Heatsink material

Status
Not open for further replies.
I'm trying to design and build a 50W Class-A amp. One of my constraints is that I want to be able to run biased for a 6 ohm load, with no more than a 30 degree C temp rise. This works out to about 200W of heat (at idle - bias 3 1/3 A).

This looks like 4 heatsinks, each having 1000 sq. in. of area, per channel. I can do that with stacked aluminum plates, but I'd like to use less.

Does anyone know what would happen if copper were used, instead of aluminum? I know, the weight will rise, and the cost will skyrocket, but I loose too much enclosure space to the aluminum stacked plates. Also, I don't like the tight spacing, and thin vanes I'd need with aluminum.

Thanks.
Thoth
 
OOPS. Bias should be 1 2/3 A, giving about 100W disipated (I knew something must have been wrong). This gives 4 heatsinks of about 270 sq. in. each. I think I can do this.

Biasing to a 4-ohm load should still be within the temp safety margin of the transistors. A 3-ohm bias would be too hot. Oh well; 4-ohm minimum.

It would still be interesting to know the heatsink area formula for copper.

Thanks.
Thoth
 
The information for the Patriot amp is gone from the LC Audio web site. Currently, the basis for my heatsink design is at 'http://home.earthlink.net/~lotusblossom/'.

LC Audio did have information on their other amps, which use extruded heatsinks ($50 each). Of particular interest was their use of Sanken transistors. I checked up on this, and it looks like these transistors will allow me to half the number of output transistors, and related components.

Thanks for the pointer.
 
well the cooling design of the patriot amp is like this:

I I I I
I I I I
I I I I
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
I I I I
I I I I
I I I I
I I I I
I I I I
I I I I
I I I I MOUNTING SIDE FOR TRANSISTORS
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
I I I I
I I I I
I I I I

do u understand what i mean?

&/mcke


DO U S
 
The dots dont exist..

I....I....I....I
I....I....I....I
I....I....I....I
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
I....I....I....I
I....I....I....I
I....I....I....I
I....I....I....I
I....I....I....I
I....I....I....I MOUNTING SIDE FOR TRANSISTORS
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
IIIIIIIIIIIIIIII
I....I....I....I
I....I....I....I
I....I....I....I
 
Thoth,

I think the difference between cooper and aluminium is basically the thermal conductance:

for pure cooper: 385 W/(m.C)
for pure aluminium: 202 W/(m.C)

where C is temperature (Celsius or Kelvin)


The book Heat Transfer - J. P. Holman, McGraw-Hill have this kind of information to help you on heat sink designs.

Regards
 
OK. Copper has about twice the ability to conduct heat as aluminum. I think this means that with a given heatsink design, the C/W for copper would be half that of aluminum. This, in general, would mean that to get a given C/W for a heatsink, the copper heatsink would have only about 1/4 the area. Thanks.

It's easy to see from the description of the heatsink for the Patriot amp, that ASCII art has limits. I'll E-mail LC Audio, and ask for this information.

Thanks for the ideas, and info.
 
Thermal conductvity is not the whole picture

Aluminium has better "emissivity" (new word?) which means that a give surface area of Alu is better than that of copper all other things being equal. Therefore, a Copper heatsink can be made better if engineered properly, but this may not be tha case if it is the surface area favours Alu and conductivity in the metal is not the issue. For the Patriot, you should be able to use thinner sheets, and more of them to enhance thermal performance.

Copper is also very very dense (reads heavy), and quite costly. I have some busbars that I am planning to use in a power amp so I know what I am talking about! A given volume of copper will thus set you back perhaps as much as 10 times as much as that of Aluminium.

Machinability: Copper is much harder to work with.

Colour: Probably not much to choose between them except Alu can be anodized black with about 10% increase in performance.

Availability: Alu is much easier to find

Magnetic properties: Copper is better than Aluminim. This is probably because Alu is diamagnetic and as I recall Copper does not have this limited liability.

Sound: Due to magnetic properties above, Copper is generally considered to allow for better sounding equipment (ie cabinets made out of ...)

Looks, corrosion etc. -- consider for yourself what is best.

