Sup Forums's experiences with delidding?

Just purchased a delid tool, thermal grizzly conductonaut liquid metal tim, and will be delidding two of my processors soon. Will also be relidding it afterwards, too.

What am I in for?

Other urls found in this thread:

amazon.com/dp/B0002BBX3U/
loctiteproducts.com/tds/CNTCT_SILICONE_tds.pdf
youtu.be/Bhe85LLQEfw?t=3m33s
rbt.asia/g/search/text/"infinity glue"/
overclocking.guide/the-truth-about-cpu-soldering/
twitter.com/SFWRedditImages

So I'll be delidding my Skylake 6700k and Devil's Canyon 4690k.

Originally just wanted to do this on the Skylake 6700k, but I think I'll go ahead on my other processor, too. 6700k is barely stable at 4.8GHz at 1.44v, but kinda feel the voltage is a bit excessive and moving to a liquid metal TIM should help me control temperatures and therefore be able to use lower voltages a bit without affecting the overclock. It can get up to 80C on an x264 loop stress test (AIDA64 Extreme reaches 73C) on a H100i V2 with two good Maglev fans.

In my head it makes more sense to prolong the life of the CPU by foregoing Intel's shitty TIM, rather than sticking to Intel's TIM and letting my CPU cook itself to death over the next few years. At least by delidding it, I can prolong its lifespan by having better control over thermals and lower voltages.

Reason being is that Corsair Link reports the H100iV2 as being 40C under load, but the CPU package itself is 80C under load. If perfect thermal transfer happened, these two temperatures would be equalized, but the fact of the matter is that there's a huge temperature gap since the shitty TIM is letting the CPU cook itself.

Also want to do it to my 4690k running at 4.5GHz because it's currently in a mini-ITX case (portable LAN/HomeTheater PC) with a small cooler, and controlling the thermals would be a big boon for it too.

You're in for possibly 2 dead CPUs

I guess we'll see. The delid tools on the market seem to have good reputation for delidding, barring the poorly manufactured 3d printed plastic ones that still require a hammer/vice to delid.

...

I like this.

>heatsink sensor reports 40C
>meanwhile Tcore is 80C
Is the TIM really that bad on newer Intel CPUs?

Give me a few moments while I get a stress test going and let the H100i V2's water temperature hit stabilization.

This is how it looks at the start, notice the huge differential between 80C package and 25.6C H100i V2 water temp.

Scraping all the black glue off takes a while. I'd recommend using an automotive gasket sealant to reattach the heatspreader but just use a very small amount

I dunno why people always recommend using specialized gasket sealant. I just got this since it's a lot cheaper and a lot more precise.

amazon.com/dp/B0002BBX3U/

It works from -60C to 200C, too according to their spec sheet.

loctiteproducts.com/tds/CNTCT_SILICONE_tds.pdf

I can understand the sentiments against using super glue type adhesives, though, which is exactly why I went for a silicone adhesive by Loctite instead.

Why even seal it back up if the paste is going to degrade? Just making shit that much harder.

>amazon.com/dp/B0002BBX3U/

So you can handle the CPU normally and so that it degrades more slowly since it's now in a sealed vacuum rather than being exposed to extra oxygen.

Since I'm purchasing the delidding tool, I can just delid again and relid it again without a problem. But yeah, going through the same razor blade/vice method twice is just stupid.

I thought liquid metal TIMs didn't degrade.

All non-solder TIMs degrade, including your normal Arctic Silver and liquid metals. Liquid metals do "last" longer than the normal pastes since they start very thermally conductive and it takes a long while for it to get to a point where it's hampering thermal conductivity, though.

How long for liquid metal vs whatever crap Intel is shoving in there?

autism only needed by gaymer children to get few fps more

according to intel autists the shit under the IHS in intel CPUs is dow corning stuff. Its shitty thermal paste in terms of conductivity but "should" last as long as the cpu (10 years) as long as its not exposed to air

Believe it or not, Intel shitty 7700K NEEDS to be delidded to maintain sane temperatures

t. dumb intel jew from israel

I wish I had a folder full of these.

And it does last a long time, but at the same poor thermal conductivity.

The reality is that the product will be used for ~3 years then changed out to the brand newest thing.

i dont use jewish faulty products so i wasnt aware, lmao at goys who fell for the shit paste cpu, FX-8350 doesn't have this problem

What's wrong with the 3d printed ones? They seem more efficient since they are plastic and not metal against metal. I rather the plastic give instead of the CPU. And a vice as more force that's easier to control.

>he took the meme seriously
What the fuck.

Wrong, don't use anything besides the liquid metal Tim and the hole down bracket of the mobo to hold the heat spreader down

Lots of wrong information in this thread. The man issue with Intel chips is the back glue they use to keep the heat spreader down cures to be a mm or so to thick and leaves the tim to make up to big of a gap

1) These tools all have bad tolerances depending on the accuracy of the printer being used

2) It still puts pressure on the substrate from the side.

3) People still have to supply their own vice or hammer, in either case risk is involved with both.

