Where to wire analog input cable shield

Thread Starter

alexfrey

Joined Feb 23, 2019
22
I am new in automation industry and trying to learn fundamentals. In a project we have several analog inputs. My friend told me to wire cable shielding to earth terminal in one end. I know it is advised to wire it to earth but what happens if I wire it to power supply common ? What is the difference or it may make same effect ?
Thanks.
 

Wolframore

Joined Jan 21, 2019
2,609
I have seen positive ground in old British cars. Many issues related to that design. Just use ground for shielding. Its much safer...
 

Bernard

Joined Aug 7, 2008
5,784
Connect to equipment ground. Best if 2 signal wires encased in a shield. Might look at
some of the comments below.
 
Last edited:

Ya’akov

Joined Jan 27, 2019
9,069
I am new in automation industry and trying to learn fundamentals. In a project we have several analog inputs. My friend told me to wire cable shielding to earth terminal in one end. I know it is advised to wire it to earth but what happens if I wire it to power supply common ? What is the difference or it may make same effect ?
Thanks.
The reason for connecting the shield to one side is to prevent ground loops. If there is a potential between grounds on the two devices, current will flow on the shield. The function of the shield is to prevent induced interference, it shouldn't be used to carry a signal. The connection to a good earth ground on one side allows it to function, doing that on both sides allows it to be a signal path.
 

Ya’akov

Joined Jan 27, 2019
9,069
There is a tendancy now to connect both ends of shielded cables, especially when equi-potential bonding has been applied.
See Ch6 of the PDF.
Max.
So long as the equipment is all well bonded to a proper common earth, there's no reason to protect against ground loops that can't happen.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
Years ago I remember that everyone stressed the need for only connecting shield to one end, since the suggested practice of equi-potential bonding, ideas have changed.
This is what I started doing in my installations.
Max.
 
Last edited:

Ya’akov

Joined Jan 27, 2019
9,069
Years ago I remember that everyone stressed the need for only connecting shield to one end, since the suggested practice of equi-potential bonding, ideas have changed.
This is what I stated doing in my installations.
Max.
In recording studios, particularly in the past, it was extremely difficult to ensure no potential among grounds so even tough I always used a star ground, with my own earth system when possible, I still grounded the shield on the input side only. I often had to remediate hum in studios we bought time in.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
The OP mentions he is working in the automation industry which is the one I worked in and the PDF tends to apply to.
The intimation by the OP of the power supply common not being connected to earth ground is one option method, I always subscribed to the opposite method of earth grounding all power supply commons where possible.
Max.
 
Remember a couple of things:

1. Shielding reduces RFI
2. Twisting reduces EMI

So a lot of stuff is twisted pair shielded.

That said for the shield to work, it should be at earth potential. If you have the option to pick which end, use the sourcing end.

Grounding the shielded at one end helps eliminate ground loops. Believe me, you want to design a ground loop out of your circuit and not find and fix later.

I was using instrumentation in a lab and it's easy to find multi-channel pseudo differential voltage measuring things.
Now what happens when that instrument wants a setpoint too. The easiest way out is to use a current source and a resistor at the device's voltage input. Nothing really has to be isolated and no ground loops will exist providing there is a path to ground.

Now take something messy, 4-6 channels of Mass flow controllers that require +-15 to operate. One power supply for 4.

Now each controller is off in subspace somewhere. For potentiometer setpoints the pot is referenced to the MFC. There is a 5V supply to the MFC.

Everything works UNTIL you want a computer to supply this 0-5V setpoint signal. Yep, I missed that.

That particular MFC box was in-house made and was used successfully manually. I added calibration pots and decimal selects to make calibration easier. I also added a switch that set the readout to full scale which made calibration a snap.
Just missed the ground referenced input signal. Pseudo-differential readouts would have worked.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
Grounding the shielded at one end helps eliminate ground loops. Believe me, you want to design a ground loop out of your circuit and not find and fix later.
What comments do you have regarding the mention of it in the Ch6 PDF, if read.?

