Customer being shocked in a properly bonded pool

[brantmacga]

Customer called and reported they had felt electric shock in their swimming pool.

The pool is bonded per NEC; I checked water to deck, water to handrails, deck to handrails, pumps are bonded, lighting niche is bonded, etc.

I performed a battery of tests around the pool mostly checking for voltage between surfaces. The customer said before I came they had the power company out also. So I called the engineer to discuss what they had checked.

The short story is, the lighting niche has a lower resistance than the deck/water/handrails. The customer only felt electric shocks when swimming past the light. They did not feel any shock when entering/exiting the pool. I was only able to read 2v to ground, engineer said the day he came they read 7v to ground in the water around the light. This wasn't to the deck, but to plugged into an extension cord leading back to the house. Water to deck its 0v.

The engineer I spoke with said he and a 3rd party engineering firm have been conducting a variety of tests on pools in the area where customers reported electric shock over the last several months. He told me their conclusion is that all newly constructed pools should have a 4x4 or at least 6x6 wire mesh under the concrete. We talked for about an hour, and discussed a lot more in detail, but this was the short version of the story.

Their recommendation to customers has been A) tear up your pool deck and install a wire mesh/grid, or B) float the pool bond from the electrical system ground, and remove 120v lighting fixtures.

This is the first time I've encountered someone being shocked on a pool w/ what we assume to be proper 4-corner NEC bonding underneath. The resistance from water to deck is pretty consistent around the entire perimeter of the pool but it is about 10 ohms lower at the lighting niche according to my meter.

 

[tom baker]

Mike Holt once said he would not swim in a pool with 120 v lighting
Interesting comments about bonding

 

[wwhitney]

I'm not super familiar with the details of these installations, but a comment and a question:

If a person feels a shock while in the water and not touching anything else, that means there has to be a voltage gradient within the water itself. At which point it's imprecise to refer to, e.g., the voltage from "water to deck" without specifying where in the water.

In what way is the data you collected inconsistent with some sort of fault within the wet niche? If you kill power to the wet niche, do the non-zero voltage measurements become zero?

Cheers, Wayne

 

[brantmacga]

Mike Holt once said he would not swim in a pool with 120 v lighting
Interesting comments about bonding

its certainly made me reconsider having a 120v fixture in my pool. I think maybe its time to upgrade.

 

[macmikeman]

This may well be something that can be traced to a broken or loose neutral in a neighbor's service. The equal potential bonding grid works in theory , but as we saw when Mike Holt took a dip in the pool, it isn't a 100% of the time effective, he could feel current while in there.

 

[brantmacga]

I'm not super familiar with the details of these installations, but a comment and a question:

If a person feels a shock while in the water and not touching anything else, that means there has to be a voltage gradient within the water itself. At which point it's imprecise to refer to, e.g., the voltage from "water to deck" without specifying where in the water.

In what way is the data you collected inconsistent with some sort of fault within the wet niche? If you kill power to the wet niche, do the non-zero voltage measurements become zero?

Cheers, Wayne

The voltage present in the pool isn't coming from the fixture; we don't know the source. The utility disconnected the house from the transformer and the voltage was still there.

The issue with the fixture is that the niche has a lower resistance to ground, and due to the difference in potential, current can pass through the swimmer to the niche. There is most likely a bonding lug on the outside of the niche that would be inaccessible as its underneath a pool deck, which is why the utility engineer is recommending to the float the pool. The 120v fixture would need to be removed because the bond would be disconnected. There is also an EGC to the fixture. But the bonding lug on the outside of the niche usually connects to one or more of the trim screws so that the metal ring around the light is bonded.

 

[mopowr steve]

I haven’t spent a lot of time with wet niches but it seems to me that the concept is for current leaking from the terminals/wire where they attach to the light bulb should travel to the grounded shell of the wet niche, right?
I mean did they really expect the niche to not get water in it?
So, is there a lost EG on the wet niche shell?

 

[brantmacga]

One other thing I forgot to mention; the customer has a deep well with a steel shaft, and the head is about 5' from the edge of the pool shell. There is a handhole enclosure in the pool deck that covers it.

 

[brantmacga]

I haven’t spent a lot of time with wet niches but it seems to me that the concept is for current leaking from the terminals/wire where they attach to the light bulb should travel to the grounded shell of the wet niche, right?
I mean did they really expect the niche to not get water in it?
So, is there a lost EG on the wet niche shell?

No the EG is not lost. The niche has a lower resistance to ground than the surrounding water/shell.

 

[macmikeman]

The voltage present in the pool isn't coming from the fixture; we don't know the source. The utility disconnected the house from the transformer and the voltage was still there.

The issue with the fixture is that the niche has a lower resistance to ground, and due to the difference in potential, current can pass through the swimmer to the niche. There is most likely a bonding lug on the outside of the niche that would be inaccessible as its underneath a pool deck, which is why the utility engineer is recommending to the float the pool. The 120v fixture would need to be removed because the bond would be disconnected. There is also an EGC to the fixture. But the bonding lug on the outside of the niche usually connects to one or more of the trim screws so that the metal ring around the light is bonded.

