Problem with Pool Lights/GFI Breaker

[Vince67]

Hi All,
I am working on a new pool that has a total of (12)120v LED light fixtures with a total 7.2amps. Six lights on one side, six on the other.
Being supplied from at 20a Siemens GFI bolt-on breaker.

Breaker supplies a switch and from the switch it controls the right and left sides. The switch is approximately 100' from the breaker.

My problem exists on the left bank of lights which is approximately 100' from the switch at the last connection point. When the switch is activated the circuit trips 100% of the time.

The right bank of lights works 100% of the time and the furthest connection point is approx. 50 feet from the switch.
The entire circuit is supplied with #10 THHN with no voltage drop.

I have replaced the breaker and the wires from the switch to the junction boxes and I can get a combination of the 6 pool lights to work. And then when the switch is turned off the same combination will trip the circuit when turned back on.

I have isolated the right bank from the left bank and have applied each individual light (6) from the left side onto the energized circuit without the GFI breaker tripping. I have also had the right side bank energized and applied the left bank lights one at a time and the circuit holds.

The left side lights trip with the rush of incoming current when the switch is applied or even when I bypass the switch and comes from the breaker.

I've had 6 brand new lights in their boxes, applied to the circuit and the GFI breaker held. I re-tested the lights in the pool and the circuit tripped. So, the problem is got to be a combination of the lights in the pool.

We did another test, cut an extension cord and applied all 12 lights and plugged it into a GFI outlet and it held. Maybe it's not the pool lights?

The Siemans breaker has a 5 Milliamp trip rating which is not handling the in rush of current. The GFI outlet has a 4-6 MA trip rating and is holding when plugged in.

Would the total footage of the circuit affect the 5 Milliamp GFI breaker at 3.6a and cause it to trip?
There's a 30 milliamp breaker available and this is my next option, I'm just not sure that this will solve my problem.
Thanks
Vince

 

[Little Bill]

GFCI's don't trip for in-rush as far as the GFCI function is concerned. They trip when there is an imbalance of around 5mA between the hot and neutral.
You may have some leakage in several lights, but any one light may not have 5mA. But adding more lights with a small leakage could add up to be over the threshold.

Is there not a light controller for these lights?
Does each light have it's own LED driver?

Welcome to the Forum!

 

[Another C10]

The Siemans breaker has a 5 Milliamp trip rating which is not handling the in rush of current. The GFI outlet has a 4-6 MA trip rating and is holding when plugged in.

If all 12 hold with the GFI receptacle, scrap the breaker and integrate a simple receptacle upstream of the switch. It seems like you mentioned the breaker does not handle the design well.

 

[synchro]

If all 12 hold with the GFI receptacle, scrap the breaker and integrate a simple receptacle upstream of the switch. It seems like you mentioned the breaker does not handle the design well.

Assuming it holds when using a GFCI receptacle this way, unfortunately you still wouldn't know how much margin it would have before it trips. And so you might even consider putting two deadfront GFCIs ahead of a 2-pole switch, and have a GFCI for each bank so that they'd each be seeing a reduced amount of cummulative leakage.

 

[kwired]

How much normal circuit conductor has been bypassed when using the GFCI receptacle vs the GFCI breaker.

Longer the circuit run after the GFCI device the more capacitive current leakage is a possible contribution to the problem here.

If you have a long switch leg then long run to the lights it may leak enough to trip the GFCI. But if you placed the GFCI near the pool location in a temp method you may have bypassed enough normal circuit conductor to lessen capacitive leakage that the GFCI would see.

 

[Vince67]

Thanks for the welcoming Little Bill.
No controller just a 20a single pole switch from the breaker. The drivers are in the light and sealed from the factory. I know the GFI's doesn't trip on in-rush, but I had to ask.

Another C10 and synchro- Due to the design of this building, intercepting the circuit with GFI receptacle may be difficult and not an option.

kwired- All the circuit conductor were used except the feed from the breaker to the switch (approx 100'). That was not backfeed because it works the right bank of lights all the time with no issues.
This is why I am wondering if the overall length of the left bank switch leg conductors is having an effect on the 5mA GFI breaker. And is there enough margin in the leakage on the circuit so that a 30mA GFI breaker can handle it. (my million $$ question)
I have 2 more test that I'm planning on doing today and I'll see where I go from there.
Thanks for your responses, I'll keep you posted on the solution.
Vince

 

[synchro]

And is there enough margin in the leakage on the circuit so that a 30mA GFI breaker can handle it. (my million $$ question)

A 30mA GFI breaker is for protection of equipment, but for a pool a 5mA GFCI breaker for protection of personnel would be required.

All the circuit conductor were used except the feed from the breaker to the switch (approx 100'). That was not backfeed because it works the right bank of lights all the time with no issues.
This is why I am wondering if the overall length of the left bank switch leg conductors is having an effect on the 5mA GFI breaker.

I don't think the capacitance of the wiring would likely be more than 100pF per foot, or 0.02μF for 200 feet. That would have a capacitive reactance of 133K ohms, which would draw 0.9mA at 120V 60Hz. And so the wiring capacitance might be roughly 1/5 of the problem on the left bank, but the lights themselves probably contribute most of the overall capacitance. There is likely some line-ground capacitance inside of the lights for EMI filtering, and that would add up as you connect more lights in parallel. Another factor that might be involved is the inductance of the wiring, not for 60Hz where the inductive reactance of the 10ga wire would be negligible compared to its resistance, but the inductive reactance could be significant during fast transients when the switch or breaker is actuated. As a result there might be higher frequency ringing in the LC circuit that's formed, which then causes some relatively high frequency common mode currents through the wire capacitance mentioned above, and then that trips the GFCI. Of course this is speculation without seeing a waveform on a 'scope. It's just a hunch, but perhaps you could try putting an incandescent light at the end of the left bank of lights if possible, or at the switch. That might add some damping if such ringing is occurring. I would do this with the right bank disconnected, at least initially.

Is the THHN wire routed in PVC conduit? That would contribute less line-ground capacitance than metallic conduit.
By the way, is this installation in a 120/240V system or in a 120/208V one?

 

[augie47]

Anecdotal at best, but I have seen strange interactions between GFCIs and LED lighting especially foreign LED. If the receptacle is working, I'd agree with Post #3

 

[Vince67]

A 30mA GFI breaker is for protection of equipment, but for a pool a 5mA GFCI breaker for protection of personnel would be required.



I don't think the capacitance of the wiring would likely be more than 100pF per foot, or 0.02μF for 200 feet. That would have a capacitive reactance of 133K ohms, which would draw 0.9mA at 120V 60Hz. And so the wiring capacitance might be roughly 1/5 of the problem on the left bank, but the lights themselves probably contribute most of the overall capacitance. There is likely some line-ground capacitance inside of the lights for EMI filtering, and that would add up as you connect more lights in parallel. Another factor that might be involved is the inductance of the wiring, not for 60Hz where the inductive reactance of the 10ga wire would be negligible compared to its resistance, but the inductive reactance could be significant during fast transients when the switch or breaker is actuated. As a result there might be higher frequency ringing in the LC circuit that's formed, which then causes some relatively high frequency common mode currents through the wire capacitance mentioned above, and then that trips the GFCI. Of course this is speculation without seeing a waveform on a 'scope. It's just a hunch, but perhaps you could try putting an incandescent light at the end of the left bank of lights if possible, or at the switch. That might add some damping if such ringing is occurring. I would do this with the right bank disconnected, at least initially.

Is the THHN wire routed in PVC conduit? That would contribute less line-ground capacitance than metallic conduit.
By the way, is this installation in a 120/240V system or in a 120/208V one?

I'll be able to input an incandescent light and see where that leads.
Yes, for the most part the THHN is routed through PVC conduit. The conductor run through a NEMA 4x JB with ridgid nipples through the concrete slab. The switch is mounted in 4 sq box in a fully grouted CMU block wall. The pool light junction Boxes are Hubbell brass mounted in CMU block.
The system is 120/240v.
Thanks again.

 

[Little Bill]

Thanks for the welcoming Little Bill.
No controller just a 20a single pole switch from the breaker. The drivers are in the light and sealed from the factory. I know the GFI's doesn't trip on in-rush, but I had to ask.

Another C10 and synchro- Due to the design of this building, intercepting the circuit with GFI receptacle may be difficult and not an option.

kwired- All the circuit conductor were used except the feed from the breaker to the switch (approx 100'). That was not backfeed because it works the right bank of lights all the time with no issues.
This is why I am wondering if the overall length of the left bank switch leg conductors is having an effect on the 5mA GFI breaker. And is there enough margin in the leakage on the circuit so that a 30mA GFI breaker can handle it. (my million $$ question)
I have 2 more test that I'm planning on doing today and I'll see where I go from there.
Thanks for your responses, I'll keep you posted on the solution.
Vince

I've done a lot of pools and pool lights. I would be willing to share my experience for airfare and lodging!

 

[hillbilly1]

What brand of breaker is it? Square D QO and QOB are much more sensitive to inrush than others. Have you tested it with a temporary regular breaker? If it does, and if it is a Square D, you may have to use a QOHM, with a GE two space sub, with a GE gfi breaker.
Target used cheap led lights over their cosmetic cases, and the stores that had Square D gear, would constantly trip the breakers. They ended up running quite a few circuits out there for a small load.

