Thumb rule for grounding shielded cable

Hi there
Actually dont know whether shielded cable being used should be grounded on both sides or only at one side.

Is rule same for both power shielded cable and signal shielded cable.

Google came out with this. Good one and it seems that thumb rule for signal shielded cable where ground current possibilities are zero go for grounded shielding at both ends and for power shielded cable where there can be large ground current go for ground shielding at one end end.

Ground one end

I have always heard and practiced grounding on one end only. Usually the end that is supplying the operating voltage.

A signal cable should have only one point of ground - at a loop power supplying device or its panel. Grounding at both ends may create some current trough the shield wire, if difference in potentials is exist between field device and a control panel.
I'm not sure about power circuits, but would assume it would be the same.

Our whole site has Profibus PA devices which all have screens connected at both the device and control panel ends.
I have also seen devices which are terminated at the device, junction box gland and the control panel, but with this type of installation a seperate ground should be run to each junction box back to the same source of earth so that the potential is the same.
Hope this helps

It all depends...

For instrumentation 4-20ma signals, the accepted practice is to ground the shield at one end only. The reason for this is that current flowing through the shield *WILL* affect the signal. It's a current signal and induced current is *BAD*. In an industrial plant the odds are very high that ground potential will *not* be the same at both ends of the cable. I've had more than one case where a cable jacket got scraped off inside a conduit while being pulled and made a "second" ground. Lifting the intentional ground at the panel made the signal return to normal. Replacing the cable, fixed the problem.

Some plants insist on unshielded twisted pair for 4-20ma. The reason, that an improperly grounded shield is much worse than no shield at all. This actually works very well in most cases.

4-20ma signals are not very susceptible to RF. It is way outside the range of the "signal" and is usually simply filtered out. This approach is not feasible for networks where the "signal" content is in the same range as RF. Then blocking external RF becomes very important. By the same token, an induced DC or low frequency AC current is easily blocked. (That's what those capacitors are for on the link in Manmeet's post)

I agree with Mellis. I allways ground the screen at the panel or DCS end, from there it's a simple matter to lift the common ground and test with an Ohmmeter

Power (VSD) cables must have shield connected at both ends. The shield is used in this case to keep the noise in. Recently did a vsd job where drives where tripping on earth faults, found only one end of shield connected, after connecting both ends no more problems

This indeed is very interesting topic. We also take care that signal cables are grounded on one side also. As a rule all signal cables are grounded inside DCS cabinets. Steel outer wire is wrapped around outer PVC insulation and connected to the metal frame of a rack. Usually when someone suggest to inspect "proper grounding", we just check if there is a good contact between metal frame of cabinet (which is grounded) and shield steel wire.
Don't know if there are better ways.

OK....
So I think actually bottom line is not to follow rules but to solve the problem.
So first connect the shield an one end only and see if the problem is solved, if not then connect shield at both ends !! Also keep praying.

But thing is problem due to RF interference and leakage currents is random and mysterious like, things are hazy and not obvious, thus problems arises sometimes. It may also happen that when one particular machine is not working then there is no problem and when it starts then then we have the problem !!

Here is my view:
- Grounding at both ends is the best possible "shielding" and should be used when noise is a problem. (i.e around VFDs or high AC voltage)
- Unfortunately, grounding at both ends completes the circuit between two seperate parts of the machine and can create a "ground loop"
- A ground loop is current (AC or DC) flowing through the shield because of a difference in ground potential at either end of the cable.
- The possibility of ground loops is why everyone only grounds at one end (even if they don't know why they are doing it!)
- So here are my two "rules":
1. 99% of the time only connect the shield at one end. This should be the end inside the cabinet so there is a common ground to all equipment. Does it matter which end is grounded? No. But you never know when someone will forget to reconnect the shield at the 'working' end of the cable. Having them all in the cabinet makes troubleshooting easier.
2. If noise is a problem, then ground both ends BUT you must also run a seperate ground cable from the ground strap at both ends of the cable. This allows both ground potentials to maintain an even balance by sending the majority of the stray current through your big copper ground cable and minimize the current through the shield of the cable where it can induce additional noise.

TheStarr said:

If noise is a problem, then ground both ends BUT you must also run a seperate ground cable from the ground strap at both ends of the cable. This allows both ground potentials to maintain an even balance by sending the majority of the stray current through your big copper ground cable and minimize the current through the shield of the cable where it can induce additional noise.

Click to expand...

TheStarr, I'm not really sure I understand this. if noise is a problem, then it's needed that both ends of a cable be grounded. That's OK, but what do you mean by separate graoud cable?
Can you draw some simple sketchup in Paint and attach it?

Thanks

2. If noise is a problem, then ground both ends BUT you must also run a seperate ground cable from the ground strap at both ends of the cable. This allows both ground potentials to maintain an even balance by sending the majority of the stray current through your big copper ground cable and minimize the current through the shield of the cable where it can induce additional noise.

Click to expand...

TheStarr great point. But again is not good that always connect both ends and then run separate ground cable to neutralize potential difference??

Pandiani i thought to explain u through words rather then drawing.
If there is potential difference between point A and B and if shielded cable is running between these two points which is grounded at both ends then current will flow through shielding.

If we run another copper wire which is connected to ground (0V) then the potential difference at both points will fall to 0V and then no current flow through shielding.

So, what you suggest is to use one regular copper wire from cable for this purpose? For example if cable has 8 wires + one for grounding, in total 7 wires will be available for signal transmission, right?

No, the ground wire needs to be external (outside) of the cable that you are trying to shield from the noise.

If you were to run the ground wire inside of the same cable, then any current running through the ground will cause a magnetic field to radiate from the ground.

This radiated magnetic field will then induce noise on the other wires in the cable through induction.

I would run the ground in a seperate conduit if possible or at least as far away from the signal cable as is feasible.

Ok, I understand now.
Thanks for the information.