GFCI to break ground loop

[rhollan]

Re: GFCI to break ground loop

hbiss quoted another as quoting me: 'The poster stated "the neutral is bonded to the chassis of the sub-woofer amp".

I seriously doubt that he is correct in that assumption.'

I was indeed wrong on this point. The chassis of the sub amp is connected to the grounding pin on the AC recepticle, and NOT the neutral pin. My bad.

hbiss: Your understanding of my setup as you described it is correct. I hadn't considered that the loop might be as you described it, but this makes sense. I will investigate tonight. Thanks to all!

 

[dereckbc]

Re: GFCI to break ground loop

Originally posted by hbiss:
The poster stated "the neutral is bonded to the chassis of the sub-woofer amp".

I seriously doubt that he is correct in that assumption.

...I would then use a surge reference equalizer device

You got me there. You are not talking about those isolated power systems that use a center tap on an isolation transformer 120v secondary to provide a ground are you? If not, I'm not familiar with what you are talking about.

You are referring to a balanced power system. Indeed it would work, very expensive, but that is not what I am talking about. There are many surge protection strips on the market. Some come with I/O ports for Coax and Telephone lines. They bond the EGC to the coax shield with a SPD connected from center pin to ground(some also come with Telco line protectors, common for PC applications). What it would do is make a SPG and reference all grounds to a single point. It does not matter how musch potential difference there is between the sat ground and power-amp ground.

But an IGR in the way you describe is not nessecary or code compliant. In a wood frame house all branch circuits are isolated. The surge refference equalizer provides a SPG for about $35-$90 price range. I have seen this many times, and it always works. It also offers some surge protection.

 

[hbiss]

Re: GFCI to break ground loop

...In a wood frame house all branch circuits are isolated.

Correct, but in this instance we want to separate the equipment ground for the receptacle box (and AC cable armor if used, etc) that lands on the neutral bar in the panel from the ground terminal on the receptacle which would then be referenced to the same point as the sat dish ground. I believe 250.46(D) would allow that. The other alternative is to move the sat dish ground over to the neutral bar. I don't think that is allowed or even a good idea. I agree with you that there should be a single point ground but how do you do it in this case when: a) the sat dish and distribution equipment is located away from the audio/video equipment and b)the sat ground needs to be bonded directly to the grounding electrode conductor?

I believe the surge suppression strips you are referring to merely bond all the shields coming into it to the green equipment ground on the receptacles and its line cord. It doesn't interrupt the ground so it can't break the ground loop and might also introduce additional paths for current to flow.

 

[rhollan]

Re: GFCI to break ground loop

hbiss wrote: "I agree with you that there should be a single point ground but how do you do it in this case when: a) the sat dish and distribution equipment is located away from the audio/video equipment and b)the sat ground needs to be bonded directly to the grounding electrode conductor?"

Well, in Ontario, Canada, this is a bit easier than it seams: code requires both the service panel and the antenna ground to be bonded INSIDE the house to a cold-water pipe entry. Yes, this is contrary to NEC (which requires an external antenna ground rod bonded to the grounding electrode system), and the rest of Canada. I think the idea is that if lightning strikes a roof-mounted satellite dish, the resulting fire and downstroke won't give a fig whether there is a ground rod a foot away on the outside of the building. Personally I side with the NEC on this one, but when in Ontario...

Now, as the service panel and satellite dish have to be bonded to a cold-water entry, that's exactly what I did for the satellite dish ground -- two inches away from the the service panel ground conductor that is bonded to the same cold-water entry pipe, hoping to avoid a ground loop with this "single ground point". Alas, twas not to be.

 

[dereckbc]

Re: GFCI to break ground loop

Hbiss, I think you meant to say 250.146(D) didn't you? I do not know how rhollan's electrical distribution is constructed, maybe he can clarify. But I suspect he has type NM cable in plastic boxes. IGR is not necessary with that construction technique. Even if it is a metal receptacle box, if type NM is used, still no problem if box is mounted on wood framework.

