At 9:26 PM -0500 3/6/00, jpm wrote:
>I dunno, having read the various replies I still think the foil is the
>better way to go.
I hope to disabuse you of that notion, if I interpret what you want
to do correctly!
>Heat sink goo is fine I guess, but I can't imagine it's going to draw as
>much heat off of the chip than a really snug contact with the heat
>dispersing metal framework that's in the bottom case, which is why I
>proposed the foil. Got the idea form a fellow duolista who opinions always
>made sense to me.
>
>The problem with the unsticky rubber heat sink square is not so much that
>it's not sticking to the chip, but that in replacing the board you have to
>flip it upside down and lower it into the case; so then it falls off.
>Anything benign that will keep it in place for those five seconds (spit
>maybe?) will do because as the chip heats up, the unsticky glue will no
>doubt become sticky again and then adhere to the chip. Any stray air
>bubbles will likewise disperse due to the pressure from the case as well as
>the heat cycling.
>
>After having several G3 cards burn up I used the foil square to help
>disperse the heat and it dropped my cpu temp about 10-15 degrees, but your
>mileage may vary. . .
>
>Tastes great! Less filling! <grin>
I think I missed the start of this one. This is about a 2400, right?
For clarity, here is the normal stack of things cooling the CPU in a
2400. (Best in a monospace font)
CPU
------------
Foam Heat Transfer Pad
------------
Aluminum Block Heatsink <-------> Heat Pipe
------------
EMI Shield (aluminum sheet covering bottom except battery bay)
------------
Plastic Case Bottom
The heat pipe is that sealed tube going out from the aluminum block.
If anybody wants I'll explain how they work, but basically think of
it as a very low resistance conductor for heat, much lower than plain
aluminum. The heat pipe snakes along the EMI shield and is kept in
tight contact with it along the whole length. Since the pipe
presents a low thermal resistance to the heat entering it from the
aluminum block, it is heated along its whole length and transfers
heat to the EMI shield. Thus, the CPU's heat ends up being
dissipated over a large area on the bottom of the computer.
So, the goal to shoot for is a good thermal connection between the
CPU and the aluminum block. The heat pipe will take care of the rest.
Aluminum foil between the CPU and the block will not do you much
good. The reason for that foamy rubber stuff rather than something
solid is that the computer case can flex and it's good to have some
give and take in the heatsink interface when that happens.
But the major problem is that folded aluminum foil is not a very good
heat conductor. The thing about heat conduction is that it works a
lot better conducting through a solid than through air, and folded
foil has a lot of air between the folds. Good thermal connections
between a chip and a piece of metal are made by using interface
materials which can deform and eliminate air gaps between the
surfaces. The original foamy/spongy heat transfer pad is one such
material, thermal grease is another.
The only place I recommend using aluminum foil in the 2400 is between
the EMI shield and the aluminum block. Sometimes the CPU isn't close
enough to the block, and thus the pressure compressing the foam pad
between the CPU and the block is not high enough (most thermal
interfaces work a lot better under pressure, especially those foam
pads). In 2400s which suffer from this, you should put a shim
(whether it's foil or something else) between the heatsink block and
the EMI shield to raise the block a little bit closer to the CPU.
(Apple's own 2400/240 models have a sort of a spring installed in
this location.) It doesn't take much of a shim to make a big
difference.
So, to conclude this long-winded narrative, if you tore the foam pad
off and you're having troubles getting it to stay, you could get it
to stick on just by putting a tiny amount of thermal grease on the
CPU and then putting the pad on. The surface tension of the grease
ought to hold it on quite well.
If you actually tore the foam pad into a couple pieces (easy as
they're so fragile), I recommend trying to get a new one from the
accelerator company. If you make it clear what you're asking for,
you might be able to get them to send you one, or they may allow you
to send the whole accelerator back and have a new pad fitted. The
reason is that while it may look OK lying on the CPU before you
assemble everything, once it's all together and there is enough
compression for the pad to be a good thermal interface, the tear may
open into a gap, and that would allow the CPU die to develop a
hotspot on the gap.
IMO, 2400 accelerators should ship with something like the Apple
2400/240 spring to install... but not all of them did. I think
that's why so many people end up burning them up.
Tim Seufert
----------
Duo/2400 List, The friendliest place on the Net!
A listserv for users and fans of Mac subportables.
FAQ at <http://www.themacintoshguy.com/lists/DuoListFAQ.shtml>
Be sure to visit Mac2400! <http://www.sineware.com/mac2400>
X-Router | Share your DSL or cable modem between multiple computers!
Dr. Bott | <http://www.drbott.com/prod/MIH120.html>
PowerBook Guy is | Click here!
Everything PowerBook! | http://www.powerbookguy.com
Midwest Mac Parts ][ <http://www.midwestmac.com>
After-market parts for Macs. ][ 888-356-1104 ][
MacResQ Reader Specials: 2.5GB Seagate SCSI: $119, 4GB IBM SCSI: $199,
Norton Util. 4.0: $29, Mac Parts, Systems & Repairs <http://www.macresq.com>
