The mystery of the intermittent fault on flourescent light

[The Ghost]

So I was called again to look at light in unit that has gone off. Of course when I got there it was on (again) The circuit is very straight forward, supply from Merlin Gerin DB 10a MCB supplying four twins. Checked all the connections for being tight including switch (wiggled and pulled) and securely fixed into termination points, all ok. Checked cable down to light from klix rose, ok. The fitting is HF with TriPhos tubes. The light stays on during this process no flicker.
So I have checked the cabling from the MCB to first light (which is the one with intermittent fault) and checked onward to the last light, four in all. No loose connections, all lights are working except this one, occasionally does not. As you may divine from the picture below it is not a situation (20 feet up) that lends itself to popping up with a tester and having investigations without a tower which is considerable expense. The cable (singles) is in plastic trunking changing to 20mm conduit as it progresses to each light. There is no sign of damage to the cables I can access, and have fought shy of putting up a tower due to cost and the landlord being a very very frugal person to put it charitably. Any ideas of a brilliant nature welcome.

 

[marconi]

I remember that the very early tubes had an earthed conductive strip along their length to aid starting. I suppose the same capacitive effect is achieved now by internal coatings and the proximity of the tube to its earthed fitting. So check too the connection of the cpc to the fitting, the continuity of the cpc back to the DB and that the reflector is electrically connected to the cpc.

 

[Jim90]

Tad overkill for the one light?
Remember that this bloke is connected to the steel structure on a metal ladder and (within our firm) has to maintain three points of contact on said ladder. Throw a loop test into the mix and a RCD protected circuit then we're in for a show!

 

[The Ghost]

a. Does the fitting sometimes not illuminate when first energised and stay off? Or does it later illuminate?

No it works for 3 months the goes off for a few hours and comes back on totally at random. Then goes off after a few more months and goes back on in a few days etc.

b. Does the fitting sometimes illuminate immediately and later go off and stay off?

No See a.

c. Is the fitting bathed in warm sunshine? The picture seems to show skylights in the corrugated roof.

No it is not in any sunshine.

d. At the first fitting at the end of the drop what is the L-N voltage before and after turning these lights on?

Same, 246v

e. Any high amperage loads on the phase?

No there are no high A. equipment it is a kitchen with gas stove.

f. At any time is the fitting operating outside its designed temperature range? ie: first thing in the morning unit is cold.

No the unit has central heating which keeps the temperature ambient

g. Has the drop to the fitting been clamped excessively tightly deforming the live conductors? (I discovered this once on my wife's iron plug - an MK screwless clamp).

No there is no mechanical damage to any part of the lighting installation

h. Is the Klix rose faulty? Change it if a check shows it is plugged in correctly.

That is the only thing that has not been changed and thought to do that next.

i. Is there anyway someone could be playing mischief with the fitting or Klix rose.

Not a chance

j. Are the four fittings wired as a daisy chain or otherwise?

Daisy chained one after the other

k. Is is a type B or C mcb?

Type B

l. what is the length of the run from mcb to first fitting and what is the guage of the live conductors?

1.5mm² 7 m.

m. Could you swap over complete fittings say 1 and 2 and see if the intermittent illumination moves with the fitting?

A lot more trouble than it is worth, might as well rewire it if I am going that far

 

[The Ghost]

reflector is electrically connected to the cpc.

This whole complex has recently had an EICR including this unit and it has been checked and passed I have read the results of all of it and this unit has R1 R2 plus Zs all ok.

 

[Lucien Nunes]

Unlike a choke, which is simple and usually either works or doesn't, an HF ballast is an electronic switched mode power supply, containing many dozens of components (some of which have finite lifespan) and connections. Intermittent faults are not only possible but likely. If the lamp is new and good and the endcaps undamaged, I would have changed it by now.

 

[The Ghost]

Change the ballast or whole fitting. Simples

Already done.

 

[The Ghost]

Get suitable access equipment

Absolutely agree there! I have a youngman tower and will be using it for any further access. Just used the ladder wedged at the bottom for a quick look at the fitting adjacent to the faulty light.

 

[The Ghost]

If the lamp is new and good and the endcaps undamaged, I would have changed it by now.

Maybe you meant if the lamp is not new? Anyway the lamp is brand new and so are the tubes. The puzzling thing is that no other lights are affected by this phenomenon on the circuit.

 

[The Ghost]

Would be easier to do that off your ladder

No way! you have one hand to do all the work there is no way you could carry up equipment and tools and then take off the cover, it is a tower job.

 

[Lucien Nunes]

No sorry I missed the 'new fitting and tubes' comment. So what you are trying to discover is whether there is an air-gap in the circuit, or whether some local condition such as temperature is causing any good fitting installed at that point to misbehave. My money would be on an air-gap.

On a related-but-not-the-answer-to-your-problem note, I have experienced 12V DC HF fluorescents misbehave, where turning on one fitting would make another flicker or go off, or it would not light when switched on. They would interact in unpredictable ways, but would also work for days without incident. The fault was life-expired electrolytic capacitors in the inverters, causing them to modulate the DC supply with their switching frequency. As soon as I replaced the caps all were stable. But this is not likely to be the cause here for any number of reasons.

 

[static zap]

Any signs of moisture , dripping ?
( Imagining -COLD)

 

[elsparko]

are the klik roses in a loft area? quite easy for these to get kicked/broken if folk are walking in the loft area and dont look where theyre putting their big clumsy feet

 

[elsparko]

ignore previous comment, picture just loaded lol, quote for a rewire and new fittings, sod messing around with fittings that you didnt even install

 

[ruston]

ignore previous comment, picture just loaded lol, quote for a rewire and new fittings, sod messing around with fittings that you didnt even install

It's ok lol, we will ignore your finger trouble.

 

[elsparko]

It's ok lol, we will ignore your finger trouble.

surgeon says i may get to keep half of it if im a good boy in surgery