Worse temps after delid 4770k cpu

[peche]

i have never lap or something similar, you just have to clean all the mess, black silicone, thermal paste left overs and whatever you might find there, leave the processor die and IHS clean, thats it, and sit back processor carefully without the IHS, that one will be the last thing followed by closing CPU tray back

 

[Vario]

i have never lap or something similar, you just have to clean all the mess, black silicone, thermal paste left overs and whatever you might find there, leave the processor die and IHS clean, thats it, and sit back processor carefully without the IHS, that one will be the last thing followed by closing CPU tray back

Probably your IHS just happens to be closer to the die, his isn't.

 

[peche]

Probably your IHS just happens to be closer to the die, his isn't.

generally none of all done delids, thats incluide several i7's like 3570K, 3770, 3770K, 4570k 4690k 4770k & 4790K and a petium G series, none of them have been glued back, and the best results are for 3770 & 3770k's samples,

 

[Vario]

generally none of all done delids, thats incluide several i7's like 3570K, 3770, 3770K, 4570k 4690k 4770k & 4790K and a petium G series, none of them have been glued back, and the best results are for 3770 & 3770k's samples,

Yeah I never glued my old chip back either. Just put the lid back on after thermal paste the die, and clamped it with motherboard's retainer system.

 

[peche]

Yeah I never glued my old chip back either. Just put the lid back on after thermal paste the die, and clamped it with motherboard's retainer system.

same here, glue-in back its not bad at all, i haven't done it before cause i have never decided what to use for glue-in back the IHS

Regards,

 

[yogurt_21]

Now that's hardcore.

it used to be the norm. Forgive me for showing my age here but why the F would you put the IHS back on after removing it? One of the main benefits of delidding is getting back to direct contact between the die and your heatsink, just like we did in the althon xp days. The IHS was put there to protect the manufacturers from noobs who fubared their cpus by either cranking the heatsink down too tight or just plain mishandling it. A sheet of metal isn't adding anything to the equation that your heatsink isn't already doing.

 

[Vario]

it used to be the norm. Forgive me for showing my age here but why the F would you put the IHS back on after removing it? One of the main benefits of delidding is getting back to direct contact between the die and your heatsink, just like we did in the althon xp days. The IHS was put there to protect the manufacturers from noobs who fubared their cpus by either cranking the heatsink down too tight or just plain mishandling it. A sheet of metal isn't adding anything to the equation that your heatsink isn't already doing.

Because the modern heatsinks clamp tighter then the Socket A heatsinks and the CPU itself is designed to be protected by the IHS, therefore it is more fragile and susceptible to damage from over clamping. The Socket A also had little rubber dampeners on the corners.

 

[yogurt_21]

so you put it back on to avoid spending 30 cents on rubber?

 

[Kursah]

same here, glue-in back its not bad at all, i haven't done it before cause i have never decided what to use for glue-in back the IHS

Regards,

From what I've read a lotta folks use RTV sealant, which I happen to have a tube from due to resealing work on my wife's car. Something like this should work and is cheap.

For me I'd prefer to keep the IHS on rather than running naked, but then delidding is not something I commonly do. I'd also like to glue the IHS back down and keep the CPU stock in appearance and build, with an improved TIM between IHS and die. That's my goal when I do decide to delid my 4790K.

I can see those seeking maximum performance wanting to keep the IHS off...but inversely I see keeping it on being useful for protecting the die, not rigging up rubber bushings to keep the HSF flat against the die when the IHS can get that job done. I'd also rather seek a long-term maintained improvement at a slight cost of cooling performance rather than short term maximum improvements, this is definitely a preferential thing for users.

I do see where some folks were using shims or the MSI plate on certain main boards. That might be an option for some...but doesn't seem widely available beyond a few select boards.

 

[peche]

i as thinking about loctite thermal resistant adhesives, they perform great and also gasket silicone, 2 pretty affordable and always available options at every hardware store, also since i dont know how they will behave, if chemicals or ingredients of this silicone will divide due heat or produce vapors that could affect temps, so thats why im not using it, also arctic alumina adhesive its another interesting option,

For me I'd prefer to keep the IHS on rather than running naked, but then delidding is not something I commonly do. I'd also like to glue the IHS back down and keep the CPU stock in appearance and build, with an improved TIM between IHS and die. That's my goal when I do decide to delid my 4790K.

the moar you pause delid, the higher temps you will be facing, delid its a pretty easy process, all you need its a sharpen razor blade and quite much patience, it takes less than 5 minutes, cleaning takes a bit moar but still part of the exciting process!

 

[Kursah]

Well I still gotta order the TIM that I plan to use between die and IHS. I am taking your suggestions and running with em though!

