LED downlight reliability?

[nomisr]

Hi,

Does anyone have any views on the reliability of LED downlights? I'm talking about the integrated ones rather than LED GU10s for standard fittings.

I bought 60 of these things for a major extension and refurb, and 6 months later 7 have failed one way or another. One flickered (off more than on), one died completely, and the remaining 5 lost at least one of the LEDs. I'll not name the make as the manufacturer has been very good about replacing them and is currently investigating the problem.

My main concern is the time it takes to swap them out, not to mention the damage it does to the ceiling no matter how careful you are. And it can't be right that so many should fail (from 3 different batches) in such a short time, especially considering they are guaranteed for 10 years. I'm beginning to regret not using something a bit more traditional.

Cheers,
Nomis

 

[ericmark]

Be it integral or a bulb so much is down to build quality. We see again and again how some sample has come from China has be accepted but once production has started the quality falls. I have watched Utube videos of the simple GU10 lamp and how some have poor cooling it only needs the heat sink compound to run out on the production line for lamps to, in use, over heat.

I have worked on production lines and when there is a fault, finding when the fault started and at what point the product should be rejected is not easy. There is also the problem with the rejected product still arriving on the market place. In one firm I worked in some one realised they could get the rejects and sell them I am sure it happens all the time.

The 5 year life of the old compact fluorescent was it seems based on how long before only 50% were still working. I fitted 16 globe bulbs Philips and within a year we had lost 4 or 5 so wife found some cheap replacements, for 6 of the lamps used in one room. As the Phillips bulbs continued to fail the other room had all 10 replaced with LED bulbs. But some 8 years on the 6 cheap bulbs got by my wife are still going strong. It would seem I was just unlucky with the Phillips bulbs. Taught me a lesson I never buy expensive bulbs now.

 

[bernardgreen]

As Eric has said the quality varies, generally the cheapest ones are not worth buying. The best advice is to pay a bit more and buy from reputable suppliers and avoid Internet based companies and especially companies who are totally web based. There are many UK based companies with on line catalogues from which to choose your lamps. You can buy on-line from most of these reputable UK companies via their web sites and then if things do go wrong you have UK telephone contacts with a sales desk / help desk to get things sorted out. These companies have reputations to maintain. A person selling on internet auction sites can un-load a heap of defective items and get 99% bad feed back. He or she doesn't care about having a bad reputation, they start up again with a new "trading name" and sell another heap of c**p to a new set of buyers.

 

[ban-all-sheds]

The 5 year life of the old compact fluorescent was it seems based on how long before only 50% were still working.

That is the way the lighting industry, or at least the reputable part of it, has always specified lamp life.

 

[alan333]

I fitted 4 cheap ebay GU10 led's to my outside lights about 5 years ago and they're still working. Granted they're on a PIR so they're not on for long periods.

 

[ericmark]

As Eric has said the quality varies, generally the cheapest ones are not worth buying.

I wish that was the case. However my experience it was the expensive lamps which failed.

Yes buy from good UK outlet who have shops you could visit even if not in your area. But I bought all my LED bulbs from Lidi, B&M Bargains and pound shop the latter the 12 volt ones failed but at £1 what do you expect but the Lidi and B&M Bargains lamps were cheap but have not failed OK only been in about 2 years but touch wood they have done well.

 

[nomisr]

Thanks for the insights so far.

These fittings are by a leading brand based in the UK (though no doubt made in China) and were bought from a reputable bricks-and-mortar wholesaler, and not exactly cheap at £32 a pop (inc. VAT). To see almost 12% fail in under 6 months does not instil confidence. At least I had an easy time of getting them replaced.

 

[ban-all-sheds]

To see almost 12% fail in under 6 months does not instil confidence.

It may be just one of those things, and not at all in conflict with their stated expected life, as that's the time by which 50% will have failed.

 

[JohnW2]

To see almost 12% fail in under 6 months does not instil confidence.

It may be just one of those things, and not at all in conflict with their stated expected life, as that's the time by which 50% will have failed.

Indeed - and one mustn't forget that, even in the case or reputable/conscientious manufacturers, those estimates of life expectancy are inevitably little more than 'intelligent guesses' / extrapolations as regards the long-term reliability under normal service conditions. No manufacturer of a new product is going to delay marketing it for several years whilst they do 'proper' long-term testing - and any estimates of life expectancy based on what happens over just a few weeks or months (or based on 'accelerated' testing) are going to involve all sorts of assumptions, (m)any of which may not necessarily be correct.

A big problem in our world of rapidly evolving technology is that many/most technological products will become obsolete long before we have a really solid handle on their true in-service reliability!

Kind Regards, John

 

[ban-all-sheds]

That's been a problem in IT for as long as I can remember. There is (or was) a US Govt or military standard for calculating reliability. I believe the term "MTBF" is frowned on.

 

[bernardgreen]

MTBF ? Either Mean Time BEFORE Failure or Mean Time BETWEEN Failures

but has also bee used for Minimium Time Before Failure

and by a cynic Maximum Time Before Failure ( implies it WILL fail within a known time )

 

[JohnW2]

MTBF ? Either Mean Time BEFORE Failure or Mean Time BETWEEN Failures

The latter is rarely used in relation to single items, since the concepts are most commonly applied in situations in which 'failure' is terminal/irreversible - so, once a failure has occurred, the same item cannot fail again. If, unusually, 'failures' are reversible, then one occasionally sees the concept used.

The concept of 'mean time between (individual) failures' in a large group of items, at a particular point in their lifetime, can be operationally valuable, but is a totally different concept. If one has, say, a bank of 500 hard drives, or lamps, or whatever, it can be useful to know that, say, after they have all had X years in-service, one can expect that, on average, one will fail every 14 days. However, one needs to understand what that 'mean time between failures, after X years' of 14 days actually means, since (depending on the value of X) it is not incompatible with the 'mean time before failure' of an individual item being, say, 10 years!

... but has also bee used for Minimium Time Before Failure

... unless qualified or restricted, that is a daft concept - since, without such qualifications/restrictions, it will always be zero!

Kind Regards, John

 

[nomisr]

Well, since I posted this just over a week ago, another fitting has developed a fault, bringing the total to 8.

Here are photos I took of a couple of them before they were replaced.

Image7

• nomisr
• 22 Sep 2015

Image5

• nomisr
• 22 Sep 2015