Ground currents blowing up my CM choke

Did you hook an AC ammeter in line with the ground on the RS232 to measure the AC current? I've seen some jaws drop when I did. But bridge the choke first or do it to chassis because if there is 1/2A or so the Murata choke would go *PHUT* in no time.

Problem is, it might all be ok now. Until the janitor plugs in the vacuum cleaner and hits the power button.

Doesn't matter which connector, if both sides ground their RS232 to chassis that ain't no good, no matter how you turn it. Definitely not for something shipped internationally. You really need to design an isolated solution.

--
Regards, Joerg

http://www.analogconsultants.com/

put a

ground

managed

to

or

parts

worth

you're

installation.

Yes, my estimate was designed to be pessimistic with regard to my assertion. It was an estimate which you seem unable to comprehend. So, reduce the 1000 feet to 100 or 10 feet. It lowers my estimate by 10 or

100. So, the voltage is 1/10 or 1/100 of my original estimate. Noting your intellect here, you probably need those in decimals rather than fractions which are 0.5V and .05V respectively. In other words, Keith, by complaining about a *BIG* lab, you have worsened your objection and improved my assertion.

In what way? So, make it one wire. That makes my estimate 3 times worse. Okay, then 15V at most is what you get.

Got a clue yet?

No relevant technical content.

Rather than insulting me, come up with your own scenario that fits the OP's description and show how someone can be fried, DumbAss. You have yet to show superior intellect to Sloman or, for that matter anything close to his, so I suspect your forte is insulting people and that requires no technical competence.

10-50 A depending on the mains fuses used :-).

Look at

especially at the TN-C diagram and add a large single phase load to the right between L3 and PEN wires. The large return current flowing in the PEN wire, will cause a voltage drop before reaching the "consumer" (and of course more losses, before reaching the transformer). Connecting the RS-232 cable signal ground between the big load at the end and the "consumer" will add an additional path for the big load neutral terminal to the transformer, though the "consumer" signal and frame grounds. The current share is directly proportional to the conductances (or inversely proportional to the resistances) of the PEN conductor and the signal ground/cable shield.

With a good signal cable shield, this current can be several amperes. I have seen dumb terminals with melted PCB ground traces due to return currents from motors and fluorescent lamps.

These days with a lot of SMPS loads, the phase current contains a lot of harmonics. While in a balanced three phase system, the N or PEN

50/60 Hz current cancels out and no current flows in the N/PEN conductor. However the third harmonics 150/180 Hz does not cancel out but adds up and in the worst case be larger than a single phase current.

There are still a lot of TN-C and TN-C-S installations in the world, so you should really prepare for this.

If you really insist of using RS-232 for anything "long distance" communication (outside a single equipment rack), at least use galvanic isolation between the RS-232 transceiver and the rest of your equipment. In this case you can connect the signal ground and cable shield into the floating transceiver "island", thus no large return currents will flow in the signal ground, only very small currents due to stray capacitances.

Balanced communication systems such as RS-422/485, since these allow about 3 Vrms voltage drop in the PEN connector due to the big load further away from the transformer. A more or less standard practice is to ground the cable shield at one end only and at the other end to the local ground through a 100 ohm 1-3 W resistor. This will limit the big load ground return current to 30 mA at normal operation (forcing the main part of the current to go through the PEN conductor) and does not fry if momentary overvoltages occurs.

What certification do you need ? Just CE or DNV ? There is quite a lot of difference :-).

You really need to float (500-2500 V isolation) your transceiver "island" from the rest of your system, in order to comply with all wiring practices around the world.

You are just asking for troubles.

I have worked with industrial communications for quite a while and my rule of thumb are:

• RS-232 is good for connecting devices (modems, converters) in the same equipment rack

• Non-isolated RS-422/485 is good for connecting devices in the same equipment room

• Floating RS-422/485 is OK in a single building

• Only optical fibers should be used between buildings, since a lightning strike to one building and into the local grounding electrode can cause a huge ground bounce relative to the other building, exceeding the typical 500-2500 V isolation of most standard devices.

ground

managed

worth

you're

installation.

I doubt seriously that a CHUMP like you really knows what is required to initiate fibrillation,regardless of the circumstances that may or may not bring it on.

You are about as stupid as it gets. Fuck you, John S

3 x 22 Ga x 1000 ft is just under 47.5 ohms, idiot.

As if ANYTHING you EVER posted here EVER represented 'the technical/engineering community'.

You sport some SEVERE character flaws yourself there, SPORT!

Bwuahahahahaha!

I have rolled my eyes a few times in the thread already.

Not difficult to make your own discreet, single path opto coupling.

Just exercise discretion?

:-)

--
"For a successful technology, reality must take precedence 
over public relations, for nature cannot be fooled."
                                       (Richard Feynman)

Absolutely. True, all of it.

put a

ground

managed

to

or

parts

worth

you're

installation.

A single 22 ga wire is 16 ohms/1000 feet. Why would anyone put 3 22ga wires in series (3000 ft) to span two systems 1000 ft apart?

You have a comprehension problem, jack snipped-for-privacy@cox.net . Crawl back into your hole.

Many times more than your comments devoid of technically accurate arguments. You don't even know how to calculate the flight time of photons between two planets, remember?

And, obviously, you don't know how to analyze the situation this thread is all about. If you were knowledgeable about it, you would post some meaningful estimates yourself.

You really need to get that cough fixed, jack snipped-for-privacy@cox.net or it'll be the death of you. On second thought, never mind. Don't get it fixed.

To bad his parents weren't fixed, before they met.

--
You can't have a sense of humor, if you have no sense.

Too bad their parents weren't fixed, before they met. *Etc.*

Indeed.

3 in parallel would be about 6 ohms, but you forget the return. It isn't magic.

Nobody would use 3 pieces of #22 in ANY configuration to span 1000' either, idiot.

Snip.

Stay on topic, pussy boy.

I remember that it was not about photons.

I also remember that it was about a lack of recall on my part, not a lack of actual knowledge.

Nice try though, asswipe.

Declaring the CRAP you posted as being meaningless was meaningful to the thread.

You do not want my meaningful estimate of you aired. Your self image is piss poor enough already.

Bwuahah... Bwuahaha... Bwuahahahahahaa! Go ahead, dumbfuck! Call it a cough! I'll keep laughing, bitch.