.

I am having an issue with connectivity between multiple devices connected to a managed switch.

I have:


1. Managed Switch 192.168.110.5
2. PLC - Micrologics 1500 192.168.110.10
3. HMI - Red lion Graphite 192.168.110.15
4 HMI - Red lion Graphite 192.168.110.20

When I plug all the above devices into the switch I have no problem communicating thru the switch to each device. I can ping each device from my pc which is also on the same network listed above.

I have another PLC with a network address of 192.168.36.100 which I know is not on the same network as the other devices. When I plug that into the switch it actually fubars the switch somehow. My understanding is the the switch does not care about network addresses and plugging multiple networks into the same switch should have no effect. This is a level 2 switch.

The functionality of all the devices below the switch still works...plcs and hmis on the same network can communicate...the hmi's can pull and push data to the plc.

But whats weird is I can not connect to the switch or thru the switch to devices from my pc...cannot ping the devices.

That does not seem right...when I pull the odd ball network plug from the switch everything goes back to normal. I can then ping each device again.

any thoughts or ideas on why this may happen?

Any help is greatly appreciated

Frst question: Is there any possibility of there being a network loop when you add the extra device?

My gut says that your problem is not related to the IP question, but is related to something else.

Second question: What kind of switch is it? That could very much make a difference. Someone else might be familiar with the quirks of the specific switch you are using.

In theory, a layer 2 switch doesn't care what devices are plugged into it, you are correct. However, the switches are also not intended to be used for multiple subnets at the same time. A switch could very easily have been designed too smart or too dumb to be able to handle this case, which was probably not a priority for its designers.

Personally, I do that all the time (multiple subnets on a layer 2 switch), and I've never seen the issue you describe, However, I've also never done that in a production system, it has only been for commissioning/testing.

Not sure what a network loop is....

The plc that has the different network address (192.168.36.100) is plugged into a dumb switch, the dumb switch is plugged into the Stride managed switch with all the other devices. All the other devices are plugged directly into the managed switch.

I agree with you that the IP address should not matter. I actually set this up as a test rig like you have with no issues.

The only difference I can think of is the test rig used a different PLC for the odd ball network address(192.168.36.100). The test rig I used a 505 and the production rig uses a compaq logix.

I cant imagine that being the problem either...one thing that popped into my head was the possibility of POE coming from the compaq logix? Im not so familar with compaq logix to know if thats even possible.

a network loop is when the ethernet topology wis wired so that a frame can come back to a place it's already been. This is especially bad with broadcast traffic, which can cause what's called a "flood", where it just builds up until no bandwidth is left.

Essentially, say you have 3 switches: A is connected to B, B is connected to C, and C is connected back to A. This forms a loop. You can use protocols like RSTP to protect against this, and there are also many ring management protocols for when you specifically want redundancy.

If you plugged Switch A into a port of Switch B and then plugged Switch B into Switch C then plugged Switch C into Switch A you would have a loop. Or more simply, if you had two cables connecting Switch A and Switch B. When you get a loop like that there is a possibility of what is called a broadcast storm where a broadcast message keeps getting re-broadcasted around the network eventually consuming all bandwidth.

POE would be power supplied by the switch, not by the CompactLogix.

Where is your computer plugged in? Are you connected to the managed switch or to the dumb switch?

MK42 responded while I was typing (with a similar example) and mentioned a loop management protocol that steps in to prevent loops from crashing the network by shutting down one of the connections causing the loop. Perhaps it has shut down your connection preventing you from communicating with anyone.

OG

My PC is plugged into the network from across the building...

http://www.cisco.com/c/en/us/support/docs/lan-switching/spanning-tree-protocol/5234-5.html

This is what everyone is talking about. If you have a loop somewhere in the network it could flood the switch and cause problems.

Ok..problem solved..

was a network loop...

the switch connected to the PLC with the different network was then plugged back into the network...

So as soon as I unplugged the extra connection all is well.

thanks for pointing me in the right direction!:site: