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

@ABarbarian What kind of light switch do you have? Is it purely mechanical or electronic? Does it have a glow in the dark light inside?

Do inverters actually have zero volts on the neutral?

AntronX said:

What kind of light switch do you have? Is it purely mechanical or electronic?

Click to expand...

In my case it's just a simple snap switch

This is what I was thinking is the cause.

ABarbarian said:

See pics, yes when glowing there is 0.09 Amps passing through

Click to expand...

It could be that case of the either the driver or fixture is grounded to whatever they are mounted to but it doesn't look like that's possible since they both appear to be plastic.

But math and Watts law tells us the you've got current flowing somehow. V x A = W If either V or A = zero then no watts. No watts = no light. We know you've got 115 volts on the neutral.

Regarding your measurement technique: That is a clamp style meter and can't be used to make this measurement. It works by putting the clamp around the wire and most of them are only accurate once they get above 2 or 3 amps.

To get accurate measurements of less than 1 amp (mA range) one needs to use a DVM and run the circuit through it in series. Move the red lead from the V to the A on the meter then select A or mA if your meter has two ranges. I've done a bad job explaining it but I'm sure there's plenty of YouTube videos.

I've tried to show it in the bottom picture.
Cut the wire on at the blue x then use you meter to complete the circuit.

OzSolar said:

We know you've got 115 volts on the neutral

Click to expand...

Just a quick reply: 115 V is Neutral to Ground, but no ground wire is coming into the lamp, roof is cement and conduit plastic so I am confident there is no grounding of the lamp, and current passing across N-G's 115V is not applicable to the glow.

No glow when the circuit is on direct grid supply makes me feel the switch is probably not leaking current through line, and there is no ground, so it feels like there is no traditional circuit here causing Amp flow.

I like efficientPV's theory that the "inverter is putting out some HF noise and the capacitance of the cable", although my understanding of this is near nothing.

ABarbarian said:

Just a quick reply: 115 V is Neutral to Ground, but no ground wire is coming into the lamp, roof is cement and conduit plastic so I am confident there is no grounding of the lamp, and current passing across N-G's 115V is not applicable to the glow.

No glow when the circuit is on direct grid supply makes me feel the switch is probably not leaking current through line, and there is no ground, so it feels like there is no traditional circuit here causing Amp flow.

I like efficientPV's theory that the "inverter is putting out some HF noise and the capacitance of the cable", although my understanding of this is near nothing.

Click to expand...

So you do have a 2-terminal device? Either your switch isn't isolating enough (try detaching the wires from your switch, cleaning the insulation and ends, and separating them by a few inches, to have a really open circuit), or there's something to the RFI scenario.

Looking at pictures of the lamp, why are the probes connected? Current is thru the clamp. Connect brown and blue together and also connected to one leg of power. That would determine if there is leakage to concrete.

Reversing brown and blue would also be an interesting test for leakage.

Per my link, I remember reading this elsewhere years ago and always thought this was the reason. My particular setup bonds the neutral and ground in the multiplus but I don't have an actual ground rod to earth. Is it possible to power the LED via induction through 1 conductor? I mean we are talking milliamps of current.
I would test by removing the ground if I was actually at the cabin where this stuff is located but I'm not and won't be for another month.

Probably because one of the ways to make an inexpensive inverter output stage requires the output to be floating.

With a normal AC sine wave of the sort you get from the wall, the “line” terminal swings between -170 and +170 V (for a 120 V RMS circuit), while the “neutral” terminal remains close to the earth or ground terminal - because neutral is actually connected to ground at one point.

You can recreate this in an inverter by generating +170 and -170 V power buses internally, and connecting the output alternately to one of +170, -170, or neutral. With this design, neutral can remain at ground, but you need both polarities of voltage.

A cheaper design is to have just a single 170 V supply, and a floating neutral. During the part of the waveform where line is supposed to be +170 with respect to neutral, transistors connect line out to +170 and neutral out to internal power supply ground. During the portion of the waveform where line is supposed to be -170 with respect to neutral, line out is actually connected to output supply ground, while output neutral is connected to power supply +170. This requires only one polarity of 170 V power supply, so it’s cheaper, but means that output “neutral” is actually at 170 V with respect to ground for part of each cycle of the output.

It is also possible to create AC at the battery voltage, using only a single polarity of supply, using either a single primary winding and a 4-transistor H-bridge (the 2nd design above), or a center-tapped transformer and 2 transistors. Then the voltage is stepped up to line voltage by the transformer. Since the transformer secondary is isolated, it can be connected with the neutral output grounded. But transformers cost money and also add extra weight.

And that’s probably why you measure voltage on neutral on your inverter: It is probably using an H-bridge type design at line voltage, without a step-up transformer."

