Kaby Lake crap TIM , WHY ?

[Armenius]

So far everything you linked to is for extreme overclockers that will do anything for a few degrees as they try to set world records. Stress cracking of indium is only a issue when using LN2 due to the extreme temperatures, under normal operation a processor is never exposed to that extreme of a temperature shift. I don't believe this is a mainstream thing being done in the overclocking enthusiast community, the hard core community might have a few people crazy enough. You can try to spin it all you want, but they use paste to save cash and it works good enough, it simplifies manufacturing and those savings add up on mass produced items.

You're moving the goal posts...

Umm you cant delid a processor with a solder heat spreader as it will destroy the chip trying to remove it. I have never heard of anyone doing that and having a working chip afterwords. So I would like to see a example of this great result after removing a solder on heat spreader. Starting to wonder how much Intel pr pays you.

 

[thesmokingman]

Obvious is obvious, dont get your panties in a wad
You didnt consider that my core flexes a little when hot, no magic.

The point is that Intel need to leave a gap with tolerance for problems, I gave a major reason.
It explains why they dont solder.

I detailed my experience so that others might not screw up their own kit after delidding.
When I first apply paste my load temps are low 50C at 4.7GHz on all cores.
After a few days of the core flexing under load, this changes and I have to reduce the overclock to 4.6GHz.

Right now, 2 of my 8 cores are 25C higher than the rest under Prime95.
That is the best I have managed to get it so far with only a very mild tightening of the heatsink.
Shrug.
Its still better than before the delid, my VCore is slightly lower with the same clock speed.

It reads more like you've done something wrong. Did you glue the IHS back on or no?

User error doesn't have any bearing on why they use paste, nor why they solder on their 1.6K cpus.

 

[Nenu]

It reads more like you've done something wrong. Did you glue the IHS back on or no?

User error doesn't have any bearing on why they use paste, nor why they solder on their 1.6K cpus.

I applied too much pressure which may have deformed the socket a little.
But that highlighted why the lid is used for stress relief.

 

[Gideon]

You're moving the goal posts...

I already said I was wrong on the removal of the solder on IHS and not being able to do it without destroying the chip. However I was right and pretty much no one does that as there is almost no gain to doing so. Solder on IHS is the next best thing to just going bare core for temps. Intel uses paste just to save money just the same reason AMD started doing it as well. Shintai made it sound like it was common and there were big gains from it and I struggled to find much of anything on it and the few I did didnt recommend it as the gain was minimal. No one worries about solder cracking unless they are doing extreme cooling so I dont think I shifted the goal posts, someone else did. I used to do phase change cooling and never had this problem with solder on IHS, LN2 guys are the only ones that have to worry about that.

 

[SomeGuy133]

Ugh, you never want a gap between your IHS, TIM and heat element. The TIM is not the cooling apparatus, it is merely there to fill in the micro gaps between the two surfaces. I repeat again, you never want space between the TIM and heat element. If as you theorize that the IHS compresses onto the TIM each time a mount force is applied, THERE WILL BE NO TIM LEFT the next time a heatsink is applied. Each time a force is applied the TIM will be spread out, and each successive time there will be no TIM left to push down on because its already been spread out. This theory is counter to how TIM is applied and used. There'd have to be a magical force that pushes the TIM back each time for your theory to work.

more so if that was the case there would be massive air bubbles and it would get worse after each use.

 

[SomeGuy133]

Delidding have been done successfully with soldered parts for over 10 years. Then you can make up whatever excuse you want. That failed delids are more attractive to the internet shouldn't surprise you.

If you notice, TIM tends to be used on smaller dies for a good reason. And its not just about money.

I am calling 100% bullshit. TIM is used on smaller dies because they are not enthusiast/server chips. TIM was used for cost cutting. Intel even said that in the beginning.

Also Broadwell E is probably the same size as HW. I can't find broadwell E 6 core die size.

