LEDs and Temperature

[Martin Brown]

The LEDs themselves are incredibly reliable provided that you don't abuse them. I find dipping the old dim ones in LN2 to demonstrate the improvement in efficiency with a cooled more rigid crystal lattice tends to do for them after a while, but indicator LEDs I installed when they were new and novel in the mid 70's are still going strong. Unlike the old low voltage indicator bulbs they are almost indestructible.

MTBF of bare LEDs driven properly is very long.

The electronics in consumer LED lights is by far the weakest component. I have seen just one lamp failure due to a single defective LED in a long chain going down the others were due to cooked PSU capacitors. I look at all the failed ones (three so far) to see how they failed.

[The Natural Philosopher]

So fas so good...

???WTF does '4 hours a day' mean if it doesn't mean 'duty cycle'

30,000 hours is about 4 years if you leave it on all the time.

Around 1960 my late data rigged up a fluorescent tube in the garage, he died a couple of years later, but that tube was still there working, though it flickered badly, when the house was sold in 2004.

It probably was on for less than an hour a month.

[dennis]

If you sold 10 million of them and 0.01% claimed each year you would need quite a few people to keep up with the claims. It just wouldn't be economical on the margins they make.

One car manufacturer offered lifetime warranties for a while but they stopped doing it.

[dennis]

You would never use a fluorescent tube if the life was that poor.

[Dave Plowman (News)]

Yes - same here. But once you start driving a LED hard to use as a light source, the life comes down. I've had a couple of early ones explode. And they claimed a very long life too. Given their high cost put me off them - especially since the light they produced was terrible.

[Dave Plowman (News)]

Are you then saying none ever fail within the existing warranty period? Most things are likely (percentage wise) to fail early on. If not, last the design life before failing. The bathtub effect.

How long a car lasts depends on lots of things. Like type of use and servicing, etc. That doesn't apply to a lamp.

[Roger Hayter]

That the duty cycle affects the actual life in years for a given duration of switched on life is triivially true. The original claim was that a reduced duty cycle, or rather on time, increases the actual switched on life. This suggestion is not made in the quoted material. Given the likely heating rate of a light fitting it is likely that the on time would have to be reduced to less than a few minutes to afffect lifespan. It is very unlikely that 4 hours would be better than 24 hours in terms of heat damage.

[The Natural Philosopher]

THAT I agree with.

except that of course many heat/cool on/off cycles will affect the kit through differential thermal expansion, and that in itself causes problems.

[Capitol]

Not LEDs, but have Just worked out that some of my CFLs are > 35 yeares old. I shudder to think what their light output is.

[Roger Hayter]

I agree that would be a factor, and already is for incandescent bulbs. However, an on time measured in minutes would also render LED lights pretty useless for any practical lighting use.

[whisky-dave]

No thicko. I've not brought a new dimmer since I replaced my faulty one with a working one.

No why would I, as it has nothing to do with it. You just have no idea have you, it's the temperature that that's the key. We're running a lab on that today. Where students 'roast' an LED. The ide is they hold a solderign iron up to the LED to make it hot and to destroy it by just heat no voltage applied. Make : Electronics Second edition by charles platt Experiment 13 Roasting an LED Page 121.

Get this book and learn something just for once.

we brought about 60

[whisky-dave]

Why ask for a receipt ?

if yuo donlt understand the specs you wonlt be able to claim anything on an y warrenty.

[whisky-dave]

But it doesn;t stop them doing it because they know few will attempt to cla im and those that do you cabn ask have you ever had them on for more than 4 hours at a time. If they asnwer yes (which most peolpe will have had) then the warrently is invalid just as it would be if the do not remove sticker was removed.

[whisky-dave]

types of use does. Hence the 4 hours per day, or in hot confined envioments. Thsi is also why capactitors have temerture ratings, and resistors have rat ed wattages. No suprise yuo know notheing about this though.

Even computers Apple stae theres as I've ready them they have temerature an d altidue limits (really a pressure limit)

Even trams have limits on what the are meant to do under certain situations . Sure trams can go 50+ MPH but that doesnlt; mean you can take sharp corners doing those sppeds that is OUTSIDE the specs.

[Dave Plowman (News)]

But true thicko DIYer's such as me repair dimmers. What does that make you?

Sorry for assuming you had some idea of the sort of power supply used in a mains LED. I'll know better in future.

Hope it is better than the LadyBird book you seem to have got most of your 'knowledge' from.

[whisky-dave]

Clever than you because I knew fixing the dimmer would not make it do what I wanted which was to dim LEDs

you still haven;t got this basic fact righ thave you.

It's really simple but you get it wrong everytime.

Dimmable LEDs require a trailing edge dimmer if they are to dim properly.

I know far more than you do.

[John Rumm]

Yup early LED lights were fairly horrible in terms of the colour and quality of the light - quite often as bad or worse than CFLs.

The modern ones however have got vastly better - to the point now where I find the LED "filament" style lamps produce a light quality hard to distinguish from a real filament lamp (both in terms of light quality and also directionality). (the non filament ones can also do decent light, but have a different illumination pattern from a traditional GLS).

[John Rumm]

LED hut offer a 5 year warranty on theirs. They honour it as well, on the couple of occasions I have had failures a quick email has been enough to have a replacement at my door next day.

Lifetime is also rather difficult to understand exactly what is on offer. I remember a disk company offering such a warranty on their floppy disks once. Turned out when you read the small print the "lifetime" was defined as the period from new until the media developed read/write errors!

[John Rumm]

It does to a fair extent... with LEDs its particularly effected by how hot it is allowed to get.

[John Rumm]

Its simply attempting to give an example of the likely replacement schedule based on "typical" use, for some value of "typical" which they then go on to define for the pedants ;-)

Indeed.

The box for the LED candle I have beside me, makes the claim that it has a 12 year lifespan - but that is based on its 20,000 hours being used

4.5 hours / day.

The life will scale down (and possibly up) for longer or shorter daily run times pro rata.

ISTR there was mine somewhere where it was lit by similar tubes, which had been running 24/7 for decades. The fact that they were in a nice stable temperature and never turned off meant they had exceptional run times.