Ways of LED attachement to PCB

[RobEU]

1. Usual way - oven soldering. But thats quite a stressful for leds.
2. NanoBond with NanoFoil:
http://www.rntfoil.com/site/applications/electronics-assembly/hb-led-attach

I`m trying to explore all possible ways of led attachement to PCB.

Any other?

 

[DM51]

Moving to LED...

 

[BurmaJones]

"1. Usual way - oven soldering. But thats quite a stressful for leds."

Not so much, if done per specifications.


"2. NanoBond with NanoFoil: . . . "

Odd that they compare their process with reflow while they compare thermal conductivity with adhesive (worse than any kind of soldering). And, how would you possibly pass an electric pulse thru an electrically isolated thermal pad?

For hobbyist and small production runs, here is a simple and inexpensive way to approximate oven soldering with good control:
1) get an old skillet made from thick aluminium
2) get a cheap (~$15) electronic oven thermometer
3) get some large metal tweezers, needle nose pliers, etc.
4) attach the thermometer probe to the skillet, inside on the flat area
5) place the skillet on your kitchen stove or other controllable heat source
6) slowly bring the skillet to soldering temperature as specified by the board and LED manufacturers
7) place the circuit board in the skillet
8) apply any kind of solder you want
9) position the LED(s) on the solder
10) observe wetting within the specified time limit
11) carefully remove the assembly to cool

This seems simpler than simultaneously positioning the LED and preforms while applying pressure and electricity, no?

 

[lolzertank]

I guess large packages like the CST-360 could be attached with thermal grease and screws.

"Mechanical with thermal grease"

 

[mudman cj]

In case anyone is interested, let's compare the thermal performance with solder. They claim a 6-10x lower thermal resistance than adhesives. I think it is safe to assume that alumina-filled epoxies would be at the 10x end while silver-filled epoxies would be at the 6x end. Silver-filled epoxies are between 20 and 40 times worse than solders, depending upon the composition and assuming RoHS compliance. Using the solder recommended by Cree for use with XP series LEDs, the solder would be 40 times better than silver-filled epoxy. So, in this case the foil would have 40/6=6.66 times higher thermal resistance than the solder.

Now let's consider if this is a process that can be performed by hobbysists. The LED test report linked on their website discloses an 86% success rate using a manual application method. This is the success rate of the scientists that developed the product! I doubt a hobbyist is going to match that success rate. Then there is the issue already raised by BurmaJones: how are you going to pass the pulse through the pad? It requires 100 Amps of current while applying 1000 psi of pressure! Presumably, the scientists that achieved an 86% success rate built a special machine designed to do only this task. Feeling lucky?