New Science: Ground both ends of instrumentation cable?

Has the thinking changed on the grounding of shields on analog instrumentation cables?

A friend recently attended some company training on grounding & noise. Their expert demonstrated how it was best to ground the shield on both ends, specifically on 0-10VDC potentiometer circuits.

Apparently he used some sort of noise inducing device to demonstrate the effectiveness of grounding the shield on both ends. His company uses a lot of 0-10VDC stuff.

My experience was mainly with 4-20mA instrumentation & we always only grounded one end. Never had any problems. Of course 4-20mA circuits are known to be more tolerant if noise but still …….???

Any comments appreciated.

Did they use a device to induce a ground loop and see what the effects were?

I have a vague recollection (from a long time ago) that the important thing was not so much whether you grounded things on one end or both, but whether you were consistent.

You've got two types of devices to consider - those that generate noise (e.g. VSD's) and those that need to be protected from noise (e.g. 0-10V instrument circuits).

If I'm remembering rightly, the important thing was that you picked one approach for the noise generating circuits, and used the other approach for the noise sensitive circuits. So you can either ground both ends of all your VSD's, and one end of all your instruments - or both ends of all your instruments, and one end of all your VSD's.

Conventional wisdom has it that we ground one end of instruments and both ends of VSD's, simply because that's the easiest philosophy to follow practically. It's easy to earth a VSD cable both ends because the both the drive and the motor have a solid earth to connect the screen to. For many instruments, you'd have to run a separate earth. But it's super easy to ground all the cable screens at the MCC using the right terminals or cable clamps, and then just cut and heat shrink the cable in the field.

Of course, this information is based purely on a vague memory from a long time ago, so don't put too much weight on it until someone who has actual verified information comes along to tell me I'm either correct or full of s***

The reason for avoiding grounding instrument cables on both ends is that ground potential can be different at different locations, thereby inducing current in your shield conductor.

Maxkling said:

Did they use a device to induce a ground loop and see what the effects were?

Click to expand...

I wish I had more info about this noise class & what was actually demonstrated. My friend just said the expert had some wiring laid out and used something to introduce noise.

Sorry, wish I had more detailed info.

For low frequencies an overall screen, grounded only at one end, provides good shielding from capacitively coupled interference but none at all from the magnetic fields, because this can only occur if current flows in the screen. To shield against a magnetic field, both ends of the screen must be grounded. This allows an induced current (IS) to flow in the screen which will oppose the current induced in the centre conductor. The same principle applies when shielding a conductor to prevent magnetic field emission.

To minimize low frequency magnetic field pick-up, one end of the circuit should be isolated from ground, the circuit loop area should be small, and the screen should not form part of the circuit. You can best achieve this by using shielded twisted pair cable with the screen grounded at only one end. The twisting minimizes magnetic coupling, and the screen will reduce external capacitive coupling.
So, yes ground screen one end only, should not be part of the circuit i.e. the return conductor, This has proved to be the case many times in my experience, although both connected or only one end, in my years of experience it's not the magnetic field emission that causes the problem ( generated & radiated fields) it's the low frequency magnetic pick up by the cores that is the problem i.e. interference on the signal, so you have two scenarios, one where it can generate interference to other equipment two it can pick up magnetic interference & cause instability of the analogue signal you are measuring, this is probably the most prolific, as small level analogue signals are unlikely to generate high level magnetic fields.

If you ground both ends, then you have to make sure that everything is grounded properly with proper grounding wiring.

Can be quite huge task..

For fieldbus cables that is mandatory.

ASF said:

I have a vague recollection (from a long time ago) that the important thing was not so much whether you grounded things on one end or both, but whether you were consistent.

You've got two types of devices to consider - those that generate noise (e.g. VSD's) and those that need to be protected from noise (e.g. 0-10V instrument circuits).

