Voltage on the Roof???

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Electrician
I was working at a house and one of the painters said he was getting shocked off the roof. It was a metal roof and hot outside. So I thought he must have just thought he got shocked. Then another one came to me and said he also got shocked. So I went to check it out.

The part of the house where they reported the shock is an addition that was added some time ago and has a separate service. There is a single meter but dual feeds to the house & addition. The meter is on the pole about 20'-30' from the addition.

So I put a lead on the GEC or the conductor that grounds the service at the transformer. I then went up the ladder and checked the edge of the roof. Sure enough there was 120V on the roof.
Now the only thing on in the addition was a feeder breaker to a subpanel for the pool equipment. I had someone turn the feeder breaker off and checked the voltage again and it was gone. Now the only load on the subpanel was the pool pump motor.

Since I was there to do other work I didn't spend a lot of time searching for the problem. But what I did do was run a bare #8 from the EGC bar in the subpanel to the metal roof. I put an alligator clip on the end up on the roof. I had the painter clip it on the metal for me. When he touched the metal it sparked a good bit but I told him to go ahead and clip it on.
I checked the voltage again after connecting the jumper and the voltage went away. Painter was brave enough to touch the roof with his bare hand and said he didn't feel anything.

Now I know I didn't fix anything, just made it so they didn't get shocked. The HO wants me to check this out in full when I can so I thought I would share it here to get some ideas.

What I don't understand is why the subpanel or pool pump would cause voltage up on the roof. I didn't find any of the equipment with voltage, just the roof. Since nothing in the subpanel has anything to do with inside the building I'm not seeing how the roof got energized. Especially since no loads in the house were on.
Any ideas what would cause this and how to track it down?
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
That appears to be some type of connection between the metal roof and an energized conductor. You have provided a path for the current to flow, and since not breakers have tripped, it appears that the connection between the roof and the energized conductor has a fairly high impedance. Now that you have provided a path for the current to flow, you are creating heat at the point of connection. I would remove that jumper ASAP. If you don't there may be a fire.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Electrician
That appears to be some type of connection between the metal roof and an energized conductor. You have provided a path for the current to flow, and since not breakers have tripped, it appears that the connection between the roof and the energized conductor has a fairly high impedance. Now that you have provided a path for the current to flow, you are creating heat at the point of connection. I would remove that jumper ASAP. If you don't there may be a fire.

Why didn't I read any voltage after I connected that jumper?
 

JFletcher

Senior Member
Location
Williamsburg, VA
Why not just shut off and lock out/remove the breaker until the problem can be properly fixed? If that jumper gets loose, that's a shock from a rather high height.
 

MD84

Senior Member
Location
Stow, Ohio, USA
Why didn't I read any voltage after I connected that jumper?
Because you made a closed loop. You would only have measured the voltage drop of your jumper.

Try turning off the circuit causing the problem and reading resistance from the branch connection to your equipment ground. From here you can start isolating circuit components until you find your ground fault.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Electrician
Why not just shut off and lock out/remove the breaker until the problem can be properly fixed? If that jumper gets loose, that's a shock from a rather high height.
The jumper was just so the painter could finish without getting shocked.
Pool pump was the only load and they are cleaning the pool and needed the pump running.
I will either turn the pump off tomorrow or feed it from another circuit from the other service.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Electrician
Because you made a closed loop. You would only have measured the voltage drop of your jumper.

Try turning off the circuit causing the problem and reading resistance from the branch connection to your equipment ground. From here you can start isolating circuit components until you find your ground fault.
Seems like since the pool pump was the only load and it's on a GFCI breaker that the GFCI should have tripped if the pump was the problem. The crazy thing (to me) is I can't see how anything from the pool pump could reach the roof. Since no other load was on I don't see what else could energize the roof.
 

dfmischler

Senior Member
Location
Western NY
Is the roof supposed to be bonded to ground? A little googling seems to says this varies by jurisdiction.

Is there any way to see if there is wiring right under that roof? I know that if it was installed properly a roofing nail shouldn't be able to touch a cable, but...
 

MD84

Senior Member
Location
Stow, Ohio, USA
Seems like since the pool pump was the only load and it's on a GFCI breaker that the GFCI should have tripped if the pump was the problem. The crazy thing (to me) is I can't see how anything from the pool pump could reach the roof. Since no other load was on I don't see what else could energize the roof.
I missed that it went away when the feeder breaker was turned off. The feeder could be the circuit with the ground fault and this seems more likely considering the only other circuit is ground fault protected.

Check to see if the problem persists with the branch circuit turned off. That will isolate the feeder circuit as the problem.

You could throw a current clamp around your feeder conductors to check for the ground fault current. There are a few configurations which could provide useful measurements. If you can get the clamp around both lines and nuetral then it will act like a ground fault breaker somewhat. If you see current in that configuration it should be showing you the ground fault current. Check the current on your jumper and see if it is the same magnitude. Check each line to see which one is faulted. Lift the jumper to see if the fault current goes away.

If you should find a ground fault on one of your feeder conductors then you will want to trace it's path and look for the problem. Maybe a roofing screw penetrated the feeder cable. What path does the feeder cable take?
 

mgookin

Senior Member
Location
Fort Myers, FL
Is that a furnace vent in the 2nd photo? If so, you may be able to measure from that to neutral downstairs to keep you from having to go up on that roof every time you throw a breaker. You might even find your problem down there.
Have you turned off all breakers and turned them back on one at a time to see which breaker, if any, resurrects the problem?
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Electrician
Maybe a roofing screw penetrated the feeder cable. What path does the feeder cable take?
The panel is in the garage that feeds the subpanel. It goes below the panel and straight along the wall and out, about 15'. It doesn't go up towards the roof at all.

Is that a furnace vent in the 2nd photo? If so, you may be able to measure from that to neutral downstairs to keep you from having to go up on that roof every time you throw a breaker. You might even find your problem down there.
Have you turned off all breakers and turned them back on one at a time to see which breaker, if any, resurrects the problem?
Only load or breaker on was the pool pump motor. I didn't turn it off but did have someone turn the feeder breaker off, which stopped the voltage.

Just to be clear, I wasn't called there to check why there was voltage on the roof. I was there doing other work and the painters came to me asking why the roof was shocking them. So I only got to check enough to do/find what I described here.
 
Last edited:

mgookin

Senior Member
Location
Fort Myers, FL
Maybe with the garage feeder breaker open check for continuity from hot to ground at the subpanel with pump breaker open and then again with it closed. That will hopefully tell you what side of the breaker the problem is on.
 

dfmischler

Senior Member
Location
Western NY
Just a really stupid possibility: a nail in the flashing where the addition's roof meets the wall of the original building has touched a cable inside, and the flashing is in contact with the addition's metal roof...
 

FionaZuppa

Senior Member
Location
AZ
it was asked if metal roof requires grounding. yes for certain installs where pool stuff is in the mix.

but a 10ft earth spike tied to a metal roof isn't a bad idea. bonding metal roof to service GND is a questionable practice. you really do not want a lightning strike to traverse down into service GND and neutral wires.
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
Why didn't I read any voltage after I connected that jumper?
For the same reason you don't read voltage between the neutral and the earth. The only voltage to read would be the voltage drop on the bonding conductor that you installed. The current was not enough to open the OCPD, so I would not expect a significant amount of voltage drop on your bonding jumper.
 

domnic

Senior Member
Occupation
Electrical Contractor
HOT ROOF

HOT ROOF

You have a problem with a grounded conductor touching the roof.
 
Top