DH-485 Problem

First off, I want to apologize to Paneltech for highjacking his topic. I have been very frustrated by trying to make the communications link below work.

It actually does work most of the time. However, after a power cycle the DH-485 link does not come back live. If the terminal strips on the AIC+ or the Link Controller are un-plugged and then re-plugged, in any random order, after a couple of iterations the comm link re-starts and all devices resume communications.

Allen Bradley's has offered two solutions:

1) Only use Allen Bradley components (please note that EXOR has "certified" their interface's DH-485 drivers through A-B)

2) Pay $150 an hour for an A-B guy to show up and maybe or maybe not fix it.

Any help would be appreciated.

I suggest checking noise, proper shielding, separate grounding for DH-485 (away from ac grounds).
Are you using PLC 24VDC to power up AIC? If so please check inrush load of AIC, I suggest using a power supply.

I will ask some dumb questions

I doubt this is it since you said the network works but I noticed 2 Micrologix 1000's. Is bit S:0/10 set to one (1) in both of them? Is the SLC 5/03 using Port 0 and it configured as DH485 Master? Is the cable Belden #9842?

Told ya they were dumb, except for the issue involving a laptop being connected, I think with a PCMK card, I can not offer much more.

No problem Tom, now Ive seen what you are dealing with there, you need to unload sometimes!

Thanks, guys. A little more info:

Shahn:
We are using shielded cable. I don't think noise is a problem, as the panels are sitting on a desk in the office for pre-shipment testing. We are using as separate 24 VDC power supply for the AICs, not the SLC or MicrLogix power supplies. We have tried various combinations of ground, no ground, etc. We are tying back to the panel ground, which is the earth ground associated with our 120 VAC wall outlets - not the best but all that is easily available and it should be adequate. Do you have any suggestions specifically related to grounding - if that is what it takse I can drive a rod in the ground outside the office door!

Ron:
Not dumb at all. The PC connected or not connected doesn't influence the problem. We are tying in the PC through the DF1 RS-232 programming port, not through the DH-485 link. The bit S:0/10 is set in both MicroLogix units. The SLC 5/03 is set as the master. Our supplier is out of the Belden #9842, so we are using our "standard" RS-485 cable. We will get the 9842 tomorrow, and I'll let you know if that helps.

Other info: we have the max address set to 5 in the EXORs and the SLC 5/03. The MicroLogix have this parameter locked at 31(I think), as it can't be changed.

Also, the SLC 5/03 to MicroLogix communciations never drops out, even on power cycles or unpluging an EXOR ETop 12 unit. The communications loss is always between the EXOR ETop 12 even though they are tied into the RS-232 poer on the AIC+ modules. When the "phoenix" connector on the AIC+ is unplugged, the EXOR ETop 12 picks up communications to the associated MicroLogix. When either of the EXOR ETop 12 uits is uplugged, the other one will always be OK. (and as indicated in the sketch, the three EXOR's all have different adresses.)

The best guess I hve gotten so far is that the DH-485 driver in one of the units is holding on to the communcations token. Can anybody shed some light on this?

There is supposedly an "edit resume" bit that needs to be dealt with in the driver to pass the "token" back to the SLC. Does anyone know anything about this part of the driver?

Does the system just lose comms to the exor 12 panels only?
And does it do this only on power-up or after it has sat running a while.

From what I've read so far it sounds like the exor panels might be powered up and ready to go before the rest of the system .

I'm going to make a guess and say maybe if that is whats happening the exor might be seeing some "noise" generated by the processors . That is causing them to become confused and basically lock up the comms on the units.

I have never worked with an exor panel and haven't done much dh485 networking but I had to chime in.

Drewcrew6

Tom, thanks for that detailed wiring diagram. Too many guys post "it doesn't work please help me" but you provided all the details you have at hand.

I can't be certain this is the source of your problem, but try connecting up those COM terminals with a third conductor to provide an isolated return path for the A and B signals. Page 2-5 of the 1761-6.4 manual shows the DH-485 trunk connection as follows:

[attachment]

There's a great non-product-specific tutorial on 2-wire RS-485's need for a third conductor at www.robustdc.com. That taught me a lot about the theory behind the DH-485 network and the AIC isolators it uses.

It does sound to me like the token does not start getting passed when you power down, but I dunno why. I don't work for the technology transfer department so I can't say unequivocally, but I don't agree that Exor's DH-485 implementation is "approved" by A-B; we almost never license technology to direct competitors.

Try also changing the setting of the Baud Rate rotary switch from Auto to 19200 baud. I have never had to set it to anything but Auto, but something in your network might be sensitive to network timing.

drewcrew6 said:

From what I've read so far it sounds like the exor panels might be powered up and ready to go before the rest of the system .
I'm going to make a guess and say maybe if that is whats happening the exor might be seeing some "noise" generated by the processors . That is causing them to become confused and basically lock up the comms on the units.

Click to expand...

We had a similar situation with a 1394 Turbo GMC Servo Controller connected to a SLC backplane. The SLC would power up before the 1394 and expect to see the 1394 sitting there on the expansion cable (the 1394 looks like a 4 slot rack to the SLC) and sometimes fault out.

We had to put a time delay on the power to the SLC to insure the 1394 was up and running before the SLC looked for it. Maybe this is similar?

Tom

Here are some screen shots from a network manual (1770622) that may help. IMHO i don't think the AIC COM terminal should be connected to earth at each node (but I may be wrong)

Notice the difference in this picture

It had me confused for some time banghead

I strongly agree with NDZIED1 and DREWCREW6 that the problem lies with either the EXOR are coming up before the PLC or after the PLC.
I suggest to check this by powering PLC separatly than EXOR. Turn the Exor on first then the PLC. Then try turing the PLC on followed by EXOR. One of this combination may help. If so than you may have to delay one power circuit.

Thanks guys!

Thanks to all for the suggestions. We will try them today, and I'll get back with the results.

Update

Here is the status based on the suggestions to date:

1) Tim had already configured the wiring per Ken's and Ian's suggestions, and that didn't solve the problem.

2) Switching the AIC+ from Auto to 19.2 didn't help

3) Delayed power on of the various devices in all kinds of different combinations was tried to no avail.

4) We used our handy dandy AM radio EMI/RFI detector - the only noise we picked up was from the office overhead flourescent lighting. Lights out testing made no differentc (except the LEDs were easier to see!)

5) we replaced the generic comm cable with genuine Allen Bradley distributor supplied 9842 - no help.

The EXORs are the device that looses communications. Our next step is to take the EXORs out of the 22 port on the AIC+ modules and put them into the RS-485 link. This will be accomplished Monday, and I'll report back then. Any other suggestions will be entertained gratefully.

SUCCESS!

The system is now up and running, thanks to the efforts of EXOR tech support and my main man, Tim. The trick was to take the EXOR panels off the RS-232 ports on the AIC+ and tie them into the RS-485 network. No other changes were made. The system that worked is shown below.