3 watt LED mounting..

[sndscientist]

ok my question is this, has anyone ever tried to use the simple 2 part epoxy to mount a 3 watt led. There is a lot of talk about the arctic silver and similar special thermal epoxies. during my digging i could not find a single spec sheet showing the thermal properties of simple 2 part epoxy. i have tried various manufacturers like JB and Gorilla but no mention of thermal conductivity. one forum that i found mentioned that i should use GE silicone-II adhesive, which i know for a fact is an insulator. it seems like people are grasping at straws with this and nobody really knows. I'm tempted to grab a 3 watt star epoxy it down and probe the temp after 20 mins. it has always seemed that smooth solid, non porous surfaces should have good thermal properties maybe not the best but this is a 3 watt led

 

[chopnhack]

ok my question is this, has anyone ever tried to use the simple 2 part epoxy to mount a 3 watt led. There is a lot of talk about the arctic silver and similar special thermal epoxies. during my digging i could not find a single spec sheet showing the thermal properties of simple 2 part epoxy. i have tried various manufacturers like JB and Gorilla but no mention of thermal conductivity. one forum that i found mentioned that i should use GE silicone-II adhesive, which i know for a fact is an insulator. it seems like people are grasping at straws with this and nobody really knows. I'm tempted to grab a 3 watt star epoxy it down and probe the temp after 20 mins. it has always seemed that smooth solid, non porous surfaces should have good thermal properties maybe not the best but this is a 3 watt led

Give it a go! LOL - If I recall correctly the JB Weld has steel in it so it may be somewhat conductive.

 

[sndscientist]

after posting last night i applied a drop of gorilla 5 min epoxy and pressed a 3 watt star on it. i applied pressure until the epoxy oozed out from all sides and smeared the star around a tad to make sure i had no air bubbles. this morning i soldered on a 18650 battery holder and popped a battery in it. after 15 mins i can feel the heat-sink getting warm (above room temp) but not hot. at the same time the stars aluminum feels the same temp. granted I'm only a short time in, this seems to be working OK actually

~Edit.. i checked the battery voltage. it seemed a tad dimmer than i expected. the battery was nearly dead at about 3.5 volts O.C. quick swap now i'm cooking with gas. 3.8 volts across the led everything is getting warm and i'm getting excited, it seems like this may work. after all the magic smoke didn't come out yet and it's been almost 10 mins since the battery change. the plate is decently warm

 

[wdariusw]

I do not know, but maybe you can use TEG as with heatsink.

 

[BobK]

You really don't want to running a 3W LED by connecting it directly to a 3.7V Lithium battery. You need a constant current source.

Bob

 

[(*steve*)]

The good news is that the LED seems to be a similar temp to the heatsink.

You did exactly the right thing by using a small amount of adhesive and then ensuring that the excess was squeezed out.

Now you need to work on a constant current driver as indicated above by Bob.

 

[sndscientist]

this was just a mounting test. i just didn't want to have to drill holes in places that i can hardly get a drill. i can build a constant current driver easily. the purpose of me attaching the 18650 is it keeps it near it's limit. it was drawing about 630ma for a good half hour from that battery before the voltage dropped. i just wanted that little darling to heat up as much and as fast as possible in a short amount of time.the voltage across the star was about 3.9

The thing with not knowing the thermal properties of the epoxy is if it was a very slow gradual heat, there is no doubt the epoxy would transfer the heat and i wouldn't know if it was a good thermal conductor or not. remember even plastic takes heat it just takes a while.

i had been looking into this for a while, trying to find a data sheet that listed the thermal properties of 5 min epoxy, which if it's there i can't find. so like i said before i just glued a star down to a flat to220 heatsink i had laying around.

on a side note. technically with these a 3.7V lithium works just fine. there is no need for a resistor or any other electronics. i have a few of these that i use for power outages etc that are just a battery/switch and 3watt star/heatsink. these diodes have a voltage drop of about 3.6 volts so the 3.7 is fine. .1v - .3v up in heat. now if i wanted these to have a long happy life and last for tens of thousands of hours that would be different and i would take better care of them. but these cost less than 20 cents each.

