Shielded Cable And Why?

Yes, cable grounding does get confusing. One of the reasons this gets confusing is that often it does not matter if either end is bonded to earth at all. Thus inconsistent cable shielding makes no difference in many observable responses. To make matters more confusing, sometimes a "bad" shielding technique will cancel an interfering signal coupled at a transducer. This is why a turntable has an optional ground strap on the pickup cartridge and in connecting to the phono amplifier input. Sometimes it helps, others it doesn't and sometimes your speakers cannot reproduce 60 Hz, so who cares.

It also depends greatly on the frequency of the signals considered to be in band and those that are out of band. By definition a VFD will produce a wide collection of frequencies with considerable to modest power. Keeping these frequencies inside the cable shield is often easier than shielding high gain circuitry.

Then there is the most baffling aspect of bonding, ground loops. Cabling from point A to point B with both points having there own Earth ground makes a ground loop. The ubiquitous power distribution signals of 50/60 Hz now have a loop antenna or parasitic transformer loop to induce an interfering or canceling current.

See Figures 2-27 and 2-28 on pages 72 & 73 in Henry W. Ott's book, ELECTROMAGNETIC COMPATIBILITY ENGINEERING, Wiley:

• http://s1.downloadmienphi.net/file/downloadfile5/192/1388769.pdf

There are two ways in which "noise" can be coupled into a circuit. An electric field will induce charge into the conductors. The best way to prevent this is to enclose the circuit completely with a grounded conductor, a shield, grounded completely (both ends).

The second way noise can be introduced is via magnetic fields. If there are currents flowing in the ground (ground loops), grounding the shield at both ends can induce noise in the circuit from currents flowing through the shield. In this case, better shielding is accomplished by grounding only at one end, generally the signal source end.

So, which method is better depends on the situation, the noise environment, the noise frequencies present and the noise frequencies which are likely to interfere with the signal.

https://en.wikipedia.org/wiki/Talk%3AShielded_cable

Oh I wish it were true there were only two ways for noise to be coupled into a circuit. Noise can be mechanically coupled through microphonics. Then there are all of the really esoteric thermal noise pathways of Shot (Schottky), and Johnson noise. Don't forget the extremely rare ionizing radiation. Then there's my favorite, the intermittent switch known as a bad contact.

Noise just comes from so many possible places and pathways.

Oh I wish it were true there were only two ways for noise to be coupled into a circuit.

This is true, but I suspect that none of the sources of noise listed will be much affected by whether you ground the shield on one or both ends...

Yep, "triboelectric" noise resulting from active mechanical deformation of dielectric (coaxial cable) materials, especially a problem with accelerometer cables.

All good answers really, but I like to sum it up for the non-engineering crowd this way:

It's all about context.

1. If you are talking about INSTRUMENTATION and COMMUNICATION wiring, you want to prevent stray signals from getting IN to the conductors and creating errors. So grounding only one end reduces that likelihood.
2. If you are talking about POWER cables, i.e. VFD output cables, the output conductors are like radio frequency transmitting antennae (FM radio = "Frequency Modulation", that's what a VFD is doing...). So in that case, you are creating a "Faraday Cage" around the cables to keep the nasty stuff INSIDE.

I agree, and add the following.

Grounding one end only of a high voltage power cable shield is fraught with danger as voltage drop caused by induced currents in the shield flowing to earth at the ground point can create a dangerously high voltage to earth to be present at the ungrounded end. A person coming between the ungrounded end of the shield and earth could be fatally injured particularly during fault conditions.

There is also the concern with very long runs of HV cable that the impedance of the shield becomes sufficient that high voltages can build up in said shield resulting in corona discharge through the outer sheath to any external earth, this can severely damage insulation. Grounding of both ends of the shield mitigates this to some degree. Very long runs of HV often have numerous shield grounding points along their length to prevent electrostatic, inductive and capacitive voltages from reaching dangerous values.

It falls to the installers of any adjacent communications cables to protect those from interference, and this is generally done with one end only of the sheath grounded for reasons that have been provided by others here.

For aircraft I was taught to ground only one end of the shield of a cable, preferably to the equipment it was plugged in to. This was to prevent a voltage potential differential of an ungrounded cable and a possible static discharge, yet not provide a closed circuit (by grounding both ends) that could set up a electron migration corrosion situation.

Hooker

PS - not sure I got the terminology of that last part right, but, oh well.

This is a very confusing topic with as many opinions and theories as you can stand.

We work to customer specs and almost every customer spec says to ground the shield on instruments at the control cabinet (PLC-I/O cards) Process instrumentation works fine with this as does the VFD cables most time. (see my last paragraph for the exception)

Basic accepted tenet is to ground closest to the signal source, which would mean the instrument on AI circuits and the output card on AO signals. so most customer specs violate this, with no ill effects.

Another commonly accepted belief is ground loop theory that says ground at only one end, or you will get a ground loop. However, I had this thoroughly dissected and rejected by a true "ground loop" expert about 6 years ago on this forum. (he had experience in ultra-high frequency high power circuits and shielding of same)

That said, we have used IFM shielded cables that ground the shield to the instrument through the M12 connector nut. Never had a complaint. And we have used Turck cables with the shield connected at the control panel via a drain wire that touches the shield. No complaints with these either. Add to that, we have used both of the above cables on the same project with no problems. So even the cable manufacturers cannot agree where to connect the shield, but build the cables such that only one end is grounded.

Currently we are testing a decent sized process skid with an interface cabinet (just terminals) between the skid with the instruments and the control panel (PLC) which will be remote mounted on site. In between we have temporary cables for testing, and have chosen not to connect the shield, so our instrument cables are essentially not grounded anywhere. This is working fine in our shop with a large bank of TIG welders using high frequency starting for the arc just yards away.

However, I have also experienced the effects of incorrect shielding. We used a VFD controlled by Profibus on one project, and could not get the Profibus to communicate to the system until the VFD manufacturer told us to ground the shielded motor cable at both ends. Quite skeptical was I until viola - instant success after following these directions.

Without asking why <...we...> do what <...we...> do and why <...they...> do what what <...they...> do, it is difficult to know how to proceed with this thread. Have the questions been asked of <...we...> and <...they...>?

The principal reason for earthing one end of the shield only, is to avoid the prospect of circulating currents in the shield that might affect the signals in the conductors inside it. Further, circulating currents can enhance corrosion, which is the last thing one wants in an <...aviation...> environment!

It is quite possible that the <...we...> is getting it wrong. Which is why the questions must be asked, so that the <...I...> is no longer <...confused...> and the <...we...> can amend its standards and training, so that circulating currents and corrosion become avoided.