7700K for Delidding!

Creepin_D said:
How safe/unsafe is it to place a water block directly on the core?
Click to expand...

mmm
Well first you need to take care of the difference in height

that's probably the worst

the DIE reminded me of my AthlonXP and Duron days

should be fine if handled with care
if you get the height right

also a custom cooler would probably be better for that

AiO use a pump on the CPU block
adding weight, as well as strain from the hoses

using a custom loop you remove the weight of the pump and you can make the hose or pipes fit to you're case
removing any pull from them
 
MV75 said:
So run without the lid? Or are there certain tolerances on the heatsink mounting that prevent that? I haven't mucked around with hardware on this level in years, I just build and go these days. :)
Click to expand...

I'm in the exact same position, and wonder what the point of re-applying the IHS actually is, is it because the core will be crushed when you apply a cooler to it without the IHS, or is the IHS thick so that the core is recessed and does not touch the cooler after IHS removal?

I'm really thinking of doing this to my old Haswell, as it's core temps are insane (Prime95) at even 4.2GHz, and I know just how bad the IHS is.

Also, does anyone know of a delidding tool that's sold in Europe?
 
Last edited:
Stimpy88 said:
I'm in the exact same position, and wonder what the point of re-applying the IHS actually is, is it because the core will be crushed when you apply a cooler to it without the IHS, or is the IHS thick so that the core is recessed and does not touch the cooler after IHS removal?

I'm really thinking of doing this to my old Haswell, as it's core temps are insane (Prime95) at even 4.2GHz, and I know just how bad the IHS is.

Also, does anyone know of a delidding tool that's sold in Europe?
Click to expand...
There's a height difference
cooler wouldn't make contact
Edit:
for instance

https://forums.anandtech.com/thread...d-3770k-an-h100-and-9-different-tims.2285595/

Overclocker UK
or Caseking in Germany

both the same tools, as they are partners

it's out of stock now (here in germany) since they make new ones for kaby lake

difference in Temps of naked vs with IHS is, what I found on forums, something like 5 degrees
 
Creepin_D said:
How safe/unsafe is it to place a water block directly on the core?
Click to expand...

Pretty safe if you had a naked mount kit example: I have an ivy bridge naked kit by ek but I need a certain water cooling block by ek. Ek need make one for this processor and WC CPU block.
 
Hitti2 said:
Pretty safe if you had a naked mount kit example: I have an ivy bridge naked kit by ek but I need a certain water cooling block by ek. Ek need make one for this processor and WC CPU block.
Click to expand...

Yeah those would be nice

however I've even seen different statements if the same mounting kit actually fits for skylake (no info as far as I can see on Kaby)

so mmm

sure would love some custom mounting from EK for each generation
 
Crixus said:
Delidding the CPU is essentially just that. You are taking the heatsink ( lid ) off of the CPU. Once properly removed you must wipe off the old inferior Thermal Interface Material (TIM) installed by Intel and reapply new, more awesome TIM. Before new TIM is added, it is absolutely recommended that one polishes both of the interfacing surfaces to reduce porousness and maximize surface contact. This allows for better Thermal Conductivity between the hot surface of the CPU and the heatsink, giving you what is hopefully a lower operating temperature while slightly raising your thresholds for overclocking.
Click to expand...

Does TIM become less effective over time similar to thermal paste? I've never done this before and may try it out on my 2700k.
 
schlitzbull said:
Does TIM become less effective over time similar to thermal paste? I've never done this before and may try it out on my 2700k.
Click to expand...

You're 2700k is soldered to the IHS
there is no paste
while it is possible to delidd it, we're looking at a decrease of temps around 5 degrees for you


might give lapping the IHS a try

soldered parts also tended to have more uneven IHS

P1070921%20%5B550x%5D.JPG
 
Peter2k said:
There's a height difference
cooler wouldn't make contact
Edit:
for instance

https://forums.anandtech.com/thread...d-3770k-an-h100-and-9-different-tims.2285595/

Overclocker UK
or Caseking in Germany

both the same tools, as they are partners

it's out of stock now (here in germany) since they make new ones for kaby lake

difference in Temps of naked vs with IHS is, what I found on forums, something like 5 degrees
Click to expand...

