DIY heatsinks


I am in the process of building a ZV4. As an alternative to buy the heatsinks I came across the idea of making my own. The schematics are in the figure below, where it is seen by the front. 3mm aluminium plates are joined with square aluminium tube around 20 mm face. This is all bolted to give a height of 161 mm with 7 plates/tubes. Bottom is stainless steel plate 2mm and top SS 1 mm. The front will be around 7 mm aluminium and the back will be SS 1mm. I did some calculations and each side will be around 0.3 ºC/W for a 400mm deep case, which is sufficient for ZV4. The MOSFETs will be mounted on the plates that go inside the amp.


  • casework.jpg
    28.6 KB · Views: 1,305
The calculations you made, I believe are valid if the structure would be SOLID not bolted. On every plate-to-plate surface contact, the heat transfer increases. The square tubing is also a big problem: heat transfer to fins travels only on tubing walls, too thin to do a proper job. If you want to go ahead, I would at least suggest to use square BARS, not tubing. :bawling:

I'm afraid it will not work. I tried to make heat sink ones too, but it was barely enough for one channel of Zen. The size is 330mm by 90mm. I used 1/2" flat bar between, compound and mounted devices in two places. Even then the center was noticable warmer than the ends. Using square tubing is not acceptable, have to be solid bars, but even then the amount of fins you suggest is low. Also mounting devices in one spot, on fin extension inside, is not a good idea; heat transfer will be very difficult.

I have seen nice heat sinks yesterday in Trading Post for $35. Save yourself trouble and possible disappointment and go with a real thing.;)


  • heat.jpg
    45.2 KB · Views: 1,211
Although it has been done before successfully

I think that for all the work involved and if aesthetics are a concern then it is worth buying a pair of "real" heatsinks.

I have been on a heatsink quest for some time now and IMO the Conrad sinks being offered by rlim on the Trading Post are an exceptional deal and are a known quantity having been used by mefinnis in his Aleph 5.

Where are you located Miguel?

Tony D.


2002-03-22 10:15 pm
i have built the "lottus blossum" sinks for an aleph 5. i used the same design as you have shown, except using 18 large fins per channel. i don't have the equip to measure heat at the device junction, but the sink heat is stable at 55-60 degrees C. the heat is uniform across the entire sinks. However, and this is why i am changing to "made" sinks, I don't like the idea of the many contact points inherent in this design, it doesn't seem as efficient as actual sinks. and the time to, it took me something like 40 hours to drill all those holes.
i'd think long and hard about doing that again.

two cents for the cause.

diy heatsinks


I own the "lotusblossom" webpages. Being a carpenter, I have learned a great deal of patience. My design works well, but, it is labor intensive! I like the concept of total DIY in my amplifier projects. The more I can do the more the amplifier feels like an extension of my ego.......

My current chassis is black anodized and has huge fins. It is beautiful. It will house either the Aleph 2 or the Aleph X. I have the pcbs and components for both. When things slow down, this winter, I'll get the projects finished.

My hat's off to those with the patience to complete my design.

John Inlow
Hello all,

I found my heatsinks at a scrapyard, they take apart all sorts of electronics and make their money off the aluminium (or the copper, stainless steel or plain steel, depending on what they take apart.
The guy at the scrapyard told me that he came across "things" like that often.
Well, from the looks of it I think they are from some sort of power supply or controller.
Size: 200mmx370mmx40mm (2 pieces). I had to pay 3 Euros per kilo. 2 Heatsinks weighed 7 kilos, so that's 20 Euros (I did some haggling, not very much though).



There are a lot of ideas here. I have to say that this idea come after seeing the "lotusblossom" webpages (thanks Carpenter). I just turned it 90º to have bigger fins and not to drill so many holes. I agree with the questions you raise about surface contacts lowering the heat transference. But I find hard to accept that the square tubes between the fins don't do the job. Aluminium has a very high heat conduction and the limiting process should be in the surface contacts - that is why I chose 2 cm tubes. The heigh of the chassis can also be increased to get more heat dissipation (under reasonable limits...).
But I never done heatsinks before so your opinions so I guess I 'm biting my nails:scared:

BTW I am from Portugal.
bars vs. tubes on DIY heatsinks

I think of the spacers between the fins as a pathway for the heat to travel. The trapped air in the center of the spacer tubes acts as an insolator and restricts the heat's path to the outer perimeter. The bar stock is capable of transferring more heat. This could be likened to pedestrians trying to cross a fenced court yard on a small, surrounding sidewalk. If the center of the court yard is accessable more people can cross.

I'm not an engineer and this is just the way I view things.

On the flip side: A chassis on Pass Labs was constructed using tube spacers and seems to be working well.

John Inlow
Peter: A heatpipe is generally a hollow tube, often copper, filled with a liquid that vaporizes at relatively low temperatures. Basically one side of the pipe is where the liquid becomes a vapor. As it absorbs heat it becames vapor, and moves to the other end of the pipe, where it condenses, releasing heat on that side of the pipe. So you would mount one end on your transistors, where it would asorb the heat, then carry it to the other end where your heatsinks are, where it would release the heat. The concept has been around for a long time, often in heating/cooling system, however the decade or so they have been using the concept in high-end servers, and lately the computer ethusiest market, making it easy to get heatpipes of varies sizes and types these days.

And no, I can't spell, and my grammar sucks. :)

Hmm, neat idea... just the opposite of how a refrigerator works. I guess the downside is that there's a slight efficiency loss, since the heatsinks will always have to be cooler than the transistor mountings in order for it to work. Ultimately, it would still be better to mount transistors directly to the heatsinks, but a heatpipe sounds like a good way of transporting heat around...