I am testing Luxeon K2 1500 ma 3,6 v led no matter what I do
it gets real hot 50 degree C in 30 minutes this is what I do.
1. I use constant current module 700ma 3,6 v " I am giving half power"
2. 1,5 x 1,5 x 1/4 thickness piece of aluminum
3. I used white compound grease in between heat sink.
when I talked to future electronics engineers one guy said it is not
suppose to get that hot then other said the heat sink should be 3 times
bigger then today some one else said what I use is very bad heat sink.
My question is does any one used K2 leds and know how to dissipate
heat? I want to use 3 of them in 3-4 inch round aluminum heat sink.
The thickness can be higher if I needed. any info will appreciate.
Does your heat sink have any fins on it? Sounds like you are just
describing an aluminum plate. Fins equal much more surface area which will
do a much better job of cooling the LED. There are also much better heat
sink greases, but they won't make a lot of difference without an adequate
heat sink.
I assume the heat sink itself is getting that hot, indicating the thermal
resistance between the LED and the heat sink (the grease) is not
significant, but the thermal resistance between the heat sink and ambient
is significant. About all you can do is increase surface area and increase
air flow. In a project I did (http://www.denmat.com/lights/allegro/),
there's a copper heat sink with lots of fins. Under normal use (a few
seconds at a time), heat capacity of the heat sink is enough to keep the
LED cool. With continuous operation, though, we have to force air by the
copper finned heat sink. The heat sink is machined out of a block of
copper and is pretty heavy. We have a thermistor sticking up into the heat
sink. It adjusts the fan speed based on the (PIC measured) heat sink
temperature.
> Does your heat sink have any fins on it? Sounds like you are just
> describing an aluminum plate. Fins equal much more surface area which
> will
> do a much better job of cooling the LED. There are also much better heat
> sink greases, but they won't make a lot of difference without an adequate
> heat sink.
>
> Brian Kraut
> Engineering Alternatives, Inc.
> http://www.engalt.com
>
>Hi to all engineers,
>
>I am testing Luxeon K2 1500 ma 3,6 v led no matter what I do
>it gets real hot 50 degree C in 30 minutes this is what I do.
>
>1. I use constant current module 700ma 3,6 v " I am giving half power"
>2. 1,5 x 1,5 x 1/4 thickness piece of aluminum
>3. I used white compound grease in between heat sink.
>
>when I talked to future electronics engineers one guy said it is not
>suppose to get that hot then other said the heat sink should be 3 times
>bigger then today some one else said what I use is very bad heat sink.
>My question is does any one used K2 leds and know how to dissipate
>heat? I want to use 3 of them in 3-4 inch round aluminum heat sink.
>The thickness can be higher if I needed. any info will appreciate.
Work out what temperature you want to run at. Let's say 15 deg rise above ambient = 40 deg at 25
ambient.
Work out the wattage - input power is 3.6x1.5 = 5.4w
Not sure what the efficiency is but assume, say 5w of heat to get rid of
You need a heatsink thermal resistance of 3 deg.C/watt for 15 deg rise at 5W (15 deg /5w=3) - take a
look at heatsinks of this sort of rating in a catalogue to get an idea of the sort of size/surface
area needed. It will be somewhat more than a 1.5x1.5 inch square.
3 of them on a 3-4" sink is going to need either some very long fins on the back or a fan.
An example can be seen in the Lamina catalogue
www.laminaceramics.com/docs/AccessoriesDatasheet.pdf
Page 4 - The Atlas devices are similarly rated to the K2
> I am testing Luxeon K2 1500 ma 3,6 v led no matter what I do
> it gets real hot 50 degree C in 30 minutes this is what I do.
I did some testing on Luxeon V starts, with built in heatsink. They
get HOT. 15 mins, solder starts smoking. Heatink with grooves,
roughly Lucozade Cap (sorry been down pub) sized keeps them 'safe'.
Cant be too far off what a K2 uses.
