RS-485 protection

You can't assume that a few hundred volts is worst-case. More like a few kV, possibly more briefly while ionization paths develop, if there is a somewhat nearby lightning strike. Results can be impressive, and very location-dependent.

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

It's been a while since I looked, but I don't remember seeing a lot of RS-485 interfaces spec'd that far up. They seem to be designed to withstand industrial control stuff (24 VAC or VDC) and maybe telco battery (48 VDC) but not full line voltage.

There is probably a way to do what you want at the component level, but I don't know what that is. And, of course, you need it to fit in half a sugar cube, draw no power, cost a nickel, last forever, etc. Having said all that...

How about designing your gadget with a plain old RS-232 port, and then using a module from somebody like B&B Electronics or Black Box to turn that into RS-485?

Cons:

Money, space, power, etc.

If some really clever person sees the RS-232 port there, and if they happen to have RS-232 available at the far end, they will think "Ah! I don't need this complicated RS-485 stuff!", and proceed to string an RS-232 cable around the factory. This ends in tears when ground *here* is 10 volts different than ground over *there*, but they will still call you to complain.

Pros:

If somebody decides that the RS-485 interface needs to ride the lightning, the external module is likely to bear the brunt. If your gadget is expensive, replacing the module is a relatively cheap way to get up and running again.

You can easily plug a laptop into your gadget, or into the RS-232 side of the remote line, for troubleshooting.

If the BOM fairy is smiling down upon you, you can even put a DPDT toggle switch on the TX and RX pins on your gadget, so they can forget to bring a null modem cable for the laptop if they want.

Alternative suggestion:

If you do it yourself out of transzorbs, FETs, caps, etc, put all that stuff on a small I/O daughterboard that is easily replaced. Okay, this adds connectors, which is bad if you have moisture or high vibration. But I think some day somebody will try to blow up one or more of the interfaces, and it's nicer to swap a $20 board than a $500 one.

Matt Roberds

I have one DMM that takes a goofy fuse and another that takes a 15 or

20 amp ceramic 0.25x1.25" fuse, the same as every microwave oven on the planet. I have spares for both, but I'm more adventurous with the one that takes the microwave fuse.

If you look at the specs, the 15 A or 20 A input often has a really silly short duty cycle, like "15 sec on, 5 min off". I don't know whether they are worried about the fuse, the shunt, the test leads, or what.

Put 1 mA through it to make sure it's not open circuit.

Put rated voltage through it at low current to make sure it doesn't arc.

Put maybe half rated current through it and check the voltage drop.

Depending on the particular fuse, you should be able to put the rated current through it for some amount of time before it blows. Do that and check the voltage drop and self-heating.

If you're picky, get out the calipers and check its dimensions against spec. Maybe the plating/material on the end caps, too.

As has been said, blow a few per lot on purpose to make sure they resemble their specifications. Disassemble the blown ones to make sure the fuse body material and fuse wire resembles the spec.

Matt Roberds

Definitively use galvanic isolation.

Also be very careful about signal ground and cable shield connections (ground shield at one end only).

If you interconnect two buildings fed from different distribution transformers or even two buildings with separate grounding electrodes for lightning rods, you can get a huge ground potential rise in one building.

Look for at least 2500 V isolation, even better use fibers.

Thanks. I'm mostly doing a sizing for what a really bulletproof RS-485 would cost, plus I think that there's probably value in some iteration of the depletion FET/polyswitch combo.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Have you considered straight transformer-coupling, with a NRZ or other self-clocking modulation scheme? Tiny little transformers just saturate when the signal input is 'way too high; keep the wire from melting, and it all works fine afterward.

No, after they saturate there's nothing but the copper resistance to set the impedance, and they do fry.

A series capacitor can protect against DC and 60 Hz overload, as long as you use a modulation scheme that dorsn't mind the resulting frequency response.

A transformer and a polyfuse might work, too.

Use all RJ45 connectors. Nobody is going to put 120 VAC on an RJ45. PoE, possibly.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation