Ground Loops

[Gun Geek]

Need some help techie electrical types:

Subject is AC power distribution wiring, problem is ground loops causing a mild shock to people climbing in and out of a lake onto a floating dock. This dock has existed for 3 years.

Scenario:

House at the top of the hill - standard modern US house wiring - grounded with a ground rod driven into the earth close to the foundation (water pipes are PVC). Triplex wiring (2 hot and 1 neutral) taken from breakers in the house box running down the hill to a dock floating on a lake. Dock is steel frame construction, floating on styrofoam - it does not touch the bottom. Some of the framing, and an aluminium ladder (bolted to the steel) protrude down into the water. Neutral wire is connected to the frame of the dock. Other end of neutral is connected to the grounding point. If one of the hot wires gets loose and touches the dock, it will short, tripping a breaker. This does not allow the dock to become electified so that a hapless sunbather could get between the frame and ground - getting electrocuted.

When a person swimming in the water and in an upright position (more like treading water) approaches the ladder (or a piece of steel that goes into the water), a mild shock is felt (sensation kinda like putting a 9v battery on your tounge). If horizontal, little or no shock is felt.

Trouble shooting actions:

1) Disconnected the "hot" wires from the dock (opened the breaker) - no change in the shock

2) Disconnected the "hot" and the "neutral" - shock went away.

3) Measured between the ladder rail in the water 500mv DC and 1.5vAC with a cheap digital VM.

First action:

The 3 wire set-up is not exactly code. It is safe as it is, but it has the potential to cause problems if the neutral wire breaks loose. Code is that a separate ground wire should be run, and the neutral not touch any metal (in this case on the dock). The separate ground wire should connect to the frame, and the ground lugs in the outlets.

I ran a separate ground wire and made sure the neutral does not touch the frame of the dock.

Result: No change in the shock!

Second action:

Connected a copper wire and a piece of copper sheet metal 18" square from the frame of the dock and sunk it about 10' into the lake.

Result: shock was less, but not gone.

3rd action:

Isolated the ladder from the steel. The ladder is also bolted to some wood support members, I took the bolt that were through the steel and into the ladder out.

Result: no shock on the ladder, same shock around other steel members.

My diagnosis is that the lake and the grounding point up the hill are at different potentials and running a "ground" wire down the hill & connecting to the frame of the dock (to make the regular AC power safe) brings those 2 potentials into close proximity. The swimmer completes the circuit and thus gets shocked.

This year has been very dry here (grass is brown, dirt is hard and dusty when it is usually moist and soft). I think the dry soil makes the "ground" point at a higher potential (the soil is an insulator) than the lake.

We tried wetting the ground (trickle of water for days) around the grounding point.

Questions:

1) Diagnosis make sense? I think this is a classic ground loop that plagues aircraft refeuling, etc.

2) Is this dangerous? I know your immediate reaction is YES, but I'm not so sure.

3) What else can I do? Drive more ground rods? Wait till it rains again?

This is the "Questions That 'Don't Fit Anywhere'" forum

Thanks!

[George]

DO NOT attach any wires from the house power to the dock metal parts in any way, shape, or form unless they are a Ground connection wire. Neutral can/does carry current in a situation like this.

Run four wires to the dock.

1 Ground from the ground tie point at your service box

2 Neutral from the Neutral buss in service box (neutral must not be bonded to ground)

3 Hot #1

4 Hot #2

Wire them only to the appropriate lugs on the grounded receptacles.

Use outdoor rated plastic electrical boxes if possible. Clip the tab that connects the ground lug of the receptacle to the box. This prevents the box (if metal), or the screw faces from becoming part of the return path. This will allow any ground fault within a device to dump it's load into the ground plane back at the service box ground tie point which is a much better practice. Treat the whole thing like one long extension cord and do not let there be any connection of the neutral wire to any metal part of the dock whatsoever.

--

Regards,

[shred]

Although it's called 'Neutral', it carries current and thus is a bad thing to be hanging onto at any time.

[EricW]

What Geo said.

The "ground" that is the water, and the ground rod at the house are at two different potentials. You absolutely should NOT be connecting the neutral wire to anything besides the outlet receptacle. You also should NOT be trying to ground the dock solely to the house ground. Do not connect any wires from the house electricity to the dock or ladder - as you will simply elevate the potential of the dock.

You will need to acheive shock protection from the use of GFCI breakers. Personally, I'd cough up the ching and double dip: GFCI outlet and GFCI breaker. Test the breaker(s) by plugging a hand held shop lite into the dock outlet and literally dumping it into the water. Hopefully there will still be enough voltage drop to trip the breaker. If it does, you're probably good to go. If not...

P.S.

It's dangerous as hell to let the situation continue. It only takes MILLIamps to kill. I think it's on the order of 30 to 60 millamps for a fatal shock IIRC. That's .030 to .060 Amps from a circut that's designed to carry 15 to 50 amps.

Ground loops are super-dangerous.

P.P.S.

Ground loops are a different phenomena than grounding an airplane prior to refueling. Airplanes need to be grounded because they acquire a charge flying through the air. That's why they put static wicks on the trailing edges of the wings - to help bleed off that charge. Because airplanes are totally insulated from the ground by the rubber tires, they are essentially just a big capacitor sitting there.

[Gun Geek]

You will need to acheive shock protection from the use of GFCI breakers.  Personally, I'd cough up the ching and double dip: GFCI outlet and GFCI breaker.  Test the breaker(s) by plugging a hand held shop lite into the dock outlet and literally dumping it into the water.  Hopefully there will still be enough voltage drop to trip the breaker.  If it does, you're probably good to go.  If not...

<{POST_SNAPBACK}>

Did exactly this this originally - worked great

It's dangerous as hell to let the situation continue.  It only takes MILLIamps to kill.  I think it's on the order of 30 to 60 millamps for a fatal shock IIRC.  That's .030 to .060 Amps from a circut that's designed to carry 15 to 50 amps.

Ground loops are super-dangerous. 

<{POST_SNAPBACK}>

Disconnected things after George's post last night - working the re-engineering plan today.

[JohnRodriguez]

The subject of grounding for your dock can found in the NEC 2005, article 553 "floating buildings".

The whole point of grounding is to get/make everything at a single potential. You need to change your cable out to a 4 wire setup (2 hots, 1 neutral, 1 ground) set up a j-box on the edge of the lake/pond and run the power cable into it, bond the ground to it and drive a ground rod on the edge of the lake/pond, then complete the run to the dock. bond all the metal parts on the dock to ground conductor. Eric's comment about not bonding the dock to ground is 100% wrong (sorry about that, but its true)

You probably have voltage on you neutral, and without a ground to check it to you will never know. Which could be a problem upstream of the power source. The other source could be induction on the neutral, which would be the same case above, especially on a long power run.

out here where i work (copper smelter, power house, I'm a industrial electrican) we have a set of 120/240 signal cable that are about 8 feet below a 13,800 volt run of triplex power cables. they induce about 90-100 volts dc onto the signal cables, enought to get you attention if your not careful, so your 1.5 Vdc should be about right.

[cking]

The confusing point with wiring is that the grounded and grounding are two different wires. I can't remember which one refers to witch. But the White wire is the neutral or return. The Green wire is the ground. The comment about a 4 wire setup is correct. If you don't want to bury a fourth wire. Put a ground stake at the dock and tie all green wires, outlets third prong to it. Do you need 220 volts at the dock????? If not use the three wires to provide, hot, neutral and ground. The way it setup now the GFI will probably trip all the time.

IT is true that the neutral and ground are tied together at some point, don't wire things that way.

[EricW]

John,

I mis-read his post. Sorry.

I thought he had already driven a ground rod at the dock and was still getting shocked because of the ground conductivity (or lack therof) problem. Yes, *if* you can get a solid earth ground near the dock, that will do the trick. I've seen instances where doing that was extraordinarily difficult to impossible.

My other concern would be the changing potential between the dock, it's local ground, and the water with changing seasons and water levels. If you ground near the dock, make it a deep ground so that's it's always touching wet soil.

Anyway, the moral in this is to not use neutrals as grounds. Neutrals are neutral to the accompanying hot wires. There's no guarantee they're neutral to anything else.

[JohnRodriguez]

I just reread GunGeeks post again also and saw that he had the neutral tied to the frame. Big no no. What's happening is with the neutral tied to the frame, voltage and current are trying ot get to ground or zero potential. With electricity trying to get to a zero potential, it doesn't care if it goes tru metal, water or you, it will get there. With the neutral tied to the frame it is taking the quickest path to ground and a wet person ( who's resistance is about 750ohms, pure water doesn't conduct, dirty water with minerals in it is about 5000 ohms) is feeling the current go tru them when they touch the frame. what is saving the people from getting killed is they are at equopotential with the water and frame, kind of like a power line and bird.

something on a similar note, just for info. Electricity travels in a sin wave, whether its 120, 240 480, 4160, 115kv, 325kv, or a lightning bolt. and this is in a wire or on the ground you walk on. Because of this you have what is called step potential on the peak to peak waves that travel tru the ground. If you are ever driving along and see a downed power line, stay in your vehicle especially if its a high voltage transmission line, just the act of walking could get you. There have been cases where a lightning bolt hits by a herd of cows and the cows that are standing horizontal to the bolt will survive and the one who are perpendicular to the strike will be killed (that's step potential, very dangerous) About 6 years ago we had a grounding class given by the fedaral electrical investigator out of Boulder CO ( he only investicated the electrical fatalites) and that was a eye opening class on the importance of grounding. The one tip he gave us is in a lightning storm, don't lie on the ground (your head and feet could be at a different potentials) just crouch down or stand with your feet together.

[George]

just crouch down or stand with your feet together.

That is a darned good example of why single point ground systems are preferred ;-)

My feeling on the situation he has going here is that he really should treat this as a long extension cord with a quad box at the end and not try to ground locally. Unless the box installation is done by a professional, there are too many things that can be done wrong in establishing a local ground plane.

I do a lot of temporary power installs for event staging and I am a firm believer in bringing the ground along with the power from the service tie point and NOT trying to establish a new earth/cold water pipe ground at the remote distribution point. This eliminates the possibility of multiple ground paths of varying resistance and is generally the most reliable way of having a fault return path at the remote distribution point.

Using GFCI outlets is probably the wisest choice here also. Good call Eric.

--

Regards,

[Flexmoney]

I gotta ask...why do you need/want power at the swimming dock anyway?

[EricW]

BTW, I studied power distribution as part of my EE degree curriculum, which didn't cover squat about *practical* issues of electricity (because most profs have never worked a real job in their lives). My experience in industrial instrumentation makes me tend to agree with George about the virtues of single point grounding...to a point. If you can design in a ground net so that you can tie back to a single point, that's the best possible situation. Some times you don't have that luxury.

In this case, I tend to side with John because of the metal dock. If it was a wooden dock, I'd feel differently. The problem is that a short to the dock, without being on a common ground to the house electricity could possibly leave the dock charged. You could ground the dock separately from the house, but you still would be left with the difference in grounding potentials as a risk. If you *cannot* get a solid ground at the dock, then you're back to depending on GFCI for "protection".

No matter what you do, the dock needs to have the crap grounded out of it at the shore.

As an aside, I tested a GFCI outlet once, and it still takes a hell of a zap to trip the things over. I know there's a test button, but I wanted to see what happened externally - and it ain't pretty. It's worth the scorched outlet to demonstrate to yourself the importance of proper grounding to prevent people (and pets) from becoming a return path.

[George]

If the ground conductor brought from the house tie point is at least a gauge, or so larger than the hot wires in use, it can safely be used to ground the dock and all it's metal parts. This will keep the ground system single point and eliminate any possibility of a resistive ground fault path that can happen with a local earth ground at the dock.

As far as making the ground connection to the dock, make sure the connection is made to each major part of the dock separately as corrosion and paint can electrically insulate mechanical connections to the various pieces of the dock if it isn't one solid piece of metal.

[JohnRodriguez]

hey Eric

When you tested the GFI did you take a piece of wire and go from the hot to neutral? If you did then all you did is show the system a low resistance load, and the more distance you have from the source, the harder it is to trip the breaker. The best way to test a GFI is to take a wiggy and go from ground to the hot leg. A GFI moniters the current tru a CT on the ground wire and the trip range for one is from 3-5mA, you can get a wet location GFI that will trip at 12mA but those are for the little submersible bilge pumps that will trip constantly with a regular GFI ( there's going to be a lot of people scratching there heads over this conversation).

The ground net you mentioned are called ground grids and we have to test them every year for OSHA and MSHA, but they are only installed on substations. The main reason they are installed is for fault currents and gives them a good path to ground when they do occur.

As far the dock goes, I would bond the metal with a 4/0 bare ground wire and run it in the water to the shore, just in case any equipment fell in the water especially the type that has no ground in the plug (double insulated) fell in the water.

[Gun Geek]

I knew this was an interesting topic.

Gotta underline a couple of points:

1) The dock was originally wired with the neutral tied to the frame. That neutral has a direct path back to the ground rod up the hill (this is the same point where the service is tied from the power co). This neutral is the exposed conductor in the triplex. I inherited this situation

2) The dock now has a 4th wire added. It is also tied directly to the ground point up the hill. This wire is an isulated #6 solid copper. Now the netral wire is not tied to the metal frame of the dock, but the ground wire is. This was done for exactly the reason Eric identified, the risk is that a loose hot could electrify the dock with 120 AC, which is not good. Now a short will blow a breaker - dramatic, but safe.

"In this case, I tend to side with John because of the metal dock. If it was a wooden dock, I'd feel differently. The problem is that a short to the dock, without being on a common ground to the house electricity could possibly leave the dock charged."

3) With all power removed (breakers open on the phases) you still get a shock. The only path is to the ground rod up the hill. The shock is from the difference in potential of the earth ground up the hill and the potential of the water. You do not get a shock if you disconnect both the neutral and the ground wire (did this only with the phases off) Re-connect the either the neutral or ground and the shock comes back. My theory is that the potential exists because the soil is dry and therefore the ground at the top of the hill is at a higher potential than the water. This indicates that one solution may be to drive more/deeper rods at the top of the hill. But more grounding at the top of the hill may not work either - See #4

4) Driving a ground rod at the dock is not possible. The top soil is maybe 3' - 4' at the top of the hill and zero at the dock. The geology is that of a limestone canyon. Here in Central KY, limestone lurks below everything (this is the land of Mammoth Cave and Horse Cave, and the limestone is why we have good Bourbon). This lake is a dammed up (big) ditch created by eons of erosion. Officially this formation is the Kentucky River Palisades, and this lake is a dammed up tributary of the KY River, called the Dix River. The water eroded away the topsoil and the limestone forming the ditch. The shore is limestone - there no soil into which one can drive a ground rod. Also the yearly water level fluctuation in this lake is on the order of 30 - 50 feet. It is a has been a hydro lake, but now it is primarily flood control (plant was converted to coal fired steam and coal gas fired turbine generation. The power generated by the hydro isn't enough to light the other boilers!, but I digress).

5) Flex - this is more than a swimming dock (I bet your image is a little floating affair about 10' x 10'). Not so. THIS is a party dock. 45' x 45', 3 boat slips, refrigerator, freezer, microwave, pizza oven, wet bar (cold water hose down the hill). One goes down the hill in the morning and does not go back up the hill until bed time. . There are 2 phases down the hill because of the current load - we have 2 circuits each carrying lower current so we don't trip breakers. Maybe some pics will be forthcoming.

6) All outlets are GFCI

Eric, I also have an EE degree, and, just as you said, we didn't cover this kinda real world stuff either.

I don't like the idea of not tying the dock to a ground - More research, and keep up the discussion.

BTW I have spoken to 3 electricians, 2 say ground the dock, one says "float it" just as George descibes. I get a blank stare when I say the ground is what causes the shock.

[EricW]

John,

I tested the GFCI ground to hot. It worked, but I'm glad I wasn't the conductor.

During a college field trip, we toured a substation and they mentioned the words "ground" and "net" which is my entire knowledge of ground nets as they relate to the electrical power industry. ($40K shot to hell ) What I know from ground nets is from old factories and old computer rooms. Totally archaic now, but back in the days of yesteryear, that's how they managed to equalize everything to get good, common ground references for the signals.

===================================================

GG,

I think your only real option is to disconnect the dock from the house ground then "ground" it to the H20 via multiple (stainless?) probes into the water. If the water's acting as a 5K ohm resistor per contact point, you'll want a lot of contact points to cut that down so that it's less than a person. 10 probes ought to get you down to 500 ohms (theoretically). Might want a few more for good measure.

You will just have to make sure that nothing you plug into the outlets externalizes the outlet ground, so that the ground path doesn't get closed again. Less that perfect, but I think as close to "safe" as you can get short of resorting to extraordinary means.

Leaving the dock totally ungrounded as one EE said seems like a recipe for disaster. Hot wire gets into contact and it will simply turn into a 120/240V electric fence for some hapless swimmer/boater to fry themselves on.

[Gun Geek]

John,

GG,

I think your only real option is to disconnect the dock from the house ground then "ground" it to the H20 via multiple (stainless?) probes into the water.  If the water's acting as a 5K ohm resistor per contact point, you'll want a lot of contact points to cut that down so that it's less than a person.  10 probes ought to get you down to 500 ohms (theoretically).  Might want a few more for good measure.

You will just have to make sure that nothing you plug into the outlets externalizes the outlet ground, so that the ground path doesn't get closed again.  Less that perfect, but I think as close to "safe" as you can get short of resorting to extraordinary means.

Leaving the dock totally ungrounded as one EE said seems like a recipe for disaster.  Hot wire gets into contact and it will simply turn into a 120/240V electric fence for some hapless swimmer/boater to fry themselves on.

<{POST_SNAPBACK}>

Eric,

We're on the same lambda. I was thinking some copper sheets suspended about 15' below the dock (water is 60' off the front of the dock, 25' off the back - closest to the shore).

A more "orthodox" solution would be better for the general observers, but there may not be a good orthodox solution...

[EricW]

GG,

What about corrosion of the copper plates over time? Once they turn green, they'll pretty well be insulators won't they? Or not?

I've never tested the conductivity of corroded copper in H20, so I'm totally yakking out my behind, but I sure have had to clean a lot of corroded terminals before.

And I have to say this is a *really* interesting problem. Ought to throw it in the face of my former profs and watch them squirm for a while.

[lefty45]

Have the utility co check for an open neutral on the system--Have seen a similar problem with a swimming pool. the feed to the house was bad. also a check would be to amp clamp the grounding electrode conductor going to the rod- any current could indicate a open neutral as well

[ChuckS]

Let me ask a question or two : Why do you thing the shock is due to your dock? Is there some other possible source in the lake for the current? You stated

"You do not get a shock if you disconnect both the neutral and the ground wire (did this only with the phases off) Re-connect the either the neutral or ground and the shock comes back."

This tells me that the lake is charged somehow and "discharges" into your ground line through the dock (and the swimmer). If the dock is floating electrically, it will eventually reach equalibrium with the charge of the lake so when the swimmer in the lake touches the dock, they feel no difference in charge or shock. If the dock is grounded, there is some current flowing through the lake/dock interface which probably has quite a bit of impedance. The shock comes from the swimmer shorting out the impedance between the dock and the lake with their body.

Just some late night thoughts...

[Gun Geek]

GG,

What about corrosion of the copper plates over time?  Once they turn green, they'll pretty well be insulators won't they?  Or not?

I've never tested the conductivity of corroded copper in H20, so I'm totally yakking out my behind, but I sure have had to clean a lot of corroded terminals before. 

And I have to say this is a *really* interesting problem.  Ought to throw it in the face of my former profs and watch them squirm for a while.

<{POST_SNAPBACK}>

Oh yeah, the copper will corrode (though since there is less oxygen it won't be as fast as in the air). Stainless probably won't corrode, but anything & everything will get algae growth which will make it less conductive. Gold leaf anyone?

I agree it is an interesting problem. I just wish it was somone else's!

[Gun Geek]

Let me ask a question or two : Why do you thing the shock is due to your dock? Is there some other possible source in the lake for the current? You stated

"You do not get a shock if you disconnect both the neutral and the ground wire (did this only with the phases off) Re-connect the either the neutral or ground and the shock comes back."

This tells me that the lake is charged somehow and "discharges" into your ground line through the dock (and the swimmer). If the dock is floating electrically, it will eventually reach equalibrium with the charge of the lake so when the swimmer in the lake touches the dock, they feel no difference in charge or shock.  If the dock is grounded, there is some current flowing through the lake/dock interface which probably has quite a bit of impedance.  The shock comes from the swimmer shorting out the impedance between the dock and the lake with their body.

Just some late night thoughts...

<{POST_SNAPBACK}>

You're exactly right - there is a discharge from the lake into the dock (or the other way around). That discharge is because the wire connected to the ground rod up the hill and the lake are at different potentials. If you disconnect the wire from the ground rod ("ground" wire and neutral) then the doc can reach steady state and, since the dock is floating electrically, it will reach the same charge as the lake, and there will be no shock.

The problem with having the dock float (not grounded) is that it will be dangerous if a hot wire touches it. The frame of the dock will be at 110v wrt the lake. A swimmer touching the ladder/dock will be the primary discharge path, and ZAP.

This is a well known and well documented problem around marinas (and one of the reasons you're not supposed to swim around them). Big aluminium hulled boat has some snafu with wiring - hot lead touches the aluminium, and doesn't trip a breaker cause the hull isn't grounded. Hull sits at 110. Swimmer gets close to that and goes into cardiac arrest. Death is ruled a drowning!

[Gun Geek]

Have the utility co check for an open neutral on the system--Have seen a similar problem with a swimming pool. the feed to the house was bad. also a check would be to amp clamp the grounding electrode conductor going to the rod- any current could indicate a open neutral as well

<{POST_SNAPBACK}>

Something like this will be my next step. I'm going ot go back to basics and make sure the connections are good.

After that, I'm going to explore better grounding at the top of the hill - maybe several ground rods.

I've been considering talking to the utility co, but that may open a kettle of fish that I don't want to open.

[Gun Geek]

The subject of grounding for your dock can found in the NEC 2005, article 553 "floating buildings".

The whole point of grounding is to get/make everything at a single potential.  You need to change your cable out to a 4 wire setup (2 hots, 1 neutral, 1 ground)  set up a j-box on the edge of the lake/pond and run the power cable into it, bond the ground to it and drive a ground rod on the edge of the lake/pond, then complete the run to the dock.  bond all the metal parts on the dock to ground conductor. Eric's comment about not bonding the dock to ground is 100% wrong (sorry about that, but its true) 

You probably have voltage on you neutral, and without a ground to check it to you will never know.  Which could be a problem upstream of the power source.  The other source could be induction on the neutral, which would be the same case above, especially on a long power run.

out here where i work (copper smelter, power house, I'm a industrial electrican)  we have a set of 120/240 signal cable that are about 8 feet below a 13,800 volt run of triplex power cables.  they induce about 90-100 volts dc onto the signal cables, enought to get you attention if your not careful, so your 1.5 Vdc should be about right.

<{POST_SNAPBACK}>

John:

Got a link to this NEC article?

[lefty45]

I don't mean to imply that your service is bad but it could be a utility co open neutral. Since they typically take their neutral to ground at every pole, with the earth being very dry, current is seeking the path of least resistance.

It would be interesting to talk to any neighbors to see if they are having similar problems