RS485 is bidirectional does it mean it is fullduplex?

For Steve's benefit, you should have included one more section:

Installation throughout Multiple Buildings

There are two separate installation procedures depending on the type of electrical earthing system.

MEN System at one building

In this case the installation may be made as though it exists in a single building. The cable shield will be continuous throughout the installation.

MEN System in each building

In this case there must be electrical isolation between buildings with different MEN systems. Each building is wired as a separate and complete installation with the cable shield tied to the building protective earth at one point.

That is in *each* building.

I'm certainly *not* a power engineer.

You don't seem to be a comm engineer.

As I've noted many times, telco comm cables are grounded at both ends.

Please quote where those specs discuss grounding of comm cables, or explain what ground loops are, or in support any of the other invalid pronouncements you've been prone to making.

When you deal with the MEGA-AMP power world you will get to see how much of a safety hazard connecting the screens at both ends is. Consider that, under certain circumstances there may exist 75 to 80 Volts between the two earths and that the current source is huge by comparison to the screen current carrying capacity. If we assume that the run of cable has a screen resistance of about 1.5 Ohms (a reasonable figure for quite a 150m run of decent signal cable) then 80V * 80V / 1.5 Ohms gives you a power dissipation of 4.267kW along the length of the screen. So, you had better limit the current flow in the screen in some manner or you fry the signal cable.

Many of us have and employ this sort of consideration in our daily practice.

Each installation does need to be examined and understood. I know that our earthing scheme is split three ways. There is building earth which is the main safety earth for the entire building. We also have an instrumentation earth which is a quite large (10mm) cable from an instrument earth busbar around the building and linked to building earth in certain specific locations. Finally we have machine earth which is associated with the MEGA-Amp equipment that we use in our experiments and which only connects to building earth at certain specific points. During experimental operations we can see differences between the machine earth and the instrument earth that are of the order of 75 to 80V. If you need to work on some of the equipment during operational periods we have to write a safe working procedure fill out a permit to work form and get it all signed by a large number of people before we are allowed near the equipment. Sometimes it can take a couple of days before you are allowed to repair some measurement equipment.

If you are speaking of purely telephone wires then most cables I have seen in that category are not screened. Just twisted pairs. However, as I do not work in telephones I will bow to yours and anyone elses knowledge of that aspect (I could probably ask my brother who works with telephones as well as pwer systems in his daily work).

I am certainly not in telephony.

I know that on navy ships and submarines they connect the outer screen to the feedthrough grounds as they penetrate each bulkhead. There are special feedthrough glands made for this purpose. Then, they also go through a lot of effort to ensure that the electrical ground really is as close to ship ground as possible (usually less than 250mV). A ship or submarine can be considered akin to one building with own power source, however, the cable runs are quite long.

I would hope that it is obvious to all from this discussion that the issue of connecting screens/shields is rather dependent on the circumstances that the equipment is going into. The screen connection impedance idea that I mentioned in my earlier post would probably work for the widest range of different situations but I would still countenance understanding the issues that surround each individual instalation. It can save a huge amount of grief later.

And since we are *not* dealing with a power distribution system, your point seems to be less than appropriate.

Reality check time: telephone cables all across the country and around the world are grounded every 3000 or 6000 feet.

Not in the usage of the term "grounded" that I'm familiar with. Perhaps you could give us a little more details.

Telephone cable, the stuff on the poles or buried in a trench, that goes between buildings and on down the road, has a shield. (Heck, some cables have six inches of armor plated steel around them!)

Cross connect cables between bays does, and between racks does not.

Drop wire doesn't. Frame wire doesn't.

Then you should not propose pseudo-authoritative answers to questions relating to that field.

The longest run on an aircraft carrier, or a super tanker, is short. More than 3 miles is considered "long" in general, though specifically that may not always be the case.

You've now mentioned ships and "MEGA-Amp equipment". There are many special cases, probably none of which lead to any enlightenment except when they are explicitly discussed as extremes and flagged as unusual.

