I've got GU10 LED downlighters in my extension, supplied by the builder's subcontractor (so a bit arm's length).
So far, about 8 months later, four of the twelve have failed. They start flickering for a few days, then go very dim. On about 12hrs a day.
Contacted the supplier (Meridian) who wouldn't comment on how long they were supposed to last [see below], but sent me two replacements.
The box says "2yrs Guarantee", which is 34hrs a day at the "average life" of 25k hrs.
Hmm.
If the MTBF is 25k hours that means some proportion of them are going to fail before 25k hours. If that was a median TBF it would be half of them. But they've run about 3k hours, so it's hard to think of a distribution where a third of them would have failed at that point.
I suppose this is the hazard of fitting lights with unreplaceable lamps - it only takes one of them to die to make the rest of the installation look bad.
Theo
GU10 suggests replaceable lamps rather than permanent ones. The usual problem with LED lamps is their ability to shed heat, and that is heavily dictated by the enclosure. Many GU10 fittings will leave the lamp completely exposed - and one would expect that to give adequate cooling. Fully enclosed in a downlighter - especially if covered in insulation could result in much higher run temperatures and so dramatically shorter life expectancy.
Average doesn't mean minimum. I expect 25k hrs is average under ideal conditions. Having 4 fail in under 3k hrs seems surprising, but maybe not that surprising.
Why did they only send two replacements? If 4 failed in under 2 years they should send 4.
FWIW I have about ordered about 30 GU10 LED in the last 2 years, cheapest I could find, none have failed. At about £1.5 per bulb. I wouldn't be that fussed if a few did go. Not Meridian bulbs.
Look on the bright side at least you didn't by an integral LED downlight (why would anyone???)
Not really, these things don't follow idealised statistical distributions. The conditions for statistical limit theorems don't really apply.
I thought high early fail rates are common in electrical components, after that they may have much higher expected life times.
Yep.
I have had a very low failure rate. As others have suggested, it seems to depend on whether the bulb is sufficiently ventilated. As a matter of interest, does anyone know whether LEDs situated outside last for longer?
Yes, I've had pretty LED table lamps run hot and fail, but I can't see any reason a ceiling downlight wouldn't be well ventilated, many were designed to handle halogen bulbs.
No, it shows the folly of putting LEDs in an inappropriate fitting.
GU10 downlights tend to be fittings with almost zero airflow to provide cooling and heat kills LEDs (and the power supply electronics)
But not to provide any cooling for the bulb. The fitting is designed to run hot as can be seen with ceramic bulb fitting (rather than plastic)
About 12 years ago we had an extension built with 20 GU10 downlighters with mains halogen bulbs. After doing a quick bit of multiplication and realising just how much power they were using I ordered some cheap led GU10 bulbs off of eBay. Within about two months about half a dozen of them failed. The vendor replaced them all and I?ve not had a single failure since despite these being on for many hours every day.
GU10 fittings aren?t the worst offenders IMO. G9 fittings on the other hand...
Tim
The so called rated life seems to be a lab. one, where they are kept cool. In normal use, pie in the sky. Hence only a 2 year warranty.
Scott used his keyboard to write :
I have had two fail, in the same enclosed fitting and outside over recent times. So much depends upon the ability for cooling air to cool the electronics down. I have since modified the fitting so as to allow some air flow.
I have none indoors which lack a good exposure to air flow and none have failed indoors.
Halogens give out 10 times as much heat. They need to have cooling for the bulb. Sure the equilibrium temperature for a 50W halogen may be much greater than a 5W LED would tolerate. But the equilibrium temperature for a 5W bulb will be much, much lower. I will leave one on for a couple of hours and measure the temp.
To put things into perspective a friend used halogen GU10s and was constantly having to replace them. She switched to LED four years ago and as far as I know none have blown. Even if they do go, they are so cheap they would need to blow far more often than they do to annoy me. I don't call an electrician to change a light bulb.
My suspicion is that early GU10 LED bulbs were unreliable, and manufactures seized upon this to promote a product with a higher profit margin, integral fittings. The comments you see from electricians etc, that you see online are probably from people who have read the propaganda rather than accumulated enough evidence to form a valid opinion of their own.
But WTF do I know :-)
FWIW my early bedside lamp LEDs failed (not GU10), I leave one on 24hrs a day, the more recent ones of similar design haven't.
Even 5w generates an awful lot of waste heat, so far as the electronics of LED's are concerned. I prefer electronics to run <50C, their life is reduced at much above 70C. Really, the cooler running the better.
Simple question is how warm are the bulbs when one has been left on for a few hours? That almost invariably will explain the shortened life.
If it is too warm to touch then it is almost certainly drying out its PSU capacitors which will result in premature failure. Electrolytics and heat do not get on. It might be worth your while taking it apart to check how it has failed (as I recall you are an electronics engineer).
My money is definitely on capacitors that don't any more. The other common one I have seen is single point failure on one LED die in a series string but that doesn't do the PSU flashing thing.
They are good for the rated hours if they are able to dissipate away their waste heat but downlighters in ceilings tend not to be able to do that.
The original insulated fireproof fittings intended for running very hot quartz halogen downlighters are particularly bad for killing LED bulbs.
Rather like electric car range
Just fitted 134 of them. Not my choice.
And not bad I only hit one joist when drilling them out.
I swapped my parents kitchen spots from halogen to GU10 LED in August
2016[1]. They still work. [1] I went nearly all LED prior to that. A couple of the GU10s have been replaced and so have a couple of BC lamps in pendants.
Ages ago I fitted these to the bathroom, bought from B&Q. Fittings were nice, but the supplied bulbs failed very quickly. Bought replacements from TLC which have lasted very well - several years. They will likely be replaced by LEDs when they start failing.
After 2 hours on the GU10 case was about 58C, at its hottest point, measured with a laser thermometer. I dunno what that implies about the internal temp.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required