Lamps failing too soon

[pen]

We have a persistent problem with light bulbs failing much earlier than their projected lifespan would predict. GU10s seem to last about 12 weeks, tungstens about 10. And really irritatingly, the jolly expensive Megaman low energy compact fluorescents which I have used to replace a number of GU10s are blowing too - a projected 5 year lifespan has gone down to about 6 months. There is a known problem with AC frequency here (Gibraltar - British standards) - is that enough to cause these problems or should I be looking elsewhere and giving the Electricity Authority a hard time? Is there any way I can protect my lamps, other than changing every appropriate switch to a "soft start" dimmer (yes I know that most fluorescents can't be used with dimmers).

Some of the lighting circuits are extensions off the original, 15 year old wiring - they were done by a qualified and well recommended local electrician.

The worst of it is that I have sufficiently bad vertigo to dread changing a downlighter in a 9 ft ceiling.......

[John]

Try measuring the supply voltage. I used to work in a location that had the same syptoms and the supply could climb to 260V when the load was light (no pun intended).

[Dave Plowman (News)]

Life of all lamps is in hours used.

[Sparks]

Out of interest, where do you get yours? The chepest I have found is some bloke off eBay (£8.65 for 11w ones if I but

18 of them!)

Sparks...

[Dave Liquorice]

Yep and unless you track each light bulb you can confuse yourself if you have several of the same lamp in one fitting or room. If you have 6 lamps in one room each individual lamp may last 2000hrs but you'll probably be replacing one every 300 or so hours.

[meow2222]

only clever people put lights where theyre a problem to change bulbs. I remember seeing one where not only was it unreachable, but even scaffolding was complex to erect due to the stair layout. And anyone falling off would have dropped a long way. Gotta wonder what Einstein put that there.

in nearly every case like this its down to the user not doing the calcs. If you have, and are coming up short, monitor the supply voltage with an rms meter. Frequency makes no difference. Soft starting filament bulbs has almost no effect on life.

NT

[Dave Fawthrop]

|On Sat, 25 Feb 2006 22:46:19 +0000 (GMT), Dave Plowman (News) wrote: | |> Life of all lamps is in hours used. | |Yep and unless you track each light bulb you can confuse yourself if you |have several of the same lamp in one fitting or room. If you have 6 lamps |in one room each individual lamp may last 2000hrs but you'll probably be |replacing one every 300 or so hours.

As OP states that he suffers from vertigo, and finds replacing lamps difficult, he may wish to use the Industrial method of replace all lamps at the same time, so he will have to use the steps at 1/6 of the frequency.

[pen]

I can be fairly certain that they're blowing too fast since the installations are quite new - the downlighters were fitted at the end of August last year, and we don't run the lights day and night! The GU10s are particularly irritating since they aren't even what I specified - unfortunately I was away when the work was carried out and returned home to find a fait accompli. In the end I had to get some of them changed for the original spec R80s as the light levels were far too low - this was done in mid-January this year with GE lamps, since when I've had to change 90% of them. The Megaman replacement CFs were also put in in January, about 850 hours ago - three out of 24 have failed in 2 differnt types of fitting.

One of the GU10s blew so badly it destroyed the circuitry in the dimmer switch.

The remaining GU10s are now all running off soft start dimmers (again as originally specced, but not as originally installed - grrrhhh!) I hope that this will greatly prolong their life - I know of an 18 lamp chandelier which has not had a bulb change in 4 years since it was given a soft start.

[pen]

The elctrician supplied them, at vast markup :(

I have bought them in the UK from bltdirect, but they weren't as cheap as yours.

[pen]

I was not happy about downlighters in a 9 ft ceiling but hubby and I couldn't agree about any other ceiling lamp so I agreed to R50 downlighters controlled by soft start dimmers. What got installed were GU10 downlighters off bog standard dimmers. The dimmers have now been changed after I obtained the correct ones. As I mention in another post, my experience of soft start is that it does extend filament life

- though I'll let you know if I'm proved wrong.

Our flat is a duplex, and one (outward opening) window has been set on the stairwell 10 feet above floor level. As it is also 6 stories up you won't be surprised to learn that washing it is not a priority.

[jimwatt]

Problem is that the life of incandesant lighting decreases exponentially with voltage.

In Gibraltar this is 240v rather than the UK 230v nominally but can rise to 250v for a considerable time (according to the logs of my UPS)

The solution is to fit the low energy bulbs available from Morrisons.

However, its also worth checking your installation as a neutral fault can cause higher voltages - I've seen 320 in one place, and quickly switched everything off !