Your assumption of the surge arrestor is correct. It bonds all the grounds, and signal grounds together forming a pseudo SPG for the A/V equipment, not a true SPG at the power source. None the less the effect is the same, loops are broken after the bond. True there may be loops ahead of the surge arrestor, but by bonding all the EGC's and signal grounds together at a single point all voltages are equalized and in phase, therefore eliminating any potential differences.

Rhollan, brought up something very interesting by stating the sat receiver and service ground are made within two inches of each other. IMO that is as good as it can get. It leads me to believe his sub-woofer is powered on a daisy-chained branch circuit with something ahead of it injecting noise on the EGC circuit.

Rhollan do you have the expertise to temporarily run a bonding jumper from the sub-amp EGC to the sat receiver coax ground? If so, you might try it to see if it eliminates the problem. If it does the fix is easy and cheap.

And if the GFCI without a ground works, wired/labeled per code, economical, and your satisfied with it, go for it. I am just a train driver and tend to be curious and find the root cause.

[ January 06, 2004, 07:37 PM: Message edited by: dereckbc ]

 

[rhollan]

Re: GFCI to break ground loop

dereckbc wrote: "I do not know how rhollan's electrical distribution is constructed, maybe he can clarify. But I suspect he has type NM cable in plastic boxes... Even if it is a metal receptacle box, if type NM is used, still no problem if box is mounted on wood framework."

This is correct. NM to metal receptacles, grounded via the safety ground, mounted on wood wall studs.

dereckbc also wrote: "Rhollan, brought up something very interesting by stating the sat receiver and service ground are made within two inches of each other. IMO that is as good as it can get. It leads me to believe his sub-woofer is powered on a daisy-chained branch circuit with something ahead of it injecting noise on the EGC circuit."

I tried plugging the sub amp into other AC receptacles to no avail. I am starting to think that there is a current flowing on the service panel connection to ground being picked up. I suppose I should check the integrity of this connection -- not by breaking it, of course: that would be foolish without a bypass, but rather by measuring the resistance across it from the service panel to the same cold water pipe that serves as the whole-house ground. The distance is about 20 feet, and I can easily compensate for the resitance of my leads (well, I'd run a #10 from the pipe back to the panel, and measure there, with house power off, lest I poke something where I oughtn't to, difficult though that might be).

Still, suggested by dereckbc: "Rhollan do you have the expertise to temporarily run a bonding jumper from the sub-amp EGC to the sat receiver coax ground? If so, you might try it to see if it eliminates the problem. If it does the fix is easy and cheap."

Yes, I can run such a jumper. I have a bunch of #10 single conductor lying around. But, thinking along the lines of a resistive connection between the service panel neutral and house ground at the cold-water entry, I measureed the resistance between the pipe and the safety ground connection of a nearby AC outlet, and found it 0 Ohms for all intents and purposes. What would a jumper in parallel with that accomplish?

Finally: "And if the GFCI without a ground works, wired/labeled per code, economical, and your satisfied with it, go for it. I am just a train driver and tend to be curious and find the root cause."

Satisfied? Well, yes, I am satisfied that it is acceptably safe, but, as others have chastised me, it is not as safe as a proper safety ground, with the ground loop eliminated. I consider it a temporary solution. I too, would like to track this bugger down.

 

[dereckbc]

Re: GFCI to break ground loop

Originally posted by rhollan:
Still, suggested by dereckbc: "Rhollan do you have the expertise to temporarily run a bonding jumper from the sub-amp EGC to the sat receiver coax ground? If so, you might try it to see if it eliminates the problem. If it does the fix is easy and cheap."

Yes, I can run such a jumper. I have a bunch of #10 single conductor lying around. But, thinking along the lines of a resistive connection between the service panel neutral and house ground at the cold-water entry, I measures the resistance between the pipe and the safety ground connection of a nearby AC outlet, and found it 0 Ohms for all intents and purposes. What would a jumper in parallel with that accomplish?

Finally: "And if the GFCI without a ground works, wired/labeled per code, economical, and your satisfied with it, go for it. I am just a train driver and tend to be curious and find the root cause."

Satisfied? Well, yes, I am satisfied that it is acceptably safe, but, as others have chastised me, it is not as safe as a proper safety ground, with the ground loop eliminated. I consider it a temporary solution. I too, would like to track this bugger down.

It is just a simple test to see if using ground equalizer will work or not before throwing any money into it.

If you are satisfied with the GFCI solution, by all means do it. I did not read anybody beating you up for the idea, just pointing out that if you were to have a fault to the chassis, there is potential danger.

I share the same curiosity, of why, what, where? If you are sure the sat antenna ground, and service ground are properly bonded, try looking at the routing. Could be that one is routed over a lightning ballast. Check to see that there is no current in the MBJ. Could be any number of things. Just keep eliminating possibilities until you find it, or use some method already mentioned to break the loop. Good Luck.

 

[rhollan]

Re: GFCI to break ground loop

dereckbc: I too am curious. While no one "beat me up" over the issue of using a GFCI, I do acknowledge the dangers of a hot short to chassis: (1) shock hazard if there is no path to neutral (which there is via the signal shields); (2) the fire hazard of returning a fault current via the signal shields if the GFCI does not trip. I consider the risks minimal, and acceptable, but do note that a GFCI is no guarantee that someone won't be electrocuted, just a mitigation of the risk.

Still, two things would have to go wrong to have a bad outcome: a fault condition, and either failure of the GFCI to trip or electrocution of a particularly sensitive individual. Given that GFCIs are acceptable per NEC and CEC in situations where safety grounds are not available, I again conclude that the risk is acceptable. I also drive a car and cross streets.

Now, I measured the AC voltage between the signal shield at the sub and the AC receptacle safety ground, and low and behold, saw what appeared to be about 50 mVAC. Of course, full scale on my multimeter was 5 VAC, and the needle movement (yes, it's an old analog meter) was barely half the smallest division, but the needle DID budge. I made sure to not touch the probe tips lest the stray field on my body was being picked up. The little mirror on the face of the meter to ensure one is reading it squarely was handy .

I did note some oxidation of the copper at the point where the service entrance is grounded to the cold water pipe entrance so perhaps this needs to be rebonded (though I measured no appreciable resistance at this point between the ground wire and pipe).

I did notice that if I grounded my meter where the service is grounded to the cold water entry pipe, and touched the other meter lead, I'd pick up several hunderd mV AC. This didn't surprise me: there are AC lines all over the place through the first story floor joists (I was in the basement). But, it makes me wonder if I'm seeing 60 Hz AC induced on the loop I've got from ground to sat entry to sat rx to tv to receiver to sub amp to safety ground to service panel to ground, currently broken by way of the ungrounded GFCI.

Then again, the house neutral and safety ground are essentially in parallel between the service entry and power transformer on the pole (assuming a really good earth ground). It is unlikely that both phases of the 240 VAC entry carry the same current, so there will be some current in the neutral (and safety ground) return from the service entrance to the utility transformer. I'm told a few amps on the safety ground is not unusual.

So, we have 20 feet of #6 safety ground from the service entrance to the cold water pipe entrance, and effectively accross this, the safety ground to the sub AC recepticle, signal lead shield, and a 10 foot run of #10 from the satellite entry. IIRC, 20 feet of #6 has 7.94 milliOhms resistance, and to sustain a voltage across it of 50 mV requires a current of 6.30 Amps. Doesn't strike me that it's out of the realm of possibility. Again, assuming the same resistance in the earth ground and neutral path from the service entrance to utility transformer, that's a total imbalance of 12.6 Amps on a 200A service. Gee, someone plugged in an electric kettle on a 110 VAC line, or turned on the microwave. Sounds plausable to me,

Bonding the satellite antenna entry at the service entrance instead of the cold water pipe entry would eliminate this problem, but would be less safe and not code complient.

If my analysis is correct I will either have to stick with the GFCI solution, or use an isolation transformer on the sub power supply (good to 250W), or audio isolation transformers on the signal lines.

 

[hurk27]

Re: GFCI to break ground loop

rhollan
Did you get a chance to check all the receptacles on this circuit for a neutral/ground connection. The other thing that I would check for If you have a amp probe amp meter check the ground at the panel to see if someone has used it for a neutral in a hidden place if there is any current on this ground then there is something that pulling neutral return power through it. try disconnecting the ground at the panel and see if there is voltage between it and the ground bar. this will also let you identify what it is when it don't work any more but be careful as this might cause voltage to appear on all the grounding of this circuit. another thing is shut off this breaker and check to see if the neutral still has current on it. if so someone has boot legged a neutral off this circuit then just turn off each breaker until it is gone then find the neutral connection and connect it to the correct neutral. Any one of these can cause a ground loop.

 

[rhollan]

Re: GFCI to break ground loop

hurk27 wrote: "Did you get a chance to check all the receptacles on this circuit for a neutral/ground connection."

Yup. The all show ground bonded to neutral, at the service entrance panel, I'd expect.

Also: "The other thing that I would check for If you have a amp probe amp meter check the ground at the panel to see if someone has used it for a neutral in a hidden place if there is any current on this ground then there is something that pulling neutral return power through it."

I'd have to stick a conventional amp meter in series with panel and service entrance ground, and I am not eager to lift the ground at the panel as this would likely raise the neutral voltage (with respect to earth ground) if any current was flowing through ground. I'd likely let an electician do this: while I'm confortable with what I'd be doing, I prefer to let sleeping ground bonds lie, as it were, in fear of not remaking them properly (i.e tightly enough).

But, I still don't follow your reasoning: off course there will be some current through the service entrance ground: the 240V, center-tapped (and grounded) entrance is never perfectly balanced with regard to the current in both 120V legs, and the neutral return and earth ground path from the service entrance panel to the utility transformer are in parallel. Ergo, there will be some relatively small current (less than ten amps) between the sercice panel and earth ground. Given that I have a 20 foot run of #6 providing this ground, my calculations show the 50 mV I measured effectively across this would be impressed by a 6 amp or so current through it. I see no way of avoiding this small imbalance: I think it's that damn 20 foot run of #6 ground wire (some 7 milliOhms) from the service panel to cold water pipe entry that's the culprit. If the panel were grounded to a cold-water pipe closer, or with a hevier guage, the problem would be less severe. Ontario electrical codes... go figure... apparantly such a long run from panel to earth ground is O.K.

Now, if you mean that there might be something on the SAME AC receptacle circuit that is impressing a neutral return current on the shared safety ground back to the panel, that is another possibility: I am definately seeing 50 mV AC coming from SOMEWHARE. Worth checking.

And: "try disconnecting the ground at the panel and see if there is voltage between it and the ground bar."

Let's be clear: temporarily disconnect the circuit safety ground to the AC receptacle powering the sub, or ground bonding for the entire panel? Either could carry a neutral return causing the problem (and, given the long run of the service entrance grounding wire, I suspect the latter).

Continuing:"this will also let you identify what it is when it don't work any more but be careful as this might cause voltage to appear on all the grounding of this circuit."

O.K. It looks as if you're talking about the individual circuit ground and not service entrance ground (I STILL suspect a small current imbalance in the latter). That I'm comfortable temporarily disconnecting.

More: "another thing is shut off this breaker and check to see if the neutral still has current on it. if so someone has boot legged a neutral off this circuit then just turn off each breaker until it is gone then find the neutral connection and connect it to the correct neutral. Any one of these can cause a ground loop."

Neutral or safety ground on the affected circuit? Though either would be a wiring error, I can't see a bootlegged neutral (only a bootlegged ground) causing this loop.

You've given me some good ideas, hurk 27. There definately appear to be 50 mV AC of ground noise SOMEWHERE in this circuit. I just have to found out where: anything that returns current through the safety ground could be the culprit. My bet is on the minor (is 6 amps minor?) service ground imbalance and the 20 foot run of #6 from the service panel entrance to the cold-water pipe entry. A way to test this, albeit temporarily as it breaks code, is to lift the satellite entry ground from the cold-water entry, and make it at the service panel ground, or nearby safety ground on an AC recepticle (though that could have ground current of it's own if miswired). If the loop is gone, it has to be in the service ground, and if not, it comes from elsewhere. That strikes me as a much safer test than lifting the service entrance ground for any reason.