 

[KrisBlueNZ]

If this is a test to prove the suitability of epoxy for use later, you may need to consider the matching of the two surfaces. If the surfaces you used in this test were both very flat, or happen to be deformed in the same way, the epoxy layer between them will be very thin, and this may contribute to the good performance you see. This may give you a false impression of the suitability of epoxy, and if you can't consistently repeat that surface matching in future, epoxy may not continue to be as effective as it was in your first test. Just something to consider.

 

[(*steve*)]

Just another note. The idea of heatsinking is to have metal to metal contact between surfaces.

In real life, even flat surfaces have imperfections, so even when bolted down, a device will touch in some places and not in others. This means that the effective contact area may be significantly smaller than what you think it is.

Air fills the gaps between the metal surfaces. Air is a very poor conductor of heat.

When heatsink compound is used, the idea is to use enough to maintain the same metal to metal contact, but to fill the remaining air pockets with the compound. The heatsink compound itself is also a poor conductor if heat! However it is many times better than air.

Epoxy is probably not a great conductor of heat, however, of you can use a sufficiently small quantity that it simply fills the voids between the metal surfaces it will both adhere the two surfaces and fill the air gaps, probably doing a reasonable job of acting as a heatsink compound.

If possible, I would keep pressure on the joint until it cures.

It is worth remembering that there are about a bazzilion heatsink compounds, many with extraordinary claims about their performance, but a test has shown that you can get 90% of the performance of the best of them with toothpaste.

There are a whole host of reasons why you wouldn't use toothpaste though. You have no idea how it will change over time, whether it will react with the metals, dry out, etc. However you do know these things with epoxy. The largest downside with epoxy is that it sets solid -- but that's what you want in this application.

 

[sndscientist]

here i will give you the reason for the epoxy, this isn't commercial it's personal. i have a tremendous (around 180) of these things laying around from past projects. the past projects were outdoor spot lights, flood lights and i think i made a few ceiling cans. in those they were greased and clamped. 2 nylon screws set 180 degrees apart with stainless nuts on an aluminum backing plate. fast forward about a year (2 weeks ago) digging through my bins i found them and remembered how many i had purchased so i started trying to find other uses for them. automotive kinda stuck with me since the price of the 120 volt E27 LED bulbs dropped. In the vehicle there are a few places that gaining access to the rear of the plate are impossible, and thanks to the wiring harness self tapping screws are a no go. which leaves me with my choice of adhesives.

granted the epoxy isn't ideal, yes i know that a flat metal to metal with a thin layer of grease would be perfect in this case it's not possible. the only way i could do that is to screw it to a piece of metal and zip tie it up. the negative of that is zip ties don't like the PA temp extremes. they snap or randomly fall off if they are allowed to move at all. great for wires bad for anything else. i did order a cheapie tube of the thermal plaster from e-bay to see how or if it works.

so having that said and knowing that a 3 watt led aimed at the floor of a vehicle from less than a foot away probably doesn't need the full 650Ma to be outlandishly bright. i was thinking of regulating it down to 400Ma and basically just need to keep it from burning up

 

[Merlin3189]

You're right about there being limited info on epoxies. Presumably they think it's of interest only to industrial scale buyers.

But if it's of any help, ALH make NP2001 water clear fast cure resin with specimen thermal conductivity of 0.2 W / K m (cf their high thermal conductivity resin at 1.3 W / K m )

After that I can only guess, because there are so many unknowns.

A 3W led I looked up has 10 deg/W chip to metal base, with only 0.36 sq cm base.
If you said 0.1 mm thickness for the glue, you get 14 deg/Watt.
(When I measured a bit of Araldite between two smooth steel plates, it was about 0.1mm with finger pressure. On warming to liquify the glue I could get it to less than 0.05mm , which would give 7 deg/W )
So chip to heatsink could be 17 to 24 deg/W , ie. chip 51 to 72 deg above heatsink at 3W.

As for heatsinks your guess is as good as mine when it comes to random bits of metal.
(This gave some hints http://www.giangrandi.ch/electronics/thcalc/thcalc.shtml)

Anyhow, you're probably as well off trying it empirically as you are!