OddOne:
I put eight Rebels onto a single 1" dia. round double-sided PCB and have no thermal issues with them as long as the board is bolted to a heatsink of some sort.What I did was design the board to put six 0.02" dia. vias as close to directly under the thermal pad on each Rebel as I could, and use silver solder paste and my [homebuilt reflow soldering griddle setup to reflow the LEDs onto the PCB. I then carefully backfilled the vias with solder (taking care to not heat the individual LEDs enough to make the lead pad solder let go of the LED and let it slip or fall off) and smoothed out the backside of the board as best I could.
Thermal transfer seems to be more than adequate for continuous operation (albeit only for a couple minutes at a time as the little bastards are bright!) with eight 100lm white Rebels at full power when I stick it down to a 4" square of half-inch-thick aluminum bar stock.[/quote]
What’s this PCB for? That’s some serious light output density… Did somebody ask you to build a flashlight with infinity candles output, and this was the closest you could get to that?](http://www.e-f-w.com/community/content.php?cid=5ad25bd584990c62a8215040ee6c491e)
I’ve considered building small boards to light nano-aquariums, and was dreading the MPCB route. Thanks for this thread, maybe my other project will get off the ground now.
NleahciM:
So I’m a bit out of the loop here. Some of these high power LEDs have thermal slugs on their bottoms that are neither connected to their anodes or their cathodes. Are they completely electrically isolated from the LED? Or are they just connected to another part of the LED? I mean like - if you have an array of high power LEDs - can these slugs be tied together?
The Luxeon Rebel series LED’s are built on ceramic and have an electrically insulated thermal pad…
There are 2 things here. One is that the pad is electrically neutral and may be connected to each other or a grounded heatsink at another potential- within reason. That last part I feel compelled to mention because there’s gotta be some kind of limit the mfgs have not characterized. If you put a bunch of them in a 110v string with a grounded heatsink, the top one has 110v from die-to-sink and that sort of thing might destroy it. We don’t know what strength dielectric each mfg built in there because they don’t say. Luxeon, I don’t know if they use a metal oxide layer or ceramic or what between the die and the outer pad, all we know is it’s an insulator.
The other thing is that the pad is solderable. The older stuff like Luxeons were aluminum, which can’t be soldered, and they could only be attached with thermal epoxy. Couldn’t use thermal grease on that type either because there’s no place to screw down the emitter to provide the pressure on the board. To make life even harder the Star boards were often not electrically insulating either and the problem became even more complicated.
The electrically neutral/solderable pads rock in theory, because 1) they can be reflowed with a heatgun, 2) solder has a real low thermal resistance, 3) with no electrical insulator needed we can get metal-to-metal heatsink connections which have the potential to have much better thermal resistances. Low thermal resistances have a lot to do with how much power you can get from an LED before the die temp rises and efficiency and lifespan are lost.
Now, though, they keep making boards for Rebels and Crees and stuff which already have unnecessary electrical insulation in the device itself on electrically insulating MCPCB. Which sucks, because the thermal resistance of the electrical insulator is fairly high compared to thermal via construction. The thing is this is the way MCPCB is made and it’s not easy to do anything else.
OddOne: how did you determine you have no thermal problems? Without a proper testing procedure it is very difficult to know what your dies are doing. But high temps will definitely reduce the power output and prematurely degrade the die whether you can see the difference or not.
NleahciM: because a solid piece of copper put through the board will not mate up well with either the top or bottom surface, may not stay in place during reflow, and may experience thermal or mechanical stress. If there’s not a high degree of flatness between the copper lug and the board and pads then the device may not solder correctly. Then if the bottom is not flat it will increase the thermal resistance to the heatsink.
Trying to bolt down the board on the sink may compress the surrounding FR4 while this solid copper stud does not compress and instead punches upward, stressing the LED’s body. Could lift one of the electrical pads or break the LED. Thermal stresses may do that too. Thermal expansion of the FR4 could actually make the perfectly spaced solid copper stud, soldered to the LED, pull away from the sink as the FR4 expands and it overheats the device instantly.
Could it work? Well, maybe. Maybe you could drill holes around (but not under) the LED thermal pad, jam in some tight-fitting solid copper wires into the holes, and cut the protruding tips off flat with a milling machine. This… well, it’s a lot of work and hard to say if it’ll work at all much less whether it’s going to perform properly and work in the long term.
FYI GP minimum hole size is 0.075mm, it’s on their website.
NleahciM:
What’s this PCB for? That’s some serious light output density… Did somebody ask you to build a flashlight with infinity candles output, and this was the closest you could get to that?
A prototype I’m working on. The goal was to put the absolute max in terms of lumens in the smallest possible footprint while making the works able to actually be usable in a practical sense. I figure ~800 lm total @ 350mA/3.4VDC per LED is pretty much the limit for the current tech and packaging.
Oznog:
OddOne: how did you determine you have no thermal problems? Without a proper testing procedure it is very difficult to know what your dies are doing. But high temps will definitely reduce the power output and prematurely degrade the die whether you can see the difference or not.
I have “no thermal problems” in the sense that when I bolt the PCB to an aluminum plate (2" x 2" x 1/2") the temp for both plate and the portions of the Rebels I can get my probe to touch climbs up to ~40-45C in room temp air and with all eight LEDs taking 3.4VDC @ 350mA each, and both temp and light output are stable (as much as my decidedly hobbyist-level gear can resolve such variations) over time. (“Time” in this case being 30 minutes continuously.)
Shy of spending pointless amounts of $ to purchase more thorough testing gear to quantify things more precisely, I’m satisfied with what I’ve been able to come up with and how well it seems to work.
Hi, first post. But I’ve been looking into the same issues for a new daytime readable LED backlight for LCDs. Since I was going to prototype this with a home etched PCB, I don’t have thermal vias available to me. But I realized that soldering a solid piece of copper wire between the top and bottom layers will work as well if not better than a via.
researching this idea a bit, I came across this:
http://www.ruwel.com/_upload/2008_38/Ku … 008_EN.pdf
Basically, a copper plug to pass heat directly from the top layer to the bottom. And for our LED purpose, we can do the exact same thing ourselves. For the thermal pad of the LED, create a 0.05" plated through hole. Then plug the hole with a 16 AWG solid copper wire. The trick would be laping the ends flat. The 16AWG wire should be just big enough to press into the hole and stay. Now you can solder the thermal pad right to this copper plug and pass the heat directly to the bottom layer and the heatsink. It should even be possible to pre-cut the copper wires so that minimal laping is needed.
The idea should still work without a plated through hole as long as you can drill a 0.05" hole to press in the copper wire. It will still pass heat from top to bottom, but heat transfer to the copper layers will depend on solder to fill any gaps.
Oznog, thats some interesting numbers you have there. The real question though, theory aside, has anyone actually tried a side by side comparison of a decent MCPCB vs an equivalent size FR4 with thermal vias…
brenhale:
Oznog, thats some interesting numbers you have there. The real question though, theory aside, has anyone actually tried a side by side comparison of a decent MCPCB vs an equivalent size FR4 with thermal vias…
Lumileds application note AB32 shows a comparison of thermal resistance for MCPCB and FR4, as a function of LED spacing. (http://www.philipslumileds.com/pdfs/AB32.pdf See page 17.)
For example, for Rebels spaced 2mm apart:
MCPCB: 11 K/W
FR4 open PTH vias: 7 K/W
FR4 filled and capped vias: 3 K/W