>I assume the heat sink itself is getting that hot, indicating the thermal
> resistance between the LED and the heat sink (the grease) is not
> significant, but the thermal resistance between the heat sink and ambient
> is significant. About all you can do is increase surface area and increase
> air flow. In a project I did (http://www.denmat.com/lights/allegro/),
> there's a copper heat sink with lots of fins. Under normal use (a few
> seconds at a time), heat capacity of the heat sink is enough to keep the
> LED cool. With continuous operation, though, we have to force air by the
> copper finned heat sink. The heat sink is machined out of a block of
> copper and is pretty heavy. We have a thermistor sticking up into the heat
> sink. It adjusts the fan speed based on the (PIC measured) heat sink
> temperature.
>
> Harold
I am assuming you are using some sort of light pipe there. There is no way
you are cooling an LED out at the tip of that thing!
>> 3 of them on a 3-4" sink is going to need either
>> some very long fins on the back or a fan.
>
> It will need more than a fan, probably some form of liquid
> cooling to pull the heat away to be radiated elsewhere.
Why? A high-end desktop CPU dissipates close to 100W and uses a heat-
sink not much bigger than that (with a fan.) While you want the LEDs
somewhat cooler than a CPU (?), you're only talking about less than
20W...
(Old CPU fans tend to be pretty cheap and easy to find, and are a
good way to experiment with high power LEDs. Carving things down to
a useful size for something like a flashlight is a separate problem :-)
>Why? A high-end desktop CPU dissipates close to 100W and uses
>a heat-sink not much bigger than that (with a fan.) While you
>want the LEDs somewhat cooler than a CPU (?), you're only
>talking about less than 20W...
OK, maybe I am reading it wrong, I read it as 0.75", i.e. attempting to put
three LEDs very close together, possibly to make a small multicoloured light
source.
Sure, if he is able to mount them on a finned piece the size of a CPU
heatsink, then a fan should be fine.
I don't get this. We're talking 2.5W of dissipation. The design power for
CPU chips is sometimes over 100W, and often over 50W.
You should certainly be able to adequately cool this thing with a reasonable
heatsink. A 1.5"x1.5"x0.25" chunk of Al is *not* a reasonable heatsink.
I get to thinking about good 'ole class "A" audio power amps. Again, the
heatsinks weren't all that amazing. Sometimes just an aluminum back panel
for the enclosure, although often with fins (sometimes inside the case with
appropriate ventilation holes. And, short of the really juicy ones, they
were convection cooled. No fan needed.
I find this discussion very interesting because I calculate that the
landscape lighting I use to illuminate my American flag all night (flag
etiquette says it is fine to fly a flag at night: if it is lighted) adds up
to a good part of my power budget. When I go out and point my Luxeon
flashlight at the flag it looks pretty good (I'd need more), and that puppy
sure aint pushin' 700ma.
NOW: 100W * 12 hours * 270 days = 324 kWh
(I take it down in the worst weather)
I pay about US$ 0.17 per kWh (even though I live just about next door (20min
drive) from one of the largest hydroelectric plants in the world).
324 kWh * $0.17= $55
[plus bulbs at about $10 a pop]
So, I was thinking a 2 or 3 K2's or such would do the job nicely. I even did
the math on the heat, but I must have slipped a decimal point of something,
because it certainly didn't look to horrible to me.
> I find this discussion very interesting because I calculate
> that the landscape lighting I use to illuminate my American
> flag all night (flag etiquette says it is fine to fly a flag
> at night: if it is lighted) adds up to a good part of my
> power budget. When I go out and point my Luxeon flashlight at
> the flag it looks pretty good (I'd need more), and that puppy
> sure aint pushin' 700ma.
>
> NOW: 100W * 12 hours * 270 days = 324 kWh
Some people <http://www.darkskies.org> would prefer you to rely on sunshine, it's
free, after all.
The global warming people might have a say as well.
