Please Try to understand my origional HS idea?!?!

Well i was thinking about HS designs. First off measure the core on top of the chip to the most exact measurement that you can. As far as each of the 4 sides and from top to bottom (which will only be millimeters). Then using the totally lapped HS, find out where the chip sets on the HS when locked in the socket as normal. (using as to see the chip imprint)

Then take a tool (probably a router) to dig out a square from the bottom of the HS that is the EXACT size of the die on the chip (or as close as possible). Then use AS2 and reapply HS, if done correctly, the HS will not only transfer heat from the TOP SURFACE only, but from the 4 sides as well.

maybe it wont lower temps, but i dont see a way it could raise them, the four sides dont really have any air moving over them. THe only downside i see is cuttin a hole the wrong size and get a chipped core. But take ur time and go slwo and it shoudl work jsut be careful when putting it on.

I will be trying this soon with a Glaciator. Will post results.

this is a pic

I'll be very interested to see the results. Theory sounds fine, although the thermal transfer via the very small area on the sides will IMO be minimal. What really interests me is exactly how you are going to get a surface sufficiently flat to make decent contact with the core. Many people spend many hours lapping HS's to get that perfection and there is no way you are going to lap the interior of a recess that small.

well i have access to a drill press, id just lock the HS in the clamps and lower the bit into the recessed area and lock it there, and just move it side to side, if it is locked i canot go up and down to make it unflat just start at the lowest point and shae everything else off to make it equal flatness. Maybe alot of work, and maybe the results WILL be minimal, but a 2c drop id still be happy with.

I know the sides total area is not much but take the total space taken up by the sides. Combine that and youll get about 1/2 the total chip surface area (of the top), in essence if the 4 sides equal half the surface area of the top of the chip, then cooling the sides, would be the same as having 50% more surface area on top of the chip...wouldt it?


I know that if we had 50% more sufrace area on top it would drop ALOT of temps

If you try it and it doesn't perform positivly, all is not lost....

Fill in the cutout area with smelted silver and re-lap and the heatsink will be better in performance guaranteed or double your money back.

I hate to throw a monkey in the works, but the other components on the board are very close to the same height as the core (not quite, but close). This would limit you the either making cutouts for all the components, and leaving sufficient room for the rubber pads to work good, or only having about .020 touching the sides of the core (which I personally don't think can be done exactly) which will not offer the increase you're looking for.

Another thought is that the core has a small amount of epoxy on it's edge from the mounting process, which will inhibit any cooling, and make precision nearer to impossible.

This is all just my opinion, and I don't intend to discourage your Pioneer Spirit, (in fact I hope you could make it work). I just wanted to offer a couple of things to take into account.
Great luck to ya!

maybe fromt me a small bit of silver so i can if i fail? that way im not stuck witha crapper HS when im done?

thanks digger ill try to get rid of teh epoxy if i can , adn ill be taking the black pads off. Ill take a good look and see if it can actually be done, but the more yall shoot it down is good, it gives me thinks to take in to consideration that i didnt see b4

Jason, PM me your address


-eom-