Instrument Shield Grounding

Shield wires generally are earthed at one end so as to avoid ground loops - try Wikipedia for further explanations on that.

If the transmitter is leaking current to its body, then it has an internal fault that warrants its replacement.

do you have to be so stringent on the answers. and how to make it simple in lamen terms for someone to understand. As you can see, he still doesn't get it.

The shield wire is always grounded in the control panel and not in the field. It is true that if grounded on both ends, you will have a ground loop. But primarially is acts like an antenna when grounded on both ends and can pick up noise. which will give you erroneous readings on your milliamp loop control circuit.

hope this helps you. I've been working on controls and electrical for the past 30 years. I really love what I do for a living. So I must be good at it. not to sound big headed!

Actually, shields can be grounded at EITHER end (field or panel), but not BOTH at the same time.

For convenience's sake, grounding's usually done at panel because then all the cables tied to a panel are grounded at a single point; no need to find where the instrument'S located to check it, no need to go from one instrument to another when commissioning, etc.

DZ

It is merely common practice as to ground them in the panel because to trouble shoot your ground loop. It is easier to have them all together to elliminate which one is bad. And if you have a bad cable you can find it easily.

Agreed that connecting both ends to the ground will cause loop current to flow and may cause error in readings. but the confusion still persists!!

i wanted to ask that if i am not connecting the instrument shield at all (not to the ground for sure), then how can it be called as connected to ground as now not even a single end is connected.

Sorry if I explained myself poorly.

Grounding IS needed; but make sure that a cable's not grounded at BOTH ends. Grounding at the instrument or at the panel is OK.

And good practice: always do the same thing at a given place. If a factory has all grounds connected to panel grounds, always ground in panels.

DZ

To really confuse issues:

From EDN, September 3, 2009, "Tales From the Cube", by Jeff Fries (a principal engineer and
technologist for global signal technology at GE Transportation):

"I came across an old textbook stating that, when the cable length exceeds one-twentieth of a wavelength, you should ground both ends of the cable shield instead of just the receiver end."

Not being a Sparky, I don't know if this is correct, but according to the article, it solved the noise issue he was facing...

Interesting thing, this wavelength thing. That being said, it applies to frequencies higher than those for control and power systems (so communications, yes; and networks, perhaps). The reason for shielding is twofold in this case: toprevent outside radiation to be superposed on the signal on the wire/cable, and to prevent the signal from radiating.

The wavelength of a signal equals: W = c / f, where c = the speed of light (which is 300,000,000 meters per second in vacuum) and f = frequency.

For 4-20 mA instrument signals, those are DC so zero-frequency by definition when in steady state. The idea then is prevent signals from outside to be superposed on the 4-20 mA signal.

For power signals, f = 50 or 60 Hz, and c(cable) = a significant fraction of c (for example, 0.66 c), so wavelength is say around three million meters (so 3,000 km). Thus, shielding at both ends for power cables in a plant isn't needed since the lengths of the cables are nowhere near one-twentieth (say 150 km) long.

Some sort of protection for power cables of long lengths may be needed for power transmission over long distances (Hydro-Québec here in my province in Canada) added some systems to compensate and protect against signals induced by solar magnetic storms ... hydro power is generated in Québec's Far North, transmission lines ar elong, and H-Q's grid was knocked out back in the 80s during a big magnetic storm).

So: in a plant, connect cable and wire-pair shields at ONE end. And the 1/20th of a wavelength thing doesn't apply for power and instrument signals. For comunnication signals, it might.

DZ

Usually your piping is grounded or at least connected to a grid. All metal materials should be protected to slow down the corrosion process. But this is another subject.

I agree completely with Vkeyinc, but in reply to your second question. If your shield is terminated to a system that is not grounded, the shield floats right along with your ground system, but it still works as a drain to outside interference. If your system is not grounded it could lead to a whole new set of problems. I am not going to bring up because it leads to grounding, grounded, shields, bonds, etc.

I would like to note that the with some new BACnet and other Comm controls there is an awful lot of grounding the shields at the units rather than the base and some with no grounding of the shield period.

I suggest you stay within the design specs, they are supposed to know what they are doing.

Shield wires should always be grounded on one end. Typically the ground is made in the control panel. Grounding on one end and not both ends reduces the chance for ground loops. If both shield ends are not grounded the transmitter may be grounded to the surface it is mounted to provided this surface is metal and has it's own earth ground. While this is not a good idea I have seen it done from time to time. If this is an intrinsically safe installation then the shield must be grounded in the panel.

Hi all,

I am going to throw another spanner in the works here. With intrinsic safe loops - you are limited to the energy you may put into the unsafe area. If you ground the shield at both sides, you can induce stray currents that will surpass your energy limitation - thereby making that loop unsafe.

Another thing I have picked up many years ago in refineries, is that when they grounded the shield at both ends - the instrumentation became susceptible to lightning strikes that would blow the instrument, cable and most of the DCS away. This is because lightning looks for a potential. So hence - only ground the shield at one side and not both.

Kind regards,

Craig

well ,what I experience is that transmitter body has a different ground and end up to less than 10ohm earth pit .

what I feel is the signal conditioning ,

minimum 12 vdc is sufficent to run the transmitter but to have adequate DC supply at transmitter end is 24 VDC , small interference will not affect the dc current signal as long as voltage is sufficient, but if the level of voltage goes below 18VDC than this small interference also sufficient to vary the current signal(without shield protection), here it necessary to have shield earth (less than 1ohm) in order to shield the interference into the current signal . this is in general if not earthed properly ,we found in the loops with passive barriers or with long length signal cables where the loop voltage drops considerably between 14 to 16vdc .

we can also test it by giving fix signal at transmitter end and check on scope the variation of signal at dcs end or source end with and without shield .

regards

Adil