RS485 Qustions

You can get almost anything to work if it runs slow enough. I'd drive the 4 runs using 4 drivers, and keep the branches as short as possible. A lot depends on the way you want to work the system. If you need a fast fixed- time poll rate it's a lot harder than if you are checking every few minutes.

Useful article:

How are yoou getting power to the remote units? How much do they take? Is this the caravan site it sounds like?

Paul Burke

I go pretty much go as slow as I like, refresh interval time for each slave, can be as much as five minutes or so.

Can't do 4 drivers, not possible with the cabling routes available to us.

Decided to power each slave individually, as mains power is available at each point. You have a good memory.

Mark, Definitely take a look at:

I haven't worked with 485 since a "nightmare" remote camera installation project many years ago (pan/tilt/zoom, etc..)

But the above link has lots and lots of ideas, hardware, and the like.

-mpm

Is there a reason why you are going half duplex? With a single master arrangement there is no need for everyone to listen to everything. If you are going to use off the shelf CAT5 cables you have plenty of conductors and if you go full duplex you can make hubs along the backbone to run off to different areas. Just take advantage of the failsafe of the Maxim receivers and it is pretty easy to make active hubs with a pass through for the backbone. Then each "leg" can handle its' own termination and so on.

Jim

I'm suspecting you will need a few repeaters or drivers since you can't daisy chain.

The best app note I know of for understanding RS-485 wiring is TI's SLLA036. Unfortunately TI does not have this anymore. I googled for it and found it here

TI seems to have another document SLLA173 now. I have not read it so can't vouch for it. It appears to deal with High Speed issues only and not basics. also might want to look at

Hawker

On 3/27/2007 3:48 AM, The digits of Anti-Spam's hands composed the following:

Seems you've nothing untowards there. I've done RS485 a few times and the "Art and science of RS-485" PDF at ...

is a really good write up. In a kind, sensible world, we'd be using RS485 instead of that USB nightmare.

--
Posted via a free Usenet account from http://www.teranews.com

">

Up to about 56K you could probably get away with no terminators and run in a Star....

you could do a folded bus

(view with fixed pitch font) T O R | detail 1 | U O | N | branch |~~~~~~| -----------o + K \\=-=-=-=-=-=-=-\\|DEVICE| -------. /| /=-=-=-\\/--=-=-/|______| \\/ | | |O(2) (1) /\\ | O O| -------' \\| | |O .... sub branch -----------o - O O| | | |O O .--\\/--. detail 2 | | | ------------------ O |device| ------------------ | `------' O ------+ +------- | ----+ | | +----- O | | | | .-|--. | | | | |[R] | last device has | | terminating `----' resistor O | and -=-=- represent a twisted pair 0 | ---- and | etc in the detail diagrams | represent individual conductors power wires are not shown, but if you use 4-pair cat5 cable there's three unused pairs on the trunk and two on the branches that can be used for power.

Bye. Jasen

I agree with TT_Man. The only reason to have line termination is to eliminate reflections that can distort the waveform. If the bit rate is low enough, you can show that reflections will not interfere with the signal significantly. A lot of our RS-485 applications ran at

19200 and 38400 and they never needed termination. Also, without terminators, there is less load on the drivers which means you can have longer runs or more slave devices. What is your bit rate?

Lou