i7-6700k runing hot

[Vario]

Nice work!

And I wouldn't bother with the delid.

Delid helps more with temperatures than voltages, and your problem is voltages.

Delid is a big pain in the ass because if you move your PC you can unshift the lid and then have to pull the whole thing to repaste it, the paste you need to use to get any benefit is Coolaboratory Ultra which is another pain in the ass as well because its a liquid metal that dissolves aluminum and some other metals as well as it can bond with copper alloys. Then you get into the risk aspect of removing the lid, which is more risky on the current Skylake then it was on the Ivy Bridge.

This is just from my personal experience with a really badly binned i5 3570K, much like your i7 6700K, it used a lot of volts stock. It didn't overclock any better after delid even though it dropped 15 to 20*C after delid, but the temperature wasn't really any benefit to getting a higher clock speed or lower voltages. And whenever I moved the case, I had to be careful because it could shift the lid slightly and then require repaste. Then the CLU ate the top surface of my H100i because I used it both under and on top of the lid. At most I got another 100 Mhz out of the chip, which matters fuck all when it comes to any real world aspects.

How about you just buy a nicer air cooler instead, like a NHD14 or NHD15? Or you could do Phanteks PH-TC14PE or Cryorig R1. That will give you the same lower temperatures of a delid and looks awesome too!

 

[infrared]

My experiences with de-lidding are the same as Vario's, it's not beneficial at all if you aren't overclocking, and even if you are the gains are very minimal (+100mhz tops is about right), despite being a substantial temperature improvement. I wouldn't worry about it @IBRAHIM_007 Good job on finding a stable voltage.

Edit: For those interested, I opted to scrap the IHS all together and used a metal spacer made by aquacomputer to protect the processor, THIS.
I think it's a nice solution but not infallible, if your heatsink/waterblock has a very convex base then there's still a chance it could chip the die. I lapped my waterblock flat to be safe, also the waterblock was touching the corners of the socket and making poor contact with the chip so I had to shave them off with a razor. Now all that's out the way it's easy to work on, but it didn't help temps over de-lidding and keeping the heat spreader so it was all for nothing really. Gave me something to tinker with I suppose.

 

[Grings]

good post vario

I dont think delid is worth it with skylake unless you are looking at unrealistic (for day to day use) overclocks way over 5ghz

The temps of the things can be lowered to hedt/sandybridge levels with liquid metal on the chip, but its not hot to the point its a problem like on ivybridge and non refresh haswell anyway

 

[RejZoR]

Delid thats it!

are you nuts? or at least do you have an idea of the purpuse or reason why people take delid as a cooling improvisation ?

Erm, no. I had Athlons that were naked and the edges of the core started chipping as I was re-mounting cooler several times. And those were big chips. Skylake is tiny in comparison due to small manufacturing process, meaning such edge chipping can be fatal. I'd certainly not recommend it. Voltage and LLC control is safer path for such people.

 

[peche]

and its sad that intel use crappy thermal paste between the cpu and IHS in a $329 cpu

all intel user after sandy bridge are in the same point....

Delid is a big pain in the ass because if you move your PC you can unshift the lid and then have to pull the whole thing to repaste it, the paste you need to use to get any benefit is Coolaboratory Ultra which is another pain in the ass as well because its a liquid metal that dissolves aluminum and some other metals as well as it can bond with copper alloys. Then you get into the risk aspect of removing the lid, which is more risky on the current Skylake then it was on the Ivy Bridge.

i have moved my rig on several lanparties... no problems so far, i have replaced and repasted without minor problems, also you could paste or glue back processor's lid with arctic alumina thermal glue....

How about you just buy a nicer air cooler instead, like a NHD14 or NHD15? Or you could do Phanteks PH-TC14PE or Cryorig R1. That will give you the same lower temperatures of a delid and looks awesome too!

this is a quite great idea, AIO watercoolers are great choice sir..,.,

Erm, no. I had Athlons that were naked and the edges of the core started chipping as I was re-mounting cooler several times. And those were big chips. Skylake is tiny in comparison due to small manufacturing process, meaning such edge chipping can be fatal. I'd certainly not recommend it. Voltage and LLC control is safer path for such people.

respect you point, but im talking about the coment on delidding your LGA2011 chip that you stated before...