RTV is a high temp sealant, I've read of A LOT of delidders using it to keep the IHS in place...and as an auto mechanic I used it a bunch on engines, transmissions, differentials, etc. for various sealing surfaces.

 

[phanbuey]

http://www.performance-pcs.com/thermal-grizzly-conductonaut-high-performance-thermal-grease-1g.html

 

[Kursah]

http://www.performance-pcs.com/thermal-grizzly-conductonaut-high-performance-thermal-grease-1g.html

Doesn't seem to really be in stock anywhere nor available on Amazon with the rest of Thermal Grizzly's stuff. Makes me wonder if it is discontinued?

 

[Sasqui]

Just run it direct die...

That's how my 3570k sits now. Honestly, there was about zero difference vs. aftermarket TIM. The hard part was getting even pressure on the die. I worked on it for about an hour before I could get the system stable, the pins weren't making contact evenly.

 

[yogurt_21]

That's how my 3570k sits now. Honestly, there was about zero difference vs. aftermarket TIM. The hard part was getting even pressure on the die. I worked on it for about an hour before I could get the system stable, the pins weren't making contact evenly.

a common de-lid complaint. Its never really guaranteed to do anything and if the factory used good TIM, it especially won't do anything.

but for me if you already are bothering to de-lid, and then suddenly you're afraid of running without it...maybe you all should steer clear of this to begin with.

might as well be the guy who bought a phase change cooler and then was too afraid to actually use it. He knew he was going to have to deal with condensation when he bought it.

 

[peche]

Well I still gotta order the TIM that I plan to use between die and IHS. I am taking your suggestions and running with em though!

did ya already decided the thermal p you are going to use?
my votes its to go Coollab ultra + MX4, despite Thermal Grizzly stil tentative!

RTV is a high temp sealant, I've read of A LOT of delidders using it to keep the IHS in place...and as an auto mechanic I used it a bunch on engines, transmissions, differentials, etc. for various sealing surfaces.

will make a try on 2 damaged older processor i have, to see how ihard to apply and glue back will this process could be, thanks for the cheer up!

 

[Kursah]

did ya already decided the thermal p you are going to use?
my votes its to go Coollab ultra + MX4, despite Thermal Grizzly stil tentative!

I haven't decided TBH. I would like to try the Thermal Grizzly Conductonaut, but it seems to be hard to find and cost more than it is worth when you do.

From reading it seems that Ultra trades a little efficiency for being far easier to use and clean, while still being far more effective than any other non-metallic solution out there, even trading blows with Pro in some comparisons.

I have a bunch of MX4, though there's some other options I was looking at... Thermal Grizzly Kryonaut was one of them and Gelid GC-Extreme was another. Though really, I'll probably stick with my big tube of MX4, and maybe get another tube of the Noctua stuff that's treated me so well that came with my U14S.

At this point if I want to get this going...Thermal Grizzly is out of the picture, which doesn't bother me much. I'm still on the fence if I wanna kill my warranty yet or not too. But I'm getting eager. Maybe I'll delid my server's 4790K and swap it in my main rig to see how that changes things with it, since it was the worst out of my two chips when used out of box.

 

[Sasqui]

a common de-lid complaint. Its never really guaranteed to do anything and if the factory used good TIM, it especially won't do anything.

but for me if you already are bothering to de-lid, and then suddenly you're afraid of running without it...maybe you all should steer clear of this to begin with.

might as well be the guy who bought a phase change cooler and then was too afraid to actually use it. He knew he was going to have to deal with condensation when he bought it.

No pain, no gain. I'm still running the system with no IHS after some 4 years. If I could do it all over again knowing what I know now ...I wouldn't do it all over again.

 

[peche]

I haven't decided TBH. I would like to try the Thermal Grizzly Conductonaut, but it seems to be hard to find and cost more than it is worth when you do.

yes, compared to other brand they charge premium,

From reading it seems that Ultra trades a little efficiency for being far easier to use and clean, while still being far more effective than any other non-metallic solution out there, even trading blows with Pro in some comparisons.

coollab ultra is the easiest paste to apply, to remove and one of the best rated! for delid purposes...

I have a bunch of MX4, though there's some other options I was looking at... Thermal Grizzly Kryonaut was one of them and Gelid GC-Extreme was another. Though really, I'll probably stick with my big tube of MX4, and maybe get another tube of the Noctua stuff that's treated me so well that came with my U14S.

well at this point if you have a big bunch of MX4 just go ahead and order the missing liquid metal for the intended use, dont waste time with other if arctic meets all your requirements, as it does for me!

