Electrical Wiring - Gremlins (Help Required).

[Dibs-h]

I have a hallway light (was just a simple pendant) controlled by a pair of 2 way switches. 1 is a traditional 2 way switch, with a common, L1 & L2 connections. The other is an old (period) toggle switch with 4 connections. This arrangement has been working fine for years, but replaced the 2 core & earth from the switch to the pendant and the 3 core & earth between the switches as squirrels had been in the loft once upon a time and eaten some of the outer insulation.

The cables were replaced and the wiring identical, wire for wire.

Worked perfectly fine with a pendant with a CFL lamp in the holder. Decided to replace the pendant with a pot light, fitted a normal halogen GU10 and works fine. Switch on from either end and switch off from the same end or the other end and it works as expected.

Replaced the bulb with an Osram LED and this is where it goes weird. Switches on fine from either end and switches off fine from the same end. If you switch on from one end and then switch off the other end - this is where it gets weird. The LED lamp flashes for 0.5s every 10 seconds.

My Megger inductive tester shows no votage present at the fitting\wires. I've checked other Live cables, so the tester isn't at fault. So I put a multimeter across the Live & Neutral and was expecting zero volts. But I get a reading of 70 or so volts, even with no bulb in the pot light.

This is where, I'm stumped! Plan A is to remove the wiring from the old toggle switch and transfer it to a new std 2 way switch and see what happens, i.e. count or discount the toggle switch. I'm wondering if the toggle switch is faulty. Problem is I'm loath the replace the toggle switch for

1. it's period
2. it's not 1 switch in the plate, but 3 (2 for other lights that work fine).
3. there isn't really the space.

Any ideas?

Cheers

Dibs

 

[EssexChris]

Does it show the flashing fault if all other lights and circuits in the building are turned off? Wondering if it’s an induced voltage on the cable charging a capacitor inside the LED light maybe.

Chris.

 

[Dibs-h]

Does it show the flashing fault if all other lights and circuits in the building are turned off? Wondering if it’s an induced voltage on the cable charging a capacitor inside the LED light maybe.

Chris.

I don't think so. The upstairs lighting circuit is on it's own fuse\circuit and all other lights in that circuit were switched off.

Cheers

Dibs

 

[RogerS]

Job for AskTheTrades

 

[ankledeep]

Get a sparky in...get him to test (I think) R1+R2 and Earth loop impedance

question...are you on PME or do you have an earth spike ...or...horrors of horrors by some mis chance..BOTH at the same time which is possible if for example your house is PME but your shed had to be as they call it TT because it has external sevices in it (IMO TT for a shed is vital since you can stand on true earth outside and touch the case of a machine IN the shed...BUT this has to be done very carefully the earth from the house must NOT be connected to the earth provided for the shed via the earth rods.) NOW not trying to be rude...but if that was gibberish to you.....GET A SPARKY.....

 

[nev]

just to confirm - no dimmer switches, no transformers?

 

[Dibs-h]

No dimmer or transformers. Plain old 240v pot light. The earth on the house is from the supply (no earth spike) - if that makes sense.

Dibs

p.s. Might stick the multimeter on the toggle switch and see where this 70v is coming from. Failing that, replaced a 1 gang 2 way switch earlier today, due to it looking cruddy (paint on it etc.) so might just wire that in, instead of the toggle and see what happens.

 

[ankledeep]

No dimmer or transformers. Plain old 240v pot light. The earth on the house is from the supply (no earth spike) - if that makes sense.

Dibs

p.s. Might stick the multimeter on the toggle switch and see where this 70v is coming from. Failing that, replaced a 1 gang 2 way switch earlier today, due to it looking cruddy (paint on it etc.) so might just wire that in, instead of the toggle and see what happens.

In that case i suggest you deffo get a sparky to check R1+R2

since in a pme supply neutral is connected to earth...neutral should be at earth potential at all times....SO either your neutral has gone high resistance at some point or your earth/neutral bonding point has done similar...either way it needs looking at..... try your multimeter across N/E

 

[RogerS]

Essex Chris is correct. Somewhere along the cable run it is adjacent to another cable..could be ring..whatever. Anyway, it is picking up induced voltage (either capacitive or inductive). That is the usual expanauion for other types of lamps. No idea how LEDs are configured but suspect that is the cause. Cure? 100k resistor across the LED.

Not convinced it has anything to do with neutrals etc ...He only changed the bulb!