This is exactly what I was speaking about. I bet it is stray current from the neighbor who has lost his little sheep er. I mean neutral...

 

[macmikeman]

Forget all the stuff on your property. Once the utility disconnected you and you still have voltage issues, it is coming from outside the property. The utility should have taken steps to find it using " the Beast " tool right then and there cause there is still two issues to deal with. Your pool isn't safe, and the neighbors house might well burn down from no good return path .

 

[mopowr steve]

So, is there voltage present when the niche circuit is off?

 

[wwhitney]

The voltage present in the pool isn't coming from the fixture; we don't know the source. The utility disconnected the house from the transformer and the voltage was still there.

Ah, OK. So there's an earth voltage gradient being imposed by the POCO's equipment or someone else's service.

But I take it there's no perimeter bonding grid? Or is there a grid, but metallic wet niche conduit extends outside the grid? It seems to me that a perimeter grid that encircles the pool and encloses the full area of any metallic wet niche conduit would ensure that there's no spatial earth voltage gradient across the pool itself.

So it seems like this is an example of why the NEC requires perimeter bonding, and 4 point bonding alone isn't sufficient.

Cheers, Wayne

 

[Buck Parrish]

When the power cpmpany cut power. Did they leave the neutral connected ? I'm thinking like Mike, probably the neighbors poor neutral connection.
Your neighbors house may find that your pool grounding, well grounding, pool equipotential is some what a better then the utilities return path for their house..

 

[qcroanoke]

Mike Holt once said he would not swim in a pool with 120 v lighting
Interesting comments about bonding

I concur with Mike Holt

 

[Little Bill]

Does the pool just have a #8 bare encircling the pool, then jumpers to 4 points? I saw a video demonstrating that the alternate method of just a #8 bare conductor didn't give the proper bond, and a grid of 12" sq was far better. Can't remember if Mike Holt offered the video for viewing, but I don't think it was one of his videos. Whoever made the tests was going to try to get the code changed to require a grid pattern and do away with the alternate single #8.

 

[Buck Parrish]

Customer called and reported they had felt electric shock in their swimming pool.

The pool is bonded per NEC; I checked water to deck, water to handrails, deck to handrails, pumps are bonded, lighting niche is bonded, etc.

I performed a battery of tests around the pool mostly checking for voltage between surfaces. The customer said before I came they had the power company out also. So I called the engineer to discuss what they had checked.

The short story is, the lighting niche has a lower resistance than the deck/water/handrails. The customer only felt electric shocks when swimming past the light. They did not feel any shock when entering/exiting the pool. I was only able to read 2v to ground, engineer said the day he came they read 7v to ground in the water around the light. This wasn't to the deck, but to plugged into an extension cord leading back to the house. Water to deck its 0v.

The engineer I spoke with said he and a 3rd party engineering firm have been conducting a variety of tests on pools in the area where customers reported electric shock over the last several months. He told me their conclusion is that all newly constructed pools should have a 4x4 or at least 6x6 wire mesh under the concrete. We talked for about an hour, and discussed a lot more in detail, but this was the short version of the story.

Their recommendation to customers has been A) tear up your pool deck and install a wire mesh/grid, or B) float the pool bond from the electrical system ground, and remove 120v lighting fixtures.

This is the first time I've encountered someone being shocked on a pool w/ what we assume to be proper 4-corner NEC bonding underneath. The resistance from water to deck is pretty consistent around the entire perimeter of the pool but it is about 10 ohms lower at the lighting niche according to my meter.

Did the power company disconnect the customers neutral when they cut power? IMO that would need to be done first before trying any thing else.

 

[synchro]

A suggestion is to disconnect all conductors going to the lighting niche (i.e., hot, neutral, EGC, bonding conductor) in the junction box or other closest connection point. Then measure the resistance between the bonding conductor and the pool water near the light to verify that the bonding connection to the niche is at least somewhat intact. If there is a metal trim ring around the light in the pool, it would be desirable to measure the resistance between the metal ring and the end of the bonding conductor to confirm that there's a solid bonding connection to the lighting niche.
Also measure the resistance from the other conductors to the bonding conductor to see if any leakage resistance is present.

 

[don_resqcapt19]

This sounds like elevated neutral to earth voltage on the utility primary, but even if that is the case, there should be no shocks occurring if everything is correctly bonded. With everything correctly bonded, the only voltage to drive a shock would be enough current flowing on the bonding system to create a voltage drop that results in a shock.

 

[shortcircuit2]

Does the pool just have a #8 bare encircling the pool, then jumpers to 4 points? I saw a video demonstrating that the alternate method of just a #8 bare conductor didn't give the proper bond, and a grid of 12" sq was far better. Can't remember if Mike Holt offered the video for viewing, but I don't think it was one of his videos. Whoever made the tests was going to try to get the code changed to require a grid pattern and do away with the alternate single #8.

Electric Power Research Institute (EPRI) in Lenox Massachusetts did extensive research on pool bonding and concluded the single #8 ground ring was insufficient.
EPRI-

StrayVoltage::Recommendations for Safe Pool Design

In Massachusetts there is an amendment requiring the copper grid for inground pools and the cooper ring is not used.