Ground loops

[bthielen]

A while back I was wiring a new addition onto my home. Part of this project included the addition of a new hallway from an existing room to a new outdoor entrance. I added a light in the hallway on a 3-way switched circuit. This 3-way circuit was fed from the old existing room's lighting circuit. The 3-way switch near the exterior door shared a duplex device box with the new switch for the exterior light, which was fed by the new lighting branch circuit. What surprised me was that the inspector required that all the ground wires within the duplex box be twisted together. Since these grounds were fed from two different branch circuits, doesn't this provide a potential ground loop? I raised this question and was told it was not a concern. I find this surprising with todays use of computer systems and other electronics.

Any thoughts?

 

[don_resqcapt19]

Re: Ground loops

First, the code requires this (250.148) connection and second, where is the source of this current that would be circulating in the EGC?
Don

 

[bthielen]

Re: Ground loops

A ground loop occurs when there is more than one ground connection path between two pieces of equipment. The duplicate ground paths form the equivalent of a loop antenna which very efficiently picks up interference currents. In our business where we use VFDs, Servo systems, etc., electical noise is becoming a more serious problem everyday. The concern I have is the introduction of radio frequency noise onto the EGC via ground loops which could potentially damage equipment in our homes. The european community (CE) has begun addressing this issue in their electrical codes and I'm curious why we, in the United States, haven't been doing likewise.

 

[don_resqcapt19]

Re: Ground loops

bthielen,

The concern I have is the introduction of radio frequency noise onto the EGC via ground loops which could potentially damage equipment in our homes.

Can you cite any documented cases of damage to equipment that were caused by two or more equipment grounding conductors being connected together? Almost all of the "electrical noise" on power distribution systems is caused by the very same equipment that is sensitive to such noise. The European's have rules that limit the amount of noise that any equipment is permitted to put onto the power distribution system.
The lack of bonding the multiple EGCs to the box and to each other can increase the impedance of the fault clearing path due to the inductive reactance of an EGC that is not run with the power conductors. If you have circuits A and B in a box and only terminate the EGC from circuit A to the box and there is a fault from circuit B to the box, the impedance of the fault clearing path is increased.
Don

 

[dereckbc]

Re: Ground loops

Since these grounds were fed from two different branch circuits, doesn't this provide a potential ground loop?

NO, not from an AC branch circuit distribution point of view.

Ground loops and bonding is the most misunderstood subject in the industry. Generally speaking it is impossible to prevent ground loops in AC distribution systems. For example in industrial and commercial applications you have metallic raceways supported on by the building concrete or building steel. Then the equipment is mounted on concrete floors, in contact with adjacent equipment and support infrastructure. All this forms incidental contact with ground, which forms more loops than you can count. So its almost impossible to prevent incidental contact, unless extreme measures are taken.

In a wood frame home it is possible to obtain an isolated ground by using a dedicated circuit, by using NM cable to a plastic box mounted on a wooden frame member.

There are very few industries, which even attempts to prevent ground loops by using an “Isolated Single Point Ground Plane”. The only two industries I am aware of are Telephone companies, and cellular/radio sites. But they still intentionally form loops on the ground electrode side of separately derived systems to lower the overall high frequency impedance to earth on DC powered systems. But all the AC powered systems are installed in what is called the “Integrated Ground Plane”, and would never try to isolate as it would impossible to police to prevent violations of an isolated ground plane.

 

[roger]

Re: Ground loops

Hello Dereck, I would like your opinion of what is loosely refered to as "isolated ground systems".

The following question is an extension of your post.


Lets take the average metal stud facility, install an iso ground receptacle. In this facility we have metal piping, conduit, structural steel, the studs, and duct work physically touching through out the structure, how did the concensus come about, or maybe I should ask , why do some still think that having a separate conductor will negate the "noise"?


Roger

[ July 21, 2003, 09:05 PM: Message edited by: roger ]

 

[dereckbc]

Re: Ground loops

Roger, a little test question for me. OK I will bite. Isolated ground receptacles (IGR) are only intended to be used for a possible means of obtaining common-mode electrical-noise reduction on the circuit in which it is used. It has no other purpose. Its effects range from no effect, worse noise than standard solid grounding, to the desired effects in that order of success.

I have seen it used with limited success in recording studios and A/V applications. However, most professional studios are use a better and more successful approach using balanced power systems (120/60).

