Is there a place I can get pics of the heat sinks used in the Dual 400,450,500,533's, Quicksilvers, and MDD PowerMac's. I already found the one for the 1.42Ghz.
That's a strange design for a heat sink. Plates separated by air aren't very efficient. The idea of a heat sink is to increase the surface area of what is to be cooled. Flat-horizontally oriented plates don't increase the surface area directly. Think of what an air cooled motor cycle cylinder looks like with fins oriented at an angle so that you generate more surface area to the piston chamber. Don't quite get why they used that flat plate design?
I'm sure they've done tests on vertical finned vs. horizontal finned, probably found the horizontal way to be better. Though I do not see how it would matter anyway, either way you get more surface area. Maybe the way it is rotated matters.
Don't take this as gospel - I have a very limited understanding of thermo-dynamics.
Conventional "finned" or "spiked" heasinks are designed to work in passive environments, where there is no fan and the heat is dissipated by the heatsink by means of convection, and thereby the source, by conduction.
If you're going to add a fan to equation, the heatsink should be designed so that the airflow is as uninterupted as possible. In this environment, convection doesn't really matter as the air is being forced through the surface area far faster than convection would allow anyway.
Most of the heatsinks you see follow the finned or spiked approach with the fan simply bolted on top. This is perfectly practical for the majority of applications. My guess, is that it's cheaper to simply machine a block of aluminium, than to assemble the various plates and spacers of this new design.
These copper plates have a massive surface area compared to their mass. It looks as though the airflow through the heatsink is virtually uninterupted. I don't think you would be able to remove anywhere near that amount of material from a conventiional aluminium heatsink.
The whole thing looks as though it's been designed with maximising conduction in mind (including the choice of material). What's up with the "elevator shafts" holding the horizontal plates together? They look as though they have a function other than simply holding the copper plates apart. Perhaps they are designed to accelerate the conduction in some way?
Please correct me if I'm wrong! I find this stuff fascinating, and would love to learn more!
<strong>That's a strange design for a heat sink. Plates separated by air aren't very efficient. The idea of a heat sink is to increase the surface area of what is to be cooled. Flat-horizontally oriented plates don't increase the surface area directly. Think of what an air cooled motor cycle cylinder looks like with fins oriented at an angle so that you generate more surface area to the piston chamber. Don't quite get why they used that flat plate design?</strong><hr></blockquote>
The position of the chips is on one side of the heatsink, so your described design would mean much cooler fins on one side of the heatsink. Two of the heat pipes are probably located very close to the processors. When you fold up the case, the fins are vertical and the heat pipes are lateral.
3rd EDIT: Well, I'm at a loss over these big, thin copper plates, too. Unless they are getting *very*, *very* hot, I don't see how the temperature can really conduct from the heat conduction poles to the extremities of the plate for good heat transfer to the air. Most of the heat will pile up near the poles and dissipate in the local areas of the plate. So I imagine a great deal of surface area toward the outside (and thus material) goes unutilized. Considering it is copper (not a cheap material to use indiscriminantly), the liberal use of material is even more baffling (no pun intended).
Ah well, you can never really tell for sure until you do some thermal FEA on the design.
Comments
<a href="http://www.ocf.berkeley.edu/~ceugene/photos/heatsink2.jpg"; target="_blank">And another.</a>
<a href="http://www.ocf.berkeley.edu/~ceugene/photos/heatsink3.jpg"; target="_blank">And another.</a>
<a href="http://www.eugenechan.com/gallery/newG4/P2040043"; target="_blank">A QuickSilver 2002 Dual 1 GHz heatsink.</a>
Thats the haetsink inside my dual 533
[ 03-06-2003: Message edited by: Mount_my_floppy ]</p>
<strong><a href="http://www.ocf.berkeley.edu/~ceugene/photos/heatsink1.jpg"; target="_blank">Another</a> 1.42 GHz heatsink pic.</strong><hr></blockquote>
Blue motherboard? Does that mean anything special like the red prototype a while back?
they seem a bit absurd
<strong>has anybody actually measured how hot the heatsinks get? let alone the processors w/o heatsink?
they seem a bit absurd</strong><hr></blockquote>
The fan stays put when you open the case, so the new G4s have to be sufficiently cooled without it.
Conventional "finned" or "spiked" heasinks are designed to work in passive environments, where there is no fan and the heat is dissipated by the heatsink by means of convection, and thereby the source, by conduction.
If you're going to add a fan to equation, the heatsink should be designed so that the airflow is as uninterupted as possible. In this environment, convection doesn't really matter as the air is being forced through the surface area far faster than convection would allow anyway.
Most of the heatsinks you see follow the finned or spiked approach with the fan simply bolted on top. This is perfectly practical for the majority of applications. My guess, is that it's cheaper to simply machine a block of aluminium, than to assemble the various plates and spacers of this new design.
These copper plates have a massive surface area compared to their mass. It looks as though the airflow through the heatsink is virtually uninterupted. I don't think you would be able to remove anywhere near that amount of material from a conventiional aluminium heatsink.
The whole thing looks as though it's been designed with maximising conduction in mind (including the choice of material). What's up with the "elevator shafts" holding the horizontal plates together? They look as though they have a function other than simply holding the copper plates apart. Perhaps they are designed to accelerate the conduction in some way?
Please correct me if I'm wrong! I find this stuff fascinating, and would love to learn more!
<strong>That's a strange design for a heat sink. Plates separated by air aren't very efficient. The idea of a heat sink is to increase the surface area of what is to be cooled. Flat-horizontally oriented plates don't increase the surface area directly. Think of what an air cooled motor cycle cylinder looks like with fins oriented at an angle so that you generate more surface area to the piston chamber. Don't quite get why they used that flat plate design?</strong><hr></blockquote>
The position of the chips is on one side of the heatsink, so your described design would mean much cooler fins on one side of the heatsink. Two of the heat pipes are probably located very close to the processors. When you fold up the case, the fins are vertical and the heat pipes are lateral.
3rd EDIT: Well, I'm at a loss over these big, thin copper plates, too. Unless they are getting *very*, *very* hot, I don't see how the temperature can really conduct from the heat conduction poles to the extremities of the plate for good heat transfer to the air. Most of the heat will pile up near the poles and dissipate in the local areas of the plate. So I imagine a great deal of surface area toward the outside (and thus material) goes unutilized. Considering it is copper (not a cheap material to use indiscriminantly), the liberal use of material is even more baffling (no pun intended).
Ah well, you can never really tell for sure until you do some thermal FEA on the design.
[ 03-09-2003: Message edited by: Randycat99 ]</p>