UKC

Bit of a random question...

We have a light on our stairs which is wired in on a two way switch - one switch at the top of the stairs, one at the bottom. It's recently been installed and is working fine.

We then installed an LED light is one where you can adjust the light with a remote allowing you to adjust the brightness and set it to warm/cool white etc. 

You turn the light on and adjust to the settings that you want. If you leave it 10 seconds before turning it off at the switch, then the settings are saved. This works fine if you use the same switch (i.e. always the one at the bottom of the stairs), but if you use the other of the two way switches it resets the light back to default settings. 

Does anyone know enough about this kind of wiring to guess why the light might be resetting itself?

That shouldn’t happen.  In wiring terms the switches should both look identical to the light.

The only conceivable explanation I can think of is switch bounce. Even that is grasping at straws.  Changing the offending switch might fix it.

Does it matter which way you throw the offending switch to turn the light off, or does it happen in both possible directions?

That is bizarre. One of those weird issues that makes me want to get the tools out and work out what is going on.

One thing I can think of is if it's still getting power by leaking to earth when the neutral is disconnected. This would be surprising, as these sorts of bulbs are normally double insulated with no earth connection. With a two way switch setup and the light off, either both lamp terminals are connected to neutral, or both to live. With a live connection and some leakage to earth it could stay powered one way, but not the other.

As above and also grasping at straws but does your lighting wiring include a neutral? (an actual neutral not live and switch live)

My money's on a dodgy connection, open everything thing up and redo the terminations.

Silly question maybe, but are both switches plain physical switches, or are either of them electronic ones (eg. time switches, etc.)?

If I had to, I'd open with a guess of it's been wired badly so that each switch can provide either live or neutral to a lamp terminal. Probably to save installing 3+E cable (or to re-use some existing cable). It's possible then that if the circuits are different top and bottom of the stairs or even from different ends of the same circuit then under some conditions (heavy loads, the I part of V=IR, or a loose connection, the R part) there may be enough voltage difference between the upstairs live and the downstairs live (or neutral and neutral) to sustain volatile memory (<2V would probably suffice) in the lamp but not to illuminate it. It also requires a rather particular behaviour from the lamp's ACDC logic supply but that's not impossible. 

That wouldn't always give the effect you observe, if it was the case then you should be able to get the same effect from either switch by positioning the other in the 'right' state to begin with but you may somehow have missed a combination.

If you have a multimeter you could check what voltage is present across the lamp holder under all different switch states. Be careful, if my guess is right the semi-exposed bayonet part of the lamp may be live under some conditions. Or just get a sparky in, it's not right.

jk

I'm with Wintertree. As a test, replace the offending switch with another (preferably new). Maybe there is a 'return the lamp to default mode' by switching it on and off a number of times and its seeing the bounce as the reset command.

Thanks for the responses - it is good to know it shouldn't happen. 

It happened in both directions.

Unsure on the neutral. It is all brand new with 3-core cable & new switches on it's own circuit back to the fuse box. It will be whatever current regs require.

Just normal switches (one is a double, one is single)

I had another play around with it at lunch and this time it did remember the setting. One difference was that I left it on longer, and then left it longer between pressing the other switch. 

One thing which I hadn't done before was flick the same switch on and off a few times without any real pause. What happened with this today, was the light flicked between the different lighting styles (warm, white and blue?) with each on/off. 

If I left it 30 seconds between each on/off then the next time I pressed it it retained the same light as before.

So it's quite possible that the light has some kind of setting (not detailed in the manual) which allows you to change the light settings without the remote they supply. And if that's the case then it is probable that when we have been testing it, we have been moving between the two switches too quickly which has then altered the setting, rather than reset back to default which was my presumption. 

So maybe in "normal" use where you would turn it on, leave it for a while and then turn off later it will work properly. I'll have another play with it later to see if that is the case...

That makes sense as some headtorches work like that. Keep pressing to cycle through different light settings, but as soon as you leave any setting on for more than a few seconds, pressing the button again will then turn it off, regardless of which setting in the cycle it's on.

As a programmer, been there, done that. Still got the bruises.

Never underestimate the ingenuity of idiots.

An LED bulb in one room of my last house used to have a brief, dull flash every 30 s or so when turned off.  It was only visible when my eyes were dark adapted.

I figure it was some kind of parasitic power extraction, perhaps from the small voltage on the neutral line produced by loads within the house (although that would have been v. low with only the quiescent draw of the small gas boiler and of the oven) or leakage through the switch giving enough juice for the SMPS to charge the output capacitor enough for the LEDs to light up and drain it every so often. 

That was a great house in terms of electrical history; 1880s red brick mining town mid terrace, and the loft had two previous generations of power cables in it.  The oldest was twin core in lead sheath, it came in through one dividing wall and went out the next with a Bakelite junction box where a ‘T’ was implemented by twisting the three conductors together (separately) for each of the two wires.  This was made safe and secure by something that looked like a ceramic thimble, which had an internal spiral and was twisted over the twisted wires, further twisting them and insulating them.  Going even further off topic, as a left handed person it’s very frustrating working with cables twist-jointed by a right handed person…

Still got these two small ones.

Stuff like this is weird but it does happen. I wonder whether if it's possible the input filter is somehow rectifying low frequency radio energy coming in on the floating live wire to charge a capacitor?

jk

When the said light was put in, the electrician's Makita DAB radio caused some strange interference with the LED light. They definitely didn't like each other.