This is an LED lamp design for a big honking array of pluggable modules. There will be one control module for supplying power and signal, and then a number of LED modules can be plugged alongside. The object is to make a system that can have replaceable modules, and can be expanded by adding more modules and using a bigger power supply.
I’m sticking to the formula so far, Cree XP-G (or XP-C, or XP-E) LEDs, CAT4101 drivers, hardware PWM, and plenty of ESD protection.
There is a control module, one is required per lamp:
Then there are lamp modules, this is the Cree XP-G module:
The board-edge connectors are pretty big in comparison to the boards, the dimmer is 25 mm wide and the lamp modules are 40mm wide each. They will be 100mm (about 4 inches) tall.
This lamp is the big cahuna, the momba mutha, the big mofocky. I want a wall long array of LEDs, I want it brighter than downtown Las Vegas. The major problem here is that if I run it at the same average temperature as downtown Las Vegas, it’ll burn out like a $5 roulette bet on double zero. It will need a serious thermal solution, and but it won’t could be solved with conventional methods. I need a fully scalable industrial strength solution, invented in a basement. So obviously, this also needs to be simple, this is not nuclear science, it’s more akin to building a decent ice house using the refrigeration technology of 1881. I need some kind of thermal transfer using downtown hardware store materials.
The copper heat pipes being used in laptop computers and the high end CPU coolers (that I’m trying) are basically copper tubes with some type of wicking inside, plus some thermal transfer fluid. Several technologies for wicking the internal fluid are named and researchable. The fluid in the pipes is likely water, maybe with some ammonia, but I’m also not willing to chop one into bits to find out, especially after paying $40 for a used one (one of the copper ZeroTherm BTF95 coolers came from a resale site, I had no problem at all with a purchase from NextStep Recycling in Eugene, Oregon). I think it’s just water in a partial vacuum. The pipes are flattened in some forms, and I don’t know if the pipes are reshaped before or after the filling, that might make the vacuum theory weird.
I won’t redo heat pipe function here. I would like to be able to make heat pipes myself using copper water line, heat proof fiberglass stove gasket material for wicking, and a fluid with a boiling point between room temperature and the max I expect to run a lamp, which would be from 75 – 85 degrees Celcius. Making the pipes myself, and probably using some type of blowtorch to seal them, it would certainly help if the fluid was non-flammable.
Acetone: would be easy except it explodes if you look at it cross-eyed, very flammable. Same goes for pentane (one atom up from butane, several down from octane). Methanol and Ethanol, both flammable, might work, but obvious problems. There’s stuff called perfluorohexane, which looks interesting, but hard to get and $16 for 25 grams if I could get it.
So I found something. It was hard to get, but I got a bottle of it. I can’t say what it is, as I’m not entirely sure I should have it. It’s safe for me to test with, but not for production, so let’s say I have some just to see how well it will work. Yeah, that’s my story and I’m sticking to it. In comparison the The Unnamed Agent, I’ll also try distilled water.
First, I make them. Hammer flat one end of a 3 foot section of cheap copper water line, solder it, shove in some fiberglass rope, add fluid, hammer other end, and solder that end. It’s probably best if the pipe is hot while sealing it, as a pipe sealed cold will have more pressure in it than one sealed hot when it is later heated. So maybe have the first end in hot water when sealing it.
To test, stick one end in boiling water, and run away. 100 degrees Celcius is more than I ever want the circuit to get, so that should be enough. Maybe. I might have to try torching one just to make sure it *will* explode at some point, need to know what that looks like too. Ya know, to make sure it never happens. Uh, and maybe not. I’m free thinking right now. When I say explode, all I really mean is that it blows out the solder weld at one end. That would be a fail, and a different manufacturing technique or ingredients is required. I iz sew skien-tiffic!
And if it doesn’t ‘splode? What kind of existing aluminum structure can I use to transfer the heat? I have to clamp the tubes to it. I have a foot long piece of aluminum 4″ wide and 1/2″ thick which should fit this circuit nicely.
Would water cooling be easier? How about a closed convection system filled with oil using a small radiator? Many, many questions, easy to ask from the drawing room, much harder to answer without real work. Well, no more workie today, need to play game a bit and then sleep.
This project is definitely behind the other CPU cooler projects, if the circuits on those work, then this one may start to look interesting.
28-NOV-2010 I’ve made a stub board for terminating the ends of an array. It’s not required, but it’s a nice touch. $267.55 total for 5 boards of one tester, one stub board, two lamps, and one dimmer (each). Estimated costs of board parts: $250
24-DEC-2010 I decided to skip trying to get these perfect every single time, and I ordered 6 combination boards from Silver Circuits’ prototype service. They’ll be green and silvery instead of white and gold, but hey, it’s fine. Here’s the image of the panel, I’ll be sawing these apart myself:
That’s a dimmer module, two lamp (Cree XP-G) modules, a stub module, and a tester on the right end. Parts are here, Silver Circuits is making 6 of them for about $130. The parts are here, I combined this Digikey order with parts for the 315 Radio. I got 2700K LEDs for the lamp modules, they will be even a little bit more yellow than the Butterfly Sconces.
Copyright (c) 2010 MDVE.NET. Some rights reserved.
This design is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License .