The perfect heatsink

Yep, another heatsink post. But as everyone knows, heatsinks are often the hardest parts to find so there's a lot to talk about.

I'm building the Aleph 2. Therefore I need to dissipate 300W of heat per monoblock (You can see my concept below)

I've spoted a heatsink but I need a double check from the pros.

Here it is

I'm planning to use 12 of them per monoblock as shown below (3" high each).

Here's what I think:
> 300W of heat overall: 25W of heat per transistor/heatsink
> For 3" high heatsinks, thermal resistance = 0.82ºC/W
> Therefore each heatsink has a temperature raise of 20.5ºC [25W*0.82ºC/W], which sounds very good to me.

Furthermore, Aavid sells this extrusion profile in bar of 8' which is very handy (could have my heatsinks up to 4" high).

What do you guys think?



  • a2.jpg
    97.2 KB · Views: 4,535


2000-11-24 12:02 am
25W per 0.83C/W sink will not be enough. I think you are forgetting to add the thermal resistance from the insulation plate (mica, kapton, ceramic, silicon or whatever). You must isolate every FET from each sink. This will add anything from 0.5 to 1.5c/w of thermal resistance. This will depend on the type of insulation plate.

You can find a lot more info on this, if you check the older threads.

Hi Gabriel,

Nice Drawing!

A couple of considerations:

a) The internal heatsinks will run hotter than the ones on the outside of the case. The older PS Audio 200C amps had a similar design and I know from talking with the designer that they regretted that design...

b) I agree that you also have to take into account thermal resistance from the insulator etc.... Figure on ~ 1.0 c/w from the insulator if using mica...

c) Thermalloy is very expensive as to purchase from. They seem to have no interest in supplying us small guys... You might consider M & M Metals, VEP or R-Theta...

d) Finally, if you increase the heatsink length be sure to use the proper multiplier for the change in thermal resistance....the dissipation change is not linear... The models on the Thermalloy website work well for this...

Hope this helps!

Good luck,


This is more a question than an answer.

Please correct me if I'm wrong cause im new into this
I calculate the final temperature like this: x=(P*(a+b+c+d+e))+amb.

x= junction temperature
p=power in watt
a=junction to case(C/W)
b=case to insulator(C/W)
d=insulator to sink(C/W)
e=sink to air(C/W)
amb=room temperature

example: TIP3055 with good insulator, 25W and the sink as in the above post I end up with this figures:

Final temperature X=86.8 C

Am I on the right track or??

Hi folks.

I'm starting to lose faith in that project since heatsinks don't stop to grow larger. My first idea was to build an Aleph 4 in the current box design, but I realized I had to go for the Aleph 2...splitting the heat in two boxes.

Now that we talked about numbers, I would like to talk about real life comparisons.

Here's an example:
Mark Finnis built an Aleph 4 (here) using 12"x14" heatsinks for each channel (168in² of heatsink surface per channel).

Using my design (4" high heatsinks) I get 240in² of heatsink surface per channel (per Aleph 2 monoblock). Furthermore, the heat would be more evenly distributed on many heatsinks (instead of all transistors at the center of a huge heatsink).

Even if my heatsinks inside the casing will get hotter, the larger heatsink's surface should counterbalance this.

Since this is not a commercial amp, I don't mind having it hotter than usual class A amps. My goal is really to have slim design.

If you tell me that the Aleph 4 generates less heat, maybe I could simply split the Alpeh 4 in two casings...

What do you guys think? Maybe Mark Finnis himself could comment.


R-Theta heat sinks are excellent. I have placed small orders with them for custom heat sinks and they have no problem with these.

R-Theta also has a nice online software package to help with heat sink design and device placement so that all devices will operate at near identical temperatures. The software will also calculate junction and heat sink temperature profiles. In the past they had a few bugs in this software but those have likely been fixed.

The type of fins you specify will depend on wether you are using conventional air flow or forced air. I use conventional along with very slow rotating fans. Forced air fins have closer spacing. Base thickness can also be specified. I would rate R-Theta at the very top of the list of heat sink suppliers.

Some individuals find it hard to find heat sinks. I find that heat sinks are really not that hard to find they are just expensive to buy. For the a power amp dissipating 300 to 500 watts per channel expect to pay about between $1.50 and $2.00 per watt for a two channel amplifier. Thus spending $600 on good quality heat sinks for a two channel amplifier dissipating 300 watts per channel is not unreasonable.

In many commercial products the case and heat sinks can easily add up to half or more of the total parts costs.

John Fassotte
Alaskan Audio
Hi guys,

Here's where I have a bit of experience.

I echo the pricing info of some posters but disagree with others. There are commercial suppliers that have very high prices for small quanitities. I got my sinks from M&M Metals in Irving Texas. Use the R-Theta calculator though it is great for sim'ing your sinks.

My sinks were under $20 each and I used 2 per side of the amp. So that's $160 total for sinks for a pair of amps that cost north of $1200. You can easily pay $1000 for similar sinks though, and the highest price I got was about $2000 for similar sinks. So you can see that there is basically an order of magnitude range on heatsink pricing.

On the bolts to hold the front panel. Interesting thing, that. My buddy actually found them. They are chromed cap head bolts from a custom biker shop !!! If you can find a place with more tattoos and leather per person than you've ever seen before in your life, then you're in the right place. I bet any Harley shop will have similar for 10x the price...... THey cost me about $1.00 each. I drilled a hole that would clear the threads, then drilled another hole, larger, that would clear the head, but barely. Its tricky as you want the holes to be perfectly concentric. You have to do it on a drill press. Control the depth of the larger hole so that you have the same depth for all the holes - also allow enough material under the shoulder of the bolt so that the panel doesn't pull off because the material is too thin !!!!! I like the chromed look also, though I can't take the credit for the idea.... It adds a lot to the front of the amp IMO, but its a taste thing. I'm really glad my pal found them.

More on bolts.... the pic on the website was taken when I didnt have the tops bolted down. The bolts I used there are the black grade 8 button head bolts. They give a "softer" appearance than using cap head bolts in that position. I may try some cap head bolts there one day to see how it looks.

Amps outside dimensions are 9" high, 12" wide, 14.625" deep.

I doubt you can make the case thinner as my toroid barell clears the sides as is.... be careful here. Use a CAD program to model your amp before you start getting case parts bought. That way you will be less likely to scupper a bunch of expensive parts.

Thanks for the kind words on the finished amps.
Heat sink costs

I forgot to mention that the cost of heat sinks is affected by whether or not you can utilize a stock extrusion or not. If you have to have a custom type or perhaps a custom length then cost will go up substantially. So attempt to find a stock extrusion that will work to save cutting and other engineering charges. Also anodizing requirements may also add to the costs. How you intend to mount devices also effects on the type of heat sink that can be used

The cost I mentioned in my prior reply are for custom heat sinks and are thus quite a bit higher than using stock items purchased in small quantities. They are however realistic. I have also purchased heat sink extrusions by the pound. This really reduces the cost if you order say 1000 Lbs at a time. Get a lot of individuals to combine a order and lots of $ can be saved.

John Fassotte
Alaskan Audio


2001-02-04 4:23 am
The thermal resistance of 0.82*C per watt is for a 100*C rise.With a power input of 25W the resistance will be 1.134*C per watt, or 28.3*C above ambient.This is a common mistake.This profile is best for forced convection.Under moderate convection the thermal resistance drops from 1.134*C/W to 0.310*C/W, the poor performance under natural convection (and good performance with forced convection) is due to the close fin spacing.A length of 7" of this profile would have a thermal resistance of 0.825*C/W with a 25W input under natural convection.