grounding both sides or just 1 side?

Should I ground both sides or just one side?

I've been recommended both ways.

What I am interested in is reasons/arguments pros/cons if you were to say for both.

No fighting please.

Maybe one discussion for motor leads, another for control wire shields?

Instrument cables only get drained to ground at one end. It's to keep potential external noise out of the cable. I normally ground the end where the signal is being measured, but I have run into cases where not grounding the shield or grounding the opposite end yielded better results.

VFD cables get grounded at both ends, to drain the large amounts of induced noise that are going to be present inside the shield.

That has been my practice.

Take a look at wikipedia , ground loop:
https://en.wikipedia.org/wiki/Ground_loop_(electricity)

OkiePC said:

Instrument cables only get drained to ground at one end. It's to keep potential external noise out of the cable. I normally ground the end where the signal is being measured, but I have run into cases where not grounding the shield or grounding the opposite end yielded better results.

VFD cables get grounded at both ends, to drain the large amounts of induced noise that are going to be present inside the shield.

That has been my practice.

Click to expand...

That is the way I have always been taught. But I have heard dif. via techs over seas. They seem to ground both ends of the signal wire.
I'll have to get more info, but I seem to remember a customer of mine grounding the signal wire at both ends to help a situation, but I also seem to remember his electrician doing something stupid like running the signal wires in parallel w/ the motor leads. I will have to check on this.

Same as OkiePC only ground one end.
Usually in the I/O enclosure.

I'm not so much looking for the answer than I am looking for the reasoning behind the two methods of thought.

widelto said:

Take a look at wikipedia , ground loop:
https://en.wikipedia.org/wiki/Ground_loop_(electricity)

Click to expand...

This is what I was taught

Signal wires should be shielded and the shield connected to earth at the source. This is bonding, not grounding, technically, but we all call it grounding. This is sometimes confusing, but use PLC analog as an example. I run a cable from the PLC analog output to the input of a VFD, I ground the shield at the PLC. If I'm running a cable from controls cabinet to an analog sensor, I would ground the shield at the power supply, or as close to as practical.

Wires that feed motors should be grounded at both ends due to the larger currents available, and because any "noise" that gets induced into the wire is unlikely to cause any sort of unintended operation of the motor. A few extra milliamps on a 1 HP motor isn't going to change the speed, but a few extra milliamps on a 4-20mA signal sure will skew things!

again, I keep getting the answer is A. Why is the answer A?
yes, there was a wiki article that I still need to read.
I assume it is because electrio-magnetic lines of flux may be induced in the shield if it is grounded on both ends?
How would making sure there was a least resistant path to ground hurt us here? That is the part I am confused about. Especially since the East recommends it and I've seen it help. Sorry for being stupid on this.

I'll save you a click:

"In an electrical system, a ground loop or earth loop occurs when two points of a circuit are intended to have the same ground reference potential but instead have a different potential between them.[1] This can be caused, for example, in a signal circuit referenced to ground, if enough current is flowing in the ground to produce a voltage drop and cause two ground points to be at different potentials.

Ground loops are a major cause of noise, hum, and interference in audio, video, and computer systems."

I see your point. dif. in potential = voltage, voltage path = ground loop current.
What I don't see, is how this ground loop current would induce more noise.
As it is obviously apparent in sound equipment, may it differ in larger induced emf where it would just pick the shortest path to ground?

By grounding at one end only, you determine where the stray or induced energy can go. Creating two grounded points and the possible imbalance in voltage at them, will cause a current to flow. The amount of current that could be induced is related to the strength of the magnetic field inducing the voltage. I suspect it would take an incredible magnetic field to produce an appreciable amount of current. In practicality, I have seen plenty of signal wires where the shielding was grounded at both ends and no problems with the signal on them. If things are wired and bonded correctly, the potential at all ground points should be the same.

We must differentiate between Grounding, and Shielding.


In the majority of cases a connection to GROUND is assumed to be a zero-ohm connection to a common potential point (we all know that a zero-ohm connection is improbable), and is used as a SAFETY connection. We use grounding to ensure that exposed conductive objects cannot be above ground potential, making them safe to be touched by someone standing on the GROUND. Therefore, GROUNDING must always be done LOCALLY, so that means bonding to ground at BOTH ends of a cable run.

Conversely, a connection of the SHIELD of a cable to GROUND provides for a degree of noise immunity in conductors carrying low-voltage or low-current analog signals, i.e. the SHIELD connected to GROUND is NOT done for any degree of SAFETY, but for convenience. Grounding a Shield cable at one end only prevents any imbalanced currents flowing in the shield, inducing noise into the low-power analog circuits. Think of a SHIELD connection as a DRAIN for unwanted noise.

That's my 2c take on it, prepared to be shot down though.....

EDIT, lots of further posts since I started this

OkiePC said:

Instrument cables only get drained to ground at one end. It's to keep potential external noise out of the cable. I normally ground the end where the signal is being measured, but I have run into cases where not grounding the shield or grounding the opposite end yielded better results.

VFD cables get grounded at both ends, to drain the large amounts of induced noise that are going to be present inside the shield.

That has been my practice.

Click to expand...

I second that.

I have always terminated the ground for shielded control wiring (like analog signals) at the cabinet end and trimmed the shield at the device end in the field. This allows the prevention of ground loops which just circulate currents and defeat the purpose of shielding your signal cable. (This is usually for low voltage signals)

The ground for VFD's and motors and other devices are always terminated at both ends, and this is very important to allow shorts to ground to travel back to the panel and trip the breaker. Otherwise if that voltage does not have a solid ground it will find it's way to ground maybe through you or something else.

This is a bit of a minefield, generally the "screen" component of a cable should be grounded at one end (normally the equipment at the other end is bonded by an earth, the screen should not be relied on as a bonding (earthing) conductor.
For example: a signal cable from a sensor is a two core screened cable, the screen should be connected at one end (normally the control panel) any junction boxes in-between the screen should be carried through as an isolated connection. The equipment should be bonded to earth (where applicable) using a earth designated cable capable of withstanding the fault current.
In other words, either use a 3 core screened cable (2+E+screen) or external bonding. This eliminates the possibility of a large fault current using the screen as a path, this is what I was taught in my early days.
An example of this is when I commissioned a oven using SLC500 and RIO, the RIO contained analogue & digital signals, the installer connected the screens at both ends, the remote panels were also bonded to the main panel with the appropriate earth bonding cables, we were experiencing quite large fluctuations on all analogue signals, disconnecting the earth bonding(not by me but the customers own engineer cleared the most of the errors), I re-connected the bonding cables & disconnected the screens at the RIO, all was good. This also applies to high speed transmission lines, perhaps read up on "Reflections in transmission lines", this may also apply if noise is present.