Is this a known AMD64 problem?

[pHaestus]

CPU seems to get stuck to bottom of hs/wb and rips out of socket. This often bends pins. Does anyone else have this problem? If so how'd you solve it? I'm contemplating ripping the IHS off and seeing if that would settle things...

[jaydee]

Originally Posted by pHaestus

CPU seems to get stuck to bottom of hs/wb and rips out of socket. This often bends pins. Does anyone else have this problem? If so how'd you solve it? I'm contemplating ripping the IHS off and seeing if that would settle things...

I imagine it is a problem. Using TIM and then compressing it like that over a larger area probably creates a "suction" effect. Odd it rips it out of the socket though. I wouldn't take the IHS off though. That would make some mounting systems messed up such as Swiftecs as it lowers the hight of the CPU quit a bit.

[Tempus]

Yeah, its a problem. I haven't bent any pins (yet) but I've had the section effect occur 4-5 times now.

If you figure out a good solution let me know.

I've wondered it maybe taking it off "warm" will work.

[j813]

pH please do update us if you did remove the IHS Especially if have Temp monitorings documented.

This is the case w/ all IHS Procs, Intel, etc?

[eander315]

I've had it happen to several Intel P4 processors on Dell GX260, 270, and 280 PCs. In every case the CPU came out with the heatsink without any bent pins. I assumed it was caused by the gray goop they use as a thermal compund, but perhaps it is a more widespread issue.

I think it's caused by the combination of the large heat spreaders combined with the apparently weak sockets taht must be used due to the tiny, fragile pins found on modern CPUs. Older CPUs that had big heat spreaders never came out, so it must be the pin size/socket design that's causing the problem.

[Cathar]

Common issue.

Need to sllliiidde.

Heck, even used to be an issue for non-IHS CPU's at times. Seen a number of reports of people pulling their Barton AMD dies out of the socket still attached to the base of the heatsink.

[Tempus]

True, but the A64s and P4s have that nice mounting frame that prevents you from being able to twist or slide it off.

[j813]

Amount of TIM?

[killernoodle]

This was very common with p4s and ceramique.

[Ls7corvete]

<--- Been there with P4, stock thermal pads are sticky little SOBs.

Havent heard of damage though....just bent pins

[pHaestus]

I can slide when wb testing but it means taking the mounting posts every time. :/

[mad mikee]

I don't have that problem since I can't stand IHS, and remove them as soon as I see the chip works ok

Pull it off ASAP you turn off the PC and First twist slow and slightly until it moves fairly easy, then angle it up SLOWLY.

[Tempus]

mad mike - - what are your ram timings with a 310 FSB. And what divider are you using?

[plywood99]

I took the ihs off my a64 3200 and the temps dropped 7*c. The ihs is frequently cupped.

[jman1310]

i had it happen once on a dell 400sc
thought i had broken the vapor lock enough with twisting but when i gently pulled it off. the socket was empty

slightly shocking
but no damage

since then i twist a lot

Originally Posted by pHaestus

I can slide when wb testing but it means taking the mounting posts every time. :/

you can leave the posts in - there is a slight leeway in the mounting holes, just wiggle it more times
if that doens't give you enough spin, try just loosing the mounting posts a turn or two which should make them very floppy and then it's easier to put them back

[alexwai]

Originally Posted by jman1310

you can leave the posts in - there is a slight leeway in the mounting holes, just wiggle it more times if that doens't give you enough spin, try just loosing the mounting posts a turn or two which should make them very floppy and then it's easier to put them back

Quite similar method I used. After that, I twist the waterblock left to right a bit. Repeating this process for 4 to 5 times, the waterblock can be easily pulled off.

[Spot]

It's always during the my twisting and sliding when the whole chip get rip off.

[mad mikee]

Originally Posted by Tempus

mad mike - - what are your ram timings with a 310 FSB. And what divider are you using?

TCCD + good winnie, now jumbling things around and the winnie is burning in some Gskill BH-UTT, and the Sandy (Wha?) is playing w/ some other TCCD @ 10x270 so far while I'm redoing the WC w/ a 2-342 (presently using BIX + modded mcp350 + Maze4 acetal GPU + cascade whilst waiting for Cathar to send the G5).
Look over @ xtremesystems.org & dfi-street.com for timings, past and future

[Tempus]

Originally Posted by mad mikee

TCCD + good winnie, now jumbling things around and the winnie is burning in some Gskill BH-UTT, and the Sandy (Wha?) is playing w/ some other TCCD @ 10x270 so far while I'm redoing the WC w/ a 2-342 (presently using BIX + modded mcp350 + Maze4 acetal GPU + cascade whilst waiting for Cathar to send the G5).
Look over @ xtremesystems.org & dfi-street.com for timings, past and future

Well I figured your setup was like that.
I was just wondering what your ram timings are. I'm getting fairly ugly timings when I go over 260 FSB - thats about the limit of my 1:1 with this set of HyperX. I've got another pair that will go closer to 275. But thats all at 3-7-7-12 or something around that ugly.