Athlon IHS Removal and the "CAM"

This goes back to a concern people had in the socket 7 days when removing the IHS on a K6-2/3 CPU. I know I was one, because I pulled the heatplate off of every K6 series CPU I owned .

Also those with large flat heatsinks on socket 7 who then upgraded to socket A and tried to re-use the heatsink (FDP-32 as an example) found that CPU temps skyrocketed under load. Removal of the heatsink showed the same problem as discussed, with thermal paste being uneven and just filling in the gaps. Alot of people cracked their CPU this way too. After I bought my A7V and duron 650, I didn't have the cash for an FOP32/38 heatsink, so I made my FDP-32 usable by grinding a step into the bottom. Temps went down, to say the least.

So this is really old news, but it's good to reiterate to the lastest generation of overclockers that didn't have this experience with past platforms.

You have any pix or steps to remove the IHS since you have done it so many times? Sure would help out alot of pplz

one tip: razor blade. There is adhesive around the edges of the IHS. cut around the edges, through the adhesive and pry it off.

(I got a k6-2 333 to 600 with that trick.)

For the K6 series it was easy enough. Just slide a blade in at each corner until you feel a little resistance giving up per corner, then grab the top with some needle nose pliers and pop it off. On the other hand, you could use a screwdriver to pop it off, but you risked breaking a corner of the package.

I have not popped an A64 top, so I can't comment on that, but I imagine the same rules apply with more consideration for the package since you can actually cut through this one. The K6 series package was ceramic, so it was of no concern, you just had to avoid going too deep under the IHS or you could slice a surface-mount resistor.

Single edge razor blade. If you lack patience, don't do it. Let the blade do the work, feeling its feedback as you go. Took me about 30 minutes, going slow and cautious. Now that I have a more accurate thermal diode reader, it would be nice to do it again and really get a better idea of how much it gained me. I seem to recall, with the flawed MSI reading system, it got me like 1C lower at idle and 2-3C lower at load.

Hoot

If they use the same stuff as K6-2s to goober it down with, then "sawing" at it with dental floss might work rather well. It will take a while but should have the advantage that the edge of the core or any other surface mount components will probably fray the floss rather than be damaged by it. Might take you several pieces to get through all the glue though. Another approach might be to lasso it with a loop of floss and then twist it real tight with a pencil and let the pressure pull through the goop. Doing that though you'd want to be real careful you weren't exerting a bending/breaking force on the packaging with the pencil. One way around that would be to use a short lenght of narrow tube, like a pen top, that you could feed the ends of the floss through, then use that to keep a purchase on the edge of the IHS, without pressuring the packagin and use the other end as a kind of pivot for your winding implement, so you'd be winding it up through the short tube.

regards,

Road Warrior

Edit: BTW razor blade technique that works best for me on K6-2s is to rock the blade around the corners rather than trying to saw or slice with it. As if you were using it to cut through a pencil or tube by rolling it.