Make your own... heatsink! · Jun 9, 04:34 pm by Max
Hi again; sorry for the delay in writing this. I had a lovely article ready on building an amplifier, unfortunately, I wrongly adjusted the gain (or perhaps a component shorted out - I'll never know) and, long story short, a 1/4W resistor ended up having about 50W put through it, thus exploding the resistor, and the shrapnel somehow shorting out a couple of nearby transistors as well, (also managing to put 30V DC through a test speaker, shorting that out as well, and - worst of all - shorting out a channel of my iPod which I was using for testing it and which I hadn't bothered connecting in series with a resistor; so now my iPod only plays through one channel). Anyhow, enough ranting...
When building my 800W amplifier, I obviously needed a rather large heatsink (or, rather, four of them). I thought you might want a little (and I mean little!) tutorial on how to build your own while I try to fix the amp. A heatsink is, basically, dissipating the heat efficiently from a component (such as an IC or a single transistor in my case). It should have as high a surface area as possible (you can actually mill grooves out of the fins themselves if you want to increase it that much); however, a little like electricity, larger blocks of metal conduct heat much better than smaller ones; thus, if you have the component joined to the rest of the heatsink by only a piece of thin wire, the heat dissipation will (understandably) be fairly low.
Basically, there are two main methods, the former producing a better quality heatsink but taking a little more time and money (£60 for my 4 175x75x60mm heatsinks if I were to have made them that way), and the latter producing a slightly worse heatsink but taking up less time, being cheaper, and wasting basically no metal. I'll cover both ways in this article.
Method 1: Subtraction
I call this method 'subtraction' because it involves starting out with a solid block of aluminium and milling grooves out. This does give better results, but is much more expensive.
Firstly, you must start out with a block of aluminium of the overall dimensions you want your heatsink to be (Figure 1). You could either order it in or cast it yourself (though you'll need to make your own mould if you choose the latter even though it may be marginally less expensive).
Figure 1: A solid block of cast aluminium.
After having obtained your block of aluminium (with the conductive surface [the one facing the component] as flat as you can make it) you should mill grooves down either horizontally (Figure 2)...
Figure 2: The block with holes milled in 'horizontally'.
...vertically... (Figure 3)
Figure 3: The block with holes milled in 'vertically'.
...or even both, one of the advantages of this method (Figure 4)
Figure 4: The block with holes milled in both ways.
...and, hey presto, you've got your heatsink! (N.B. if you've never milled aluminium before, the trick is to do it in small 'layers' of about a millimeter at a time, and go back and forth. This method is very time-consuming, especially for thick heatsinks.
Method 2: Addition
I've called this 'addition' simply to be the opposite of subtraction. The method sort of is addition; it involves milling smaller grooves into a base block and sticking fins into the grooves.
There's nothing really more to say about this. Take a block of aluminium (Figure 5). This should be fairly thin (much thinner than the block you used for the first method, anyway).
Figure 5: A base block of aluminium (no, really, it is aluminium =D).
Then, mill grooves (horizontally or vertically) into the block - these should go through roughly half the depth of it (Figure 6)
Figure 6: The base block with grooves milled in.
Finally, stick fins inside it. These should also be made of aluminium and should be of the thickness of the hole you milled (in my case 5mm). They *should* fit nicely in (friction fit). If they don't, you'll have to weld the aluminium together (if you have such capabilities, tightly screw it together, or abandon that groove completely.(Figure 7)
"KH" (see comments) has kindly given us this tip: "If you mill the slots for a tight fit you can preheat the base so that it stretches slightly and insert the fins then allow it to cool down for a very tight fit."
Figure 7: The base block with the fins stuck in.
I've actually had a thought. You could perhaps create an (iron?) cast using the second method and cast an aluminium heatsink inside. If anyone tries this method and has any success with it, please leave a comment!
Anyway, that's about it... good luck with your heatsink!
Max,
I’ve done aluminum (err…sorry…aluminium!) casting, using Very Simple Equipment and can offer a few pointers.
1) It’s VERY easy, and safe if you’re smart. However, you are dealing with 660°C/1220°F metal, which is a whole new world of HOT. Learn about casting safety first, and I’m not responsible, etc. etc. ALWAYS WEAR EYE PROTECTION FOR EVERYTHING. Stereoscopic vision is nice, I think. Keep it.
2) You can easily build a “one-shot” furnace with a small (40-liter/10gal) garbage pail or similar, a sack of sand, 2 tin cans, a sack of charcoal, a bit of pipe, a clay flowerpot and a hair dryer. Oh, and I use the hacksawed-off bottom half of a propane canister as a melting crucible (make SURE the canister is TOTALLY empty!. Look at backyardmetalcasting.com, or google “hobbicast” for a good Yahoo! group. Also check out “The Charcoal Forge”, by Dave Gingery (lindsaybks.com or Amazon) for full instructions.
3) I’ve had better luck than some at using “beercanium” as my source of metal. You’ll want to pound the cans as compactly as possible. It’s best to have a puddle or “heel” of molten metal to then add cans to; if you have a piece of more solid aluminium to start with (old window frames, a car piston, etc.) it will go more easily. WATER WILL CAUSE MOLTEN ALUMINIUM TO EXPLODE, causing Grievous Bodily Harm. Molten aluminium is actually less viscous than water – it will go RIGHT through tiny holes, including in thick canvas. Moral – leave the cans out in the sun for a few hours or days, first. (You do have sun in your green and pleasant land, right? >;-)
4) Pour your molten aluminium into a sand mold. (An easy method I’ve not used is to simply cut a chunk of foam to size and put it in common fine sand – this saves mixing up casting sand and making a mold.) Wait a LONG time before shaking out and touching the part – I once picked up a casting after 2 hours – it was still well over 200°C and I had a shiny pink palm for a week. (Fortunately, it was my left hand. >;-)
5) Aluminium sands very easily with normal sandpaper – “wet-dry” from an auto shop, used wet, seems best. Aluminium is ‘gummy’ and quickly clogs sandpaper – you’ll go through a bit, but you’ll have a beautiful shiny block. I use my wood-shop belt-sander, or a random-orbit hand-sander. Wear a mask – trust me.
