Snow l.e.d. lifetime?

[font=&quot]So, I've become something of a fan of PeakLedSolutions. I've read here that the snow led's don't survive well. I read there was a test done by a CPF member, I am unable to locate it. The snow led will degrade to 50% in less than 20 hours!? I have a AAA that has many (more than 12) lithium AAA's run through it, it doesn't appear to have dimmed. Are they really that bad of a lifetime? [/font]

I read the same thread and am curious to know as well. I recently purchased 2 custom 3 LED HP Snow Kilmajaros and a 3 LED HP Snow Matterhorn and love them all! One of my KIlmajoros I tested up to 40 hours (still having some decent light). If the post was on the mark, then theoretically, my light has already lost some of it's luster. However, I have not noticed anything with the naked eye.

I guess we will have to see if MJ of Peak jumps in to clarify.

C
C
Chimo's test data from THIS thread.

I'm the one who posted that and it seems I had it wrong- the LEDs do start dimming right away, but it takes much longer for them to reach 50% than I had thought. Sorry to cause any undue concern. Also, remember that the Peak circuit doesn't drive the LED at 40ma for very long, unless you compulsively change out the battery, so you probably don't have to worry about it (by the time you do, there will be something else out that you'd like to buy- like a Peak luxeon AAA ). The only way it might matter is if you only use lithiums and let the light run for long periods of time (intermittent use causes less heat to build up).
I'm going to go back and edit the other thread with this info.

Edit* I've edited the other post and included a link to this one. While I wouldn't say the Snow LEDs are a much of a concern, it is telling to look at the chart for the Nichia at 60ma vs the Snow at 40ma for the same period of time.

Here is the thread I was following for the last 6 months on LED life. I believe Chimo summerised all his findings in the 2nd & 3rd posts.
http://candlepowerforums.com/vb/showthread.php?t=89759&highlight=Head2Head

Chimo's charts should be taken more as a comparator between LEDs then as strickly "how many hours such-and-such LED has to 50%". He runs his LEDs with no cooling (not surprisingly, as that would basically require mounting each one in a torch), at constant current and all at once, so what he sees isn't what we will see.

The official line is 1,000 hours for ultra-power and 6,000 hours for high-power for the Snow29 LEDs. My McKinley Ultra hasn't shown any dimming yet to the meter, so even if you figure that Peak's numbers are wrong by a factor of two we got a while.

I think archangel is correct.

The chart is misleading, as NO one runs their lights for days at a time.

I think if you did the same thing to many of our luxeons (constant on, constant current) we'd find horrible LED life.

Thanks. I have been known to run my led lights for days, but as mentioned, it maybe runs at full current for an hour or so. Assuming you don't change the battery out every 90 min. But even with a dozen batteries through, say five hours each, that's only 60 hours, likely a year or two of use for a "normal" person....and maybe 12 hours of full power drive at that.

Same response as here.

FWIW, I agree that for typical flashlight purposes, most people will not put the amount of stress on the LED as the constant current marathon burn-in and will not approach the cumulative hours. As was stated above, the charts are good for comparison. They are also good for non-flashlight applications to help choose drive levels.

As far as cooling during the burn-in, I believe the test setup would be as good as, or better than, a flashlight for the following reasons:
1. The LEDs are exposed to ambient airflow instead of being enclosed in the head of a flashlight.
2. The leads were not trimmed (more direct heat radiation surface)
3. The power source was separate from the LED, so its heat generation did not affect the LED.

In any case, it's great that questions are being asked and people are putting some thought into the matter.

Paul

[/QUOTE]
As far as cooling during the burn-in, I believe the test setup would be as good as, or better than, a flashlight for the following reasons:
1. The LEDs are exposed to ambient airflow instead of being enclosed in the head of a flashlight.
2. The leads were not trimmed (more direct heat radiation surface)
3. The power source was separate from the LED, so its heat generation did not affect the LED.
[/QUOTE]

While I agree generally with your assessment of the utility of this experiment, I think that the test set up described is actually inferior to the heat sinking provided by a metal flashlight. A few examples;

1) If you put your hand in the ambient airflow over the burner on your stove, does it burn as quickly as if you put your hand on the burner? Direct contact is always the most efficient form of heat transfer, and the metal head and body of a flashlight have a significant thermal mass. This thermal mass would have to be overcome before the internal LED temperature could rise much.
2)For short bursts, heat would be moved away from the chip very rapidly by the large thermal mass of the light body. I doubt the thermal mass of the lead wires amounts to much. In longer bursts, the metal body, with it's high conductivity, would move the heat into your hand or the air much more rapidly than the low conductivity epoxy and two wires could into the air alone, so in either case the flashlight would be better off.
3) This very well might be true, depending upon how efficient the circuit is and how well it moves it's own heat to the body of the light. This is a definite design variable, but it doesn't overcome the huge difference in thermal mass between the air around this test set up and the metal body of a torch.

It may be futile to post to a dead thread, but I love a good thermodynamics conversation. :rock:

As far as cooling during the burn-in, I believe the test setup would be as good as, or better than, a flashlight for the following reasons:
1. The LEDs are exposed to ambient airflow instead of being enclosed in the head of a flashlight.
2. The leads were not trimmed (more direct heat radiation surface)
3. The power source was separate from the LED, so its heat generation did not affect the LED.
[/QUOTE]

While I agree generally with your assessment of the utility of this experiment, I think that the test set up described is actually inferior to the heat sinking provided by a metal flashlight. A few examples;

1) If you put your hand in the ambient airflow over the burner on your stove, does it burn as quickly as if you put your hand on the burner? Direct contact is always the most efficient form of heat transfer, and the metal head and body of a flashlight have a significant thermal mass. This thermal mass would have to be overcome before the internal LED temperature could rise much.
2)For short bursts, heat would be moved away from the chip very rapidly by the large thermal mass of the light body. I doubt the thermal mass of the lead wires amounts to much. In longer bursts, the metal body, with it's high conductivity, would move the heat into your hand or the air much more rapidly than the low conductivity epoxy and two wires could into the air alone, so in either case the flashlight would be better off.
3) This very well might be true, depending upon how efficient the circuit is and how well it moves it's own heat to the body of the light. This is a definite design variable, but it doesn't overcome the huge difference in thermal mass between the air around this test set up and the metal body of a torch.

It may be futile to post to a dead thread, but I love a good thermodynamics conversation. :rock:[/QUOTE]

Good comments. I believe they would apply well to a Luxeon type LED application. I have not personally seen any 5mm LEDs in flashlight applications with the effective 5mm LED lead to flashlight body thermal conduction you mention above. Many manufacturers seem to think those 5mm LEDs don't generate enough heat to warrant any thermal releif.

Paul

I doubt if 30 ma is going to heat up the led much. It's just 0.1 watt. It might be interesting to attach a tiny thermal probe to the LED lead and see what happens. The lead temperature near the envelope should be pretty close to the die temperature.

paulr said:

I doubt if 30 ma is going to heat up the led much. It's just 0.1 watt. It might be interesting to attach a tiny thermal probe to the LED lead and see what happens. The lead temperature near the envelope should be pretty close to the die temperature.

Click to expand...

You are correct - 30mA is a pretty safe current for most 5mm LEDs. However, many folks here are driving them at currents between 50-100mA. The power dissipation at that current level is around 200-400mA (depending on the Vf). I have noticed the 5mm LEDs get warm at the higher current levels.

BTW, ArcAAA current is around 50mA (Pd ~ 200mW).

Cheers,

Paul