 

[lolfail9001]

I already said I was wrong on the removal of the solder on IHS and not being able to do it without destroying the chip. However I was right and pretty much no one does that as there is almost no gain to doing so. Solder on IHS is the next best thing to just going bare core for temps. Intel uses paste just to save money just the same reason AMD started doing it as well. Shintai made it sound like it was common and there were big gains from it and I struggled to find much of anything on it and the few I did didnt recommend it as the gain was minimal. No one worries about solder cracking unless they are doing extreme cooling so I dont think I shifted the goal posts, someone else did. I used to do phase change cooling and never had this problem with solder on IHS, LN2 guys are the only ones that have to worry about that.

No gain? Roman reported like 10 degree drop in Prime95 back in a day. That's a plenty in my book.
Also, Intel do have to worry about solder cracking, they have dollars to lose on every cracked die. Meanwhile little to no CPU has died from mild overheating in last whatever years. Your personal experience in this case is mostly irrelevant since you are not soldering your CPUs.

I am calling 100% bullshit. TIM is used on smaller dies because they are not enthusiast/server chips. TIM was used for cost cutting. Intel even said that in the beginning.

Also Broadwell E is probably the same size as HW. I can't find broadwell E 6 core die size.

Did Intel?
Also, Broadwell E's largest die is like 450mm^2 for 24 core against 600+ mm^2 for Haswell-E 18 core. Broadwell 10 core is most certainly much smaller than Haswell 8 core.

 

[SomeGuy133]

No gain? Roman reported like 10 degree drop in Prime95 back in a day. That's a plenty in my book.
Also, Intel do have to worry about solder cracking, they have dollars to lose on every cracked die. Meanwhile little to no CPU has died from mild overheating in last whatever years. Your personal experience in this case is mostly irrelevant since you are not soldering your CPUs.

Did Intel?
Also, Broadwell E's largest die is like 450mm^2 for 24 core against 600+ mm^2 for Haswell-E 18 core. Broadwell 10 core is most certainly much smaller than Haswell 8 core.

I was referring to BW E 6 core being probably the same size as a 4 core HW, which makes the others guy point about small dies needing TIM instead of solder BS.

 

[lolfail9001]

I was referring to BW E 6 core being probably the same size as a 4 core HW, which makes the others guy point about small dies needing TIM instead of solder BS.

BW 6 core does not exist, it is a cut of 10 core die that is ~246mm^2 [1] in size. For comparison, Haswell 4+2 die is 177mm^2 [2]. Oh, and Skylake is a whooping 122.4mm^2 [3]

[1] http://hothardware.com/reviews/inte...dition-10-core-cpu-review-broadwell-e-arrives

[2] http://www.anandtech.com/show/7003/the-haswell-review-intel-core-i74770k-i54560k-tested/5

[3] http://www.anandtech.com/show/9505/skylake-cpu-package-analysis

 

[Shintai]

I am calling 100% bullshit. TIM is used on smaller dies because they are not enthusiast/server chips. TIM was used for cost cutting. Intel even said that in the beginning.

Also Broadwell E is probably the same size as HW. I can't find broadwell E 6 core die size.

Please link where it was done for cost cutting. If I remember correctly there was a podcast about it and cost wasn't part of the issue.

Broadwell-E up to 10 cores uses the same die at 246mm2.

http://www.overclockers.com/3step-guide-to-overclock-intel-haswell/

Speaking of temperatures, surely people will be saying “Haswell has the same problem as Ivy Bridge,” referring to the thermal paste issue. Yes, Haswell has thermal paste, but from Very Authoritative People, the TIM is not the problem. As has been posited across the net when people de-lid Ivy Bridge chips with great results, it’s really the black adhesive that’s the culprit. When you cut out that adhesive, it allows the IHS to sit closer to the CPU die, meaning there is less thermal paste through which the heat has to travel, leading to significantly lower temperatures. Intel’s TIM is really quite good, but the manufacturing process leads to that glue being just a little too thick, which is why you see such temperatures.