Regards,

 

[RejZoR]

I haven't said anything about deliding. Except it's very risky either way and in case of LGA2011, also impossible due to soldered IHS.

 

[infrared]

Not impossible, but very risky for sure.

Sorry to OP, I guess the convo kinda went off topic

 

[slozomby]

personally i'd say upper 50's at full load on a 212 is just fine. some light overclocking wouldn't bring temps or volts up too much either.

 

[Grings]

I actually didnt pay too much attention to op's last few posts as we all derailed it with the delid thing, but i see you ran into another bluescreen

Have you tried some lower llc?, i found with my asrock boards (a z68 and z87) that while the top 2 levels are usually too aggressive, as you initially found, the ones around the middle work quite well, perhaps give 4 a try

 

[Mussels]

I know, but mine still has 2 physical cores more. And it's made on older manufacturing process that's more power hungry. Either he has absolutely the worst 6700K or his board has the worst voltage settings. In which case he should really dumb them down. Dramatically.

as the chips get physically smaller, they're transferring heat to a much smaller physical area - through crummy 'just good enough' thermal paste/TIM to the heatspreader. It's beem this way since ivy bridge and its only getting worse every generation with the die shrinks.

delidding makes huge temperature differences on these.

 

[biffzinker]

Coolaboratory Ultra which is another pain in the ass as well because its a liquid metal that dissolves aluminum and some other metals as well as it can bond with copper alloys.

Does nickel plating on copper prevent the liquid gallium from bonding?

 

[infrared]

Does nickel plating on copper prevent the liquid gallium from bonding?

This is what the leaflet says (including weird translated grammar!):
"The application on copper surfaces is a bit easier than on nickel-plated ones, but possible is both"

 

[biffzinker]

Already using the stuff under, and on the heatspreader both are nickel plated so I didn't see any reason to be concerned. Figured the nickel coating would be non-reactive to the gallium.

 

[R-T-B]

sadly my motherboard allow only -5 to -100
which mean i can’t select for example -17 or -13

on -15 with level 5 LLC , sadly it crashed (blue screen)
when i tried to run this benchmark " CPU PhotoWorxx " in the AIDA64
and my motherboard won’t allow me to choose ( -14 , -13 , -12 , -11 ) it only accept -5 or -10 or -15 etc... each step = -5 all the way up to -100

i increased the voltage from -15 to -10 with level 5 LLC , and it didn’t crash no blue screen

and the temperature still good i think -10 is the sweet spot for me hopefully it never crash

Thank you very much guys , still much better than stock Auto Voltage

i don’t want to delid my cpu , its risky with razor method it could scratch or damage the cpu , and i never tried this before

i wish if i have this tool

its called " Rockit 88 Intel CPU Delid tool "

and its sad that intel use crappy thermal paste between the cpu and IHS in a $329 cpu

As someone who delidded a skylake, I don't recommend it. Yes, it's very sad Intel chose to use such a shitty TIM between the cpu and heatspreader, but I find it very diffilcult on thermal paste reapplication to keep the heatspreader still on the die and ensure a good even application there. Takes me like 3 tries and a bad application will end up worse than factory.

Then again, this is coming from the manfrog who is famous for being unable to apply thermal paste without wrecking motherboard sockets by the truckload (I have... motor difficulties, ok guys? .)

 

[infrared]

Then again, this is coming from the man who is famous for being unable to apply thermal paste without wrecking motherboard sockets by the truckload (I have... motor difficulties, ok guys? .)

The thermal paste was bad, the microfiber cloth was the final nail in the coffin , man I felt for you when I saw that though

 

[R-T-B]

The thermal paste was bad, the microfiber cloth was the final nail in the coffin , man I felt for you when I saw that though

I'm jumpy. It's the frog DNA they used to put my right arm back together... or something. Tell me, it's a frickin' party in there:

It's ok. My computer is working now... all's well that ends well, even if I have a much more budget board.

