CPU yanked everytime I remove XP120/XP90

[smakme]

Is this happening to you too? I can't remove the HS (I got XP120 & XP90) from the MB (2 DFI nF4 and 1 MSI nF3) CPU socket without yanking out the CPU (2 E3 3200 and 1 E4 3700) as well. This happens even after I run Prime95 for 2hrs. I use AS5 and I've followed the application procedure from their website.

Does anyone have any trick in removing these HS without yanking out the CPU?

[metro.cl]

i didnt have that problem maybe you are using to much as5, also try rotating the hs while pulling softly

[Ugly n Grey]

After unclamping rotate the HSF back and forth as much as the movement will allow in order to break the bond and introduce some air into the paste.

[bachus_anonym]

It used to happen to me all the time with XP-90 on Neo2 NF3/DFI NF4. But I used Arctic Ceramique which is more glue-like than AS5.

Once you un-hook the clips from a mounting bracket, twist heatsink few times, left and right, and again, and again... :) And like you did it... Warm CPU up before doing it. Also while removing, try to squeeze a tip of your finger in there so you can push CPU in while pulling heatsink up. That's what I was doing and it worked well then. I bent a whole row of pins on my Winchester once, when I pulled that XP-90 for the 1st time. My heart stopped for a sec, but I fixed it and it was all good :lol:

Good luck!

[mtb856]

Like the other guys have said, undo the clips and then twist the HS a few times.... what I usually do is try to twist while lifting one corner up to make sure that I'm breaking the seal between the TIM and the cpu.

I did the same thing bachus did.... I was pulling my SLK-948 off of my 2.4C and bent an entire row of pins and almost broke one off, I thought I was going to cry but it all worked out.

[Ugly n Grey]

Admit it Bachus, you bent those pins because you were tippling...

This is far more likely to happen with fresh paste than old paste as well. Someone else mentioned what Bachus did when i searched for the problem, which is removing the cooling when it's still warm helps...

[bachus_anonym]

Admit it Bachus, you bent those pins because you were tippling...
I can't remember but quite possibly as I have a beer next my keyboard most of the time :lol:

[uOpt]

This is the main reason I don't use AS5, it is much stronger than other pastes. I use AS Ceramique which is easier mechanically and cools almost as good.

As for the actual problem, you can try to insert a piece of thin strong plastic between CPU and HSF to "cut" through the paste. Plastic from theft-protection retail packaging is good for this.

[g-a-m-e]

Or gently ram a screwdriver between the proc and heatsink, gently push the screwdriver as far as it will go, then firmly twist it ...

god that was a lame joke ...

[Ugly n Grey]

Or gently ram a screwdriver between the proc and heatsink, gently push the screwdriver as far as it will go, then firmly twist it ...

god that was a lame joke ...

:slap:
http://www.xtremesystems.org/forums/showthread.php?t=74828

[g-a-m-e]

ok oops...sry !..

so back on topic, altough i always (have to) twist them off myself, i've heard that pins can actually break doing so ...

That true? Let me rephrase,...That ever happen to anyone ?

[Ugly n Grey]

I have yanked a few out, never broken a pin though.... I find of course that stripping the IHS off also solves THIS problem. I don't even understand what the point of them (IHS) is to anyone who overclocks.

[IYP]

use less AS5 and twist before you pull off

[PowerSlide]

i always have this problem wid CM hyper6 wid as5..even twist the proc still stick to the cooler..twist a pin harshly on my p4

dun hav prob b4 wid xp-120..cooler out smoothly

[uOpt]

Yeah, the Cooler Master heatpipe towers are so tight in the frame that you cannot twist them enough. Same problem with the coolers on some Xeons I had.

[rick_fx]

Twist the heatsink while holding the cpu in the socket (from 2 opposite corners) and voilÃ:banana: :)

[mad mikee]

Patience is an xtreme virtue here. I go throught this when I test the CPU for functionality b4 removing that POS, err I mean IHS.... :D

[uOpt]

Twist the heatsink while holding the cpu in the socket (from 2 opposite corners) and voilÃ:banana: :)

Not possible with the Cooler Master heatpipe constructions, they fill out the frame.

[IYP]

use less AS5 when you mount....thats your problem

[uOpt]

Actually more paste should make it easier to remove...

[STEvil]

more paste ensures the paste squeezes out from under the heatsink and when you remove it something has to fill that gap (causing suction).

A thin film is about all you need..

[uOpt]

more paste ensures the paste squeezes out from under the heatsink and when you remove it something has to fill that gap (causing suction).

A thin film is about all you need..

Too much paste will still leave a thicker film. The adhesion gets stronger as the the HSF gets closer to the CPU heatspreader, so the ideal thin film will make it harder to remove than a thicker film.

[STEvil]

only if you use too much..

if you use a thin film then when you attempt to remove the heatsink you will only be creating suction for a fraction of the amount of time. More means suction happens longer since TIM will be filling any gaps and/or restricting allowing air to fill as you remove the heatsink.

[FireDragon]

take the IHS off and you won't have this problem i never had this problem as my winee and venice was always nude!!!


Dragon

[mtb856]

ok oops...sry !..

so back on topic, altough i always (have to) twist them off myself, i've heard that pins can actually break doing so ...

That true? Let me rephrase,...That ever happen to anyone ?

I highly doubt that is possible..... with the number of pins on a modern cpu, the force is distributed quite a bit, so I can't see the pins snapping from twisting it unless the cpu wasn't all the way in the socket.

[smakme]

Hey everyone, since I made the OP I thought I'd give you all an update. I finally got around to taking the XP120 out and replacing it with the Ninja. It went very well. The XP120 and AS5 had been on the processor for about 2 months so the AS5 paste has had time to cure. I did the gentle twisting (which everyone recommended) on the heatsink for half a minute and off came the HS with the CPU firmly seated in the socket. I was so relieved in seeing the IHS and not the pins.

And by the way, I got an immediate 4-5C drop in core temperature with the Ninja over the XP120. I just love it when things work out as hoped. Thanks again to you all for your sage advices.

[Welz]

have any cpu's been damaged in the process?

[smakme]

In my experience, no CPU ever got damaged but I've had the 2 back rows of pins bent when the CPU gets yanked out with the HS. With repeated bending and unbending I'm sure they'll start falling out. OTOH, I can be content with whatever OC I get with MB, CPU and HS... nah.

[NickS]

After unclamping rotate the HSF back and forth as much as the movement will allow in order to break the bond and introduce some air into the paste.

That's what I do unless im in a hurry.

Then I take a deep breath, grasp the case with my left hand, heatsink with right. Then I use a TON of force quickly, and rip it right out, STRAIGHT UP & OUT!!! THIS IS IMPORTANT. :) :slobber:

[mtb856]

:hm:

[MRBIGSHOT]

this is one of the main reasons why i went back to ceramic. i cant twist much with my tdx block. ive never pulled the cpu out though, mabey a 754 locks a little better than a 939

[mtx]

That's weird, isnt your CPU locked in with the lever?

[JasonDTM]

Easy way to not yank the CPU out is to run the computer, then pull the sink off(after you've turned it off... ofcorse :/ ), smooth as butter. lol

[zenzog]

That's weird, isnt your CPU locked in with the lever?

It feels like the lever on K8 boards is weaker than the ones on K7.. at least, to me...

[mtx]

It feels like the lever on K8 boards is weaker than the ones on K7.. at least, to me...

THat is very strange. I feel the same way, the lever moves left to right..

[Xyus89]

I use a hairdryer , I just point it on my cpu heatsink on the hottest and it will come off like a charm.

xyus