DH-485 Network

Hey everyone, I am new to the site. I posted yesterday about some insight on a new vfd/plc combo we are installing. Most (all) of my experience is with Modicon running on the Modbus+ network. I started a new job Monday and we have all AB stuff!!! Anyone have some help with the trasition? hehe. I guess the most difficult thing is knowing in's and out's of each series. I do have a technical question, and I appologize for any vagueness in advance until I get used to this AB stuff. There is a DH-485 network with 9 nodes, I think it supports up to 32? There is a SLC500 controller at each node, (almost 100%sure of that) Then some ASCII magic happens, and data comes out of a lable printer. The problem we are having and it's flakey if you will, is that when we connect the end node other nodes lose comm, and it's sporatic. It's a long run but nowhere near the 4,000' max. trunk length that is in the book, I'd estimate 2,000' at the most. The last 1/2 of the trunk is run with the non-specified wire. The book specs. Belden #9842. They did not use this for the last physical 1/2 of the trunk. Could this be the reason I'm losing comm on my OTHER nodes when I connect the end one? Like I said very unfamiliar with AB and appologize any vagueness in my verbage with the AB stuff for now. Would just appreciate some ideas. (note the wire that was run instead of the correctly specified Belden is still shielded and drained) Thank you Greg

On any RS-485 link you want to make sure the last node has the proper terminating resistors - chcek that first.

Allen Bradley DH-485 is a lot fussier about capacitance and impedence in the communications wiring than most RS-485 serial communications. You don't need the magic Allen Bradley Blue Hose, but you should at least stick with the Belden wire that corresponds to the Blue Hose in properties. If you contact your local Belden wire distributor or Belden directly with the Belden number or brand and type of the wire you do have, they can tell you if it is close enough.

Thanks Tom J.

I assumed this may be the problem, so I'm going to have new wire run friday. Just wasn't HOW sure AB was with capacitance and other characteristics of comm. wiring. I'll let the group know if this solves the issue. By the way Tom, I'm also in Milwaukee!!! Where do you work? If you're not comfortable posting it here, shoot me an e-mail at [email protected] Oh, as far as end of line terminating resistor, there isn't one, and I didn't see in the AIC module book that there need be one. It had a diagram as how to daisy-chain the nodes, but didn't spec. and end resistor. And remeber this problem doesn't exist when I unplug my last or end node #9, If I simply disconnect it and do no further termination all the other nodes are happy. Thanks again Tom.

It may not be the cable

First of all, DH485 doesn't use the legendary blue hose - that's for DH/DH+/RIO and a few other obscure apps. DH485 trunk cable is 2-pair shielded, of which 1 1/2 pairs and the shield are used.

Is it possible that the troublesome node has been added some time after the original 8? In which case, perhaps the termination for one end of the trunk is at the wrong node. The trunk line should be terminated at its physical ends. The termination is simply a jumper between terminals marked "A" and "term" (5 & 6). There should be a jumper on the two nodes at the physical extremities of the trunk, and only these two nodes. Also, there should only be one link between shield and ground on the network (terminals 1 & 2).

I assume that all nodes have the same baud rate settings and there's no duplicate node numbers...

Greg,

Do yourself a favour; if you are new to SLC's, download the Instruction Set Reference Manual (1747-6-15) from the a-b website.

Refer to Chapter 13, in particular to page 6... We have problems with our network caused by the fixed I/O processors that cannot "talk" to more than 2 other nodes at a time....

Gerry

Yes, I fairly confident all of the comm setting are the same and correct.........occasionally the network will go a few days working just fine, then suddenly lose comm on various nodes. And yes the end termination jumpers are in place. (forgot to mention that sorry.)... I think I'm going to start with getting the correctly spec'd. Belden pulled for the main trunk. I'll let you guys know if that helps.

Greg

If the characteristic impedance of one type of cable is a whole lot different from the other type of cable composing the remainder of the network, reflection problems are not at all uncommon on daisy-chain networks. It probably wouldn't matter if your network were wired entirely with the incorrect cable, but mixing the two types may be where the problem originates.

I've seen this where A-B "Blue Hose" and Westinghouse "IMPACC" cable are mixed; they're both made by Belden and both have similar blue jackets, but the impedance mismatch is enough to cause ringing on the line where they meet.