RS-485 noise, wiring, and stuff

Hello all,

Customer of mine has an RS-485 Modbus network. A PLC reads and writes to about a dozen drives. He was complaining about communication errors on the network. I looked at the software and raised the com error time to 90(!) seconds but that doesn't solve the problem.

Returned with a scope to the customer site today and saw the following wave-form. The edges look very nice so I'm doubting that it's a termination issue. The cables do run all over the place and there's lots of power cables not too far away. Termination appears to be OK as does the shielding. The cable is spec'ed as an RS-485 cable.

I'll have to return in a few days to take a good long look at everything but I'd appreciate any ideas as to where the source is most probably coming from.

(And sorry that the picture is upside down...)


Thanks all!

Yosi

Shield terminated in the connector, or terminated with a big fat clamp on ground bus bar ?

Was it working before? Any new equipment installed?
How is it being wired to each unit on the bus? Try disconnecting devices from the bus and see if it can be isolated to a specific path.

High frequency noise of over 2 volts. Coming from VFD outputs. Use VFD rated cable on output leads to motor, or find a way to shield them.
Route RS485 lines as far away as possible from motor wires.
Try adding snap-on ferrite cores to RS485 cables.

JesperMP said:

Shield terminated in the connector, or terminated with a big fat clamp on ground bus bar ?

Click to expand...

With a big fat clamp on the ground bus bar.

skorp said:

Was it working before? Any new equipment installed?
How is it being wired to each unit on the bus? Try disconnecting devices from the bus and see if it can be isolated to a specific path.

Click to expand...

It worked before and it's still working now, just not very well. All of the devices eventually respond. We are going to do the standard stuff like disconnect everything and return one drive at a time. Also I've told the customer to remove the uLogix from the panel (which is a G-d awful mess.)

As far as new equipment, not that I've been told about but that doesn't mean much.

As I said I've yet to crawl the cable myself and that's way up on my list.

Thanks for the input!

keithkyll said:

High frequency noise of over 2 volts. Coming from VFD outputs. Use VFD rated cable on output leads to motor, or find a way to shield them.
Route RS485 lines as far away as possible from motor wires.
Try adding snap-on ferrite cores to RS485 cables.

Click to expand...

Thanks! Will look into the ferrite cores. Didn't think about that.

If it is working somewhat, you may need to make some tuning adjustment.

Are you giving the equipment enough time to respond to the master? Make sure that you are only sending one Modbus request at a time and get an acknowledgement before sending the next request.
What is your baud rate? Sometimes lowering it some will help.

Did you measure correctly, that is differentially, between A and B on the RS485 bus? If you did the signal is horrible - but should probably work anyway.

Read this - especially the part on failsafe-bias resistors.
http://www.ti.com/lit/an/snla049b/snla049b.pdf

It's THE most common RS485 problem that very few seems to know anything about.


PS. I wrote a much longer post but I always get logged out while writing, so I'll stick to social media length posts from now on.

.

Pete.S. said:

Did you measure correctly, that is differentially, between A and B on the RS485 bus? If you did the signal is horrible - but should probably work anyway.
.

Click to expand...

Thanks Pete!

Yes, I measured between A & B. I'd been suspecting a noisy network for a while but scopes are not as easy to come by as they once were. I don't see any signal degradation and the edges are crisp so it shouldn't be a termination issue. I will look into that link you sent.

Thanks again!