Any relevance to our situation?

OzSolar said:

To get accurate measurements of less than 1 amp (mA range) one needs to use a DVM and run the circuit through it in series

Click to expand...

I think the attached pictures fix my measuring technique.

When run in series on Neutral getting 0.058 mA
When run in series on Line getting 0.10 mA

Last picture shows if I disconnect the positive the glow remains
If I disconnect negative but connect positive, no glow

Probably not important:
If I connect the incoming negative to the lamps positive and no other connection, have glow

wpns said:

So you do have a 2-terminal device? Either your switch isn't isolating enough (try detaching the wires from your switch, cleaning the insulation and ends, and separating them by a few inches, to have a really open circuit), or there's something to the RFI scenario.

Click to expand...

Switch is a simple mechanical switch

See attached picture, if positive line is broken the glow remains, I believe that eliminates the switch not isolating being the cause?

efficientPV said:

Looking at pictures of the lamp, why are the probes connected? Current is thru the clamp. Connect brown and blue together and also connected to one leg of power. That would determine if there is leakage to concrete.

Reversing brown and blue would also be an interesting test for leakage.

Click to expand...

I was using an incorrect multimeter in my first set of readings, I think I fixed that. The probes were connected to try and read what Amps were passing through, I just made a post about the results.

"Connect brown and blue together and also connected to one leg of power. That would determine if there is leakage to concrete" - I understand you mean the incoming neutral (blue) connect to the lamps positive (brown), but what is the leg of power? Or do you mean the lamps neutral and positive connect to either incoming line or neutral? And then what do I measure?

"Reversing brown and blue would also be an interesting test for leakage" - I tested different incoming and lamp connections: if the incoming neutral connects to either lamp line or neutral there is glow, anything to lamps positive no glow.

So it only glows when neutral is connected (blue). Seems like capacitive coupling from neutral to concrete ceiling. Set your multimeter to AC voltage and measure from LED fixture metal case to neutral. Scrape off a bit of paint to make contact. Start with highest voltage setting and work your way down until you get a reading.

Checkthisout said:

This requires only one polarity of 170 V power supply, so it’s cheaper, but means that output “neutral” is actually at 170 V with respect to ground for part of each cycle of the output.

It is also possible to create AC at the battery voltage, using only a single polarity of supply, using either a single primary winding and a 4-transistor H-bridge (the 2nd design above), or a center-tapped transformer and 2 transistors. Then the voltage is stepped up to line voltage by the transformer. Since the transformer secondary is isolated, it can be connected with the neutral output grounded. But transformers cost money and also add extra weight.

And that’s probably why you measure voltage on neutral on your inverter: It is probably using an H-bridge type design at line voltage, without a step-up transformer."

Any relevance to our situation?

Click to expand...

I am not sure it is relevant, because the glow persists when inverter is set on bypass mode and there is no voltage difference between N-G, I believe in that situation the inverter is not doing much, it is passing through the grids power (however my issue suggests the inverter or inverter cabling is doing "something" because the glow disappears when powered directly from grid).

ABarbarian said:

I am not sure it is relevant, because the glow persists when inverter is set on bypass mode and there is no voltage difference between N-G, I believe in that situation the inverter is not doing much, it is passing through the grids power (however my issue suggests the inverter or inverter cabling is doing "something" because the glow disappears when powered directly from grid).

Click to expand...

How does the inverter pass through grid power? Is it a transfer relay? Is it an inverter/charger?

I guess I can look it up.

I'm now very curious about my setup. Haven't had to run the Gen but now very curious if it happens when gen is supplying power.

Thanks for sharing. In my case I'm not worried about anything but very curious why I have the glow.

The only other oddity I have is the gfci will sometimes pop when I shut the ceiling fans off.

I did some voltage measurements, disconnected the neutral cable and left the line connected, switch off, between "neutral in" other locations:

1. When on inverter mode (no N-G bond/floating):
- to LED frame screws: 120 V
- to LED neutral: 120 V
- to concrete: 20-30 V

2. When on bypass mode (grid's N-G bond):
- to LED frame screws: 240 V
- to LED neutral: 205 V
- to concrete: 20-30 V

My assessment is it is probably a small current being pushed by capacitance. I could understand the floating N-G might cause this in inverter mode but could not understand why inverter's bypass mode caused the glow, when direct grid did not.

However the path from inverter's AC Out to breaker box and loads is a lot longer plus it has some excess cable I have rolled up, this might be causing the extra capacitance needed to push charge down the neutral line which explains why any source taking the inverter AC Out route is causing the glow.

I might unravel some of the rolled up cable and see if it affects the glow amps.