48 volts with switch off!

[Jeff Wisnia]

Get yourself a "Y" socket adaptor like this:

With the wall switch OFF, screw the adaptor into the socket of interest and put an LED bulb in one side and measure the "switch off" voltage in the open socket.

Betcha it'll be a lot less than 48 volts then.

Now, screw a regular incandescent bulb into the open socket.

If the LED bulb stops glowing when the incandescent is put in then I agree with the other posters who said that capacitive coupling, probably in a "switch leg" piece of cable is letting enough current run through the LED bulb to make it glow.

You could experiment by screwing a plug adaptor like this:

into the open socket so you could easily stick the leads of different value carbon resistors into its slots to see if you could get the LEDs to stop glowing with a resistance value high enough so that if it was permenantly wired across the socket it wouldn't generate much sensible heat.

I'd start with a one watt 20,000 ohm carbon resistor and see if that works. If it does, try an even higher resistance, say 100,000 ohms. If you can get the LEDs to stop glowing with a resistor of 20,000 ohms or greater, then you could wire resistors of that size directly across the fixture socket(s).

Let us all know how you make out.

Jeff

[Percival P. Cassidy]

I don't need the "Y" adapter, because this is a multi-lamp setup anyway: three lamp fittings on a track. The 48V reading is with identical LED bulbs in the other two fittings.

Yes, with a regular incandescent bulb in one fitting, the LED bulbs in the other two fittings do not glow.

That's an idea. I'll try it when I can get to the store and buy those items.

I will.

Perce

[Percival P. Cassidy]

OK, but

(1) there are no overhead power lines within a mile. Walking around the house with that meter and a lead trailing like an antenna I see not even a 1 volt reading.

(2) It's a pain that the LED lights don't switch off completely. The voltage with the switch off is sufficient to keep them glowing dimly. Perhaps replacing one of the bulbs by a low-wattage incandescent would hold the voltage down, but that offsets the power saving benefit of the LEDs.

Perce

[Percival P. Cassidy]

LED screw-in replacements for incandescent bulbs. Our local Sam's Club has a few different varieties of them. I have a few that are claimed to be 45W equivalent in light output for a consumption of 3.5W. The light is somewhat blue, but it's fine for my purposes and works fine as a replacement for the previous 50W incandescent flood.

No timer. Just an ordinary mechanical wall switch.

Perce

[Art Todesco]

Possibly a lighted mechanical wall switch?

[Ralph Mowery]

Try taking one of the wires off the switch , or replace the switch. It may be possiable the switch has developed some leakage across it. Maybe someone tried cleaning it and got some fluid insided the switch.

[clare]

Cover the bulb - make it completely dark - the voltage will likely go away. You are LIKELY getting a DC voltage FROM the light because an LED is ALSO a photocell. An LED will produce a DC voltage when exposed to light. Green ones are Gallium Phosphide and make 1.65 to 1.74 volts each. Not sure what the white ones are and what voltage they produce but I think they are Indium Gallium Nitride - so likely around the same output.

[Percival P. Cassidy]

No light in the switch.

Perce

[Percival P. Cassidy]

It's only because I walked into a supposedly dark room that I discovered that the LEDs were not totally dark.

There was no ambient light that could have caused any photovoltaic effect.

Perce

[clare]

And remember - White LEDs also have a phosphorescent material in them that makes white light from blue - the Phospor actually produces a yellow light.

[JIMMIE]

I got the same results when I put some LED lights on a light activated switch that used a solidstate switch. The fix was to change it to one that used an internal relay. These can be hard to find.

Jimmie

[Art Todesco]

O well, I just thought of that. BTW, did you try disconnecting the lamp? Maybe the phosphors are still glowing a little .... I know, that's a big stretch!

[Jeff Wisnia]

Oopsie, I assumed 48 volts with NO bulb(s) in the socket(s).

Have you tried that (No bulbs, but with the same meter) and if so does the voltage read higher than 48?

Well, that shoots the phosphorescent glow idea down even more.

[bud--]

I agree with Jeff (and others).

A common test device is a simple neon light with 2 tests leads. If you plug one lead into the hot side of a receptacle and hold the other lead the neon light will light up (very dimly). This works if you are isolated with only extremely high resistance path to anywhere. There is a *very* small capacitance from your body to ?the world?.

The capacitance between parallel wires is far higher and could drive a small current to the LED bulbs and produce a small amount of light.

[Jeff Wisnia]

Yup, that human body capacitance is commonly accepted as being 100 picofarads for an avergage sized h*mo sapiens.

That's the capacitance which stores a high voltage charge when you scuffle across a rug, and then discharges to something else (like your cat's nose) when you get too close to it.

Jeff

[Percival P. Cassidy]

I hadn't measured the voltage in the second room, but I assumed that the glowing LEDs after switching off had the same cause. In fact this was simply a brief "decay time," probably as a capacitor discharged.

It turns out that it was a lighted switch after all -- but no light had been visible in the switch with the CFL bulbs that I replaced by LED. And even when I had the switch hanging out of the wall. the "Lighted" label was on the side away from me.

Perce