AM4 owners - have you ever pulled out CPU while removing the CPU cooler?

[jonUK76]

I was just interested in how prevalent this problem is. I've done it once in perhaps 5-6 removals from an AM4 socket. The other day I had to remove my new CPU to temporarily put the old one back in - it had only been in a few hours and was installed with Noctua NT-H1 paste which is not a "hard" paste. I loosened the cooler by rotating and and sliding the base around as much as I could - it seemed to be moving freely enough, but on lifting, it still pulled out the CPU. Luckily I got away with it with absolutely no pin damage. I've seen that others haven't been so lucky.

EDIT Sorry, I've screwed up the poll Can't edit. The 5th one should be "Yes - several times, and I've caused damage at least once". The first one should be "No - never, but I've only removed a CPU very few times." I was trying to differentiate between users who remove CPU's frequently (hobbyists, frequent builders etc), and those who rarely make changes to their PC's..

Perhaps delete and start again - it's not right.

 

[Solaris17]

Perhaps delete and start again - it's not right.

no that will make me big mad. just report them and a mod can fix it



fixed.

 

[Lew Zealand]

I have an R5 1600AF sitting upside down on my desk right now, effing glued to the Wraith cooler it seems. It was replaced with an R5 5500 which is working so no damage. That was my third AMD CPU cooler pull and with 5 AM4 CPU upgrades so far I'm at above 50% forcible removal, but with no damage to any parts.

Yet.

 

[oxrufiioxo]

I've read that this can be a problem with AM4 ages ago so I always warm the pc up a bit before removing a heatsink it has worked 100s of times over the last 6 years working with the platform. I only use thermal grizzly paste.

 

[sam_86314]

Nope, never.

Which is surprising since I used to use Arctic MX-5, which was a really tacky and sticky compound. I've since switched back to MX-4.

I own three AM4 CPUs, and I've done several cooler swaps among them.

I usually remove it when it's still warm from use and also give it the ol' twist and pull. Always works to break the cooler free without pulling the CPU out.

 

[Sithaer]

Nope tho I've only removed my CPU a few times between 2018-2022 while I had my AM 4 system with a 1600x.
I was using MX-4 at the time and a Be Quiet!Pure Rock Slim cooler, I always let my CPU heat up a bit before taking off the cooler like looping Cinebench for a few mins or so and then turn off the PC and when removing the cooler I gave the cooler a slight/gentle twist and that way it never happened.

 

[bobbybluz]

Somebody I know trashed a mobo & 5950X trying to remove an Arctic Liquid Freezer II 280.

 

[AusWolf]

Maybe once or twice, but no damage was done.

 

[ir_cow]

Every once in a while, but the trick is to run a heavy CPU benchmark. Warm the CPU up first before turning it off and pulling the cooler. Works everytime because the paste isn't rock hard at that point.

 

[Palindrome]

Many times never actually caused any damage (as far as I could tell) to the socket, retention mechanism or CPU itself. My worst experience was having a 3700X stick to the bottom of a Wraith Prism, that combo had been running and in use for about 3 years at that point. I spent a good 20minutes trying to get the CPU off using an assortment of chemicals and metal tools, because it was like they were bonded together by concrete. The coldplate now has a bunch of deep pitting but the CPU works. Wish AMD would stop adding concrete to their stock thermal compound

 

[jonUK76]

I have an R5 1600AF sitting upside down on my desk right now, effing glued to the Wraith cooler it seems. It was replaced with an R5 5500 which is working so no damage. That was my third AMD CPU cooler pull and with 5 AM4 CPU upgrades so far I'm at above 50% forcible removal, but with no damage to any parts.

Yet.

Perhaps indirect heating would help? One idea I have is take a shallow pan of water, and place the heatsink in the water upside down. Obviously water level should be well away from the CPU. Heat gently for a while and allow the heatsink to heat up thoroughly (don't allow water to boil) and then take it out of the water and attempt to slide off the CPU.

Somebody I know trashed a mobo & 5950X trying to remove an Arctic Liquid Freezer II 280.

Ouch!! Not a cheap accident

 

[Assimilator]

Serious question, why is this poll multiple choice?