Many thanks for taking the time to answer...

I have been all over OCUK site, and have used the search and all categories I can think of, and I cannot find any kind of de-lidding hardware... Do you have a link or just the make of the De-lidder? I'm sorry, I am normally the first to roll my eyes at people that can't use search, but I really can't find it! lol
 
schlitzbull said:
Does TIM become less effective over time similar to thermal paste? I've never done this before and may try it out on my 2700k.
Click to expand...

Uh... I think the Sandy Bridge ones are soldered to the heat spreader, so this method of delidding does not work on them. It might be possible with the right tools and enough heat to loosen up the solder, but chances are high on just destroying the whole thing.
 
jimbob200521 said:
I'm seeing a lot of scuttlebutt about people delidding this new CPU, guessing the TIM Intel used wasn't up to snuff? I've seen people "report" as much as a 10° difference once delidded.
Click to expand...


why are consumers doing intels work when intel should of already been doing this work for us. we shouldnt have to delid our cpus. this is just pathetic on intels part. the TIM that they use is useless. with all their r&d money they cant even put a proper TIM on their cpus. intel your useless and pathetic
 
canmanone said:
why are consumers doing intels work when intel should of already been doing this work for us. we shouldnt have to delid our cpus. this is just pathetic on intels part. the TIM that they use is useless. with all their r&d money they cant even put a proper TIM on their cpus. intel your useless and pathetic
Click to expand...

Yeah I never understood why they'd skimp on something as important as TIM on a K series CPU. I could understand an i3 or something that was intended for an office environment but a CPU meant to be OC'd? Come on now, that's first grade stuff there...
 
Stimpy88 said:
Many thanks for taking the time to answer...

I have been all over OCUK site, and have used the search and all categories I can think of, and I cannot find any kind of de-lidding hardware... Do you have a link or just the make of the De-lidder? I'm sorry, I am normally the first to roll my eyes at people that can't use search, but I really can't find it! lol
Click to expand...

only in archive

https://www.overclockers.co.uk/der8auer-delid-die-mate-hs-000-dr.html

probably making a new one for kaby

otherwise ebay might be you're friend
I found mine that way

it's just in between generations that don't fit
 
jimbob200521 said:
Yeah I never understood why they'd skimp on something as important as TIM on a K series CPU. I could understand an i3 or something that was intended for an office environment but a CPU meant to be OC'd? Come on now, that's first grade stuff there...
Click to expand...

There's actually an ongoing discussion in other boards that it's not so much crappy TIM, but too thick glue

having done the delidd myself
I have to say I was astonished how much paste there was in between

I was always taught that the job of the paste is to fill up microscopic unevenness between the 2 surfaces

since Metall on Metall would give best heat transfer, but no surface is absolutely plane and even

paste fills up the smallest places


but there was quite a lot of TIM under the IHS
looked to me like the 2 surfaces never had actually contact at all
 
I haven't really been into the "OC" crowd for awhile, but I was always a nervous wreck putting heat sinks on Athlon XPs back in the day and being afraid of cracking the die, and having to worry about shorting the f#$ker out with Arctic Silver....I haven't OC'd much in my recent builds largely because of lack of need and time to tinker like I had in my early years. How much difference are folks seeing with de-lidded CPUs? Enough to warrant the pain in the ass, or is this just something for folks playing at the extreme end of things?
 
Pusher of Buttons said:
I haven't really been into the "OC" crowd for awhile, but I was always a nervous wreck putting heat sinks on Athlon XPs back in the day and being afraid of cracking the die, and having to worry about shorting the f#$ker out with Arctic Silver....I haven't OC'd much in my recent builds largely because of lack of need and time to tinker like I had in my early years. How much difference are folks seeing with de-lidded CPUs? Enough to warrant the pain in the ass, or is this just something for folks playing at the extreme end of things?
Click to expand...

Once Kyle gets the testing done, you'll see exactly for this example. Most posts I've seen are showing a 10-12C decrease in temps. Probably not worth it IMO, unless you have free time and money to burn. The difference in temps would probably get you another 1-2% increase in your overclock, but unlikely you'd get much more than that.

I seem to remember reading the problem wasn't with the TIM itself anyway, it was the spacing between the die and the heatspreader. De-lidding and reseating the heatspreader manually just enabled you mate the surfaces better than Intel's high-volume automated process.
 
