Another Shocking Experience

[jerjwillelec]

Another service call today. This time to a resturant with an apartment above. Immediately I can tell this is a do it yourselfer with enough knowledge to be dangerous but I wanted to check it out anyway. The HVAC crew was installing an A/C unit on a (rubber) roof. When they ran the line set through the roof and happened to touch the EMT (A/C feeder) they got nailed. They proceeded and later when they touched the EMT and the rubber roof, they got nailed again.

Here is where I come in. With my FLUKE 179 (I am purchaseing a Simpson 260-8 ASAP), I read... EMT to refrigerant line: 59 VAC. EMT to metal roof boot of vent pipe: 73 VAC. EMT to rubber roof: anywhere from 59 VAC to 73 VAC. Keep in mind... The feeder to the A/C (8/3 NM cable in a 3/4" EMT) has never been energized.

I then go to the source... 120/208 VAC single phase 100 amp sub panel... Neutral bonded to the can... Feeder is a #2/3 SE cable (no ground)... A #10 AWG (in 1/2" EMT) is ran to the nearest water pipe and bonded to the EMT and the water pipe... (By now you can see it's a "do it yourselfer") All of my meter readings at the loadcenter are normal except when I read phase to any disconnected and isolated conductor in the A/C feeder, it reads 111 VAC (again, I am buying a analog meter soon)...

I then proceed to the Service Entrance and find the earth ground and water main are not bonded (the water main is not even connected). I bonded the two with a temp #14 and no change to my readings.

I did not personally "check" by getting shocked, so I am taking their word for it. Why am I getting voltage readings on a rubber roof. It (seems) obvious to me that it is a grounding issue, but only at the rubber roof. No one seems to get "hit" anywhere else. Help Please!!!

 

[220/221]

Divide and conquer. Follow the path.

Generally, energized conduit seems to come from an open neutral somewhere. It doesn't HAVE to be in that citcuit.

It will help to seperate the conduit and see which way it's coming from.

Don't leave it like that!!

If the conduit is NOT bonded the source could be the ungrounded conductors within.

 

[SiddMartin]

do you know if the the voltage is coming from the subpanel? It may not be feasible(resturant), but you can find out which circuit is causing it. If you turn off each breaker till you lose the voltage, then you'll know which circuit is causing the problem. then troubleshoot that circuit.

Maybe you have already eliminated this possibility

 

[gar]

080609-2119 EST

jerjwillelec:

A Simpson 260 or 270 ( I do not believe the 270 is made anymore) has an input resistance of 20,000 ohms/volt for DC and on recent models AC is 5000 ohms/volt and on my original, from circa 1948, AC was I believe 1000 ohms/volt. Note 20,000 ohms per volt corresponds to a 50 micro-ampere meter movement.

5000 ohms/volt and the 2.5 V range has an input resistance of 2.5 * 5000 = 12,500 and at 250 V this is 100 times larger or 1.25 megs.

A typical digital meter is usually a fixed input resistance of about 10 megohms or maybe more.

Get a 10,000 ohm to 100,000 ohm resistor and put it in parallel with your digital and you will load the item being tested more than with the Simpson on the 250 AC range. At 100,000 ohms you need a 1 W resistor.

Somewhere on this forum I discussed making ground potential measurements.

.

 

[jerjwillelec]

Thank you for your ideas. I did, through process of elimination, end up shutting down the whole building and the current flow is still there. I personally tested it with my arm and (digital vs. analog or not), there is current flow. I failed to mention this before but this building is in an old down town district where all of the buildings on the block are connected structurely. What gets me is, this seems to only happen at the roof which is rubber. The building is wood framed with wood sheating.

 

[hillbilly]

With all of the buildings being tied together, there must be conductive path(s) between them.

You could be getting return current from one of the neighbors thru (one of) these paths.

Are all of the services seperate for each building?

Can you get a amp probe on your building's GEC?

Thr rubber roof part.....I'm still pondering:-? ........static maybe?

steve

 

[hardworkingstiff]

Are the services overhead?

 

[khixxx]

I have seen up to 84v on a conductor that was induction but not sure what you would call it. I did not see sparks when I grounded it. Noise on the loop was still there so grounding one end was pointless. The overhead service or even overhead power lines might cause this induction but I highly doubt it.

You say the EMT is for the AC feeder? Is it possible they ran the EMT through another conduit when they drilled? Meaning they nicked another energized conductor from another panel maybe the main? I drilled into floors and sparks went flying conductors stayed energized. If this is a do it your self project the lad might of pressed onward and thought the sparks were cool.

The rubber roof is confusing but as material breaks down they could become conductive.

Ken

 

[iaov]

With all of the buildings being tied together, there must be conductive path(s) between them.

You could be getting return current from one of the neighbors thru (one of) these paths.

Are all of the services seperate for each building?

Can you get a amp probe on your building's GEC?

Thr rubber roof part.....I'm still pondering:-? ........static maybe?

steve

Not all rubber is an insulator. If the rubber contains enough carbon it will conduct current.

 

[hardworkingstiff]

Rubber tires are steel reinforced. Is there any reinforcement in the rubber roof and is it steel?

Are the roof joists steel or wood?

If the services are overhead, how are they attached to the building? Do they impact the rubber roof?

 

[BryKey]

Hate to dig up an older thread but this one struck me.
What are the chances that one of the neighboring buildings is hijacking power from the building in question? Look at the Meters for the surrounding buildings, Esp. if the other buildings have apartments above like the one you are working on.
Chances are someone had their power turned off and has tapped into it.