LED life versus current (at and above rated) in white light bulbs

Ah, much nicer. I misunderstood your post on HSM.

You might sneak some holes in on the sides or something. Otherwise I can't see any path for airflow at all in those pictures.

The safe thing is to measure it. Our old rule-of-thumb was that a 1in^2 surface, vertical in free air, would rise to 60C with natural convection, or ~35-40C/W over ambient per in^2. I don't know how accurate it was, but that was the Old Man's ROT.

This random sample from Aavid has 1.7in^2 of surface area, and they rate it at 24C/W. I take that as their natural convection rating, which works out to 41C/W/in^2).

formatting link

That assumes free air, which you may not have if there's no ventilation.

In your favor is that you don't have to dissipate the whole 9W. LEDs are so efficient these days that they radiate several watts as light.

Ha! :-) (I thought of that for mine too, before I made the 'beer can' heat sinks :-))

Cheers, James Arthur

Reply to
dagmargoodboat
Loading thread data ...

The Lumileds graphs used to show 90% lumen maintenance @ 50K hours, IFF you keep 'em cool.

The industry has changed to a different, standardized test regimen since then, and I haven't looked recently to see how that affected their results.

Cheers, James Arthur

Reply to
dagmargoodboat

Look carefully what that "test current" is.

IIRC at CREE it is about 30-50 % of Imax.

Reply to
upsidedown

I've built a few high power (0-5000 lumen) bicycle lights at home so I'm very familiar with this.

LEDs with low resistance had reverse voltage ESD damage from sloppy assembly. The die will slowly burn away until the damaged area consumes more current than provided. An early test for this is to apply a few microamps of forward voltage for an hour. Damaged LEDs will eventually appear to be shorted while good LEDs will stay lit.

LEDs that went open-circuit were likely cracked from thermal flexing. They may have been cheap because they were contaminated or flawed.

--
I will not see posts from astraweb, theremailer, dizum, or google 
because they host Usenet flooders.
Reply to
Kevin McMurtrie

Yes, of course, and that IS what current I am running at, the TEST cuurent, NOT Imax! Imax is SCARY, I think one Amp! It would probably burn holes in the floor, not to mention our poor retinas!

Test current is 350 mA, Imax is 1 A.

Jon

Reply to
Jon Elson

Just another data point. The first retrofit I did had 10 LEDs in it, and it has been running for 2.5 years, guesstimated at about 2700 hrs or 27000 unit-hours. The second two units have been running for about 1.5 years, but more hours/day, guesstimated at 3300 hours, or 132000 unit-hours. (Unit- hours referring to hours x number of LED components.)

They seem to be doing QUITE well. I've had 2 bad solder joints on these things, they are leadless SMT, so getting them perfectly soldered to the PC board material is tricky. When they go out, I just go down the line with a DVM in diode test mode to find the bad solder joint, and it is back on line in a couple minutes.

Jon

Reply to
Jon Elson

Earlier you said

In that case, those are "3 W" LEDs, since the power is usually quoted as Pin = Vf x Imax.

IMHO, a "3 W" LED is usable as 1 W (350 mA) light source and "1 W" as

300 mW (100 mA) light source. Also with such underrating the thermal design becomes much easier keeping Tj well below Tmax.

Sure there are "10 W" LEDs with Imax at 3 A, but light output well below 10000 lm, thus less than 100 lm/W and very short usable life time. Possibly usable as a flashlight, but not much else.

Reply to
upsidedown

Thoothu pithare Govinda !!!

Reply to
Archer

Cree is a conservative company, they do NOT recommend running them at 1A. I thought they listed these as 1W, but I can't find that description anywhere in the spec sheet, now.

it's all going the way of the 6 HP vacuum cleaner motors, 5 Hp wall-plug air compressors and the rest of that crazy consumer-rated crap. But, Cree sells mostly to large manufacturers of commercial lighting products, and won't get away with misrepresenting their products.

Jon

Reply to
Jon Elson

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.