Quote:
When is simple shielding sufficient?
In most industrial plants, simple-shielded lines are sufficient for the reliability of the
plant.
The double-sided connection of the shield to the earth is thereby the most
effective shielding measure.
When is double shielding necessary?
Double shielding is advantageous for signal transmission in a particularly
interference-prone environment.

Max.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
In the industry:
- shield of data cable is grounded on both ends
- shield of motor cable is grounded on one end
Although when wiring a VFD to motor, the shielded cable instruction is to ensure both ends connected to earth ground, also this is automatically done if using metallic conduit, flexible or otherwise.
Max.
 

jfeyes

Joined Mar 23, 2021
2
I realize this is an old thread but one of the new electricians in our assembly shop decided to not follow the prints he was given and he connected many of the analog instruments shields at both ends, when I questioned his work and told him to fix it he argued this point and he cited this forum and thread, so I read it and had to comment so others don't follow this ill advise.
The comments such as "There is a tendency now to connect both ends of shielded cables, especially when equi-potential bonding has been applied. See Ch6 of the PDF. " are misguided. The physics and principals of electronics haven't changed therefore the reasons for only tying one end of the shield to ground still prevail. The Siemens article that is referenced earlier in this thread is concerned with emitted electromagnetic interference and compliance with the IECs EMC regulations, and seems to be focused on cabling of high frequency data signals. The article clarifies by mentioning a "special case" where only one end of a shield is connected quoted here "The single-sided shield connection must be applied only for special cases (e.g. only analog systems without digital technology):" The application of shielding as discussed in this thread is this special case. If anyone is interested in the reason for this single point connection here is a very good application note on the topic from Analog Devices(AN347).
 

Attachments

MaxHeadRoom

Joined Jul 18, 2013
28,617
Your opinion. That app note quotes '70's bibliography
Early in my career it was normal and accepted practice to only bond one end of a signal cable shield to earth ground.
This mainly due to what was considered a means of reducing the effect of ground loops.
This came to be especially important with the advent of Industrial systems such as LV (5v) CNC systems mixed with H.V. VFD sytems etc.
Things have changed, Since then then in this century the likes of Siemens have explained the method of equip-potential bonding practice in order to prevent ground loops.
.
 

Deleted member 115935

Joined Dec 31, 1969
0
One thing to remember, is not to put a signal / power through the shield,
if you do its no longer a shield , its a radiator,
even if there is a DC path for the gnd, if current flows in the shield, it will radiate.
 

Ya’akov

Joined Jan 27, 2019
9,069
I realize this is an old thread but one of the new electricians in our assembly shop decided to not follow the prints he was given and he connected many of the analog instruments shields at both ends, when I questioned his work and told him to fix it he argued this point and he cited this forum and thread, so I read it and had to comment so others don't follow this ill advise.
The comments such as "There is a tendency now to connect both ends of shielded cables, especially when equi-potential bonding has been applied. See Ch6 of the PDF. " are misguided.
Regardless of the merit (or lack thereof) of your criticism, there is a more important point. The electrician should have followed the prints, not decided on his own. If he had questions, he should have raised them. What he did was wrong, possibly not for technical reasons but certainly for professional reasons.

If he is the TS on the original thread, he said himself he is a beginner. If he thought the advice here made sense, he should have raised the idea with a supervisor and done whatever he was told to do after that.

It is not at all clear if there is any technical reason to lift grounds in an equipotential environment but it is crystal clear that random folks on the Internet are not the authority for not following explicit instructions from those who not only employ you but know more than you do.
 
What would you do in this situation?

You discover that the doors of a lab being renovated would open into the major means of egress, so the door would slam into the face of oncoming foot traffic. You knew that it was not planned based on discussions. The big boss wanted eveyone to stay out of the area until renovations were complete.

I told my boss and I got in trouble for being in the room. The problem got corrected. I made a quick tour and discovered other problems. I did manage to get my boss to schedule a walk-thru before things were completed. One thing I remember being found was a valve plastered into the wall, so it could not be moved.

If there would have ever been a next time, I'd keep my mouth shut. There were more instances.

