RS485 grounding

You shouldn't care if you can guarantee in some way that A and B will always remain within the common-mode range of the receivers at both ends.

But maybe there isn't a common ground.. or maybe there is but there is a small voltage between the two (with very low source impedance). Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Ground both ends, hard, to the equipment chassis. Why would you want huge common-mode spikes on your data? Why bother with a shield if it's going to have nasty potentials on it relative to the signals?

John

You might also have some big currents flowing in the shield if the "ground" at each end is a different voltage.

--
Dirk

http://www.transcendence.me.uk/ - Transcendence UK
http://www.theconsensus.org/ - A UK political party
http://www.onetribe.me.uk/wordpress/?cat=5 - Our podcasts on weird stuff

3 different answers -- actually I've been using just an unshielded twisted pair with no problems but I have had a couple of users with some transient problems.

I'm just using SN17576 chips on everything and opto-isolated to boot.

Perhaps they're using a shield.

You're saying DON'T use a shield?....it's just inviting problems?

My guess is the users are using cat5 twisted pair.....a worst case would probably be 300 feet of wire at most.

Only 1200 baud though.

org/- A UK political party

Our= podcasts on weird stuff- Hide quoted text -

Which is why you sometimes "get by" by putting some resistance in one side of the ground connection. I see no reason to put 100 ohms at both end though.

George

The point of the resistors is to avoid nasty *currents* if the customer stupidly sets up a ground loop situation. There shouldn't be nasty potentials with 100R in there, certainly not relative to the CM range of the signals.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

So, in conclusion....

Use shielding when you can (or always?) and connect as I have diagrammed, a 100 ohm resistor from circuit ground to the shield on both ends.

What about protection on A and B as well?

Zener diodes to ground on each leg? (If so, what value woyld you use?)

I'm wondering as well if there is a plug in replacement for the SN75176B that has built in transient suppression?

Thanks very much.

The reasoning behind the resistor on each end is that cable runs sometimes get worn insulation and end up with the shield connected to earth ground somewhere in the middle of the run. The resistor on each end limits the current through the shield in such cases.

--
Guy Macon

Maybe 7V (bidirectional TVS not zener). And a series fuse of low current rating and reasonable interrupting capacity. The TVS adds some capacitance that might affect things at high data rates.

Lots claim to, especially the inherently fragile CMOS ones. Nothing on-chip compares remotely to a reasonably large discrete TVS. If you're lucky they'll survive a human body static discharge.

Worry about this a lot if your run goes on a roof, outdoors, or is fairly long (just the induced current from a nearby lightning strike can do some interesting things to solid-state electronics, and even passive compenents). Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Thanks so much.

I'm starting with the TVS diodes and just twisted pair and doing my own "field" tests. I left out the diodes in my design and that may be the greater problem.

I'll play some "Ben Franklin" with a few setups.

Conventional wisdom is to ground only 1 end. Much more importantly (as you state) is to use twisted pair cable.

FWIW, someone could probably make a small fortune (cottage industry) making RS-485 protectors for DMX-512 applications. Much higher speed data than you're using.

Either shunts, clamps, opto-isolating. whatever!! For DMX stuff, this could all be housed in a male-female XLR-type "adapter". Put one at both ends (if opto) and totally isolate the copper! Or if clamp, one (at either end) might do the trick?

Anyway, we (our church) would be a customer. For sure. I'm suprised there is not more (and more affordable) protection devices out there for RS-485. Guess it's a matter of supply & demand?

Still, it seems once or twice a year the theatrical lighting controller at our Church gets absolutely hammered with lightning. It's installed correctly. The twisted pair is even in a metal conduit, and NO ground loops. I'm probably going to optically isolate the whole damn thing, but RS-485 was always billed as more "robust" than its single-ended counterparts....

-mpm

You don't say if the connecton is between two factilities. By factilities here, I mean two structures with separate earthing. If yes, then every RS-485 wires in each cable must be earthed where the cable enter than structure. Earthed means a connection to earth ground that does not conduct signals but connects everything else. Avanchle diodes are used but typically have too much capacitance. Other solutions include gas discharge tubes, Sidactors, or special low capacitance avanache diodes.

If the connection is internal, well, get Montandon and Rubenstein's Nov 1998 paper (IEEE Transactions on Electromagnetic Compatibility) to start learing (appreciate) what you must do.

Intall a good set of lighting rods -- you may only need one if you have a steeple -- with a proper grounding line.

References:

--
Guy Macon

Be sure to route the ground wire through the Baptismal pool ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
 I love to cook with wine     Sometimes I even put it in the food

I only ground at one end mainly due to over currents that can take place if for some reason large induced currents are near which is another reason you never lye the wire in the race way with high potential or high current rungs

Doing so will create some major problems of unpredictable cross talk of foreign signals down to damage on the RX/TX end points.

If you feel you must ground both ends, do the last end via a cap.

Also, in a heavy R.F. environment I like to keep the shield not connected to anything at the equipment and only ground it at the ground source. This helps remove any R.F. from traveling that could be generated from the equipment. Which is why I use a coax shielded ground with the shield not attached at the computer end but both the center and shield attached at the ground end. It grounds my PC via the center conductor, but also keeps the switching noise out of my radio equipment from the switching PS and other devices in my computer.

"

It will be a lot more robust of you ground the shields on both ends. No RS-485 receiver likes a few kilovolts of common-mode voltage.

John

I once had to run a data link between two high rise building... around

200'. There was around 10VAC difference between the two "grounds" :-(

I "hard" grounded one end and used a resistor paralleled with a cap on the other end, and used transformer drive/receive on both ends.

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
 I love to cook with wine     Sometimes I even put it in the food

Ahh good, I hoped someone else put caps in parallel with the single resistor. Just keep the time constant well away from 60 Hz.

George

In general always connect the shield to the receiving end of something otherwise the shield will inject noise into cables from a noisy source. Keep in mind that resistors drop power and coupling the shields to ground through a resistive load with dissipate the noise as heat. Also all of this might not be necessary if you clean the tarnish off the ethernet male and female connections to ensure a low impedance connection. Also look for a star ground with lots of copper of metal for the sheild, hope this helps.