4) They're not infinitely reusable due to the softness of the plastic.

youtu.be/Bhe85LLQEfw?t=3m33s

This fatass foreigner had issues with fitment on his 3d printed tool.

Put it back on with RTV silicone and order some plastic blades to get the factory silicone off. Take your time with it and remember to leave a gap. If you do it well, it's impossible to notice the difference in case you need to RMA it.

Also, if you take notice at 4:05 this complete dumbass KEEPS TURNING THE VICE even though it was delidded and he didn't even realize it due to not hearing the sound.

The milled aluminum tools are very obvious when delid happens. The IHS pops off and clanks against the aluminum and that sound lets you know to immediately stop applying pressure.

If this idiot kept going he would've bent the entire substrate inside the shitty plastic tool.

Majority people have a vice if you're not an idiot living in your mom's basement. Also majority people here on Sup Forums only need it for once or twice use and this 3d printed one covers that just fine.

SL has yet to have any complaints with liquid metal for 4+ years. No performance loss so far.

Dow Corning stuff is guaranteed to last 10 years but the performance isn't the greatest.

An adventure into how you got jewed with dogshit stock TIM.

>DELID THIS: the thread
Weak cucks

>Everybody buying an Intel processor must buy a vice and delidding tool

>Intel Circumcision Thread

>The reality is that the product will be used for ~3 years then changed out to the brand newest thing.
Intel marketer's wet dream.

>FX-8350 doesn't have this problem
Literally garbage tier CPU

Gasket sealant is better suited and not any more expensive than what you suggested and if you work on vehicles you probably have some anyway. You should apply it with a toothpick or something very fine

You could leave it loose but it's more fragile and finicky to handle

It's not wrong it's the proper way to reattach the heatspreader but with less adhesive material. I've done this several times with excellent results so I speak from experience

I bought a Ryzen when they were first released. It's great to know my TIM is solder to begin with.

got the der8auer tool ready to go as a pre-summer project (southern hemisphere so i have a few months) for my 3770k and 6700k

fucking 6700k already hits 95C under full load

wtf. what are your ambient temps. stock clock?

If you change the paste, you're just postponing the problem: insert heatsink directly on the die or use CPU with soldered heatspreader. Not even liquid metal (between die and ihs) is comparable to soldered heatspreader or directly placing heatsink on the die.

I have Noctua NH-U12P (or the previous model, I'm not sure), directly on the die of Haswell Pentium (any cpu on lga1150 would work thus). So bin that non-soldered heatspreader.

>Is the TIM really that bad on newer Intel CPUs?

>have to circumcise your cpu to make it work acceptably well
Tell me Intel isn't genuinely Jewish

user used the rockit tool. Very safe and fast. If you delid the devil canyon, just put some electrical tape over the caps.
I still believe debilscanyon isn't a very different upon delid.
6700 still is a bitch running 4.8+ but kaby runs 5.0 on 1.3 easily.

If you use the rockit cool kit you can even relid very easily.

You should use liquid metal on die. It's not even because it has better thermal conductivity but because it's stable.
In rocking +1 year and no signs of degradation yet.
If you use just the needed anmount I bet it can rest for years without issues it's basically indium+gallium

If you have an AIO with those screw down mounts I don't see the point of putting the lid back on. Just carefully apply some TIM of your choice over the chip and carefully remount the AIO using screw down each corner a bit at a time method. Done.

works for me, not a gaymer. at least it doesnt have thermal problems because it doesnt have cum between ihs and the die. FX-8350 is based

In several years, I just saw degradation on laptops. Metal Tim should remain stable for a long time under die.
Ppl usually say shit about pastes, I had one 950i with cheap paste over ihs for 4 years and when I removed it was still okay.

The only real thing is that regular paste can't withstand direct die heat. It will degrade very fast.

Also clu is better than clp. Either use conductonaut or clu. Also don't use liquid metal Over ihs/cooler it's not going to change much and you can end with sciddrnts.

It's a shame the die guards that come with motherboards are not flat plate metal sheets otherwise you could use those.

>1.45 v for 4.5 GHz
Kys retard.
I have 5.0ghz under 1.34v using 1.45 is absolutely retardation. Stock voltage on auto is around 1.1-1.2. you can run 4.5 all cores with less than 1.15v

>hurr I won the silikun luttry y dedn't u
>hurr

You can't the socket retention clip make the die don't contact. You need to replace the mounting for a drect contact, but you risk crack the die and it doesn't make much difference.

>my chip can do [frequency] at [voltage] so all chips can do [frequency] at [voltage]

No idiot. Even shit tier 7700k don't need1.45 to run stock/4.5 .
Who even did this graphic is a shill or a retard.
The most shitty 7700k I saw required 1.25v to all cores 4.5 so you don't know what you saying.

Just stick to your inferior infinity glue © cheap shit.

Got my 6700k delidded with Thermal Liquid ultra.

Temps improved by a solid 15 degrees.

OC I'm running is 4.5hz all core, 1,27v on air with max temps under extreme torture tests not getting above 64 degrees.

Could probably push it to 4.7 ghz for non noticable gains but I'd rather take this oc with the significantly lower voltage and low temps

>go to the trouble of delidding
>run mild overclock that didn't require delidding to achieve

Really glued my IHS.