I could also mention a few odd places... the Faraday enclosure for radio and carrier equipment at a Coast Guard Loran station. Anything near a large power generating facility. The "tempest" shielding into a room where the USAF was doing super secret seismic analysis (to detect nuclear testing in Siberia). Or the operation of electronic in proximity to DEWLINE or Long Range Radar sites. For that matter a Toll Center is one of the harshest electrical noise environments around.

These are all wonderful stories and any time you want to swap tales, we can. Using them to try to explain grounding on comm cables won't help anyone understand what is involved though.

Understanding the basics comes first. Trying to extrapolate these extreme situations to be representative of the basic theory is exactly what causes so much confusion.

At every point where sections of cable is spliced, the shield on both sides of the splice is bonded, and grounded to an earth ground. That is generally a copper rod driven into the ground.

See

There is currently some discussion on AES48.

This document is on Industrial Ethernet.

This one is a facilities departments standard.

So, there is plenty of material there that states that the signal cable shields shall only be connected at one end. Like I have stated before, each situation should be carefully considered, calculations performed, safety assessments performed (yes, even for the simple audio lead between intercom units) and the complet circumstances fuly understood before you go in with a half baked notion of what is right. Circumstances can change things quite drastically. You just need to understand the underlying physics of the situation a bit.

OK, we will start by assuming that the Earth Ground potential at the left and right ends are exactly the same, just for starters. We will also assume that the Rwire for the right hand end is the same as the Rwire at the left hand end.

Current flow in the cable shield will be roughly (0.5*Rwire)/Rload. If Rload was a few kOhms and Rwire was nearer 0.01 Ohms then there is almost a negligable current flow.

Now consider the case where the left hand ground moves up to 80V away from the right hand ground, but that the impedance between the two ground connections remains less than 10 Ohms (not unreasonable in some very large buildings). I am sure that you can see why we have been advocating the grounding at one end only rule as a safety aspectg of RS422/RS485 networks, as well as almost any other low level signal measurement/management system.

Where you do need to connect the screen wire to both ends then you had better include some impedance in the connection that limits dangerous current flow and yet provides sufficient coupling to ground for the higher frequency signals. As I have stated before, all the circumstances have to be well understood to make the right choice.

I don't think anyone wants 53A being carried by the shield of a signal cable. I know I certainly wouldn't like to see that happen. This is the sort of thing that we have been trying to get you to see as a real risk for some of the systems we are dealing with. I have even seen scope leads fry due to someone not respecting the earthing scheme in place (on a high power motor drive).

If you are speaking of twisted pair screened cable (the type in very extensive use in my workplace) then I thought that we had already agreed that the twist in the cable cancels out most of the noise of a signal because the noise is a common mode across the pair. The screen, in this case, really does extend the shielding of the enclosure out to the plant. I am certain that these principles are in many of the books on telecommunications and electrical theory.

It is more pertinent than you realise. The equipment I deal with incorporates a range of measurement and control systems that includes nano-volt signals, RS485 signals and 4-20mA current loop signals passed over 150m of multi-core shielded twisted pair cable to the instrument racks. The environment that this lives with is extremely noisy, features high magnetic fields, extremes of temperature range and earth potentials that can leap well off the normal reference level (and it is not the noisiest environment I have worked with). Some of the communications signals run in quieter areas and span about 50 different buildings on the site. Each building has its own sub-station and the site has a SuperGrid power feed (it uses short bursts of extremely high power at regular intervals). So, as you can imagine, as a Systems Engineer with responsibility for just some of the equipment on this site, I have to know how the site earthing scheme works and what risks are involved. So in all respects I am quite on topic for this thread and relevent with the advice I have provided.

Last time I checked around here the telephone company cables were unshielded twisted pairs (but that was quite some while ago).

The other types of cable we use on site include unshielded and shielded power cables, heliax coaxial cables and triaxial cables. I have to know how to deal with each and every type.

This one starts out pretty good, when discussing theory, and fails terribly when discussing implementations. If nothing else, the discussion of jack fields should be enough to indicate the confusion.

Regardless, nothing there contradicts what I've said. The entire discussion is about cables within a single structure, ground system and power distribution. It also appears to be discussing single pair cables for the most part.

IEC 61000-5-2, 6.4.2, points out, "Shields of cables are bonded to the earthing network at one or two extremities depending on the signals being transmitted and on possible electromagnetic interference sources."

