32 volts....

Triggerjay · Dec 2, 2016

Unplug lights, it all goes away.

Brew12 · Dec 2, 2016

Triggerjay said:
Unplug lights, it all goes away.

Yup, looks like you have a fault in your lighting and a bad ground path. I've helped quite a few people who thought their lights were shocking them where it wasn't the case. I think you are the first person who actually has a bad light!

Triggerjay · Dec 2, 2016

Here's the bracket setup..

Brew12 · Dec 2, 2016

Triggerjay said:
Here's the bracket setup..

Nice! Makes further troubleshooting easy. Odds of having identical faults in both lights is almost none. Looks like you have an open ground in your power strip and the corrosion is tracking from hot to the grounds of your lights. A new power strip should fix your situation if I'm guessing right.

Triggerjay · Dec 2, 2016

I don't understand how the V is getting into the water.. must be salt creep somewhere. The bracket is isolated from the water.

Brew12 · Dec 2, 2016

Triggerjay said:
I don't understand how the V is getting into the water.. must be salt creep somewhere. The bracket is isolated from the water.

It could be an artifact from induced voltages with your DC pumps. Since they operate by creating a new AC voltage at a variable frequency the fault voltage from the lights could read a difference compared to the induced voltage from the DC pumps.

Triggerjay · Dec 2, 2016

Brew, thank you for helping me think this through!

Brew12 · Dec 2, 2016

Triggerjay said:
Brew, thank you for helping me think this through!

Glad to help. Although I doubt you really needed it. Sometimes it just helps to talk things through.

Let me know if you run into anything else you see that is unusual if you don't mind. I learn from these threads, too!

Awesome Dennis · Dec 3, 2016

That meter should read amps through the probes but you have to put it in line with the load.

Brew12 · Dec 3, 2016

Awesome Dennis said:
That meter should read amps through the probes but you have to put it in line with the load.

I checked again and the instructions for that meter are pretty clear that it doesn't measure current using the probes.

Awesome Dennis · Dec 3, 2016

You're right my mistake

Brew12 · Dec 3, 2016

Triggerjay said:
Brew, thank you for helping me think this through!

Find anything today?

kattz · Dec 3, 2016

I'd add, toss the cheapie power strip and spring for a real GFCI power strip. Lowe's... Get a outlet tester ($5 at Harbor Freight) and check that your outlets are actually grounded; floating grounds in homes are common. Are you using an in-sump or in-tank grounding probe?

Finally, I've seen a lot of LED fixtures have issues where the diodes are surface-mounted on a backplane of aluminum to double as a cheap heatsink. After a while, salt creep, corrosion, and other issues cause the applied insulator coating to break down. Many lights are generating more volts that you'd imagine and that voltage can cross salt and moisture traces very easily. The tank is at ground potential and you brush the fixture or bracket and <ZAP>... My wife took a 75VDC shock from LED's about five years ago. Now I meter everything to ground after a tank build or something gets added, use GFCI outlets, and a ground probe in the sump tied to the outlet screw.

Meter the outlet. You should have 115VAC +/-10% AC hot to ground, 115VAC +/-10% AC hot to neutral, and 0VAC AC neutral to ground. If you get more than a millivolt range reading at the neutral to ground, you've got other issues.

TheEngineer · Dec 3, 2016

@Brew12 you're awesome! Love when there are other people like me out there. I'm a EE. You too?

My first guess here is a missing case ground strap. You can check continuity between the case and the ground pin of the power cord for the light. I'm going to guess it is open on one of the lights.

Brew12 · Dec 3, 2016

icecool2 said:
@Brew12 you're awesome! Love when there are other people like me out there. I'm a EE. You too?

My first guess here is a missing case ground strap. You can check continuity between the case and the ground pin of the power cord for the light. I'm going to guess it is open on one of the lights.

I'm not an EE but I work as one. No college but did 10 years as a Navy nuke electrician. Now I handle all the projects and maintenance for a 500kV substation with 960MVA of installed main transformers along with my industrials entire power distribution system.

I think this has to be a power strip issue. With 2 separate lights he would have to have a bad ground connection in each which I would consider unlikely. I'm guessing the ground connection in the power strip corroded and not only isn't connected to the home ground, but is using the light grounds to feed power to the case of the fixtures.

Would love verification!

TheEngineer · Dec 3, 2016

Brew12 said:
I'm not an EE but I work as one. No college but did 10 years as a Navy nuke electrician. Now I handle all the projects and maintenance for a 500kV substation with 960MVA of installed main transformers along with my industrials entire power distribution system.

I think this has to be a power strip issue. With 2 separate lights he would have to have a bad ground connection in each which I would consider unlikely. I'm guessing the ground connection in the power strip corroded and not only isn't connected to the home ground, but is using the light grounds to feed power to the case of the fixtures.

Would love verification!

Ah, a Navy nuke. Makes sense now

I must have missed a detail where you got that impression. It looked to me like the measurements were coming from the light rail to the ground and it looks to me like the lights are mounted with metal cables and caribiners. Carry on and I'll follow along too

Brew12 · Dec 3, 2016

icecool2 said:
Ah, a Navy nuke. Makes sense now

I must have missed a detail where you got that impression. It looked to me like the measurements were coming from the light rail to the ground and it looks to me like the lights are mounted with metal cables and caribiners. Carry on and I'll follow along too

It was a combination of these 2 comments combined with the picture of his system showing that his lights don't come close to making contact with his tank.

Triggerjay said:
So, I got out my trusty fluke meter, and from ground to light bracket I get 32 volts AC.

Triggerjay said:
The water voltage to ground drops to zero as well as the fixture to ground when the lights are unplugged.

I think something is grounding out his tank and that the source of the 32VAC was his ground measuring point, not the tank.

TheEngineer · Dec 3, 2016

Brew12 said:
It was a combination of these 2 comments combined with the picture of his system showing that his lights don't come close to making contact with his tank.

I think something is grounding out his tank and that the source of the 32VAC was his ground measuring point, not the tank.

Hmmm. Perhaps a combination of a bad ground in the power strip and a missing case ground is providing both the source and path. It could explain the 32vac source.

This is fun. I wish I was there so I could poke at things