LED lights ghosting when on solar/inverter, but not when on grid power

[ABarbarian]

I have some circuits powered by solar/inverter, and some circuits powered by grid.

Some circle LED lights have a dim glow at night-time when they are on the solar circuits, but if I move their circuit to grid supply there is no glow. The glow is constant while the switch is off, it does not go away over time.

This suggests it is not the switch or bulbs that are causing the glow, but the circuit's source.

Anybody have ideas how to fix this?

 

[LakeHouse]

I have some circuits powered by solar/inverter, and some circuits powered by grid.

Some circle LED lights have a dim glow at night-time when they are on the solar circuits, but if I move their circuit to grid supply there is no glow. The glow is constant while the switch is off, it does not go away over time.

This suggests it is not the switch or bulbs that are causing the glow, but the circuit's source.

Anybody have ideas how to fix this?

Just throwing out a shot in the dark here…. But do you have a N-G bond on the inverter circuit? I’d start by measuring ac voltage from neutral to ground, hot to ground and hot to neutral on a circuit that’s off.

 

[Recovering EE]

Just throwing out a shot in the dark here…. But do you have a N-G bond on the inverter circuit? I’d start by measuring ac voltage from neutral to ground, hot to ground and hot to neutral on a circuit that’s off.

I concur. Sounds like a ground loop, which can be caused by an inappropriate N-G bond somewhere.

 

[ABarbarian]

Just throwing out a shot in the dark here…. But do you have a N-G bond on the inverter circuit? I’d start by measuring ac voltage from neutral to ground, hot to ground and hot to neutral on a circuit that’s off.

Thanks for the suggestion and sorry for the slow reply.

When in inverter mode the AC out is floating, no N-G bond, difficult to fix at this time due to inverter model.

However when I change to bypass mode the AC out is no longer floating, it uses the grids N-G bond and there is no voltage difference neutral to ground, but the lights still glow.

- When I turn off the (line) breaker for the lights circuit the glow remains.
- When I turn off the 2 pole master breaker the glow disappears. To me, this means there is charge going down the neutral.

Any ideas how to diagnose?

Maybe I can hardwire a bond in inverter mode and see if the glow goes away ...

 

[OzSolar]

Any ideas how to diagnose?

Light tells me that you have a small of amount of current leaking from the neutral onto the ground when you open the switch.

FWIW, using a two pole switch will make the light turn off.

You can run the light in series with your voltmeter and you will likely see a few milliamps on the neutral when the light is off.

I ran into this at 480v factory we were installing a new LED lighting system on. One style of the fixtures used a small 480v to 277v transformer, apparently they couldn't get 480v LED drivers. It had to do the transformer and where the motion sensor was breaking the circuit.

Just throwing out a shot in the dark here…. But do you have a N-G bond on the inverter circuit? I’d start by measuring ac voltage from neutral to ground, hot to ground and hot to neutral on a circuit that’s off.

As LakeHouse already mentioned, please do let us know what these voltages are. N to G is the one I'm most curious in.

 

[AntronX]

Which inverter? Where are you located? Can you draw a diagram of your wiring?

 

[Checkthisout]

Have this same issue at my cabin and always have.

Never bothered to diagnose so curious what the outcome is.

 

[Checkthisout]

Light tells me that you have a small of amount of current leaking from the neutral onto the ground when you open the switch.

How would this "power" the bulb when one side of the circuit is broken at the switch?

If I had an LED on the bench in front of me, how would I get it to glow slightly with no power applied to one side of it?

Can you get it to glow by applying a/c power to just one side of the bulb?

 

[AntronX]

How would this "power" the bulb when one side of the circuit is broken at the switch?

Via filter capacitors inside LED power regulator circuit if there is some residual AC between neutral and ground.

 

[efficientPV]

Are these bulbs really cheap. Your inverter is putting out some HF noise and the capacitance of the cable is enough to supply power to the LED. A LED will light with a couple nano amps of current. The lamp has to have some external load to drain off this leakage. A 100K resistor or .022 capacitor of appropriate rating will stop this. I actually have a hall lamp that I turned into a night light bu putting a .1uf capacitor across the switch so it would light up dimly all the time. Depending on internal design, this will work, flash or not light at all.