>
> So, I was thinking a 2 or 3 K2's or such would do the job nicely. I
> even did
> the math on the heat, but I must have slipped a decimal point of
> something,
> because it certainly didn't look to horrible to me.
Wire one up, hold it in your hand. Then tell me that.
Welcome to Luxeon headache. If you use constant current and
biggest heat sink and apply half power to it I promise
it will cool down and if you talk to their engineers the
answer is it is all your problem the part is well tested
it gets 25c on normal operation which I never seen.
>>I assume the heat sink itself is getting that hot, indicating the thermal
>> resistance between the LED and the heat sink (the grease) is not
>> significant, but the thermal resistance between the heat sink and
>> ambient
>> is significant. About all you can do is increase surface area and
>> increase
>> air flow. In a project I did (http://www.denmat.com/lights/allegro/),
>> there's a copper heat sink with lots of fins. Under normal use (a few
>> seconds at a time), heat capacity of the heat sink is enough to keep the
>> LED cool. With continuous operation, though, we have to force air by the
>> copper finned heat sink. The heat sink is machined out of a block of
>> copper and is pretty heavy. We have a thermistor sticking up into the
>> heat
>> sink. It adjusts the fan speed based on the (PIC measured) heat sink
>> temperature.
>>
>> Harold
> I am assuming you are using some sort of light pipe there. There is no way
> you are cooling an LED out at the tip of that thing!
>
> (especially in my mouth!)
>
> -- Bob
Yep, there's a light pipe! In your mouth, though, we could do vapor phase
cooling...
>> I find this discussion very interesting because I calculate
>> that the landscape lighting I use to illuminate my American
>> flag all night (flag etiquette says it is fine to fly a flag
>> at night: if it is lighted) adds up to a good part of my
>> power budget. When I go out and point my Luxeon flashlight at
>> the flag it looks pretty good (I'd need more), and that puppy
>> sure aint pushin' 700ma.
>>
>> NOW: 100W * 12 hours * 270 days = 324 kWh
>
>
> Some people <http://www.darkskies.org> would prefer you to rely on sunshine, it's
> free, after all.
>
Well, that isn't a bad idea, but the cost of a light pipe from China to my
front yard to illuminate my flag at night might be a bit prohibitive!
And, of course, my light is powered by sunshine. At least partially by the
local hydro (can you say water cycle) and unfortunately also by nasty fossil
plants.
>> So, I was thinking a 2 or 3 K2's or such would do the job nicely. I
>> even did
>> the math on the heat, but I must have slipped a decimal point of
>> something,
>> because it certainly didn't look to horrible to me.
>
> Wire one up, hold it in your hand. Then tell me that.
No, I know that. My point of reference was the little 7-1/2 watt nite light
bulbs. I figured I had a good idea of their heat output.
> Well, that isn't a bad idea, but the cost of a light pipe from China
> to my
> front yard to illuminate my flag at night might be a bit
> prohibitive!
It would be even dearer for one from somewhere North of Kerguelen
island.
> >> I find this discussion very interesting because I calculate
> >> that the landscape lighting I use to illuminate my American
> >> flag all night (flag etiquette says it is fine to fly a flag
> >> at night: if it is lighted) adds up to a good part of my
> >> power budget. When I go out and point my Luxeon flashlight at
> >> the flag it looks pretty good (I'd need more), and that puppy
> >> sure aint pushin' 700ma.
> >>
> >> NOW: 100W * 12 hours * 270 days = 324 kWh
> >
> >
> > Some people <http://www.darkskies.org> would prefer you to rely on
> sunshine, it's
> > free, after all.
> >
>
> Well, that isn't a bad idea, but the cost of a light pipe
> from China to my
> front yard to illuminate my flag at night might be a bit prohibitive!
>
> And, of course, my light is powered by sunshine. At least
> partially by the
> local hydro (can you say water cycle) and unfortunately also
> by nasty fossil
> plants.
I've got some black cloth you can use instead, it makes the perfect flag.
Looks the same in daylight as it does at night.
rivernet.com.au>
