Connecting RS-485

[blackwed]

First off, comms is not my forte but in managing several industrial projects I have been exposed to a little. I once had an old hand tell me, "If you want your RS-485 circuit to work reliably and consistently, don't let anybody s### you--NEVER connect it in any configuration other than a daisy chain, from one component to the next. NO stubs, NO stars, just instrument to instrument." I now have a couple providers telling me it's OK to connect 4 field devices to one master by paralleling the 4 cables onto a single pair of terminals in the master. They say a couple hundred feet of cable connected like this shouldn't be any problem. Do I heed the advice of my old friend or can I believe this mfr?

DB

 

[xnuke]

I'd believe your old friend.

If you connect the instruments in series, and appropriately terminate each end of the daisy chain, you will have few to no problems. This is because you will maintain the characteristic impedance of the network wiring matched to the instruments/controller, which means little to no standing waves on the cabling to interfere with the signal.

If you connect your devices in parallel, you're changing the characteristic impedance of the network wiring, and the ability of your instruments/controller to talk to each other is seriously in question due to interference from standing waves.

xnuke

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jsummerfield]

I have daisy chained Allen-Bradley Data Highway Plus and Remote I/O using direct buried cable at long distances without any problems. DH+ is EIA 485. The application included several SLC500 processors each with Remote I/O. We tied the remote I/O to the SLCs and tied the SLCs to a PLC 5. We used an authorized AB systems integrator as the hardware supplier, packager and programming. They had no concern with the application and it worked.

John