 

[Vince67]

I've done a lot of pools and pool lights. I would be willing to share my experience for airfare and lodging!

Maybe next time my friend.

 

[Vince67]

What brand of breaker is it? Square D QO and QOB are much more sensitive to inrush than others. Have you tested it with a temporary regular breaker? If it does, and if it is a Square D, you may have to use a QOHM, with a GE two space sub, with a GE gfi breaker.
Target used cheap led lights over their cosmetic cases, and the stores that had Square D gear, would constantly trip the breakers. They ended up running quite a few circuits out there for a small load.

It's a Siemen's bolt-on 20a with a 5mA trip rating. This is the type of project that I'm not going to be able to start mixing and matching parts and pieces.
After tomorrow's continuing test that we'll be doing it's going back to the EE's and they can resolve this issue. I'll keep you posted.

 

[hillbilly1]

You may have to split it into two circuits then, either a two pole, or pull another neutral, and use two single poles.

 

[Another C10]

Another C10 and synchro- Due to the design of this building, intercepting the circuit with GFI receptacle may be difficult and not an option.

One last inexpensive solution without much effort could be this little bugger ..

 

[mtnelect]

Why are you using line voltage instead of low voltage pool lights ?

 

[Amps]

Why are you using line voltage instead of low voltage pool lights ?

My thought as well. Leaking voltage around a pool is never good.

 

[Regular electrician]

Hi All,
I am working on a new pool that has a total of (12)120v LED light fixtures with a total 7.2amps. Six lights on one side, six on the other.
Being supplied from at 20a Siemens GFI bolt-on breaker.

Breaker supplies a switch and from the switch it controls the right and left sides. The switch is approximately 100' from the breaker.

My problem exists on the left bank of lights which is approximately 100' from the switch at the last connection point. When the switch is activated the circuit trips 100% of the time.

The right bank of lights works 100% of the time and the furthest connection point is approx. 50 feet from the switch.
The entire circuit is supplied with #10 THHN with no voltage drop.

I have replaced the breaker and the wires from the switch to the junction boxes and I can get a combination of the 6 pool lights to work. And then when the switch is turned off the same combination will trip the circuit when turned back on.

I have isolated the right bank from the left bank and have applied each individual light (6) from the left side onto the energized circuit without the GFI breaker tripping. I have also had the right side bank energized and applied the left bank lights one at a time and the circuit holds.

The left side lights trip with the rush of incoming current when the switch is applied or even when I bypass the switch and comes from the breaker.

I've had 6 brand new lights in their boxes, applied to the circuit and the GFI breaker held. I re-tested the lights in the pool and the circuit tripped. So, the problem is got to be a combination of the lights in the pool.

We did another test, cut an extension cord and applied all 12 lights and plugged it into a GFI outlet and it held. Maybe it's not the pool lights?

The Siemans breaker has a 5 Milliamp trip rating which is not handling the in rush of current. The GFI outlet has a 4-6 MA trip rating and is holding when plugged in.

Would the total footage of the circuit affect the 5 Milliamp GFI breaker at 3.6a and cause it to trip?
There's a 30 milliamp breaker available and this is my next option, I'm just not sure that this will solve my problem.
Thanks
Vince

I’d just load the lights off a regular receptacle. I don’t see any issues with that solution. I could give you some science based equations to plug your numbers into, but I’d just try the receptacle, works every time I do it.

 

[SSDriver]

That is way too many LED pool lights on one GFCI circuit breaker. The leakage from the switching power applies in each add up. My primary work is commercial swimming pools. If we do more than 8 we get random nuisance trips. More than 10 and they become very frequent.

LED scoreboards have the same issue. For schools we used to supply scoreboards with 1 to 2 circuits. Now it's 4 to 6 GFCI circuits to avoid tripping GFCI breakers.

 

[synchro]

It would be best in this situation to add more individual single pole GFCI breakers rather than 2-pole ones. That's because with 2-pole GFCI breakers, as long as the leakages on each pole are within 4mA of each other, then the GFCI won't trip. And the difference between them could be up to 6mA before tripping. And so, for example, if the leakage on one pole was 20mA, the 2-pole GFCI won't trip if the leakage on the other pole is between 16mA and 24mA, and possibly even between 14mA and 26mA without tripping. Even though this may be unlikely to happen, it's not a good situation when pool lights are involved.

The reason I asked whether the system is 120/240V or 120/208V in post #7 is that 120/208V would not have this vulnerability for having leakages cancel within 2-pole GFCIs.