Petter
 
>> Copper has about twice the ability to conduct heat as aluminum. I think this means that with a given heatsink design, the C/W for copper would be half that of aluminum. This, in general, would mean that to get a given C/W for a heatsink, the copper heatsink would have only about 1/4 the area. << The *C/W is largely based on surface area.The only advantage copper would have would be if here was a single point heat source and it was a large distance to the radiating area(fins).For a distributed load, equal area = equal *C/W .Keep in mind the shape of the fin design is determined by convection requirements.The copper convoluted fin designs that have very high fin density will have horrible performance in natural convection.Under forced convection(fan),copper is king.
 
Surplus to the rescue

I guess I lucked out. I walked into the local industrial metals supplier, and started finding out how much the sheet aluminum for heatsinks would cost. Then, I saw 3 sticks of Thermalloy 62725 (unfinished) in the 'scrap' bin (priced by the pound; all 3 were ~$1.50 per pound). I promptly bought all 3, and I now have 31 6.75" long by 9.75" wide heatsinks, and 1 13" long piece with some bent fins. Now I just need to sand the edges, deburr them, and have them black anodized.

Thermalloy's heatsink calculator gives me a thermal resistance of 0.48 deg. C/watt for these. I'll be using a copper heat spreader on the back, so the effective thremal resistance should be lower (line source vs. point source).

Thanks for all the information/help.
 
Re: Thermal conductvity is not the whole picture

Yes Aluminum has a better emissivity than copper, and yet, copper has a better heat conductance. From my point of view, which is materials research at MIT, I think there is a practical compromise. One may use the emissivity of Aluminum by making a large surface area of Al, but embedding a sheet of copper across the inside of the assembly. So, the copper ( a thin sheet .025" ) will conduct the heat to a large bank of Aluminum fins with superior emissitivity. Thus, the best of both worlds...

Petter said:
Aluminium has better "emissivity" (new word?) which means that a give surface area of Alu is better than that of copper all other things being equal. Therefore, a Copper heatsink can be made better if engineered properly, but this may not be tha case if it is the surface area favours Alu and conductivity in the metal is not the issue. For the Patriot, you should be able to use thinner sheets, and more of them to enhance thermal performance.

Copper is also very very dense (reads heavy), and quite costly. I have some busbars that I am planning to use in a power amp so I know what I am talking about! A given volume of copper will thus set you back perhaps as much as 10 times as much as that of Aluminium.

Machinability: Copper is much harder to work with.

Colour: Probably not much to choose between them except Alu can be anodized black with about 10% increase in performance.

Availability: Alu is much easier to find

Magnetic properties: Copper is better than Aluminim. This is probably because Alu is diamagnetic and as I recall Copper does not have this limited liability.

Sound: Due to magnetic properties above, Copper is generally considered to allow for better sounding equipment (ie cabinets made out of ...)

Looks, corrosion etc. -- consider for yourself what is best.

Petter
 
Now that MIT materials are involved ....

Now that MIT materials are involved, it would be useful with some thoughts on material choice versus magnetic properties which should be much more of a general concern for DIY audio people than it currently seems to be.

With limited understanding of the issues, I believe magnetism of any type is a detriment to sound quality.

How do the "secondary" types of magnetism conceivably affect sound quality? (That is dimagnetic and paramagnetic. ferromagnetic properties are fairly well understood by laypeople)

I guess we can all safely assume there is some loss element involved? How do these factors work with the materials in question?

Petter
 
Emissivity improvements

Can one improve emissivity by doing a surface etch, similar to the way electrolytic cap alu is etched to radically increase capacitance (surface area?)

If yes, how should such an etch be performed on Alu or Cu?

Petter
 
Moby,

> One may use the emissivity of Aluminum by making a large
> surface area of Al, but embedding a sheet of copper across
> the inside of the assembly. So, the copper ( a thin sheet
> .025" ) will conduct the heat to a large bank of Aluminum
> fins with superior emissitivity.

Right. That's why I intend to use a copper heat spreader with the heatsink. The heat spreader I'm looking at is 3/16x3x8.5 copper for a single heatsink (3/16x3x18 for a double). It will be oriented perpendicular to the fins, and be centered on the back of the heatsink(s).

Petter,

> Can one improve emissivity by doing a surface etch,
> similar to the way electrolytic cap alu is etched to
> radically increase capacitance (surface area?)

The people I talked to about black anodizing said that with a smooth surface, the result would be slightly reflective. They suggested that the heatsink be bead blasted for a flatter finish.
 
Status
Not open for further replies.