Fluctuating voltage on a lighting circuit - advice please!

[The General]

Evening all.

I got called out to investigate a problem with a two-way switch on a landing light and I was expecting to find a DIY bodge with incorrectly wired 3-core. However, it was more interesting than that!

Customer reported that when the light was off, after 5-10 minutes it would briefly come on/flicker and then go off again and then repeat - but it also depended upon which of the two 2-way switches was off/on.

Assuming (I know, I know!) that there aren't any junction boxes with wires being swapped over inadvertently then the wiring is correct, so I wanted to check that the switches were operating correctly and tested voltage to earth from the L1 and then the L2 terminals, before flicking switch and checking again. Interestingly, when I checked voltage between L1 and L2 the reading kept fluctuating between around 180v and 220v and I don't mean that I checked it several times and got different readings, but as I was watching the figure went up,down,up,down in a regular pattern. This was repeated when I checked the L & N terminals on the ceiling rose with the light on, and it read a fairly constant 40V when the light was off.

I'm guessing that as the lamp in the light was a CFL this voltage is enough to have caused the flicker every 5-10 minutes, but I'm struggling to understand what is causing the fluctations?! With every other circuit switched off and just the lighting circuit on the voltage was still present.

Also on that circuit were smoke alarms, which I took out (no change) and was reliably informed by customer that there were no aerial boosters or similar items in the loft. Solar panels on roof, but I'd also isolated these when testing.

Any suggestions for what else I could be looking for, and (if anything) what to do about it!?

Thanks for your help/advice.

 

[Octopus]

I would dead test all the cables and compare the readings. My money is on a poor connection or 2

 

[Lucien Nunes]

If the lamp is an LED or CFL, the cable capacitance between PL and SL sides of the switch circuit (including the strappers) is sometimes enough to charge up the reservoir capacitor in the driver circuit until it fires briefly, discharges the capacitor, and cycle repeats.

Different permutations of 2-way switches may change the capacitance as the layout of cores in the strapper cable will change relative to the position of the CPC. E.g. if two cores one side of CPC are PL and SL, the lamp might flicker, but if they are either side of CPC that will almost eliminate the capacitance and stop the effect for that length of cable.

 

[The General]

If the lamp is an LED, the cable capacitance between PL and SL sides of the switching circuit will be enough to charge up the reservoir capacitor in the driver circuit until it fires briefly, discharges the capacitor, and cycle repeats...

Nope, it's a spiral CFL.

 

[Lucien Nunes]

Sorry, meant to put LED or CFL, the same is true for the CFL drive circuit. See also expanded post above...

 

[The General]

Oh, I should have mentioned - the voltage was still fluctuating even when there was no lamp in the fitting.....

 

[Marvo]

If your tester is a high impedance input digital type it will register the voltage that's being coupled across the wires. If you do the same test with analogue voltmeter you won't see the voltage. AsLucien suggests you can try changing which wire you have as the strapper or you could fit a snubber.

 

[The General]

If your tester is a high impedance input digital type it will register the voltage that's being coupled across the wires. If you do the same test with analogue voltmeter you won't see the voltage. AsLucien suggests you can try changing which wire you have as the strapper or you could fit a snubber.

Just for clarification so I'm sure of what 'strapper' means - you're talking about which core of the 3-core cable is used to the Common terminals at the two-way switches.....right?

 

[Lucien Nunes]

Hmm, I misread the bit about the voltage at the rose. You say it was fluctuating regularly with the light ON but steady with the light OFF. But the light was flickering when supposed to be OFF and lit steadily when ON. Or am I still misreading this?

 

[Lucien Nunes]

Also, you mention voltage between L1 & L2. The significance of this depends on how the switches are wired. If PL and SL are paralleled to L1 & L2 with COMs linked with 3rd core of 3+E, then the supply voltage in series with the lamp will appear across L1 & L2 with the lamp off and near zero with the lamp on as they are then linked together. If PL goes to one COM and SL to the other, and just L1 & L2 strap between the switches, then with the lamp on one strapper will be O/C and the voltage you read will have no significance, as it will just be the induced voltage vs. your meter input.

So what that reading tells you can only be understood if we know how the switch drop connects to the switches.

 

[Marvo]

Just for clarification so I'm sure of what 'strapper' means - you're talking about which core of the 3-core cable is used to the Common terminals at the two-way switches.....right?

Yep, that was what I was suggesting but was struggling to find the English to explain it

Having read Luciens last post I'm also wondering if I'm fully understanding your problem though....

 

[The General]

The PL & SL were in parallel at L1 & L2 along with the blue & red cores of 3-core. The yellow core ran to the COM terminals at both switches, with blue & red at L1 & L2.

 

[The General]

Hmm, I misread the bit about the voltage at the rose. You say it was fluctuating regularly with the light ON but steady with the light OFF. But the light was flickering when supposed to be OFF and lit steadily when ON. Or am I still misreading this?

Fluctuating regularly with the light on - and light was working fine.
Steady at around 40V with the light off - and light was OFF, but (reportedly) it would switch on/flicker/go off every 5 minutes or so. I didn't witness this!

hope that clarifies. I'm pretty sure the client wasn't making it up ;-)

 

[Lucien Nunes]

Right, so in this configuration with the lamp switched off and the meter between L1 & L2, the meter is in series with the lamp and supply, so the lamp capacitor could charge cyclically through the meter input resistance and make the reading pulse.

With R & B for L1 & L2, then with both the switches one way the COM will be the same side of the CPC as the strapper it's connected to, and with them both the other way it will be the same side as the opposite strapper, in which case the capacitance comes into play and causes the lamp to flash.

Meter in parallel with lamp when off could be discharging capacitive leakage to a level that doesn't cause the lamp driver to fire, hence no flashing and steady 40V.

Only thing not explained at all yet is the varying voltage at the rose with the lamp on. Did it vary at the origin?

 

[The General]

Only thing not explained at all yet is the varying voltage at the rose with the lamp on. Did it vary at the origin?

Sorry, didn't think to check.

 

[The General]

Apologies - gotta go - wife wants to watch 'catch up' telly on the computer, so I won't be able to answer any more questions! Will check again tomorrow - but thanks to all for your advice and guidance so far.
cheers

 

[Lucien Nunes]

Of course none of the above rules out bad connections and possible arcing across faulty accessories causing random effects, but most of the findings can at least be explained in terms of cable layout, induced voltages and the way CFL innards work.

 

[Marvo]

Sounds like two different problems and possibly two different solutions to them. I'm not entirely sold on the fluctuating voltage when the light is on, usually even a small voltage fluctuation (5-10%) will result in the lamp getting noticeably dimmer and brighter although with CFL maybe not so much. I'd be interested to know how large the fluctuations were when the light was on.

 

[Tidy Max]

I have just had to draw out Luciens explanation. Interesting stuff. I didn't know the cap would charge off the meter.

 

[MarkyRR]

Something that may help you.....have a look at the other roses on the circuit. I was called to a similar issue last week. With experiences in the past, I went for a physical look first of all.....unscrewed a couple of ceiling roses to have a quick look for anything obvious. 2nd one I remove, I discovered burned out N terminal, sheathing burn away on one N loop leg. Stripped everything back to good cable, replaced rose with new item & all perfect. No more issues, everything testing as it should.

Even though I had asked customer if shed had any work done recently (answered 'no'), turns out TV aerial guy had wired a socket outlet for his booster box into the pendant!

By the way, isn't "strapper" a term we all learned way back in college??