 

[dereckbc]

Re: GFCI to break ground loop

Originally posted by rhollan:
Now, I measured the AC voltage between the signal shield at the sub and the AC receptacle safety ground, and low and behold, saw what appeared to be about 50 mVAC.

This is the source of your noise. It is called "common mode" noise. The likely source of the noise is the different ground points of your sat receiver coax, and your service ground. Providing everything is up to code and operating properly there are a few solutions. The two best that come to my mind are:

1. Source your sat reviewer ground and service ground to the same point.

2. Use a surge arrestor with I/O ports for the coax. Again if you short the EGC of the sub-amp to the sat coax you get the same effect to see if it works before investing money in a SPD.

Of course you could still use GFCI to break the loop. But I see this as treating a symptom rather than solving the problem.

Wayne (aka hurk27) also gave you some good advice to pin point the problem, but it does not sound like you are confident to perform the procedures. It really does not matter if you do have some current on the EGC, it is normal. If the is current on the GE system and you source your various ground at different point you will have some slight differences of potential that will cause "common mode" noise. Thus the theory of single point ground to eliminate the problem. That's another story. Good Luck

 

[rhollan]

Re: GFCI to break ground loop

dereckbc wrote: "This is the source of your noise. It is called "common mode" noise. The likely source of the noise is the different ground points of your sat receiver coax, and your service ground."

Yup, yup! Sadly, the arrangement where the service panel neutral is bonded to ground, and the service panel grounding conductor is a 20 foot run of #6 to the cold water entry pipe is up to code in Ontario (at least it was when the house was built, and that part of the code is not retroactive). Furthermore, the satellite ground has to be made to the same cold water entry (per code) and NOT way back at the service entrance (service entrance: North side of home, satellite entry: South side of home, cold water entry: middle of east side of home). I'm certain that 20 feet of #6 separating the service panel ground from the satellite ground is the culprit.

Also: "Providing everything is up to code and operating properly there are a few solutions. The two best that come to my mind are:

1. Source your sat reviewer ground and service ground to the same point."

Without moving the service entrance, this is not possible and would violate code.

And: "2. Use a surge arrestor with I/O ports for the coax. Again if you short the EGC of the sub-amp to the sat coax you get the same effect to see if it works before investing money in a SPD."

Ah! Now I follow the idea!! Short out the 50 mV AC ground current between the sub safety ground and signal shield. I have a satellite-compatible surge suppression power strip for the A/V equipment but never bothered to daisy-chain the satellite signal coax through it: I just figured it placed a tranzorb across the coax and I already have surge suppressors on the satellite coax entry to thw house. I hadn't realized it bonded the coax shields to safety ground, shorting any would-be ground current in the process. The light bulb has gone off and I feel like a dufus.

Continuing: "Of course you could still use GFCI to break the loop. But I see this as treating a symptom rather than solving the problem."

Yes, and until now I thought the problem not solvable. I will try daisy-chaining the satellite feed through the surge-supression power strip designed for satellite feed-through, and if it solves the problem, the GFCI extention goes into my "hack box" for another day when I need to temporarily safely lift a SPG.

And: "Wayne (aka hurk27) also gave you some good advice to pin point the problem, but it does not sound like you are confident to perform the procedures."

Confident enough: yes. Wanting to avoid it: yes. As a matter of course, I avoid mucking with service entrance wiring if there are other ways to perform the experiment necessary to gather more data to solve a problem.

Finally: "Thus the theory of single point ground to eliminate the problem. That's another story. Good Luck."

Sigh. Easier said than done. I should have realized the value of the surge suppression strip satellite coax feed-thorough earlier, but the expression "surge supression" lacking the words "and grounding", made me oblivious to the obvious (such devices rarely come with schematics) and not think to check that they bond SPG to coax feed-throughs. Oh well, live and learn. Thanks to all.