The frequency does not wander around enough to cause problems.

-- Jim Watt

[meow2222]

What this tells me is you need to understand and apply some basic statistics.

Re soft start dimmers, these require electronics in the way to control startup, and these tronics drop some v, thus the lamps run less effiiciently and lasts much longer. As said, the soft start itself does more or less nothing to extend lamp life. If you dont believe me, go read up. Try some of the leading lighting expert sites.

Filament downlighters 9' up is a terrible combination. You've got the worst of energy inefficiency on 2 counts, the highest bulb change rate of all lighting types, and the most difficulty changing them. If I were you I'd install something more well thought out, eg cfl uplighters at a reachable height.

Re lamp statistics, if 3/24 have failed after 850 hours, whats the average lamp life? I think you'll find its good, you just havent done the maths.

NT

[pen]

I'm not an expert or I wouldn't have asked the question in the first place. My vague understanding is that lamps tend to go when they are switched on as the filament is cold and hence has a lower resistance which means a higher current and hence the increased possibility of breakage at weaker points along the filament which would function quite happily under "normal" operation. The soft start means that the voltage is adjusted to accomodate these initial conditions, thus helping to avoid the breaks. And I'd be quite happy if these lamps lasted as long as they ought to, I'm just extremely irritated by so much premature failure.

If the CF lamps are meant to last 5 years and I've only had them in a month or so, how can I possibly do the stats? Failure of 1/8 of the batch after only 1/60th of the stated life span seems pretty bloody high to me.

[meow2222]

This approach was popular in the 1950s, for the reasons you state. Since then it has been found soft starting makes almost no difference to filament lamp life. Googling can confirm this, or ask the people in news:sci.engr.lighting.

OK, you can get a rough figure this way - it may not be accurate but is easy:

Average lamp life will be apx number of hours when 12/24 lamps have died.

If 3/24 have died after 850 hours of use, we can roughly say 3 will die per 850 hrs use, thus 6 will be dead after 1700hrs, and 12 will be dead after 3400hrs. This gives the mean life at 3400 hrs, though not accurately. Normal life for a CFL.

NT

[Tony Williams]

One solution could be to power the lighting ring via a step down autotransformer, made from say a standard 240V/24V industrial transformer. ASCII art below, view with fixed-width font.

Vin ---------+ | *) '24V' ) winding ) | +----> Vout = Vin * 240/(240+24). | *) Vout = 227V when Vin = 250V. '240V' ) winding ) " = 218V " " = 240V. | N------------+---->

If the transformer's 24V winding is rated for say 5A, then the possible output VA is 227*5,= 1135. Probably best though to use a 24V winding with at least 2x the current rating required. This would be better for switch-on surges.

[Roger]

The message from [e-mail] contains these words:

I don't think that follows. If the 24 bulbs had been a representative sample failures would be normally distributed and will occur with increasing frequency as the midpoint approaches. At a rough guess the next 9 should now fail in less time than the first 3.

I am also puzzled by the 850 hours quoted by the OP. Surely that cannot be the time the lights have all been on. They weren't fitted until some time in January and elsewhere the OP said little more than a month. 850 hours is about 35.4 days so unless they have been on continuously actual usage is going to be substantially less than 850 hours.

[meow2222]

The bathtub curve effect should see that theyre not normally distibuted, so its more complex than that. It was just a simple rough approximation.

Msybe the OP is using them 24/7, some lights are used that way.

NT

[Roger]

The message from [e-mail] contains these words:

snip

I am not sure what you mean by a bathtub curve. You could of course mean a bell shaped curve but that *is* the shape of a normal distribution. The distribution pattern you were suggesting, if plotted against time, would be a horizontal line.

Not outwith the bounds of possibility but not exactly commonplace in a domestic situation.

[meow2222]

> Msybe the OP is using them 24/7, some lights are used that way. >

indeed. There are really too many uncertainties. As the OP didnt seem statistically minded, I just tried to offer a quick simple ballpark figure.

NT

[Roger]

The message from [e-mail] contains these words:

this response as I was away for a long weekend and have only skimmed the group since I got back.

There seem to be a good many ifs and buts in the article but the implication is that the first 2 sections of the graph should represent only a small proportion of the total population but I wonder how it is that they have 100 years worth of data for electronic components (50% of which are still going strong after that 100 years).

I can't remember if I have mentioned it before but it has crossed my mind that bulb life would be significantly reduced if our OP had been supplied with bulbs that had actually been intended for a part of the EU that operates at the lower voltage limit instead of the UK at or near the upper limit.