 

[Dibs-h]

Essex Chris is correct. Somewhere along the cable run it is adjacent to another cable..could be ring..whatever. Anyway, it is picking up induced voltage (either capacitive or inductive). That is the usual expanauion for other types of lamps. No idea how LEDs are configured but suspect that is the cause. Cure? 100k resistor across the LED.

Not convinced it has anything to do with neutrals etc ...He only changed the bulb!

The 3 core & earth and 2 core & earth that comes from the toggle switch run right alongside the Ring and also the Live line to the pot light. Wonder if I should separate them and run them along different joists?

Is it normal'ish to put a 100k resistor across the LED, i.e. the Live & Neutral to the pot light?

Cheers

Dibs

p.s. I have another LED GU10 (el cheapo) so might swap that in and see what happens.

 

[OLD]

Check the neutral is at earth potential .

 

[Dibs-h]

Put the meter across the N & E and zero volts. Didn't have a great deal of time, but given that I was a little annoyed with it, removed the 3 core & E, and had the light just operating from the toggle switch (obviously using common and L1). Then put the meter across the pot light terminals and now reads 23V - so down from 70 something. Which make me wonder whether the 3core & E is having the other 50V induced in it. The cable was new and had been sat on a reel (admittedly for some years), new nonetheless. Putting the meter across the toggle switch it reads 205V - whereas an adjacent switch - part of a 2 way switch with one downstairs - reads 240'ish. So how's that for odd?

Plan is to now pull the wiring back from the pot light and have it just "hang" in the air as opposed to thru the joists and make sure none of the wires run near each other and see if things change.

Thank God Wifey & kids are at her parents - get some right earache!

Cheers

Dibs

 

[Dee J]

Many modern lamps have problems .......

1) Anything other than basic incandescent lamps have a fancy driver circuit - generally a type of switch-mode power supply. The input of these supplies generally rectifies (turns to dc) the incoming mains and then smoothes the resultant noisy dc with a large capacitor. When there is enough voltage on this capacitor the lamp will light.

2) If the domestic switch wiring is long (either a long single run or especially a two-way setup - because it uses even more wire) then problems occur. Long runs of cable form a small capacitor. Capacitors conduct ac. So there will always be a tiny current flowing when the switch is off. Not enough to cause any prolems or waste significant energy. But it's enough to slowly charge up the big smoothing capacitor in the lamp.... that leads us back to (1) and the lamp lights for long enough to discharge the capacitor.

What effect you get depends on a combination of the lamp design and the cable lengths in your house. None of it is a fault, just an interesting combination of circumstances. Different makes of lamps display this to a greater or lesser effect. Try another lamp.

Trying to measure voltages on cables that form part of such circuits can give confusing readings. Modern digital meters have a very high input impedance and can respond to the currents cause by wiring capacitance and their resultant voltages.

If in doubt call an electrician.

Dee

BTW aren't capacitors wonderful.... conduct AC, store DC...

 

[Dibs-h]

Many modern lamps have problems .......

1) Anything other than basic incandescent lamps have a fancy driver circuit - generally a type of switch-mode power supply. The input of these supplies generally rectifies (turns to dc) the incoming mains and then smoothes the resultant noisy dc with a large capacitor. When there is enough voltage on this capacitor the lamp will light.

2) If the domestic switch wiring is long (either a long single run or especially a two-way setup - because it uses even more wire) then problems occur. Long runs of cable form a small capacitor. Capacitors conduct ac. So there will always be a tiny current flowing when the switch is off. Not enough to cause any prolems or waste significant energy. But it's enough to slowly charge up the big smoothing capacitor in the lamp.... that leads us back to (1) and the lamp lights for long enough to discharge the capacitor.

What effect you get depends on a combination of the lamp design and the cable lengths in your house. None of it is a fault, just an interesting combination of circumstances. Different makes of lamps display this to a greater or lesser effect. Try another lamp.

Trying to measure voltages on cables that form part of such circuits can give confusing readings. Modern digital meters have a very high input impedance and can respond to the currents cause by wiring capacitance and their resultant voltages.

If in doubt call an electrician.

Dee

BTW aren't capacitors wonderful.... conduct AC, store DC...

Cheers for the post Dee. Originally when it started to flash, I put the meter on the pot light terminals and watched as the voltage climbed to 70V, the lamp flashed and the voltage dropped to 20V or so, only then to climb back up and the cycle to repeat. So see what you mean about the charging & flashing.

