Powerflex 527 VFD random comms loss - Possible EMI issue? Or something else?

[Hisma]

Hey so I have an issue that is driving me insane. I have 6 AB powerflex 527 drives configured in a DLR network over CIP Safety. They go back to a EN2TR card that's configured as the ring supervisor.

What has been happening since the system was commissioned is that randomly one of the 6 CIP Safety-based STOs will randomly drop out (lose comms) for a few milliseconds, which is just long enough to trip the system those drives are controlling. Suffice to say, this level of instability is completely unacceptable.

I have some possible theories as to what it could be, but I'd like to know if others have experienced this.

My theories -

the networked STOs are using shielded CAT6 cable, but they are NOT VFD rated cable. Has anyone had issues with VFDs causing network drop-outs on ethernet networks? I mean, VFDs do put out a lot of high frequency noise, and they do make VFD-rated (high insulation) ethernet cables. The problem is I need to solve this issue ASAP, so if this IS the problem, is there any temporary solution I could use to insulate the existing cables, maybe tape or flex conduit or something?

Right now the time sync on the input is set to 10ms, which I think is the default. That's quite fast and could be causing packets to drop if the network can't keep up. I'm wondering if I should increase that time. I'd be curious if anyone has any insight on what a good time sync to work with would be. My safety task is 25ms, so I'm thinking I could increase it a bit and not have any impact as long as its below 25ms. Thoughts?

Any help would be greatly appreciated. Thanks!

 

[GaryS]

I ran into something similar a few years ago
When we added a few VFD’s and another device on the IO tree at random times we would loose communication to at least one device. Not always the same device
What I discovered is that we exceeded the number of devices that the processor can communicate with. It’s been a while but if I remember on the processor we were using the limit was 6 nodes.
In that case the only solution was to upgrade the processor they were not happy about the cost. `Fortunately for me I was handed the project to complete after the original engineer couldn’t get it working.

Dose the problem get worse when you are online with the computer ?
You need to check the manual for your processor and see what the limit is. Each IP address is a node. And don’t forget your programming computer when you are on line must be added as well
And the programming computer takes up 2 nodes. I got that information from `Rockwell support
From your description you have 6 drives the processor and add the programming the computer
That would be 9 nodes. If you have other devices on the network communicating with the processor you need to add them to the number of nodes.

If you figure it out let us know what it is.