we are changing a control panel. currently there is a marshaling cabinet where all instrument gets connected to (LH side). cabling from that TB strip (RH side) is then brought to the controller.
what we are proposing, is to maintain the existing marshaling TBs, and install another control panel that will be connected to the RH side. as of now, the existing instruments DO NOT have their shield connected at the devices end, but the shield is connected to an instrument ground at the marshaling end.
so far so good.
our proposal, is to leave the new cable's shield, the one between RH side and the new control panel, connected only in our panel, unconnected at the marshaling end.
That way, we avoid any ground loop.
the client's engineer rather wants us to leave unconnected, both end of the new cable. His argument, is that a shield should never be connected at 2 places, to which we agree. but we are indeed dealing with 2 cables, not a continuous one... (and there is no TB for the shield, that we could have used to "create" a continuous cable)
in our case, the way we want to do it, each cable involved, will only have one end connected, leaving the other end unconnected. we explained to the client, that this will prevent any ground loop.
we also believe that having no end of the shield unconnected, could potentially create issues, as there is no way to "dissipate" any noise or garbage that could be picked up by the cable ...
I'm out of arguments for the client...it's common/recommended practice to have your shield connected to one end...not both, not none!
Am I wrong in my conclusion?
Agree with you.
Unconnected shield may behave like anything, no guarantee. Even it may get short circuited with any cable and lay down with may be 110 V or higher.
In addition, I am suggesting to ground it at the marshaling box instrument ground i.e. common point ground.
A ground loop exists only when you have two ground points at different potential. E.g. ground point in the field v/s ground point in CER. You may connect a cable to ground at multiple points if you can ensure each of these points is at the same potential today and a year later. If you connect ground cable at multiple points in the same building (like in your case) you are effectively connecting to the same ground point and hence no chance of a ground loop to exist. Hope this helps and remember... The customer is always right! :)
Leaving a signal cable screen/shield not made to ground at all seems like a big no-no, - making the screened circuit vulnerable to interference pick-up - as Barindra mentions.
Are the signals isolated, or ground-referenced? Field transmitter signals are often isolated. For isolated loops there is no problem with making the screen to ground at both ends - as the screen/ground is not connected to the signal circuit and therefore any ground loops existing are irrelevant to the signal quality. But screen-ends left open at the field instrument are worrying - They can act as an aerial/antenna and induce interference into the "isolated" field instrument.
Although not exactly the same scenario as discussed, the principle is the same.
If you wanted the best method to couple noise into a circuit for testing etc. you would wrap a cable with something conductive and drive that with your interference. That's pretty much what you do with an ungrounded shield. And as was mentioned, it's unpredictable. It could be worse than no shield at all or even worse than the ground loop you are trying to prevent. Or in the absence of noise, it could work fine until there is some.
At the risk of starting a brouhaha if I understand your description correctly, you are correct and the method you are proposing follows most accepted practices for shield drain wire termination. However, some plants and sites have their own practices and methods, and in the end the Customer while not always right--but must be made to feel that way. ;-b
You might try using a drawing to aid your argument if you haven't already.
There's some good information (and a crude drawing) in this control.com thread:
You can also use the Search terms:
in control.com to look for even more related replies to similar threads on this topic. (The plus sign (+) and double quotes are a BIG hint!)
Shield drain wire grounding is a very misunderstood topic around the world, and a subject which can cause--and has caused--many heated "discussions" (arguments). I believe there are some "standards" which describe how shield drain wires should be grounded--but as is commonly said about "standards" is there there are many of them ("standards") to choose from!
In the end, if you can't convince the Customer's representative I would suggest making provisions for terminating the shield drain wires as you have suggested in the future if it becomes necessary by exposing a 50- or 65mm length of shield drain wire and then roll it around the conductors and use electrical tape to insulate it. If electrical noise causes problems the tape can be removed, the shield drain wire unwound and terminated as it should have originally been.
Best of luck--and let us know how you fare!