I am going to say most people don't have a vice, hell, I only have clamps because I do some woodworking that requires it,

>go through the trouble
It wasn't any trouble, I got it done for 25$ and now have a much cooler cpu in return and no more trouble of fans of my cpu making significant noice during heavy loads.

It's funny that you're triggered

>mfw intelcucks can't even shitpost about their DELID DIS experience without trying to make fun of AMD and making themselves look like pathetic dipshits

get a load of this 3.9ghz faggot

Holy shit. 25C heatsink temp means you're probably freezing your ass off in that room, meanwhile the CPU sits at 80C core temp.

How the fuck was this acceptable? Pretty sure that core would throttle even at stock clocks.

>mug ghz
Is this netbust all ova again?
You want 50stage bibelines back?

How does it feel to know Intel is giving up on you MUH CLOCKS faggots with Coffee Lake?

You must have one shit CPU cooler.

And I have a soldered chip, because I'm not a top goy. Just amused at your stupidity, friend.

>clock speed doesn't matter
Are you tech illiterate?

It might be soldered but you need abysmal voltages to get that piece of garbage even close to 4 ghz, it's laughable really.

What's even more funny is that you think this is something to brag about

>Mention Intel
>Cuck tries to drag discussion back to AMD

The huttburt is real.

Here, nice try faggot but I've got a soldered Sandy Bridge CPU. No jizz TIM for me.

Delid dis

The butthurt of being stuck at 2005 clock speeds must be tremendous I agree

>inferior
Soldered.
>infinity glue
Technology Intel doesn't have and wishes they did.
>cheap shit
Terrible yields on monolithic dies are quite expensive, I agree.

Why don't you own an FX 9590? Highest-clocking chip on the market and therefore the best.

>Intelcuck doesn't understand what IPC is
It's ok buddy, just fuck off back to Sup Forums where this kind of ignorance is endemic.

All that 3.9 ghz butthurt

Because it has low IPC and I can have a chip with high clock speeds AND high IPC

Nice reading skills. Get out of your mom's basement and but your own house you cuck.

Post evidence of your 7700K running at 5.4GHz or admit that the 9590 is superior.

Clock speeds >>>>>> (((IPC)))

>1.73v

Skylake have a hard cap around 4.8.
Kaby average 5.0, only shit tier are limited to 4.8-4.9. golden ones reach up to 5.3 with voltages low as 1.35.

Lots of kaby x manage to run 5.2-5.3 but I'm it's not worth costs of x299 platform.

I'm expecting coffee to see if I remain on 1151 or go hedt.

>5.4gz
Lol at speedlets

Coffee is kaby process+ so it's probably possible to reach 4.8 on a good number of CPUs with deliding. Golden ones should reach 5.0, 7800x already at mark.
It will be up to vrms.

>help me control temperatures and therefore be able to use lower voltages a bit
that's not how it works
package temperature has very little to do with how much voltage is required, it's voltages that directly affects temperatures

3.9? You need real good ryzen chips for that.
And don't even talk about the shitty infinity glue © tied to memory running below 3000 even with special kits for inferior platform.

Amd maybe become good, but now ipc frequency and latency is shitty

>max tdp: 219w

>Majority people have a vice

Recent pajeets Intel hired to damage control are the worst.
Just look at that retard!

Not entirely.
The hotter the package you increase resistance you need more voltage.
Decreasing temps also helps increase stability and lower voltage to stabilize.

but that change in resistance is so negligible that you need LN2 cooling to actually see a slight change in voltages
I've been overclocking Intel stuff for years and I've never had a CPU that needed less voltages for a certain clock after delidding and lapping (or installing an expensive waterblock).

Keep bringing up AMD you sad little turd of a human being. If you didn't have such an inferiority complex, you wouldn't have mentioned them in the first place.

...

Good luck, Intel isn't binning those chips for high clockspeeds so I doubt many will make it past 4.5 even with delid and watercooling.

Stay mad
>Muh prosumer cheap knock off
It's almost like an arm CPU lots of cheap weak cores

>20-25% less singlecore performance than Intel's absolute best = weak ARM cores
You heard it here first, folks.

You seriously really like jerking off to Intel's marketing slides, don't you? You saw that stupid shit that everyone else on the internet laughed and shook their heads at, and you said "YEAH! Screw those AMD guys with their glue!". You're mentally retarded.

rbt.asia/g/search/text/"infinity glue"/

>7GHz

Lol at speedlets

you may want to take a look at this:
overclocking.guide/the-truth-about-cpu-soldering/

Or you may want to actually view der8auer's Jewtube channel instead of reading something he wrote two years ago (about Skylake, which WAS fine). He no longer holds the opinion that thermal jizz is "fine" after Kaby Lake and Skylake-X. He even sells his own delidding tool.

Maybe you should be the one reading it, because he claims the thermal cycling required to break down the solder is -40c to 150c at 15 minutes each, which are ranges no CPU is ever going to see unless your dumping it in ln2 followed by sticking it in an oven.