The key is common-mode current flowing in ground conductors. The theory is by using an isolated ground conductor prevents any outside current from flowing in the ground conductors and developing voltage drops (common-mode noise) along the length of the ground conductor.

The best example that everyone can relate to is the hum heard in audio power amplifiers when the volume control is turned up to full gain. Lets say you have a microphone mixer, pre-amp, and amplifier powered from different outlets, and using ground referenced interconnect cabling. The slight differences in ground potential is picked up by each amplification stage and amplified and heard in the speakers. However if you use a dedicated IGR for each stage, then use ground referenced interconnect cables you defeat the purpose of the IG by creating loops which may make the problem worse. Best way to approach is to use a single standard outlet. Yes, the ground may be noisy, but from the single point reference, there is no potential difference between systems.

As you know I work for a phone company, and we use isolated ground planes on DC powered equipment. But it is not for common mode noise rejection (just an added benefit), but for preventing outside faults from flowing through the equipment like lightning. But we would never attempt to isolate the AC distribution system. As a matter of fact our standards prevent us from using IGR. The only place we would use IGR is for what we call “Co-Located Equipment Customers” (CLEC) under contact requirements. But we train drivers laugh at the CLEC’s because we know 95% of them corrupt the IGR by using signal cables with ground conductors or mount there equipment shelves without isolators in a metal equipment rack which is grounded to a “Signal Reference Grid” (SRG).

Did I pass the test?

 

[a.wayne3@verizon.net]

Re: Ground loops

OK then what about this senario.residential application in which there are 2 different circuits in one multi gang box ct.#1 is gfci protected #12 and ct,#2 is for lighting and in #14
First how can you mix the different sized grounding conductors and then there is the issue of the gfci tripping when tied into a non gfci circuit.We have encountered this same senario and there is no place to put 2 seperate boxes.We have one builder that has one model in which there is barely enough room to install and trim out a 2 gang box never mind 2 seperate boxes..........

 

[pierre]

Re: Ground loops

Allen
Are you saying that the GFCI trips when you connect the EGC of two different circuits, in the same box?

Pierre

 

[jxofaltrds]

Re: Ground loops

Connecting egc's together will not trip GFCI's. Neutrals of different circuits will.

As far as "looping", unwanted interference is usually shunted to ground. Is this causing the problem? If so please explain.

Mike P.

 

[roger]

Re: Ground loops

Dereck, you get an A+.

Its effects range from no effect, worse noise than standard solid grounding,

I have found in my experience that this is the most common outcome.

We especially have a hard time explaining (or trying to) why, with all the isolated grounding, there are problems with "noise" on ioslated ground supplied equipment .

Then on the other side of the coin there are installations of medical equipment, xray, CT scaners, MRI's, etc... rooms that are so full of ground loops (they call it redundant grounding to elliminate "noise") it boggles the mind. These grounding conductors are designed by the individual equipment manufacturers.


When you point out that their designs actually contradict what their desired goal is, they look at you with wonder, contempt, we dare you question us, or some other we don't understand expression.

Roger

 

[bthielen]

Re: Ground loops

WOW! I didn't think I'd invoke such a wild discussion. Although, I should have known better since grounding is such a complicated and important aspect of electrical systems. Great discussion! I have learned a lot.

Thanks, Bob

 

[karl riley]

Re: Ground loops

Roger and Dereck, I find this thread very educational. I do have a question, since the different ideas of how to try to eliminate noise don't always solve the problem:

do you consider as a possible source to be checked out the induced noise caused by net current magnetic fields? Is it part of your measurement techniques to eliminate this cause?

I have dealt with instrument malfunctions due to as little as 1 mG field (from a small amount of neutral on a phone line ground). Do you check out this type of source?
Karl

 

[brian john]

Re: Ground loops

Karl:

One of the first reasons given for equipment problems is grounding then harmonics (in my expierence). Depending on the type of problem, checking for net currents and utiliziing a gauss meter are some of the first things we usally check.

 

[roger]

Re: Ground loops

Karl, we don't ever get to go that far, we usually leave (we stay in the background for support as much as we're asked to) the hospital fighting the manufacturer who will eventually get it remedied on their own. These rooms are not treated as normal areas, when the medical equipment is installed we have to step back and the room belongs to the manufacturer from there on. We are shunned by the fragile egos of these people. Their attitude is "we design this equipment therefor we must be right in our electrical and grounding designs".