 

[hurk27]

Re: GFCI to break ground loop

rhollan
The ground's that I was talking about is the grounds feeding to the circuit in question. not the main service grounded conductor (neutral) as that would cause problems. and the amps meter I was talking about does not require you to disconnect any wire it clamps around the wire. Not like a VOM meter which only reads DC amps and has to be placed in series with the circuit. and lifting of the main grounding electrode to the water pipe should not cause any problem as the service should not be drawing power from the water pipe. If it is then you have a lost neutral from your service to the transformer. If there is current on the ground wire from the branch circuit feeding the audio equipment it will cause this hum. and lifting it should not cause any problems as there should not be any current on it.

 

[rhollan]

Re: GFCI to break ground loop

hurk27: Yes, I know about clamp-on AC current meters. I do not own one (though, this problem is probably a good reason to get one), and I am loath to break the whole house ground connection just to make a current measurement.

As for the neutral return to the utility transformer being broken: I seriously doubt it. Rather, the safety ground return and neutral are in parallel, and if the safety ground return via the cold-water pipe is really good, the neutral imbalance current will be split between those two paths rather equally.

The other possiblity, as you say, is that something is returning current on the daisy-chained safety ground to the AC outlet receptacle that powers the subwoofer amp. However, I have tried several other AC outlet receptacles with no difference.

You know, this bit about using cold-water pipes as safety grounds, bonded to neutral at the service entrance panel AND the utility transformer has me thinking... In some places (most, I suppose, these days), AC service is underground, likely sharing water and other facilities. What if the utility transformer safety ground bond to neutral at the transformer is, effectively, a cold-water main? Gee, then the resistance in the neutral imbalance return and safety ground return would be similar, and any neutral imbalance current would quite likely be equally split. I don't think that would be intended.

 

[hbiss]

Re: GFCI to break ground loop

What if the utility transformer safety ground bond to neutral at the transformer is, effectively, a cold-water main? Gee, then the resistance in the neutral imbalance return and safety ground return would be similar, and any neutral imbalance current would quite likely be equally split.

Well you almost have it. Consider a metallic municipal water system. Each house has it's neutral bonded to it. This water piping is then in parallel with the neutral conductor from the transformer that runs along and supplies each house. If you were to draw it schematically it would look like a ladder, one rail would be the POCO neutral, the other rail would be the water system piping and the rungs would be the individual service neutrals connecting the two.

You can see then that there normally would be some current flowing through an individual grounding electrode conductor since it would be impossible for all the loads on that network to be perfectly balanced. This current wouldn't even be dependent on your load since some current would be present even if you shut your service down at the main breaker or you were perfectly balanced.

Is this intentional? Yes! This acts as a backup should a service or distribution neutral fail.

 

[hurk27]

Re: GFCI to break ground loop

In a normal system even with current on the grounding electrode (water pipe) unless there is a lost neutral at the service or on the pole. this would not cause current between the sat. and amp. the only way for that to happen is a neutral/ground bond in this circuit or another circuit boot legging a neutral from this circuit. or this circuit boot legging a neutral off the ground. To have current on the Audio sheild it would have to be parallel with a conductor carring current, and this would be a grounding conductor as this is what is common to it. This is why I said you need to check the grounds for this circuit to see if there is current on it. What I'm thinking is there is a bootlegg neutral pulling power from the EGC in the circuit feeding the sub amp this will also place current on the shield of the audio cable because of the paralleled connection with the ground at the water pipe for the sat. and the service water ground. this is what has to be looked for. I bet a dounut that if you lift the EGC for the sub amp circuit in the panel you will se a spark from a load. this will confirm that in deed there is somthing useing the ground conductor for a neutral return. If this circuit is part of a multi-wire circuit (shared neutral) don't do it as it will float the neutral but maybe at some time in the past the neutral was lost and someone just connected it to ground to reestablish the return to the panel through the EGC.