EDIT (07/02/2017): A month later I managed to get 4295MHz - 204x21 @ 1.584vcore/1.35vtt (rly rly crappy chip), ibt, aida64 and prime95 stable, 90ºC. The chip works fine till this very day. Upgraded the CPU cooler to a CoolerMaster Masterliquid 240 in June. Using Arctic MX4.
Disclaimer: I cannot be held responsible for any damange you may cause to your hardware. Follow these steps at your own risk and, if you do, take your time and do not rush to get it done. This is for CPUs that have soldered IHS only.
So I purchased a CoolerMaster Seidon 120V this week. It was CHEAP. When I installed it I got a little (very) disappointed. My temps where about the same as with my old Hyper 212+. 90/95ºC @ 4000MHz 133x30 1.3875vcore in IntelBurnTest.
A lot of people say that it is not worth it to delid a soldered IHS, because you already have a good thermal conductivity and you could ruin the CPU (which is true if you don't take your time and do it VERY carefully) but I thought "what the hell, let's do it now that I don't have a huge cooler on top of my CPU".
So yeah, I delidded my i7 875K and wanted to share it with you.
Here's how I've done it and some pic's =]
Using gloves not to shock my CPU I got a very thin blade and started cutting the epoxy glue that holds the IHS. I started with the corners.
Then, after cutting through all the glue, I inserted some pieces of the same blade in all corners to create some pressure upwards on the IHS.
(This is an old Pentium D to illustrate) Then I put the cpu top down on the stove, making sure it was preheated, with the actual blades inserted. This ensures that when the solder melts, the IHS will separate from the die with a pop. 5~6 seconds is enough to melt the solder and not ruin the CPU.
And voilá, a delidded Lynnfield! =D
I had also to remove the retention bracket in order to fit the waterblock and make it touch the die. Because the CPU got thinner the die was below the bracket.
Apply some thermal paste (this is the one that came with the Seidon)
Had to put some springs between the screws and the waterclock retention mechanism because I wasn't getting enough pressure on the die and it was also uneven.
The thing working.
Look at those temps!!!!!!!!!!!!!!!!!!!!! Notice the CPU is working at 4088MHz not 4000! yay! I know, only about 2 minutes of stress testing, but I used to get 95º like 10 seconds after I started the test. So much for good thermal conductivity with soldered IHS...
I'll post some real screenshots later, I was testing higher overclocks and stuff. BTW this thing is extremely power hungry, can't get above 4145MHz without bumping the vcore to 1.5v which I guess isn't safe. For now I got 4145MHz 133x31 @ 1.44vcore. 80ºC. Any ideas on how to get higher overclocks much appreciated! Do you think that mirror polishing the waterblock would change things for better?
I'm curious about one thing though. Take a look at my i7 and an old Celeron D which I used to practice the procedure. The i7 doesn't seem to have solder, but some kind of TIM and it was also easy to clean off with acetone. If someone knows what that's about, please tell me ^^ Also, take a look at those ripples on the i7 die, my guess is that whatever sits between the die and the IHS, wasn't making total contact.
Thinking about shooting a video showing how to delid a soldered CPU, what do you guys think?
So that's it, thanks for reading! =P