Bandalo said:
Once Kyle gets the testing done, you'll see exactly for this example. Most posts I've seen are showing a 10-12C decrease in temps. Probably not worth it IMO, unless you have free time and money to burn. The difference in temps would probably get you another 1-2% increase in your overclock, but unlikely you'd get much more than that.

I seem to remember reading the problem wasn't with the TIM itself anyway, it was the spacing between the die and the heatspreader. De-lidding and reseating the heatspreader manually just enabled you mate the surfaces better than Intel's high-volume automated process.
Click to expand...

Suppose that makes sense. While I don't miss the risk to the die, the move to heat spreaders on enthusiast level chips never made much sense to me. I mean, why do I want a layer of aluminum and whatever junk TIM material between direct contact with a pricey HSF/AIO? Of course, the last time I had a HSF put directly on the die it was this old beast....which I thought was absolutely massive back in the day.....I'd feel less confident trying to maneuver a big tower cooler or waterblock with stiff hoses and such.
upload_2017-1-19_11-52-27.jpeg
 
Bandalo said:
Once Kyle gets the testing done, you'll see exactly for this example. Most posts I've seen are showing a 10-12C decrease in temps. Probably not worth it IMO, unless you have free time and money to burn. The difference in temps would probably get you another 1-2% increase in your overclock, but unlikely you'd get much more than that.

I seem to remember reading the problem wasn't with the TIM itself anyway, it was the spacing between the die and the heatspreader. De-lidding and reseating the heatspreader manually just enabled you mate the surfaces better than Intel's high-volume automated process.
Click to expand...

You say that now :ROFLMAO:

I was kinda surprised when my Temps on my 7600k dropped 20 degrees :D

but then mileage e may vary
as in how thick the glue actually was on you're modell
 
Hypergreatthing said:
Here, i found this article with pics. It's what i'm talking about when talking about using tape on the gpu/cpu to cover exposed parts when applying electrically conductive tim.
http://forum.notebookreview.com/thr...t-vs-cool-laboratory-liquid-ultra-pro.791489/

Any thoughts on that?
Click to expand...

We've been doing that for years now, using tape that is to boarder cpu die/ihs when applying liquid TIMs.

Bandalo said:
Once Kyle gets the testing done, you'll see exactly for this example. Most posts I've seen are showing a 10-12C decrease in temps. Probably not worth it IMO, unless you have free time and money to burn. The difference in temps would probably get you another 1-2% increase in your overclock, but unlikely you'd get much more than that.

I seem to remember reading the problem wasn't with the TIM itself anyway, it was the spacing between the die and the heatspreader. De-lidding and reseating the heatspreader manually just enabled you mate the surfaces better than Intel's high-volume automated process.
Click to expand...

It's a helluvalot more than 10-12c.

https://hardforum.com/threads/delidded-6700k.1920121/#post-1042721901
 
Pusher of Buttons said:
Suppose that makes sense. While I don't miss the risk to the die, the move to heat spreaders on enthusiast level chips never made much sense to me. I mean, why do I want a layer of aluminum and whatever junk TIM material between direct contact with a pricey HSF/AIO? Of course, the last time I had a HSF put directly on the die it was this old beast....which I thought was absolutely massive back in the day.....I'd feel less confident trying to maneuver a big tower cooler or waterblock with stiff hoses and such.
View attachment 14992
Click to expand...

Well, from the manufacturer's point of view, heatspreaders make a LOT more sense. Protect the die from mechanical damage, and provide a much better (mechanical) interface for a wider variety of heatsinks. The few degrees of temp difference are really irrelevant for 99% of users anyway.

thesmokingman said:
We've been doing that for years now, using tape that is to boarder cpu die/ihs when applying liquid TIMs.



It's a helluvalot more than 10-12c.

https://hardforum.com/threads/delidded-6700k.1920121/#post-1042721901
Click to expand...

The difference varies pretty widely from chip to chip. Which leans towards my belief that the TIM Intel used isn't the problem, it's the gap between the die and the heatspreader. There's no way you'd get a 20C difference with just a change in TIM...most TIM changes are only a couple degrees C either way. 20C implies a poor fit between surfaces.
 