 

[infrared]

Ouch! That's a lot of metal They could have at least given you some wolverine claws while they were upgrading your arm... I guess CPU sockets would be the least of your worries if you had them though. Glad your computer's back up and running anyway.

 

[Arctucas]

As someone who delidded a skylake, I don't recommend it. Yes, it's very sad Intel chose to use such a shitty TIM between the cpu and heatspreader, but I find it very diffilcult on thermal paste reapplication to keep the heatspreader still on the die and ensure a good even application there. Takes me like 3 tries and a bad application will end up worse than factory.

Then again, this is coming from the manfrog who is famous for being unable to apply thermal paste without wrecking motherboard sockets by the truckload (I have... motor difficulties, ok guys? .)

Using the Rockit 88 re-lid tool could not make re-installing the IHS any easier.

Also, the OEM TIM is actually quite good. The issue is; the bead of sealant Intel used causes the IHS to lift off the die. As I am sure everyone knows, less contact between IHS and die means less thermal conductivity, therefore, a hotter running chip.

I de-lidded/re-lidded my 6700K (using the Rockit 88), applying Thermal Grizzly Conductonaut TIM between the IHS and die, I used Thermal Grizzly Kryonuat between the IHS and waterblock.

Before, @ 4700MHz and 1.36VCore, running IBT at Maximum reached 85°C. After, @ 4830MHz and 1.39 VCore, IBT Maximum reached 71°C.

I am not recommending anyone de-lid, but my results speak for themselves.

 

[R-T-B]

If my chip wasn't such an overclocking turd anyways, I'd be looking into buying one of these relidding kits then... very interesting, thanks.

As it stands, I am just content not to mess.

 

[trog100]

the problem isnt the paste intel use.. its the assembly method they use..

a dollop of glue and a dollop of paste then press together and hope it comes out right.. sometimes it dosnt and the end result is too large a gap between the chip and the lid.. the gap just gets filled with too much paste.. the end result is less than optimal heat transfer..

doing it yourself and doing it properly (no matter what paste you use) does a better job..

trog

 

[OneMoar]

Nice work!

And I wouldn't bother with the delid.

Delid helps more with temperatures than voltages, and your problem is voltages.

Delid is a big pain in the ass because if you move your PC you can unshift the lid and then have to pull the whole thing to repaste it, the paste you need to use to get any benefit is Coolaboratory Ultra which is another pain in the ass as well because its a liquid metal that dissolves aluminum and some other metals as well as it can bond with copper alloys. Then you get into the risk aspect of removing the lid, which is more risky on the current Skylake then it was on the Ivy Bridge.

This is just from my personal experience with a really badly binned i5 3570K, much like your i7 6700K, it used a lot of volts stock. It didn't overclock any better after delid even though it dropped 15 to 20*C after delid, but the temperature wasn't really any benefit to getting a higher clock speed or lower voltages. And whenever I moved the case, I had to be careful because it could shift the lid slightly and then require repaste. Then the CLU ate the top surface of my H100i because I used it both under and on top of the lid. At most I got another 100 Mhz out of the chip, which matters fuck all when it comes to any real world aspects.

How about you just buy a nicer air cooler instead, like a NHD14 or NHD15? Or you could do Phanteks PH-TC14PE or Cryorig R1. That will give you the same lower temperatures of a delid and looks awesome too!

you can just use RTV after applying lquid ultra
then the IHS is secure the stock sealant is nothing but a silicone rtv adhesive
Permatex 80050 is probly pretty close but any old RTV will work just apply a thin layer and clamp to dry DO NOT do as some people suggest and use super glue it can degrade the sub strait

 

[64K]

If my chip wasn't such an overclocking turd anyways, I'd be looking into buying one of these relidding kits then... very interesting, thanks.

As it stands, I am just content not to mess.

If it aint broke don't fix it.