Another was during the renovation, I was told to upgrade the gas detection system, so I needed parts. While talking to the manufacturer, I found out that supplies to run the system would not be available in 6 months. This would be a $40,000 replacement, I got screamed at in his office. He mistakingly thought that I we had to install a new unit now. that was not the case.

Powers that were not in charge just asked the electrician to just "move" the gas detection systems for Hydrogen and Hydride. No consideration, that there would be 2 extra sampling lines; 3+ additional strobes, 3+ additional panic buttons, and 2 extra Hydrogen sensors. The 3A 24V power supply wasn't up for the task.

I had to put in multiple power limited class II circuits. Some strobes worked and some did not, same brand. Turned out it was an inrush current problem.

I would not finish "sections". Once I "knew" it would work, I worked on designing and ordering parts for my next phase.

The big boss threatened me with "cutting off my balls" before he became the big boss.

My boss wanted to "kiss is a$$" and show that show "x was complete".

If i didn't have it designed or the parts ordered, the entire project would suffer and not be done on time. I was not allowed to "design and build". I was building and designing on the fly, so bits and pieces of docuentation was done. Call it crypic.

We should have had a "project engineer" which I had the skills for, a a Ghant chart so that our people could work too. For the most part, construction was not in the room I needed to be in. it would have helped a lot.

I busted my ass and came in after hours because I thought the project was important. One person, a chemical engineer asked why wires had to be neet. He wanted the project to proceed faster.

One "hired help" would do his job (build a control panel) unless I was involved. I had to do the interfacing to gas control

So, just like the last time, something bad happened:
The last time, the hydrogen detectors were not installed and the gas involved in the explosion was Hydrogen. It was a failed excess flow valve combined with lines rated for about 10 PSI. A burst of hydrogen was let through that line which was normally used for pumping out the regulator. When the tank valve was reseated, a 1500 burst of hydrogen burst out of a corrugated line. The hood velocity alarms broke with the burst of Hydrogen, so the gas was shut off at the cylinder.

This time, a set of safeties were going to be installed to prevent a load lock chamber from being opened to atmosphere when the load lock isolation gate valves were open. There is a manual "pressure interlock bypass" that I needed to interlock. Never happened.

We would still probably use the mechanical valve. I did put a way so the fire alarm would not close a gate valve. They would stay in the same position. They would bend the $1000.00 load lock arm.
 
Ground loops are something that you design out from the very beginning. It's expensive to solve afterwards. Sometimesyou have to take the time option and other times, the money option.

I messed up once and the option was money. Reduce, the number of Mass flow Controllers controller to 8 at atime instead of 10. Readouts for 2 on the panel simultaneously. Use the same commercial product for 4 more channels.

Otherwise, I would have to design something that worked. * MFC's were built in-house for manual operation. I wanted to convert to analog control.

The 8 MFC's shared one 5V supply and one 12V supply, possibly -12. They could be powered easy enough, The meters required 5V. The reference had to be 5V.

The setpoint meant putting potentiometers to the MFC, 3 wires which made SP isolated in manual. The readings could be read differentially by computer or by the panel meters.

Problem, I missed, is the 0-5V setpoint had to be relative to each MFC controller ground.

I wasn't given the option of isolated current out or current out. I wanted all DAQ stuff out of the computer.

The commercial controller had a way to bring the offset to the source.

Using six 0-5 to 0-20mA isolated modules would have been too expensive and making a custom one would have been too expensive time wise.
 
My advice. Build as drawn UNLESS your one of the responsible parties that can be blamed when things go awry. donlt even point out facts "Part will be obsolete in 6 months". You will be labeled a troublemaker and your out of a job. pretend Donald Trump is your boss. You don't get paid and you don;t do what "I" want, your fired.

I helped people at work. They got promotions, not me.

Be careful abut doing things that have value, but the value can't be measured. Do you think anyone will remember that 17 years ago, you were pro-active and selected a particular method and brand for a surge suppressor and the only failures were a floppy drive and dust?

No, they won't.
 
Top