However, despite contradicting *you*, the context is not complete enough to actually know *what* they are referring to. And casual discussion, much like those on this newsgroup, are not authoritative sources.

This one contradicts your first cite above in regard to the effects of a ground loop. Shame on Belden. It is also talking only about short distance ethernet. It would make no sense at all to ground both ends of an ethernet cable. Or, more correctly, it would make no sense to attempt using ethernet over a cable where grounding at both ends was useful.

This is also talking only about relatively short runs of "control cable" all within a building, all on one ground system, and all where both ends are on the same power distribution. Grounding only one end of a cable, in those circumstances is expedient.

No, that is *NOT* what your sources above indicate. They do indicate that is some circumstances (which I've already pointed out in previous articles), the easiest means of avoiding trouble with ground loops is simply to ground at only one end. That is not a "shall only" situation at all, nor is that necessarily the best solution.

And we have yet to see *any* indication that you do.

Here, try reading these...

Here's a blurb on equipment to test the ground and shield continuity at comm cable splices (where you are saying grounds don't exist).

Here are a couple of specifications for terminating cables to building ground systems:

Here's short URL for that last one,

Here's an even more authoritative treatment of grounding,

All of these describe the grounding at *both* ends of a cable.

I'll believe that it may be the case where you are.

Except that we are talking about RS422/RS485 data communications systems here and not telephony. This is an area that, like Steve and a few others here, have a fair bit of experience in. I know that some of us are in very heavily noise ridden environments like myself. I also appreciate that some of the other contributors to this thread are in no worse than a normal office environment. So, when it comes to talk of connecting shields I think we are all well aware of the situations where to connect both ends leads to problems.

The ships themselves may not be that long but by the time you consider the routes that some of those cables take to get from one end to the other then you should not be surprised that there can be single runs that are of the order of 7km. Then you think about the number of bulkhead feedthroughs and the mind begins to boggle at the number of clamp-down points that the cable sheath has to be pared away, ferrules fitted and boots shrunk down.

In the 30 years of my career it has, to me, been the norm to work with such environments. My main field of Systems Engineering is Energy and Transportation. That includes Oil-rigs, Mines, Power Stations (Coal, Oil and Nuclear), Railway Rolling Stock and Trackside Equipment. I have also done some work in the Banking Sector and the Medical Devices Sector. I have to appreciate everything from simple mechanics and low level signal equipment through to high energy systems. I have also to deal with vacuum and cryogenic equipment. It has certainly been an interesting career.

I am also aware of some of those systems and the issues invvolved.

I explained in another reply featuring the diagrams you provided. I am sure that if you look at the maths of the situations you might get to see the consequences that most of us here are trying to prevent.

I maintain that you do the math first then you may see how much and what effect the various schemes will have. I know that the power levels and situations that I am dealing with are heavier than most others here but most of them will also have to understand the fundamental principles that I have described.

Actually, I believe the specification *is* 10 Ohms, and the target is 5 or lower. An 80 volt difference in ground potential is... so unusual that we can ignore it. Lets assume it never gets higher than 20. Or 30, if you like. (Everything I recall seeing was engineered for 20 V, max.) No doubt there *are* unusual instances were we might well see figures outside this range. And if we do, we deal with them as unusual instances...

You are again proposing unrealistic circumstances to portray the "norm".

No. The trick is to ground the cable every 3 or 6 thousand feet, so there is never get anything like an 80 volt difference.

Can't you come up with something less boorish than repeating the same thing over and over?

So how do you reconcile even a 20V ground potential difference with the

Steve

Just to be clear is what you are refering to as the shield (the bit you keep insisting must be grounded at both ends and every 3000ft) actually known to the rest of us as the armouring. If that is the case then it is no wonder that we seem to have been talking at cross purposes. Yes, armouring on cables does have to be grounded at both ends. The armour is normally a very substantial close steel wire wrap whose purpose is mainly mechanical protection. However, the rest of us, I believe, were speaking of the shield connection which is a much finer enclosing conductor and which we are all maintaing should only be connected at one end unless special arrangements are made.