And let me also add that besides the massive temp drop with delidding, the REAL key is overclocking at a high level. What happens when pushing for 4.8ghz to 4.9ghz with Skylake or 5ghz with krabby lake is that you end up going tooooo close to tjmax, think temps well over 85c at those clocks. Thus delidding drops your temp 20c, thereby making 4.8ghz on skylake and 5ghz on krabby, an easy feat to achieve.

Bandalo said:
Well, from the manufacturer's point of view, heatspreaders make a LOT more sense. Protect the die from mechanical damage, and provide a much better (mechanical) interface for a wider variety of heatsinks. The few degrees of temp difference are really irrelevant for 99% of users anyway.



The difference varies pretty widely from chip to chip. Which leans towards my belief that the TIM Intel used isn't the problem, it's the gap between the die and the heatspreader. There's no way you'd get a 20C difference with just a change in TIM...most TIM changes are only a couple degrees C either way. 20C implies a poor fit between surfaces.
Click to expand...

You have obviously not done it before yet you argue the point like you have?
 
thesmokingman said:
And let me also add that besides the massive temp drop with delidding, the REAL key is overclocking at a high level. What happens when pushing for 4.8ghz to 4.9ghz with Skylake or 5ghz with krabby lake is that you end up going tooooo close to tjmax, think temps well over 85c at those clocks. Thus delidding drops your temp 20c, thereby making 4.8ghz on skylake and 5ghz on krabby, an easy feat to achieve.

You have obviously not done it before yet you argue the point like you have?
Click to expand...

Well, to start, I've worked with thermal dynamics for years, and taken quite a few classes on the subject, specifically how heat transfers through materials. I've done my research, because according to physics, you should always see a temp improvement by removing the "lid" and having one less thermal barrier to transmit your generated heat though. But the thing is, the heatspreader material, and the TIM on the inside are VERY thin by design, and present very little barrier to heat transfer.

If you hop onto ANY big online "roundup" of TIM's done by your favorite review site, you'll see there's usually only a 4-5C difference between the top and bottom compounds. Even on sites that use crap like toothpaste or mayonnaise as a test substance. A good TIM is designed to fill the small gaps between the two metal transfer surfaces, NOT present a solid layer of material. If you have a good mount between the surfaces, the TIM becomes nearly irrelevant, and a $2 tube of thermal grease will be almost as good as a $20 liquid metal kit.

So, if you're seeing a 20C difference at these thermal power levels, that implies that you don't have direct die-to-metal contact, you have a relatively thick layer of TIM that's filling the air gap between the die and the heat spreader.

I'd bet good money if you removed the lid, scraped the old glue off and simply remounted the lid with the stock TIM, you'd see almost the same improvement in performance.
 
Bandalo said:
Well, to start, I've worked with thermal dynamics for years, and taken quite a few classes on the subject, specifically how heat transfers through materials. I've done my research, because according to physics, you should always see a temp improvement by removing the "lid" and having one less thermal barrier to transmit your generated heat though. But the thing is, the heatspreader material, and the TIM on the inside are VERY thin by design, and present very little barrier to heat transfer.

If you hop onto ANY big online "roundup" of TIM's done by your favorite review site, you'll see there's usually only a 4-5C difference between the top and bottom compounds. Even on sites that use crap like toothpaste or mayonnaise as a test substance. A good TIM is designed to fill the small gaps between the two metal transfer surfaces, NOT present a solid layer of material. If you have a good mount between the surfaces, the TIM becomes nearly irrelevant, and a $2 tube of thermal grease will be almost as good as a $20 liquid metal kit.

So, if you're seeing a 20C difference at these thermal power levels, that implies that you don't have direct die-to-metal contact, you have a relatively thick layer of TIM that's filling the air gap between the die and the heat spreader.

I'd bet good money if you removed the lid, scraped the old glue off and simply remounted the lid with the stock TIM, you'd see almost the same improvement in performance.
Click to expand...

You wouldn't get shit, maybe 5-6c like the Linus did in his video where he delidded with regular paste using stock cooler. Again, you're arguing shit in the face of evidence. And this theory of too much height under the IHS is just a theory, an assumption started by Silicon Lottery that has somehow become a freaking Gospel. No need to debate this point further lol. smh.

https://forums.anandtech.com/thread...ies-overclocks-inside-now-with-delid.2493250/
 
pretty cool little tool.

