Trying to figure out how to ground the shield of a 4-20mA loop (more details inside)

Hi all,

I need to send a 4-20mA signal from a PLC (ICLinks Everest) to a paperless chart recorder (Endress and Hauser RSG30) and I'm having a hard time understanding how to properly wire it up. Each output on the PLC has a positive output and all outputs share a common ground. Each channel on the E&H RSG30 has a positive input and a negative output, thus creating the loop. But where would I land the shield? Would it go to the common ground or is that a bad idea?

Related, I'm noticing that the other signals going to the E&H RSG30 (from analyzers, not the PLC) do not have their shields grounded either. Isn't he point of shielded twisted pair that the shield is grounded on one side?

defcon.klaxon said:

Related, I'm noticing that the other signals going to the E&H RSG30 (from analyzers, not the PLC) do not have their shields grounded either. Isn't he point of shielded twisted pair that the shield is grounded on one side?

Click to expand...

Yes, that's typical, and I typically ground the shield at the source location of the signal and tape it up (or cover it with heat shrink) at the passive end of the loop.

Someone posted here recently about a recommendation to ground a shield drain wire at both ends, and that may be fine for situations in which the ground potential is guaranteed to be the same at both ends of the cable, which, in my experience, is not always the case.

https://youtu.be/T9Snpmlmkq8

Shields should go to earth ground, not to analog (-).


I can't tell from the rear panel drawing whether that's a chassis earth ground screw circled in red just above the earth ground terminal for the supply AC voltage. That provides a tie point to some kind of multiple shield termination - terminal blocks? E&H is not the only vendor to not provide shield grounding on the rear panel.






To be honest, shielding for noise for some DI's is probably not that critical unless it's a noisy electrical environment and if the existing analog inputs don't have their shields connected and are working fine then the DI's probably don't need the shields connected.

As said above sheilds should be grounded at one end only, and isolated (taped) at the other.

A way to provide isoloation is by using a signal isolator. If you install it in the same enclosure and near the PLC, then noise and variations in the -ve terminal shouldn't be a big issue. This signal isolator can source the current output and allow it to be electrically isolated from the source. Just ground the sheild at the E&H end.

I have worked on systems where the shield has been tied to earth ground and others to the analog COM/-.

Some analog cards have POS. COM/- & SHLD - if so I use that shield terminal. Personally I don't like having the shield on the analog COM and prefer earth ground.
And grounding both ends could cause a current flow through the wire if there was a ground potential difference (It's been a while but I seem to remember it referred to as Ecm)

danw said:

Shields should go to earth ground, not to analog (-).

Click to expand...

Yeah that's what I was thinking.

I can't tell from the rear panel drawing whether that's a chassis earth ground screw circled in red just above the earth ground terminal for the supply AC voltage. That provides a tie point to some kind of multiple shield termination - terminal blocks? E&H is not the only vendor to not provide shield grounding on the rear panel.

Click to expand...

Yeah I believe that is earth ground. However, looking at the PLC I found an earth ground that is closer so that'll work.

To be honest, shielding for noise for some DI's is probably not that critical unless it's a noisy electrical environment and if the existing analog inputs don't have their shields connected and are working fine then the DI's probably don't need the shields connected.

Click to expand...

Not sure what you mean by DIs, can you elaborate?

Thanks everyone for the replies, I think I've got it sorted. There is an earth ground on the PLC not far from the analog output terminals, didn't see it at first but when I went hunting I found something that will work.

defcon.klaxon said:

Not sure what you mean by DIs, can you elaborate?

Click to expand...

My bad, midstream I confused your inputs as digital inputs, not analog inputs.