At this point if I want to get this going...Thermal Grizzly is out of the picture, which doesn't bother me much. I'm still on the fence if I wanna kill my warranty yet or not too. But I'm getting eager. Maybe I'll delid my server's 4790K and swap it in my main rig to see how that changes things with it, since it was the worst out of my two chips when used out of box.

thats a wise start, my first delid was on my main rig processor, no spare one if the operation went wrong!

Regards,

 

[trog100]

people seem to forget what tim is there for.. direct surface to surface is the ideal.. tim is there simply to fill any tiny air gaps between the two mating surfaces.. assuming the two mating surfaces are reasonably flat and properly put (pressed) together the amount of tim needed to fill the very tiny surface discrepancy air gaps is next to bugger all..

when you do the job yourself you make sure the die and spreader are properly mated together and sufficient contact pressure is applied.. this is what its all about.. not a lot else..

the motherboard clamping mechanism holds it all together there is no need to glue the lid back down..

the intel mass production dollop of tim and dollop of glue does the job well enough for what intel want.. but it can be done better if anyone want to take the time and trouble to do it.. correct assembly being the most important factor.. its got f-ck all to do with tim intel use..

trog

 

[hat]

people seem to forget what tim is there for.. direct surface to surface is the ideal.. tim is there simply to fill any tiny air gaps between the two mating surfaces.. assuming the two mating surfaces are reasonably flat and properly put (pressed) together the amount of tim needed to fill the very tiny surface discrepancy air gaps is next to bugger all..

when you do the job yourself you make sure the die and spreader are properly mated together and sufficient contact pressure is applied.. this is what its all about.. not a lot else..

the motherboard clamping mechanism holds it all together there is no need to glue the lid back down..

the intel mass production dollop of tim and dollop of glue does the job well enough for what intel want.. but it can be done better if anyone want to take the time and trouble to do it.. correct assembly being the most important factor.. its got f-ck all to do with tim intel use..

trog

True... to an extent. Some of us have painstakingly lapped to a mirror finish, but even that still requires thermal paste. You'll never get true surface to surface contact.

If I were to delid a processor, I'd clean it up as best as I could (probably with goo gone), apply a tiny amount of coollab liquid ultra between die and heatspreader, nothing else (no glue, let the retention mechanism hold it on), and mx-4 between heatspreader and heatsink.

In fact, if you're trying to glue the heatspreader back on, it could well be the glue that's causing a gap between the die and heatspreader, resulting in poor contact.

 

[peche]

If I were to delid a processor, I'd clean it up as best as I could (probably with goo gone), apply a tiny amount of coollab liquid ultra between die and heatspreader, nothing else (no glue, let the retention mechanism hold it on), and mx-4 between heatspreader and heatsink.

thats how mine is working! simplicity!

In fact, if you're trying to glue the heatspreader back on, it could well be the glue that's causing a gap between the die and heatspreader, resulting in poor contact.

not exactly, but makes sense a little, all depends how thick will be the layer of glue you use to paste it back !
i have no skills for glue it back so thats why i haven't paste back mine!

Regards,

 

[hat]

That glue will have a thickness, no matter how small. When we're talking about tight tolerances like this, there's not much wiggle room. Best to leave out the glue.

 

[phanbuey]

That glue will have a thickness, no matter how small. When we're talking about tight tolerances like this, there's not much wiggle room. Best to leave out the glue.

I would bet this is what is causing the issue. That die is not even half a millimeter thick. When I was trying to run bare on the 6700k, I was using my business card as an offset in the corners, and it was making poor contact because the business card was too thick.

 

[Kursah]

That glue will have a thickness, no matter how small. When we're talking about tight tolerances like this, there's not much wiggle room. Best to leave out the glue.

Yep I agree, but I've read that quite a few folks have been able to RTV their IHS back on after applying CLP, CLU, Cryo, etc. with success. This is the method I plan to take as I intend the CLU application of my 4790K's to be a permanent or long-term solution. Then I don't need to worry about being as careful if I do a board swap or heatsink re-seat, etc. It is easy to do a thin layer of RTV sealant...but it is also easy to go too thin and lose the adhesive properties...so finding a happy medium between that and a good IHS to die bond is the challenge.

I figure at worst, as you say, the action of re-gluing the IHS will have to be discarded. But from all the delidding response I've read, quite a few were able to re-glue the IHS with success...though there were also a few that had issues. I think it also comes down to what is used and how it is used. I've heard a thickness of a piece of paper is too thick. Luckily I have experience applying thin runs of RTV sealant. I will definitely either report back here or in my own thread when I do this task on my CPU's. But I DO NOT plan to run either CPU naked.