I posted this question a couple of months ago, and had no comments, hope to hear from you all this time, thanks in advance for your time.
Have a situation, of running approximatly 2000' of cable/conduit to connect RS485 serial between a field remote PLC and a Control room DCS. Is there any possibility of failure if I choose to use the local existing Field Analog box already connected between the two locations with multipair instrument cable? The cable is already in place with spared pairs.I've been wanted to suggest this, but I'm not sure if there is any risk involved. And by the way I have used dedicated multipair instr cable for this purpose, but never mixed with analogs.
Thanks

 

The basic question you have to look at is shielding, in my opinion. Are the wire pairs individually shielded in the cable; if so, and there is no link between the shield used for the RS-485 comms and the other shields, there should be no problems; if there is only a single shield (or none), you might have interference or crosstalk from the other signals in the cable.
Rememeber that at low speeds protocols can be very forgiving for this type of problem (Modbus, at least, can be). So why not try it on a temporary basis, for a test period? Just log all comm failures (your protocol must have some kind of failure detection) between the partners; try to have some periods where no other signals ore used on the same cable, and compare with other periods where you will use these signals at the same time you transmit data; see if here is an impact.

Note that even if you detect no impact on communications during the test period, that does not mean that you can garanty no event will ever occur because of this arrangement; it will only show that you have not found any during this trial. Only a new, dedicated cable would guaranty that.

Hope this helps,
Daniel Chartier

 

why you are taking risk in doing this connection, instead you can use a seperate cable or go for wireless connectivity. Communication problem comes and it is difficult to pinpoint where the exact problem. We had some inputs comes and dissappears for a month. We suspected the card problem, but it ends up in communication problem with the i/o panel from the controller.

 

What are some potential failure modes on existing cable/conduit or the connected equipment?

- lightning strike on or near conduit run or panel
- mechanical destruction of conduit cable (rodents, earthquake, forklift, trencher, collapsing superstructure, leaving panel door open in bad weather)
- cable degradation over time (any fabric insulation?)
- undetected short or open created during installation but not yet discovered on unused pairs
- corroded terminals in the old field box
- water intrusion resulting in soaked cable

Probability of any of the above? You got me. I'm not a spiritist and I have no crystal ball.

If youre' looking for insurance in the form of assurance from this forum participant, you're out of luck.

But, hey, if you want to save your company some money, you propose, "We've got twisted pair in place, let's try it. If it works, we're OK as long as it holds up and none of the potential hazards occur. If not, we have to run cable or fiber or go wireless."

Hook it up and try it. If the cable isn't ancient I think you've got a high probability of it working, presuming 'multipair instrument cable' is twisted pair.

Dave

 

I've seen Modbus/RTU run up to 800 metres on standard 4-20mA loop cable for batch computers on loading jetties. However the baud rate had to be cut from the planned 19200 to 4800 baud to get reasonable success rates. It was one of those sad cases where everyone agrees/promises to buy & use EIA-485 spec cable, but we showed for site-acceptance test and the installer had just used analog loop cable. This being an island in Indonesia and an estimate of 4 more months to receive a real EIA-485 cable 800 metres long without splices, so management decided better to commission the existing hack and get paid than delay commissioning.

Is there possibility of failure? Yes, assume it will fail.

Is there possibility of sucess? yes, maybe it is Ok but safer to assume it won't work until shown to work. I certainly won't even hint that 2000 metres of loop cable would work (2000 metre of true EIA-485 cable is even questionable!)

- LynnL, www.digi.com