Long term test (life) of a bulb LED Sirius

[Christophe]

Due to moving, I had to stop the long-term trial for a few days.

So I removed the bulb (a little dusty) and it looks like new: the 18 leds are still walking (phew lol), the screw has no rust (I recall that the bulb was in an outdoor lantern for months), only the alluminum part of the pellet is colored slightly in places.

Nothing to report therefore in terms of "wear", only a small strange fact: when the bulb is cut, it appeared 3 darker points centered on each led and arranged at 120 ° (they thus form an equilateral triangle centered on each LED). These points are about 0,5mm in diameter and there is no difference (visible) when the bulb is on.

I will try to make a photo before reassembly and the rest of the test.

ps: since the beginning of the test (Sunday 22 October at 18h), the bulb worked around 3800h

[Christophe]

The bulb was put back into operation on Saturday March 31 at around 18 p.m., so there was only 48 hours of "break".

[Christophe]

We are November 16 and the bulb is still on.

I think the test is pretty good, so I'm going to stop it.

Test duration: 9300h approximately minus the stops mentioned above (a few dozen hours), either more than 9200h in the end.

This test was carried out, I remind it, in an outdoor lantern directed south thus subjected to the climatic hazards (as hot as cold).

[Christophe]

This test is running from Sunday 22 October 2006 to 18h.
It is carried out on an outdoor lantern
The bulb is therefore subject to variations in temperature and humidity, in short conditions not very ideal for electronics ...

It will remain on 24h / 24 for (at least) the 6 months of winter (so rough Ardennes) a period of roughly 4000h ...

Big unearthing to alert you that, this 28 August 2013, the bulb is still on 24 / 24 ... finally the test lasts for more than 6 years ... soon 7 years ...

The 50 000h were exceeded before the 6ieme anniversary.

Here are other products based on LEDs: https://www.econologie.com/shop/ampoules-led-c-90

[Janic]

what kills the filament bulbs, it is the ignitions and extinctions: what is it for the leds?

[chatelot16]

what kills the electronics is the humidity: so it suffers much less when it is constantly lit, because it heats a little

given the low consumption of led lamps it is not stupid to light them constantly to keep it dry

if it is lit only when it is needed if it is left too long, wet and dead, the replacement of the lamp costs more than the electricity

to make the calculation, what characteristic of this lamp? power? price?

[BobFuck]

what kills electronics is moisture

especially the cost cutter that saves 5c by replacing an 105 ° C consensor with a Chinese 85 ° C ...

[Christophe]

Soon 7 years and a half and one winter more ... the bulb is always on ...

[Christophe]

That's the light bulb is going to give up the ghost after almost 9 years and a half is about 80 000 h running continuously and this in all weathers (Applies outdoors) and some lightning strikes that have done other damage in the house ...

It is not 100% HS but its light power is very low (20% -30% of the original power to the eye) ... it makes a "flash" of brightness on ignition (which seems correspond to proper operation) and then it reduces in brightness.

The LEDs are, I think, still OK but the internal transformer must be limited in voltage and / or current ... probably a component HS (capacitor?)

Do you think an autopsy is necessary?

[Obamot]

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?) ...

But finally did you put a timer or not? Because I would have liked to come back to this point, the aim of which would have been to martyr as much as possible an LED lamp (that will answer Janic). It is very simple to define in extreme conditions (less in the majority of cases), if the LED is very hot and the circuit is close, it would first have been the welds that would snap, there it is relatively easy to estimate, it would have been necessary a time delay which slightly exceeds the rise / fall in temperature (expansion-retraction) taking into account the maximum outside temperature (summer). There it depends where the circuit is placed and the power of the LED, but for it to heat well (and "hope" for a destructive expansion), you need Cree LEDs which heat to 80 ° C and the circuit close enough ... It is quite rare and expensive!

I think that for the rest, this time is sufficient for the other components (semiconductors, resistors, condos etc) since there is the power shock that counts ...!

But when it comes to testing the LEDs stricto sensu, it's useless. It is not they who would slam first (data between 30'000 and 50'000 hours, there it is exceeded by almost a third) but it would rather be the electronic components first. And in this case, during a breakdown, we could consider that the LEDs do not succeed the challenge of their longevity, which would be a false conclusion by beast paralogism ...

What makes that logically (and if I'm not mistaken ...) I would not Administrative staff suggested to put timer for testing!

(And doing two separate tests, one with and the other without would be of little use and in my humble opinion, from the moment we already know which of the two would snap first statistically speaking, but what I say so ...).