Ouch! Is it normal for LEDs to get so hot?

Fallingwater

Flashlight Enthusiast
Joined
Jul 11, 2005
Messages
3,323
Location
Trieste, Italy
I was one of the lucky few who managed to grab one of these ten-packs of generic one-watters before they got sold out.

That page says the typical driving current is 400ma. So I set my power supply to 400ma, connected it and the led lit up, nice and bright.

After about 10 to 15 seconds I thought "hmm, I wonder how warm it's running", put my hand on the bare emitter and actually scalded it a little.

I didn't get a burn, mind you, but the spot on my hand where I touched it was painful for about four or five minutes afterwards.

As far as bare emitters go I've only ever had to deal with 5mm 30ma (or so) emitters, which don't really emit any significant heat at all. I've only used high-power ones inside lights, so I don't know if such a high temperature is normal or not. Is it? Or is the emitter very quickly cooking itself to death?

Thanks.
 
Yes, high-flux LEDs will get hot at reasonable drive currents. That's why you need to heatsink them.
 
I can't heatsink the damn things. They are soldered to a small base with contacts running all around it and in the middle. Any metal touching the base would short the whole thing.
 
I can't heatsink the damn things. They are soldered to a small base with contacts running all around it and in the middle. Any metal touching the base would short the whole thing.

Perhaps the heatsink is part of the electrcal contacts. With this much current, they must be heatsinked or they will not last and will not operate efficiently at elevated temps.
John
 
Do all the Arctic Alumina, Silver, compounds, expoxies, etc, only conduct heat, but don't conduct electricty? Meaning it's safe to get it all over all the electrial contacts with out the fear of shorting things out? If so, I'm going to get myself some and go crazy. :)
 
Do all the Arctic Alumina, Silver, compounds, expoxies, etc, only conduct heat, but don't conduct electricty? Meaning it's safe to get it all over all the electrial contacts with out the fear of shorting things out? If so, I'm going to get myself some and go crazy. :)

I believe only a few conducts electricity, for most of them, they don't.
 
I think there is some confusion, the wires connect from the top, the heatsink usually is bonded to the BOTTOM.

If you have some heatsink grease, just smear a tiny dab on the bottom then set the emitter on an aluminum plate and then repeat the test.

Even without heatsink grease, if you have the bottom of the emitter in good physical contact with a smooth aluminum plate there should be a big improvement.
 
they look like 'high-power' 5mm LEDs mounted on a PCB. i doubt that they'll last long @400ma drive currents even with good heatsinking. :)
 
I think there is some confusion, the wires connect from the top, the heatsink usually is bonded to the BOTTOM.
Have you looked at the picture in the link I provided?
The connectors are on the bottom.

they look like 'high-power' 5mm LEDs mounted on a PCB. i doubt that they'll last long @400ma drive currents even with good heatsinking.
The die is much larger than that of 5mm LEDs. But the one I tried got so awesomely hot in such a short time that I too am inclined to think 400ma is too much.
 
Have you looked at the picture in the link I provided?
The connectors are on the bottom.

My apologies, I did look, but did not zoom in, and at first blush they looked like round alumna boards with the conductors on the top like the alumna or MCPCB (metal core PC Board) materials used for the CREEs or Luxeons.

Now I can see that the boards are just translucent epoxy type material.

The die is much larger than that of 5mm LEDs. But the one I tried got so awesomely hot in such a short time that I too am inclined to think 400ma is too much.

I think you are correct, there is no way those things will work reliably long term at 1w continuous output. 1 Watt is probably the max power rating at 25 degrees C die temperature. The only problem is that at 1 Watt, without any way to properly heatsink them, they won't stay at 25 C, so this kind of peak power spec is useless. If you derate them to about 1/3 watt to 1/4 watt and drive them at no more than 100ma they will probably hold together, but I couldn't swear to it. If they have reasonable luminous efficiency they still might work out for a nice area lighting project.
 
^ what he said. is the emitter portion sitting above the board itself, with legs, or sitting totally flush with the board?

the 100ma quad die 8mm leds were highly leg heat sinc dependant, if you could get the legs heat sinked they would survive, One leg in particular gets hotter.

for the 100ma quad dies, we mounted the legs themselves to a set of heatsinc rails, they worked after that without much problem, i jammed some 8 of them into a 2x4 space, with this rail folded about , and moving air through convection of its own heat.

so if you can sock the legs down to something, mabey potted in something? the legs because they are metal, and conduct the most.
then also containment, heat DOES go out the front of things, and out the sides and bottom of the epoxy, the more you contain or block the front, the less you can drive it without it dying.
 
Last edited:
So the more heat a light gives off, the less efficient it is?

I assume heat is just wasted energy sucking off the batteries.
 
So the more heat a light gives off, the less efficient it is?

I assume heat is just wasted energy sucking off the batteries.

The more heat anything gives off, the less efficient it is (heaters excepted). However, this is not to say that the hotter the outside of the light gets, the less efficient it is, with current LED technology. If the outside of a light gets hot quickly, then there's a god thermal pathway for the heat to dissipate from the immediate area of the emitter. If not, then the heat stays near the emitter, effectively roasting it.

So it's safe to say the more heat an emitter produces, the less efficient it is.
 
Do all the Arctic Alumina, Silver, compounds, expoxies, etc, only conduct heat, but don't conduct electricty? Meaning it's safe to get it all over all the electrial contacts with out the fear of shorting things out? If so, I'm going to get myself some and go crazy. :)

Some are capacitive.

I like this thermal epoxy which is neither conductive nor capacitive:

"Arctic Alumina Adhesive is a pure electrical
insulator, neither electrically conductive nor capacitive."

http://www.arcticsilver.com/arctic_alumina_thermal_adhesive.htm
 
Last edited:
Fallingwater, apart from the heat, how bright are they and what's the beam angle like?
Brightness obviously depends on the current.
At 400mah they are pretty bright, but as we've seen they'd probably not last long.
The beam is very floody. There is no clearly defined circle of light as in reflectored high-intensity LEDs or 5mm ones, it's more of a general pool of light. They would be perfect as area lights/lanterns, not so much for flashlights.

I can take a beamshot or two if you're interested.
 
Yes, I'm definitely interested (assuming they come back in stock you-know-where). Is there a good way to show their brightness? There should be some standard (and common) comparison we can use for that.
 
I think these LEDS are 8mm strawhat one (i have 50 of them), 100mA rated, 150mA tops!
8mmstraw100maLED.jpg

The light projection beam is similar to 5chip superflux
one, i will record this one later and put pic on my picasa album...
The pics are all taken with same parameters so you can judge the angle, brightness, color....
z
photo
 
Last edited:

Latest posts

Top