I think I'd simply use Noctura NT-H1 paste for the inside. Probably better than the standard Intel stuff.

I wonder if JB Weld would work or I think JB Weld has a high temp product, too?
 
thesmokingman said:
You wouldn't get shit, maybe 5-6c like the Linus did in his video where he delidded with regular paste using stock cooler. Again, you're arguing shit in the face of evidence. And this theory of too much height under the IHS is just a theory, an assumption started by Silicon Lottery that has somehow become a freaking Gospel. No need to debate this point further lol. smh.

https://forums.anandtech.com/thread...ies-overclocks-inside-now-with-delid.2493250/
Click to expand...

When you're done shaking your head, can you perhaps propose a theory as to WHY you're seeing a 20C difference? Even the worst, most generic thermal grease shouldn't have that much of a temperature delta if the surfaces are properly mated.
 
Bandalo said:
When you're done shaking your head, can you perhaps propose a theory as to WHY you're seeing a 20C difference? Even the worst, most generic thermal grease shouldn't have that much of a temperature delta if the surfaces are properly mated.
Click to expand...

I'm not going to bother because I've no way to test my theory and I'm definitely not gonna sit here and act like its a fact when I've no fucking clue.
 
They put the lids on so that when you tighten the heat sink down it does not crush the cpu die which is organic substrate. I would not be surprised if the tim they use is machine applied in a clean room with little to no oxygen. Most people use silver nitrate which bonds to oxygen molecules so the ones that have the metal to lid are likely permanently welded in place and the ones that can delided are vacuum sealed with tape/paste/glue to hold them in place when the seal is created. So the heat sensors likely do not work as well if they are no longer vacuumed sealed. Those sensors measure the resistance of the air in the sealed area and if you reseal it and the cpu has no way of telling how much air is there or what the base line resistance is you may actually be getting worse temps but the sensor can not meausre it correctly. Much like if you increase or decrease the size of the wheels on your car and the dealership does not adjust the speedometer.
 
drakken said:
They put the lids on so that when you tighten the heat sink down it does not crush the cpu die which is organic substrate. I would not be surprised if the tim they use is machine applied in a clean room with little to no oxygen. Most people use silver nitrate which bonds to oxygen molecules so the ones that have the metal to lid are likely permanently welded in place and the ones that can delided are vacuum sealed with tape/paste/glue to hold them in place when the seal is created. So the heat sensors likely do not work as well if they are no longer vacuumed sealed. Those sensors measure the resistance of the air in the sealed area and if you reseal it and the cpu has no way of telling how much air is there or what the base line resistance is you may actually be getting worse temps but the sensor can not meausre it correctly. Much like if you increase or decrease the size of the wheels on your car and the dealership does not adjust the speedometer.
Click to expand...

Whiskey Tango Foxtrot.
 
thesmokingman said:
I'm not going to bother because I've no way to test my theory and I'm definitely not gonna sit here and act like its a fact when I've no fucking clue.
Click to expand...

I honestly can't test my theory either, not without a stack of Kaby Lake samples and some good x-ray equipment. I'm just talking the facts of heat transfer. That's the fun of the discussion though.
 
drakken said:
They put the lids on so that when you tighten the heat sink down it does not crush the cpu die which is organic substrate. I would not be surprised if the tim they use is machine applied in a clean room with little to no oxygen. Most people use silver nitrate which bonds to oxygen molecules so the ones that have the metal to lid are likely permanently welded in place and the ones that can delided are vacuum sealed with tape/paste/glue to hold them in place when the seal is created. So the heat sensors likely do not work as well if they are no longer vacuumed sealed. Those sensors measure the resistance of the air in the sealed area and if you reseal it and the cpu has no way of telling how much air is there or what the base line resistance is you may actually be getting worse temps but the sensor can not meausre it correctly. Much like if you increase or decrease the size of the wheels on your car and the dealership does not adjust the speedometer.
Click to expand...

Uh, the heat spreaders aren't sealed.
 
You must log in or register to reply here.
Back
Top