If I'm remembering rightly, the important thing was that you picked one approach for the noise generating circuits, and used the other approach for the noise sensitive circuits. So you can either ground both ends of all your VSD's, and one end of all your instruments - or both ends of all your instruments, and one end of all your VSD's.

Conventional wisdom has it that we ground one end of instruments and both ends of VSD's, simply because that's the easiest philosophy to follow practically. It's easy to earth a VSD cable both ends because the both the drive and the motor have a solid earth to connect the screen to. For many instruments, you'd have to run a separate earth. But it's super easy to ground all the cable screens at the MCC using the right terminals or cable clamps, and then just cut and heat shrink the cable in the field.

Of course, this information is based purely on a vague memory from a long time ago, so don't put too much weight on it until someone who has actual verified information comes along to tell me I'm either correct or full of s***

Click to expand...

Yes the difference is the purpose of the cabling. Instrumentation cabling is not supposed to carry ground current, but vfd motor cabling should have proper grounding properties for its "shield". That is basically the ground wire.

My understanding is that grounding at both ends is hypothetically better ASSUMING you have a good quality ground at an even potential to avoid ground loops. This is apparently common in Europe, but uncommon in the US. Thus you have two sets of advice that seem to be contradictory but really just start from different assumptions.


As an example, Profinet (mainly driven by EU companies) recommends grounding both sides of an ethernet cable. EIP (US driven) recommends grounding one side.

Bit_Bucket_07 said:

The reason for avoiding grounding instrument cables on both ends is that ground potential can be different at different locations, thereby inducing current in your shield conductor.

Click to expand...

And the reason for grounding VFD cables on both ends is that the shield of the motor cables does carry current (common mode current).

mk42 is correct. Grounding both sides of a shield will always result in better noise exclusion/containment than grounding one side. But as was mentioned, if you haven't VERY CAREFULLY grounded your system you run the risk of running currents down the shield due to voltage differences between the two sides of the cable. So you have done a great job of protecting the signals lines from noise from OUTSIDE the cable only to run the risk of injecting noise into the signals lines from currents running through the shield.

Again, as was already noted, the shield in a shielded cable works on both sides. It will keep noise out but will also keep noise in. This is why VFD cable is shielded and also why VFD cable is typically grounded on both sides. Grounding both sides provides for better noise containment and the lines inside the cable are not susceptible to noise generated by currents that may be flowing through the shield due to end-to-end voltage difference.

Keith

dogleg43 said:

Has the thinking changed on the grounding of shields on analog instrumentation cables?

A friend recently attended some company training on grounding & noise. Their expert demonstrated how it was best to ground the shield on both ends, specifically on 0-10VDC potentiometer circuits.

Apparently he used some sort of noise inducing device to demonstrate the effectiveness of grounding the shield on both ends. His company uses a lot of 0-10VDC stuff.

My experience was mainly with 4-20mA instrumentation & we always only grounded one end. Never had any problems. Of course 4-20mA circuits are known to be more tolerant if noise but still …….???

Any comments appreciated.

Click to expand...

MANY decades ago, I worked for a major metals manufacturer who had the same question. They studied it, and decided that the best "..twisted pair.." cable was actually UNshielded, BUT twisted three-twists-per-inch. This twist specification was sufficient to reject or disrupt external magnetic fields. Of course, this was many decades ago, long before the mass use of VFD and other similar gear that is so common...not to mention high frequency communications.
But, it is another approach.
I've heard and taught for years to "shield at the source, and only at the source".
.
And, there is a web site for this, as with everything else in the world, it seems:
https://medium.com/@emianalyst/effect-of-cable-twisting-on-radiated-emissions-66edb16bec1c

mk42 is correct. I have a visual coming on this. Using the same plasma ball setup I did in this video, I have an upcoming video where I ground both ends of a shield and it works great. Then I add a voltage imbalance on one end of the ground and as current begins to flow through the shield, you start seeing noise on the signal wire.
https://twcontrols.com/lessons/how-...se-we-use-a-plasma-ball-to-find-out?rq=plasma