 

[efficientPV]

Via filter capacitors inside LED power regulator circuit if there is some residual AC between
neutral and ground.
View attachment 209319

Not likely at all.

 

[ABarbarian]

As LakeHouse already mentioned, please do let us know what these voltages are. N to G is the one I'm most curious in.

Thanks for all the replies, I will run some more tests and list out a more detailed post later, but in reply to this question (system is 230 V):

[When on inverter mode (no N-G bond)]
- ground to neutral ~115 V
- ground to line ~115 V
- line to neutral ~230 V

[When on bypass (N-G bond from main distribution board passed on by inverter)]
- ground to neutral 0 V
- ground to line ~230 V
- line to neutral ~230 V

Both situations above have glow, but if move circuit to grid breaker box (not pass inverter), glow disappears

 

[wpns]

ground to neutral ~115 V

Well, there you go then

 

[OzSolar]

Thanks for all the replies, I will run some more tests and list out a more detailed post later, but in reply to this question (system is 230 V):

[When on inverter mode (no N-G bond)]
- ground to neutral ~115 V
- ground to line ~115 V
- line to neutral ~230 V

[When on bypass (N-G bond from main distribution board passed on by inverter)]
- ground to neutral 0 V
- ground to line ~230 V
- line to neutral ~230 V

Both situations above have glow, but if move circuit to grid breaker box (not pass inverter), glow disappears

I think you've answered my question. 115V on the N to G x whatever mA that LED can leak (they all leak a little bit I'm told) equals glowing.

How would this "power" the bulb when one side of the circuit is broken at the switch?

If I had an LED on the bench in front of me, how would I get it to glow slightly with no power applied to one side of it?

Can you get it to glow by applying a/c power to just one side of the bulb?

Light means current is flowing somewhere somehow. If you've opened the hot then that means it's flowing between the remaining wires, the N and G. You could unhook the ground from your light and see if it goes out.

If I had an LED on the bench in front of me, how would I get it to glow slightly with no power applied to one side of it?

Not being an EE or even a very good electrician I can't explain but I bet Anton is onto something.

 

[wpns]

If I had an LED on the bench in front of me, how would I get it to glow slightly with no power applied to one side of it?

You don’t have a 2-lead bare LED on the bench in front of you. You have a 3-lead power supply (H-N-G) and you are applying 120Vac to two of the leads

 

[ABarbarian]

I think you've answered my question. 115V on the N to G x whatever mA that LED can leak (they all leak a little bit I'm told) equals glowing.

But what about when on bypass the LED still glows? In that situation there is no V difference between N-G?

 

[OzSolar]

But what about when on bypass the LED still glows? In that situation there is no V difference between N-G?

I'm not sure I'm following your question.

But I'm not surprised that it glows when there is still voltage at the N. I'm an industrial electrician by trade and we once had a situation just like this in a 480v factory. I thought we had a faulty product but in reality it met code. Still, I had to explain to my customer that some of his lights were going to glow when they were supposed to be off. Ultimately we changed to a 2-pole motion sensor and it solved that problem, well technically it just hid it because the total leakage on the circuit was well over 5mA.

Yeah... so a rough interpretation of the the below chart is that below 5mA meets code. I'd still like you to run the N in series with your multimeter and see what you get.

 

[Checkthisout]

You don’t have a 2-lead bare LED on the bench in front of you. You have a 3-lead power supply (H-N-G) and you are applying 120Vac to two of the leads

You're onto something here as it's not LED lightbulbs but LED fixtures where the LED panel or bulb are a non-replaceable part of the lamp assembly.

 

[Checkthisout]

Via filter capacitors inside LED power regulator circuit if there is some residual AC between neutral and ground.
View attachment 209319

Thank you for the diagram. I'm a little slow. Another poster made a point about it being a power supply scenario which then made it "click" in my head that the glowing witnessed in my case comes not from LED replacement bulbs put in standard lamp sockets but from LED fixture assemblies where the bulb is part of the fixture and hence the LED power supply would in fact have 3 wires as shown above and not just a hot/neutral connection like a standard bulb.

Thank you!

 

[ABarbarian]

Which inverter? Where are you located? Can you draw a diagram of your wiring?

- Thailand
- Inverter: MPP 4kW PIP-4048MS

Inverter is on an old firmware without option #38 (for allowing N-G bonding when in inverter mode)