Long term test (life) of a bulb LED Sirius

[izentrop]

Hello,
Blinking indicates that some of the LEDs are short-circuited and flashing speeds the end of the others.
It would be interesting to see what it has in the stomach.

At that time, the diet was often capacitive


In continuous ignition, it is normal that it lasts as long. It is the ignitions that make them suffer the most and reduce the life span.

An assembly of this type, with a zener and several branches with resistance by branch is better in duration.

http://www.sonelec-musique.com/electron ... _leds.html

you need Cree LEDs that heat up to 80 ° C and the circuit close enough ...

No, the less it heats the longer it lasts.

The colder it is, the longer the life of the LED is long

http://www.osram.fr/osram_fr/actualites ... /index.jsp

The light bulb of Livermore's firefighters, if it lasts so long, is that it is never extinguished and that over time, are carbon filament increased in resistance and only consumes 4 W, heater little and can last a long time like this.

[Obamot]

[...] I would have liked to come back to this point, whose goal would have been to martyrise a maximum LED lamp [...] it is very simple to define [...] if LED very hot and close circuit, it would have been first the welds that would slam, there it is relatively easy to estimate, it would have taken a duration a time delay that barely exceeds the rise / fall in temperature (expansion / retraction) taking into account the maximum outdoor temperature (summer). It depends where the circuit and the power of the LED are, but for it to heat up well (and "hope" for destructive dilation), it takes Cree LEDs that heat to 80 ° C and the circuit close enough ... It's quite rare and expensive!

[...] An assembly of this type, with a zener and several branches with resistance by branch is better in duration. [...]

if Mister Izentrop had bothered to read me correctly, he would have noticed that if in his example he presented a case for an LED to last as long as possible, I did exactly the opposite.
So I did not say less than the same thing as him ..... LOL since I still ended like this:

► View Text

[...] for the other components (semiconductors, resistors, condos etc) since there it is the electric shock of putting on tension which counts ...!

But when it comes to testing LEDs stricto sensu, it's pointless. It is not they which would slam first (data between 30'000 and 50'000 hours [...] but it would rather be the electronic components first

However it should be noted that its own reflection is not entirely accurate, as can be seen in the diagrams of CREE on the "lifespan", the operating temperature is precisely chosen for that which allows the LED to last as long as possible ... >>> yes I know, confusing!

[izentrop]

Hello

we see it in the diagrams of CREE on the "lifespan",operating temperature is precisely chosen for the one which allows the LED to last as long as possible ... >>> yes I know, confusing!

There is not even this word in your doc and I have trouble with English.
With a screenshot, I might realize, but I think the choice is due to use as standard, while with better cooling it would last longer.

[Obamot]

1) So you planted yourself on the other points that you thought and even no excuse!

2) You were researching "Life"which means [lifetime" for the full name (not even need the dico) and you would have found.

3) Now we had already talked about all this, right? >>> and you knew how to read a painting since you had answered here >>> you raise yourself there that there is an optimum temperature.

It is in the range 60 ° C - 80 ° C (but it is specific to this LED, it varies each time, for each generation.)

In short, what do you want except to quibble over every message I post?

[izentrop]

I found the pdf of the controversy: http://www.cree.com/~/media/Files/Cree/ ... _PAR38.pdf

Life time curves were made on the extrapolation of tests that lasted only 6 hours, as I understand it.

We can see on the online calculed lifetime that the service life decreases with the temperature. Same conclusion on the graph on the previous page.

It is not in the spirit of contradiction, but in relation to logic and physics

[Christophe]

wow !! pretty performance.
Changing the condos would be interesting on the informative level! And it's not difficult to do. It has all the symptoms, no?
Since it consumed 5,2W new (that you had measured with the wattmeter perhaps it would be interesting also to check the conso?) ...

Yes 80 000h is not bad ... especially in the conditions (variations of T °, ​​humidity) ...
Not re-verified current consumption but I can do it ...

Here is the outdoor lantern where the bulb was lit from 2007, facing south ...

But to finish you put a timer or not?

No, I never did automated ON / OFF cycles ... yet I have a timed relay in my lab for years!

We can always run tests ...

[Christophe]

Hello,
Blinking indicates that some of the LEDs are short-circuited and flashing speeds the end of the others.
It would be interesting to see what it has in the stomach.

It's not a flashing, just a flash, that lasts ... 1 / 10 or 1 / 5 of second on ignition.

Ok I'll open it and see if a component has heated

[Christophe]

Life time curves were made on the extrapolation of tests that lasted only 6 hours, as I understand it.

It is not rather 6048 h? Because the 0.048h is 2 min 53 seconds seems strange to me not?
Anglo-Saxons are used to put, to separate thousands ...

So, if we read these curves, these tests consider that the bulb is HS when it lost 30% of its initial brightness?

[izentrop]

Is all the leds in the light bulb flashing or only part of it?

The transparent half shell of the bulb must unclip or unscrew, make us pictures once open

Edit: you're certainly right for 6048 h and 70%, I seem to have read that elsewhere too. http://www.led-fr.net/duree-de-vie-dune-led.htm

[Christophe]

All LEDs and quite synchro ... which indicates, necessarily, a problem with the internal power supply.

Yes this bulb opens quite easily, I had already opened a few years ago.

ps: yes 6048h seems more logical (now why 6048h? I mean: why these 48h and more?)