Problem is Wifey wants pot lights in that hallway & I don't mind, but want to fit LED ones. I suspect all mains powered GU10's will have similar issues due to the drivers in them.

I wonder if do call an electrician in - can he(she) cure it? Or just charge me for explaining\confirming it?

Dibs

 

[Racers]

Hi, Dibbs

Sounds to me like you have a bad connection some where on the netural, causing it the voltage rise, check all the neturals on all the ceiling roses.

Pete

 

[Digit]

The idea of an inductive pick up from the ring is easily tested. Turn the ring off!

Roy.

 

[OLD]

Putting the meter across the toggle switch it reads 205V - whereas an adjacent switch - part of a 2 way switch with one downstairs - reads 240'ish. So how's that for odd?

So (using high resistance meter) the switch is o/c. live one side and return via the lamp 'filament other side, ok on one cct, low volts on faulty cct i would also be looking for a in line resistance which would cause a volt drop of the induction and drive this low current device but not be noticed by a ordinary bulb.
you will find this problem before any spark, draw the cct on paper and work through it.

 

[Deejay]

Afternoon Dibs

Found this in a post by a chap called Felix on DIY Not ...

http://www.diynot.com/forums/viewtopic.php?t=21025

-------------------------------------------------------------------

No doubt about it; it's capacitive coupling. If it wasn't a two way switched light I would say one of your earths was faulty. Here's why.

In a flat twin and earth cable there is virtually no coupling from live to switched live because the earth wire is in the way - UNLESS THE EARTH IS FAULTY. Check this first.

In a three core and earth cable this doesn't apply. Only one core is shielded by the earth wire. The other two are coupled. This is of no consequence with filament bulbs but if the cable is very long you could get flickering with a small fluorescent bulb.

There are two ways to wire up a two-way lighting circuit. (Actually, there are three but one of them isn't safe.)

The modern system connects all three terminals of one switch to the matching terminals in the other one. Live power goes to L1 and switched live comes from L2 - or the other way round, it makes no difference. There is no way to eliminate the capacitive problem from this set-up.

The old system had live power going to COM in one switch. L1 and L2 were connected to matching terminals in the other switch and switched live came from its COM. You can convert your switches to this arrangement with the addition of a single piece of choc block. Remove both live feed and switched live from their terminals. Connect live feed to COM instead after removing the wire that was in there. Use the choc block to connect switched live to the wire removed from COM.

To eliminate capacitive effects from this circuit make sure that the wires between L1 and L2 are on the same side of earth in the cable. Switched live will then be shielded from them by the earth wire.

Read more: http://www.diynot.com/forums/viewtopic. ... z1rkOCRuiX
-------------------------------------------------------------------------------------------------

The different wiring arrangements are shown here ...

http://www.tlc-direct.co.uk/Technical/D ... agrams.pdf
---------------------------------------------------------------------------------------

Sounds reasonable to me so worth a try, especially checking the earth continuity.

If in doubt, consult a spark.

Cheers

Dave

 

[Dibs-h]

Much appreciate the posts.

Old - what's "o\c." mean? Sorry to sound dense. Obviously know what cct means!

Work thru it - oh yeah! It's pineappled me right off - every tried showering in the dark? The bathroom lights are after this one & I removed the line out to not confuse matters! The plan is to get it working with just a std switch and then move it forward.

But will check the N's in the other ceiling roses. Would I be right in thinking that because the Earth comes with the supply (PME would that be the right term?) Neutral to Earth should register zero on the multimeter.

Roy - good idea, will be pulling the fuse on the ring and see what happens.

Did across the following

http://www.diynot.com/forums/viewtopic. ... 34#1810534

and he mentions

Capacitor 0.047uF micro Farads 250 volts AC
Resistor 100 ohms

Which come pre-packaged as a contact suppressor from RS Components

RS Stock No. 206-7847
Manufacturer Evox-Rifa
Manufacturers Part No. PMR209MB5470M100

I have fitted them in ceiling roses between neutral and switched live

There are other sources of the contact suppressors

Read more: http://www.diynot.com/forums/viewtopic. ... z1rkW3QdCu

have an account with RS, so I wonder if that will be the fallback plan!

Will post back on progress.

Cheers

Dibs

 

[Digit]

O/C is Open Circuit, opposite of short circuit, yes earth and neutral are commoned at the sub station and as necessary.

Roy.