just delid my 4790K /g/, am I in for some good temps or did I fall for the meme?
>>52432144
Yes
No.
>>52432144
wtf
>>52432144
But why?
>>52432172
get with it grandpa, its hip
delid dis dick
>>52432144
pc gore thread?
>>52432182
the idea is you replace the thermal compound inside the processor itself for more efficient cooling
here's the final result by the way
>>52432206
Now lap it to a mirror finish
>>52432206
>the idea is you replace the thermal compound inside the processor itself for more efficient cooling
no the idea is to eliminate the gap created by the black glue they used to secure the lid. There is nothing wrong with the paste on the inside except that it has probably been cooked by now and that you will have to replace when you open the cpu anyway.
>>52432222
lapping is a waste of time, end result does look pretty cool though
>>52432236
>gap created by the black glue
that's just wrong. if you place the clean heat spreader on the cpu die there is still a gap where the silicone would be. I noticed this practicing on a pentium 4 I had lying around. I sealed mine up with some black RTV silicone because I'm not a lazy idiot who lets the motherboard hold the cpu together
>>52432284
>pentium 4
I'm talking specifically about the haswell processors. Also, there is a gap but the glue is thicker than the gap.
>>52432302
this is just made up information, the silicone glue is a viscous liquid when applied, how on earth could the glue cause there to be space between the processor and heat spreader? If this only applied to haswell, care to explain why people delid other processor types then? Its cheap paste intel applies just like the cheap paste gpu manufactures apply to their coolers
>>52432284
Why the gap is there is anyone's guess(besides the fucking jews that intel who engineered this failure). But just look at the thickness of the TIM on the heatspreader in the OP . Now look at my 3570k that doesn't have the IHS glued in place.
>>52432395
At intel*
>>52432244
It should logically produce an effect. Since there's more metal-metal surface contact there should be more heat conducted.
Of course you have to lap the heatsink too or else there wouldn't be any point.
>>52432336
>>52432144
>>52432395
TIM isn't to be used to literally fill in the gap of two objects that aren't touching. It's a miracle that the TIM intel used works at all
>>52432206
Why not use one of those special brackets and apply the heatsink directly to the die? skipping the lid entirely.
>>52432395
>>52432435
Where do people come up with this stuff? I'm not taking this bait anymore.
>>52432413
On a naked surface yes. Thermal compound is very good at its job, so good it makes lapping produce almost no results. The heat spreader also bottlenecks thermal conductivity. Lapping the cooler + direct cpu die contact would produce the best possible results.
>>52432435
The shit they use is actually pretty decent quality stuff. It just can't do it's job when it's a mile thick.
Literally no surface to surface contact. The mind fucking boggles.
>>52432506
Are you actually retarded?
>>52432144
>buy overly expensive jewtel-cpu
>void warranty to fix intel's ultimate jewing
>???
>profit
And you people wonder why Intel's not using proper TIM anymore...
>>52432492
That would produce the best results but its a lot harder to apply even pressure to the much more fragile/smaller cpu die than the heat spreader
>>52432244
>waste of time
>58°C instead of 70°C for stock speeds
>68°C instead of 88°C for overclocked chips
http://www.pcper.com/news/Processors/Photos-and-Tests-Skylake-Intel-Core-i7-6700K-Delidded
Yeah, nah.
>>52432534
>implying anybody has ever needed to send their intel cpu for warranty service
>>52432551
>lapping is a waste of time
>Intel-Core-i7-6700K-Delidded
nice reading comprehension
>>52432236
>>52432395
>>52432435
>>52432520
>>52432607
You fly in the face of all the evidence and stick to the first knee-jerk conclusion everyone had the first week of Ivybridge. You're a brave little tard.
>>52432591
I had one of my 2500K chips die completely after 3 years. Sent it back to intel and got a brand new boxed chip back from them.
>>52432727
What evidence? As someone who has just completed a delid I can tell you you're wrong. They would need to apply pressure to the heat spreader to glue it to the pcb. Why would they put enough pressure to spread the silicone but stop right before contacting the cpu? The silicon is a liquid when applied and you would need to really go out of your way to leave a gap between the cpu and heat spreader. There is no gap. There is literally no source on that other than this thread.
>>52432743
>after 3 years
Warranty is only good for 3 years, anon
>>52432812
>Warranty is only good for 3 years, anon
I've had logitech replace a mouse without even checking a receipt or invoice. Not that they aren't that hard to doctor though.
>There is literally no source on that other than this thread.
Look at any pictures of a delidded haswell or ivy bridge and notice the block of gray glued to the middle of the heat spreader. I have not personally delidded my processor but people have reported up to 20C lower temps. Even shit paste applied properly will not be that bad. There's a gap.
>>52432812
>There is literally no source on that other than this thread.
Here ya go.
http://forums.anandtech.com/showpost.php?p=34053183&postcount=570
>>52432812
The fucking evidence everyone but you saw when we stopped and looked at everyone else's delidded CPUs and said "why the fuck is not a one of the heat spreaders making contact with the die?" It's as simple as opening your fucking eyes.
>>52433011
And I spent like 20 minutes compiling all those images so quit with your bait shit, we're not trying to troll you.
>tfw using based Q6600 with soldered heatspreader that runs at 54C max
Feels good.
>>52432591
all cpus after sandy bridge except devil's canyon) have the same issue, friend.
>>52432144
did you sand the lid and heatspreader with at least 2000 grit sandpaper? otherwise you fell for a meme
>>52433109
As far as cpu modding goes sanding/lapping is the biggest meme of them all
>>52433135
>void warranty for a meme
>sanding against glass to further reduce temps is a meme
>>52433159
inb4 wahwah too hard
>>52433159
>drop temps by 10C
>drop temps by 1, maybe 2C
Which one is more of a meme?
OP here. I gave the silicone some time to cure and it looks like I dropped 2C idle and 8C under load using prime95. It wasn't hard and my temps weren't too high to begin with so I suppose I can't complain
>>52433093
What's your heatsink? How high have you cranked the volts and overclock, if any?
Why would you delid a 4790k? The 4770k was the one with the bad TIM, 4790k's is fine. I'm sure you could get better temps but at that point its hardly worth the inherent risk.
>>52433309
Makes me feel a bit better...i have a 4790k sitting for awhile. i hope it's fine when i order the rest of my parts soon.
>>52433309
nope.
http://www.kitguru.net/gaming/uncategorized/anton-shilov/intels-devils-canyon-chips-ngptim-is-still-not-efficient-research/
inb4
>shillov
>>52432244
No it isn't. The surface of your IHS isn't anywhere near flat or smooth. Its generally a really porous alloy which makes it even worse. Lapping reduces the surface variation and thereby reduces the gaps that TIM would have to fill between it and the heat sink. Lapping your IHS and underside of your heat sink can lower temps by 10c.
The more perfect each surface is the less work your thermal compound has to do.
>>52434244
I regard the IHS as a really tightly attached protective cover. First thing I do is pop that fucking thing off and replace with a custom shim + piece of copper. Of course I apply decent paste as well. Next step is to throw that stock cooler in the pile with the rest of the junk coolers and put a real heatsink that weighs over half a kilo in.
