Aleph2 Heatsink from HeatsinkUSA

[kop89]

Hi,

I am thinking about using these heatsink for my Aleph2:

http://www.heatsinkusa.com/asfdas.html

this one is 8.5 inch wide (22 fins, 1 inch high). If I arrange 2 heatsinks vertically around the side of the chassis, I will get ~17 inch chassis depth, on which I will slap on 6 MOSFETs on each side.

I am thinking about using 4 pcs per mono chassis (2 vertically arranged heatsinks per side). I am just not sure if 1 feet of heatsink length will be enough (4 feet total length per chassis), or should I make it 8 feet total per chassis (2 feet pieces).

Unfortunately, the seller does not have the C/W spec is, so I will have to guess.

Anyone has ever used this heatsink before and have any suggestions how much would be enough?

Thanks!

[kop89]

OK,

So I used this heatsink calculator to estimate the C/W, with information of heatsink dimensions given on the HeatsinkUSA website:

http://www.frigprim.com/online/natco...k.html#Editing

For each 1 feet section of heatsink, I am going to attach three IRFP240/244. Assuming this is around 100W of heat dissipation, the temperature rise will be around 60 degrees Celcius according to the simulation.

If this works, then each monoblock will need 4 x 1 feet piece of heatsink (total: 8 pcs X $26 = $208), which is very reasonable price for the heatsink.

Is 60C temperature rise acceptable? What is the maximum allowable temp rise for IRFP240/244?

[woody]

I believe that is way to hot. I think Nelson recomends a 30 deg
temp rise. Besides it's not just the mosfets the power supply caps won't like that heat either.

[Bill Fuss]

1" fins are way too short for a class A amp, and the thermal efficiency also drops quickly with longer [taller case] pieces of extrusion. A 12" piece is not 4 times the capacity of a 3" piece, it is only about 2 times.
Finding heatsinks for these amps at reasonable cost has always been difficult. It's worth the time to research the rating process the manufacturers use for their published data. Check out Conrad, M&M, Aavid-Thermalloy websites and others to gain a little better grasp of what you're up against.
NP recommends in the F5 article a heatsink of 8" wide by 6" tall with 2" fins for each fet dissapating about 30W if I'm not mistaken.
When you double the fin length you are doubling the surface area, so the fin length and the width of the extrusion are more directly related to capacity than the length.

Bill

[AndrewT]

Hi,
if you double the height of a heatsink you get approximately 40% extra dissipation for the same deltaT, if the whole backplate is at the same temperature. But we know that a high backplate will have a bigger variation in temperature than a lower backplate.

Similarly, if you double the width of a heatsink you get approximately 100% extra dissipation for the same deltaT, if the whole backplate is at the same temperature.

That is why it is important to have a thick backplate for large heatsinks.

DeltaT = 60C is far too high.
For Ta=25degC
Ts=85degC
Tc~=95degC
De-rating for a 150degC plastic package ~56%.
A 150W device becomes a 66W device @ Tc=95degC

Barrredboss is very close to retailing a 10.1inch wide heatsink that will be much better suited to high dissipation amplifiers.

[Nelson Pass]

Originally posted by kop89
Is 60C temperature rise acceptable? What is the maximum allowable temp rise for IRFP240/244?

60C is too hot. Try to keep it down to 30C, even if you have to use a fan.

[Bill Fuss]

They both said it better than I did, but do your homework, and hang tough.

Bill

[kop89]

Anyone ever used catalog heatsink from Conrad, M&M, wakefield or other vendor for the Aleph2 project?

If you do, could you please tell me the catalog number and how many pieces per monoblock?

Thanks.

[Bill Fuss]

Well, looking at the power disappation of 300W per ch and spread over 12 fets, thats 25W per fet. Bundled of three fets @75W, minimum of 4 MF25-151.5 per ch of Conrad heatsinks for 30* rise. Someone please correct me if I'm in error. It's late and I have consumed a little bit of Cab.

Bill

[kop89]

Maybe I can use 4 pcs of either MF30-151.5 or MF30-2F-151.5 per monoblock.

Do you think it is better to use the flatback version and mount the MOSFET directly to the backplate of the heatsink, or use the flanged version and mount the MOSFET to the flange?

Assuming 320W heat dissipation per mono, this is ~80W per heatsink. Using this, temp rise will be:

80 W * 0.25 C/W * 1.35 correction factor ~ 27C temp rise.

So the total cost will be 8x 52 AU$ ~ $300 USD, which is quite reasonable.

Am I missing something here? Please advise.