I was just interested in how prevalent this problem is. I've done it once in perhaps 5-6 removals from an AM4 socket. The other day I had to remove my new CPU to temporarily put the old one back in - it had only been in a few hours and was installed with Noctua NT-H1 paste which is not a "hard" paste. I loosened the cooler by rotating and and sliding the base around as much as I could - it seemed to be moving freely enough, but on lifting, it still pulled out the CPU. Luckily I got away with it with absolutely no pin damage. I've seen that others haven't been so lucky.

The one time I've had this happen (no resultant damage thankfully) is when I neglected to perform the rotating/twisting motion that you described. But I've never used anything other than MX-4 and MX-5 (no stock thermal compound either since I buy pretty much everything, including coolers, second-hand).

AMD should probably have advertised a little more heavily that the AM4 latching mechanism is to mate the pins properly, not necessarily to retain the CPU...

no that will make me big mad. just report them and a mod can fix it

Aren't you always big mad? :trollface:

 

[Ruru]

Few times. No damage anywhere but now I remember to twist the cooler/waterblock. I guess I had an Intel platform for too long before AM4 as LGA doesn't have that problem.

 

[chrcoluk]

Only removed an AM4 CPU once when going from 2600X to 5600G, CPU stayed in socket.

Doesnt the latch mechanism keep it in place? If I remember right has similar latch system to intel?

 

[claylomax]

My answer is not there:

Everytime.

 

[damric]

Ripped the CPU out more times than I remember, and this problem predates AM4. Certain thermal grease causes such an extreme seal and some coolers have no wiggle room. Sometimes heating it up does no good either. This has caused bent pins, which can usually be straightened out without breaking but there have been a few fatalities.

 

[Shou Miko]

I have never even only older socket am platform really had try problem so because I am just lucky or I know how to remove it without pulling the cpu out

 

[Frick]

Never in my life, and I've used a lot of pinned CPU's. Twist and it's fine. Used only Arctic Silver 5 and Arctic Cooling MX4.

 

[Ruru]

Only removed an AM4 CPU once when going from 2600X to 5600G, CPU stayed in socket.

Doesnt the latch mechanism keep it in place? If I remember right has similar latch system to intel?

Nope. It can lift off with the cooler if the contact is too tight with the CPU.

I remember also getting S478 Pentium 4s together with the cooler back in the day.

 

[Dristun]

Yep, destroyed a 2400G while taking its stock cooler off. The latch is a lie!

 

[freeagent]

So many times. Bent 6 pins on my X3D once. That was stressful.

 

[Steve-007UK]

My mate was having issues with his pc so asked me to see if i could work out the issue.
At some point he had cleaned the pc so while i was looking at it i noticed his CPU cooler was a bit loose, went to tighten the cooler and it pulled the CPU out the socket lol.

After cleaning things up and using new thermal paste i am happy to say the pc was fixed and still works.

 

[Kapone33]

One of the worst things about the Stock Heatsinks is that they love to stick to the CPU. Those are the only ones that have given me issues. AMD's RMA process was good tome though. I did have to eat a 3100 though.

 

[eidairaman1]

Ive only pulled out a skt 478 p4 twice, after that, never again

 

[jonUK76]

Only removed an AM4 CPU once when going from 2600X to 5600G, CPU stayed in socket.

Doesnt the latch mechanism keep it in place? If I remember right has similar latch system to intel?

Unfortunately it's not the same as Intel's (and now in the 7000 series, AMD's) LGA sockets. There, a bracket locks over the top of the CPU. Basically, the socket is going to come out (which would completely destroy the motherboard) before the CPU does in those cases. AMD's sockets are just locked by the pins - my theory is as the pins have become finer (to accommodate high pin densities on modern CPU's) , the amount of force required to lift the CPU out improperly, has reduced. I do believe it would have been possible to integrate a secondary locking mechanism into AM4, such as a bracket similar to LGA sockets to lock it in place, but probably cost and/or lack of interest in the problem won.

Ripped the CPU out more times than I remember, and this problem predates AM4. Certain thermal grease causes such an extreme seal and some coolers have no wiggle room. Sometimes heating it up does no good either. This has caused bent pins, which can usually be straightened out without breaking but there have been a few fatalities.

I did something very stupid in my early days of system building, with a Super Socket 7 AMD CPU. Note to self - RTV sealant is not a suitable thermal compound...