Wonderware / AB Communications Issue

I had a application with a Wonderware 9.5 talking to a PLC540, PLC530 and a SLC500, via DH485 using a Wonderware DA server and a PKTX card. They starting getting issues of trying to write values to the PLC5's. When changing setpoints sometimes to new value would get kicked out and revert back to the old value. Entering new values sometimes took 2 or 3 tries. This problem is only on the PLC5's, not the SLC.

We then upgraded the communications to a Contrologix interface and have one DHRIO card in the chassis for the PLC5's and a seperate for the SLC; in hopes on fixing the issue we were having but primarily to get away from the KT card. All the blue hose was reterminated and resistors put where needed and made sure to follow AB's spec. The slow communications on the PLC5's still remains, any other ideas look for?

Visual inspection of terminations and resistors is a good first step, but those won't show you high traffic or induced RF noise.

If it were my system, I would start by looking at the Diagnostic features in each node.

PLC-5 and SLC-5/04 controllers have a DH+ diagnostic object you can access from RSLinx Classic; right-click in a network browse and make periodic screenshots of the error counters.

Those same diagnostics are available inside the controller data tables if you have defined Diagnostic Files for the network channel. These are often easier to write "logic traps" for, since PLC logic is a good place for timing and comparing.

As robust as DH+ is, sometimes you have to go at it with low-level tools, like a protocol analyzer and an oscilloscope. I used to make my living analyzing and repairing networks like yours, and my Cleverscope and Frontline Test Equipment NetDecoder were my heavy-duty tools of the trade.

When went through the cabling we re-terminated and used ferrules on everything. However the application is in a rice mill that's older than dirt and I don't know exactly what lies next to the cabling behind the walls. I'll take a look at the diagnostics when I go back. I'm guessing with the oscilloscope I'm just looking for noise on the line?

In reference to traffic, there are is one DA server pulling data. There are probably a dozen WW clients in the mill. Does each client poll data through the one DA server and onto the DH network? or is each client getting it's data from the computer running the one DA server? Basically what I am asking is each client polling the PLC even though there is only one computer running the DA server.

Depends on how your WW apps are set up. Are they all the same app?
Does each workstation have it's own DAServer?

Same app, only one DAserver running.

The InTouch app needs to get any IO information from a DAServer.
Thus, if you only have one DAServer, then the single DAServer is the only connections to each PLC (Typically 4 each).