bikeNomad
Newly Enlightened
The recent Zetex thread got me thinking:
Here we are, trying to get a couple of percent better efficiency out of our converters (going from 85-88%, for instance).
But then there's the temptation to mount the Luxeons inside flashlights with little or no heatsink (I've been guilty of this when trying to get something working
)
Unfortunately, there goes that efficiency we worked so hard to get out of the converter!
If you look at the Luxeon data sheet, you'll see some interesting things about thermal ratings (I'm assuming a white LS, 3.42V at 350mA, room temp 25 degrees C here):
<UL TYPE=SQUARE><LI> The junction temperature will be 20 degrees (C) above board temperature
<LI> The board will be above ambient temperature. By how much depends on your heatsink, air flow, and mounting of the LS to the heatsink.
<LI> With a perfect heatsink at room temperature, you've already lost 7% of the rated light output due to the 20 degrees difference between the junction and board. (Unless Lumileds expects people to run their LS at 5 degrees C, their rated light output is deceptive since it assumes a junction (not board) temperature of 25 degrees)
<LI> If you let the board get to 60 degrees C, you've lost another 14% off that, for a total of 80% of the rated light output.
<LI> At that same 60 degree board temperature, the forward voltage of the LED has decreased by 110mV. This is a 3% reduction.
[/list]
Why did I pick 60 degrees C? In my tests with the Luxeons in small closed plastic boxes (using thermocouples to measure the temperature), the board temperature tends to level off at 60 degrees at 350mA input. I don't know why this happens. This is probably what you're getting in a flashlight design unless you have the Luxeon mounted to a good thick heatsink with good airflow.
What kinds of board temps have you been measuring in your flashlights?
So how do we get 14% better life? Simple: get a perfect heatsink
and reduce the current by 14%. Same light output (because the Luxeon is running cooler), less power input.
I realize that we can't always get the heatsink we need, and that (at least for me) it's easier to make electronics than to machine parts.
But it seems like those of us who want the absolute brightest flashlight, or the longest run time for a given brightness with the LS, should try to spend some time on heatsinking.
Or just use the flashlights outdoors in the winter
...
It's great to see your mod pages, because I keep getting ideas for heatsinking and mounting. Keep up the good work, and thanks for sharing your designs!
Here we are, trying to get a couple of percent better efficiency out of our converters (going from 85-88%, for instance).
But then there's the temptation to mount the Luxeons inside flashlights with little or no heatsink (I've been guilty of this when trying to get something working
Unfortunately, there goes that efficiency we worked so hard to get out of the converter!
If you look at the Luxeon data sheet, you'll see some interesting things about thermal ratings (I'm assuming a white LS, 3.42V at 350mA, room temp 25 degrees C here):
<UL TYPE=SQUARE><LI> The junction temperature will be 20 degrees (C) above board temperature
<LI> The board will be above ambient temperature. By how much depends on your heatsink, air flow, and mounting of the LS to the heatsink.
<LI> With a perfect heatsink at room temperature, you've already lost 7% of the rated light output due to the 20 degrees difference between the junction and board. (Unless Lumileds expects people to run their LS at 5 degrees C, their rated light output is deceptive since it assumes a junction (not board) temperature of 25 degrees)
<LI> If you let the board get to 60 degrees C, you've lost another 14% off that, for a total of 80% of the rated light output.
<LI> At that same 60 degree board temperature, the forward voltage of the LED has decreased by 110mV. This is a 3% reduction.
[/list]
Why did I pick 60 degrees C? In my tests with the Luxeons in small closed plastic boxes (using thermocouples to measure the temperature), the board temperature tends to level off at 60 degrees at 350mA input. I don't know why this happens. This is probably what you're getting in a flashlight design unless you have the Luxeon mounted to a good thick heatsink with good airflow.
What kinds of board temps have you been measuring in your flashlights?
So how do we get 14% better life? Simple: get a perfect heatsink
I realize that we can't always get the heatsink we need, and that (at least for me) it's easier to make electronics than to machine parts.
But it seems like those of us who want the absolute brightest flashlight, or the longest run time for a given brightness with the LS, should try to spend some time on heatsinking.
Or just use the flashlights outdoors in the winter
It's great to see your mod pages, because I keep getting ideas for heatsinking and mounting. Keep up the good work, and thanks for sharing your designs!
