Bonding Mystery - Pool Light

I have an in-ground concrete pool (epoxy paint finish) with an Amerlite underwater light. The chrome bezel continuously turns black. The tarnish can be scrubbed off, but always returns. (There is also an ongoing, low-intensity battle with minor metal stains, with no known source for the metal.) Other than the niche and pool light, there is no metal in contact with the water. (No stairs, rails, etc.)

The tarnish probably indicates a bonding problem, but.................

There is a bond wire from the light niche to the bonding lug on the pump. The bond wire is #8 *stranded*, however. The wire connects to the inside of the niche and runs up the conduit with the power cord. The attachment of the bond wire to the niche is slathered in some sort of caulk. Using a wire nut in the J-box, the stranded bond wire is connected to another run of #8 stranded wire that goes back to the pump. It is a 30-40 foot run from the niche to the pump.

While the pool was partially drained I used a VOM to test, and there is continuity from the niche to the bonding lug on the pump. I connected my test wire to the screw hole where the screw holding the bezel in place is inserted, so I can assume that there is continuity from the bezel to the pump.

There are two other bond wires (also #8 stranded) connected to the bonding lug on the pump. One comes from the heat pump. The other comes from the concrete pool deck, with a bond wire from the rebar in a hot tub slab that I poured spliced into it. I have no idea of the quality or condition of the bond in the concrete pool deck.

The only real weirdness that I can see is the bond wire connection inside the niche, possibly in contact with the water if the caulking has failed. Since the goal of bonding is to keep everything, including/especially the water, at the same potential, I'm not sure this is a problem.

I had the entire pool system rewired by an electrical contractor. Beauzeax, my name for the previous homeowner who installed the pool.............and who owned a pool construction company (!), had totally screwed it up. An AFCI on the light, no GCFI's anywhere else, and the bond wires disconnected since they caused the AFCI to trip. (I wonder if there's a clue there.) The new GCFI on the pool light has not tripped.

The pool and pool light are around 15 years old. Is there any chance the power cord to the light is failing and leaking some current?

Does anybody have any ideas or suggestions on how to diagnose or fix what probably is some sort of current flow causing the tarnish on the light bezel? Can anybody think of any test points to determine if/where there is a voltage differential? Is there any way to test for leakage from the pool light power cord with the pool now being refilled?

Thanks.

(Here's another laugh. Beauzeax ran 75 feet of undersized gas line from the meter to the (now removed) pool heater............USING PVC! The undersized pipe caused the burners in the heater to run cool and moist, which rusted out the burners and baffles.)

Do you have a water bond? Perhaps the light bezel is the only contact between the bond and the water.
Is the tarnish worse when the light is used frequently? Perhaps the light power line is inducing a current in the bond wire in the same conduit.
One other thought, Isn't the Bond supposed to be #8 solid copper?

If the OP's municipality is following the International Building Code, then yes, all bonding "shall be not less that 8 AWG solid copper conductor"

I did some testing and there was no measurable voltage between the bond wires and the pump, although my VOM only goes down to a 2 VDC range. There is a slight amount of DC current (.002-.004 milliamps) between the bond wires from the light and the pool deck to the pump bonding lug. (The heat pump is currently disconnected from the pool plumbing.) There was no difference in measurements with the light on or off and/or the pump on or off.

I'm thinking of installing sacrificial anodes on the pool light and in the main plumbing, probably near the heater. Potential brand: http://pooltool.com/anodes.html.

Does anybody have any recommendations for, against or alternatives?

Also, other than code compliance is there any reason to change out the stranded #8 bonding wire for solid?

Thanks.

I'm not sure how you measured from bond wires to pump. Did you lift the bond from the pump?

As I understand it, there must be some voltage where there's current?

A sacrificial anode might be a good way to counter galvanic corrosion and would also serve as a water bond. It would almost certainly put zinc ions in the pool and cause some current in the bond. It would have to be connected to the bond to be effective.

As I understand it and as you mention, the idea behind the bond wire is to provide a low resistance path between all conductive parts in and around the pool mitigating the risk of electric shock to swimmers by equalizing the voltage of all conductive parts of the pool. I believe solid wire is required because each strand in stranded wire can corrode more quickly than solid wire causing high resistance in the wire.

Does the electrical contractor have pool experience? Pools are different than homes.

The only good advice I can give is: find a qualified, licensed electrician with extensive pool experience to look at this pool.