[ QUOTE ]
wilan said:
Hi!
I need information about a thermal management of Lamina Ceramics LED light engine BL-2000 RED. Could anyone /ubbthreads/images/graemlins/help.gif me?
How is:
1) Minimum and maximum operating temperature?
2) Dependence on life time and thermal conditions?
3) Life time of BL-2000 RED in 20C and 55C ambient temperature?
I have been traing to get these informations directly from Lamina from few weeks but by now I have unfortunately no answer. If you know answers please let me know.
Regards:
Adam Wilanowski
[/ QUOTE ]
ATTACHMENT RECOMMENDATIONS
Wavelength ? 613 - 623 nm
Voltage* V F - 6.69 - V
Drive Current I F - 2.80 - A
Power* P - 18.73 - W
Luminous Flux* ? V 168 348 - lm
Thermal Resistance T R - 2.25 3 ºC/W
The operational die junction temperature must not be allowed to exceed 125ºC and should ideally be maintained as close to 25ºC as possible.
Assuming you are driving it full power, with 18.73W, and there is a thermal resistance from the die to the back of the package of max 3 degrees C per watt, 18.73W * 3C/W = 56.19 rise from the back to the die at full power. If you had a infinte heatsink, and perfect transfer from the back of the device to your heatsink, and room temperature was 25C, the die temperature would be 81.19 degrees C.
If you look at the die temp vs. light output chart, you'll see that the typical output of 348 lumens is derated by 55%, for a actual light output level of 191.4 lumens. The same derating issues happen with Luxeons also. Derating of red LEDs is the worst out of the visible LED colors.
You will need to derate from this further, if you do not have an infinite heatsink (which is quite substantial), and also derate further for the thermal resistance of your attachment method.
Now, if you were to drive it at 1/2 power, or, 9.365W, 9.365W * 3C/W = 28.095 degrees C die temp, on a infinite heatsink. 28.095 + 25C room temp = 53.095 degrees C die temp. 348lm / 2 = 174 lumens (half power input). Derating at 53 C is about 79% from the chart. This results in 137.46 lumens.
Please keep in mind, this is with a perfect or infinite heatsink, where the heatsink never rises above 25 degrees C.
Actual heatsinks are extremely far from this.
For comparison, lets take the luxeon red-orange (about the same wavelength). From the datasheet, the typical lumens is 55 lumens, at 25C die temp and 350 mA. Typical Vf is 2.95V.
2.95V * .35A = 1.0325 W. Thermal resistance is 18 C per W.
So the die will rise 18C above the 25C ambient on an infinite heatsink, or to 43 C. Derating according to the luxeon chart is about 70%. So, .7 * 55 = 38.5 lumens output.
Again, this will drop alot with a real heatsink.
What you can do, is to put it together, and assume you have fairly perfect thermal compound (such as soldering it to a copper heatsink), then measure the heatsink temperature right by the led. It is quite typical for the heatsink or
flashlight body to hit 50C. In the case of the luxeon, add this heatsink temp, 50C to the rise we computed, which results in 68C. From the luxeon chart you see a derating to about 50% output. 55 lm * .5 = 27.5 lumens.
Now in the same conditions, a white emitter with a Vf of 3.42, 3.42V * 350mA = 1.197 W. The thermal resistance is 15C/W. So the die rises 17.955 C. 17.955 C + 50 C =67.955 C die temp (aka junction temp). Derating at this temp is 85%. Typical output is 25 lumens, so 25 lm * 0.85 = 21.25 lumens output.
Lets go one step further, as many folks are driving the White Luxeon III at 1A. Typical output is 80 lumens at 1A and 25C. Typical Vf is 3.9V 3.9V * 1A = 3.9W Thermal resistance of the package is 13 C/W. 3.9 W * 13 C/W= 50.7 rise of the die. Say the typical heatsink/flashlight temp was 50C right by the LED, and you'd soldered it to it, somehow, and maintained electrical isolation. 50C heatsink + 50.7 C rise = 100.7C From the datasheet, derating would be 75%. 80 lumens * 0.75 = 60 lumens output. In reality this would be alot worse with the heatsink/flashlight we used previously, due to the fact that we are putting about 4X the power into it, which would cause it to rise to significantly higher temperatures.
One thing we excluded was the converter efficiency which would cause a further rise in temperatures. This can be *quite* significant, especially at 1A to the luxeon. This is why nexgen converters are needed from what we have been using to date.