6) Finally, the fins. A guy on the Gingery Machines Yahoo group told me this one, and I was aghast, but tried it and it worked fine—you can use a regular table saw or router to cut aluminium. You need a carbide-tipped blade or bit, and NEVER, EVER machine metal that’s not FIRMLY clamped with a vise or clamps (cramps in the UK?). NEVER hold a metal workpiece in your hands and use a power tool on it. EVER.
I used my table saw, and a wooden holding fixture, to “mill” a smooth flat face on a 36” ((~90 cm) long casting. Scared the crap out of me at first, but worked nicely. (Don’t use an expensive blade – it’ll dull it for wood use. And don’t even think about using a handheld circular saw. Take a shallow cut, raise the blade a wee bit, and take another. Do NOT stand directly in the path of the blade, and do NOT use your hands to hold/push the workpiece – use push sticks (or better yet, attach a very sturdy/thick wood face/extension to your miter bar and clamp the workpiece to it. If you’re cutting the cross-slots, leaving ‘sticks’, go VERY carefully in very small increments. I’ve no idea whether the forces would snap off some of the sticks – I’ve never done that. Best bet would probably be to simply do it in wood, embed the mold, and cast it in finished form! Or using lost-foam casting – see http://www.buildyouridea.com/foundry.html for great (and fascinating!) photos and tips.
I’m not crazy about method #2 using glue. Cyanoacrylate adhesive (“super glue”) has a working temperature of
30°100°C, which a heatsink might easily reach. Also, you’ve now introduced a small layer of non-conductive material, which may or may not affect heat transfer significantly. The casting idea is interesting – of course, the heat of molten aluminium will melt the glue (and probably any organic substance), but it would probably still hold it in place enough as the mold fills to do the job.ANYway – sorry for the verbosity. Feel free to incorporate any of this if you care, and/or to e-mail me questions/suggestions.
Cheers!
Andrew
Reply from Max:
Wow.... there's a hell of a lot of information here - thanks a lot for commenting! I can't promise that I'll be able to reply fully in the time I've got, but I'll have a basic shot now and I'll definitely update the article with some of your suggestions soon.
Now, where to start? In reply to number 2: I'm lucky enough to have access to a small metal furnace; though I looked around the website and Google and there's a great load of information around the subject. I might get around to building one if I have the time...
As to "beercanium", thanks for the tip. The aluminium bars I use so far are (relatively) expensive and I'm sure I could save myself money by using used Coke/beer/other aluminium cans; I'll see if I can convince a few people to chuck their cans in a bin (or perhaps 'borrow' some from the local recycling centre). Oh, and yes, we do have sun here - the temperature reached 32°C a couple of weeks ago and is still fairly high ;-)
Wouldn't the sand lose its shape pretty quickly after you removed the foam? Anyway, I might try it someday. Oh, and I have had a similar experience with cast lead - nasty. Anyway, what I did was use a milling machine and cut out the grooves that way. I've heard about using a circular saw to cut grooves in that way - however we only have one with a blade for wood and I don't think I'd be able to go anywhere near it with a piece of aluminium ;-). However, I know someone who works in the aluminium industry and has a dedicated special aluminium circular saw at work - I occasionally ask him to cut stuff for me; I might try that method if I need to make another heatsink. Oh, and cramps? Nah... that's something completely different (which you might get as a result of too much leaning down over a workbench...)
When I made my heatsink using the second method, I used a 5mm router bit and the aluminium for the fins was 5mm thick. Thus it was a nice tight fit when hammered in apart from one groove which had slipped slightly, making it only a fraction of a millimetre larger than 5mm and I used CA to stick it in. My heatsink doesn't get as hot as 100°C so I haven't really experienced that - I guess you might be able to weld the aluminium together which, in retrospect, I might have done. Having said that, not everyone has access to aluminium welding gear =). Anyway, you shouldn't really need to use the glue if you do it correctly - and you might be able to find other types of glue which have a higher melting point anyway.
Anyway, that's about all I can think of for now. I don't have any qualms with you being 'too' verbose - quite the opposite in fact. ;-). I'll expand and integrate on this and my article later ;-)
-Max
— Andrew in Silicon Valley Jun 19, 04:17 pm #
You do not remove the foam from the mold. The hot aluminum vaporises the foam and displaces it (thats why its called the lost foam process), similar to the lost wax process. It helps to have metal forms to hold the sand (or coat the foam with some kind of plaster) because then you can preheat the mold and vaporise the foam before pouring the casting.
As for your addition method of adding the fins, if you mill the slots for a tight fit you can preheat the base so that it stretches slightly and insert the fins then allow it to cool down for a very tight fit.
Reply from Max:
Gotcha. I shall have a go at that if I get time (and some spare aluminium =D). Oh, and thanks for the suggestion about preheating the heatsink. I'll include that in my article...
— KH Dec 4, 08:02 am #