The irritating part of this is, in the past when I have asked for their engineers to meet with us and the building engineers, they send a salesman, who will sit through our meeting continuously saying, "I'll have to get back to you on that".

Roger

 

[don_resqcapt19]

Re: Ground loops

Brian,
Can you give some details on the types of grounding problems that you find that cause equipment to malfunction?
Don

 

[dereckbc]

Re: Ground loops

WOW, what a coincidence. I received a phone call this morning around 3:00 am informing me one of our telephone switch offices was struck by lightning, and the AC service was out and the switch was not processing calls. The switch serves all the cell sites in Eastern Oklahoma and Western Arkansas. More on that in a moment.

Karl, yes I do use a gauze meter to investigate noise problems. In a telephone office all the DC grounding conductors are run in open air and not paired with any associated power conductors. Since most of the ground systems are isolated single point, they act as antennas, and induced currents are a problem from time-to-time. I find the gauze meter useful if the induced current is from the AC power system. Occasionally some installer runs a ground conductor too close to a florescent light or motor or something and the gauze meter easily identifies the source of the induction. I do not find it useful for high frequency noise; use other test equipment for that.

Back to the lightning strike. I have spent the last 15 hours bringing one of our telephone offices back on line, whew! When I arrived I found the class C TVSS at the service entrance completely blown off the wall, generator transfer switch control board fried, transformer fuses blown, 48 volt rectifiers AC breakers tripped, UPS batteries dead, and worst of all the switch was not processing calls. The TVSS, generator transfer switch control board, transformer fuses, rectifier AC breakers, and UPS were easy to diagnose and fix. Returned service by 6:00 am.

Although the switch still had power from the battery plant it was not processing calls. Once we recovered the UPS we were not able to communicate with the switch via the terminals located throughout the building. The terminals are powered from the UPS on IGR.

After a few hours we replaced the I/O RS-232 cards that interface the terminals with the switch. I investigated the ground system on the switch to try to determine how lightning entered the switch. The switch uses a ground window and single point grounding to prevent any outside faults from flowing through the isolated ground plane. I did not find any violations with the external grounding. Then it occurred to me the RS-232 uses a ground wire as a reference for transmit and receive data. The light went off in my head. I looked at the cables and they were metallic, rather than optical, problem found.

I looked closely at the RS-232 cards that were replaced, and found the ground and power circuit board traces burnt into, problem confirmed. To enlighten you, the problem is the switch is served by DC power from a ground window where all grounds originate to form a single point isolated ground. Then the switch and all its components are kept completely isolated from any incidental contact with ground by using isolation techniques. The RS-232 cables serve multiple terminals and desktop computers located through out the facility. The terminals and desktops are powered from the UPS on IGR. This corrupted the isolated ground plane and introduced a GROUND LOOP, which allowed lightning current to flow through the isolated ground plane.

When I showed and explained the problem to operations personnel they were stunned. They assumed since they were using IGR they were safe, wrong, dead wrong. There are two solutions to the problem. 1 is to install and inverter in one of the switch frames and reference its output to the ground window, or 2 install optical isolators on the RS-232 cables. The optical isolators were ordered today!

 

[bennie]

Re: Ground loops

Dereck: Alexander Graham Bell would be proud of you.

Are you sure your name is spelled correctly and is not, Dereck Camp Bell.

Apparently you are back at work.

Good luck.

Bennie

[ July 22, 2003, 09:06 PM: Message edited by: bennie ]

 

[dereckbc]

Re: Ground loops

Thanks Bennie. Yes I found another job with another Telco. It is not the one I want, but it pays the bills untill I can find a better job. Hint, hint you A/E people.

 

[jrdsg]

Re: Ground loops

there sure is a lot of confusion out there on this subject.

we do a lot of communications and cctv. the ground loops we have the most trouble with consist of a communications cable or cctv cable [think rg59 or rg6] the shield of which is grounded at both ends. if there is a difference in potential between end a and end b then a current will flow down the shield of the cable interfering with the low-voltage signal on the center conductor. since the voltage of a cctv signal is only 1vPp, it doesn't take much of a ground loop to overwhelm it.

to complicate the matter further, many communications signals loop from one device to another such that the ground loop could be anywhere on the signal path from source to final termination. the troubleshooting can be painstaking.

isolation transformers are typically used to eliminate this problem, but they are very expensive.

[ July 23, 2003, 12:19 AM: Message edited by: jrdsg ]