When you remove the adhesive you also screw the results due to moving the IHS closer to the die. Often plain on top.

 

[SomeGuy133]

BW 6 core does not exist, it is a cut of 10 core die that is ~246mm^2 [1] in size. For comparison, Haswell 4+2 die is 177mm^2 [2]. Oh, and Skylake is a whooping 122.4mm^2 [3]

[1] http://hothardware.com/reviews/inte...dition-10-core-cpu-review-broadwell-e-arrives

[2] http://www.anandtech.com/show/7003/the-haswell-review-intel-core-i74770k-i54560k-tested/5

[3] http://www.anandtech.com/show/9505/skylake-cpu-package-analysis

anandtech has haswell 4 core also at 264mm which is bigger than Broadwell E still making the size point bullshit

Please link where it was done for cost cutting. If I remember correctly there was a podcast about it and cost wasn't part of the issue.

Broadwell-E up to 10 cores uses the same die at 246mm2.

http://www.overclockers.com/3step-guide-to-overclock-intel-haswell/



When you remove the adhesive you also screw the results due to moving the IHS closer to the die. Often plain on top.

i dont buy that adhesive bullshit. how does SL never fuck it up? They still glue it shut. It is the TIM as people have sen better results asnd worse results depending on the TIM

If you claimed earlier that intel is so smart why botch it with the adhesieve and never fixing it? I don't buy it

 

[Gideon]

No gain? Roman reported like 10 degree drop in Prime95 back in a day. That's a plenty in my book.
Also, Intel do have to worry about solder cracking, they have dollars to lose on every cracked die. Meanwhile little to no CPU has died from mild overheating in last whatever years. Your personal experience in this case is mostly irrelevant since you are not soldering your CPUs.

10 degrees with a huge risk of just breaking your $400 plus cpu.. no thanks will be the answer for almost everyone. Solder cracking is a non issue, but if you think it is then why is AMD so good at it and yet Intel has all these issues and worries. Give me a break they did it to save cash and make more profit, it's what anyone as a CEO would do. Hell Chrysler deleted a schrader valve to save .03 cents on every vehicle, just to give you a little insight on cost cutting for profit. CPU's dont die from overheating hardly anymore due to thermal throttling and onboard temp sensors. Back in the day we used to jut apply our paste straight to the die no IHS and the best way to keep your chip cool. AMD has little choice since they need maximum speed out of their chip and that is why the FX is soldered on, otherwise I am sure they would paste it and pocket the money.

 

[lolfail9001]

anandtech has haswell 4 core also at 264mm which is bigger than Broadwell E still making the size point bullshit

Except that Haswell 4+3e quad core is sold as BGA package, so yeah, there is nothing to glue it to.

i dont buy that adhesive bullshit. how does SL never fuck it up? They still glue it shut. It is the TIM as people have sen better results asnd worse results depending on the TIM

They fuck it up as hard, just not as hard as apparently Kaby did.

10 degrees with a huge risk of just breaking your $400 plus cpu.. no thanks will be the answer for almost everyone.

Ironically getting a cooling appropriate enough to drop those 10 degrees will run for more quick enough.

Solder cracking is a non issue, but if you think it is then why is AMD so good at it and yet Intel has all these issues and worries

AMD's soldered dies are literally 300mm^2+. Intel's glued ones are half that.

Give me a break they did it to save cash and make more profit, it's what anyone as a CEO would do.

So, why do they not glue Broadwell-E, then? And don't start on volume thing, Intel's server business is ridiculously big.

 

[SomeGuy133]

Except that Haswell 4+3e quad core is sold as BGA package, so yeah, there is nothing to glue it to.

They fuck it up as hard, just not as hard as apparently Kaby did.

Ironically getting a cooling appropriate enough to drop those 10 degrees will run for more quick enough.

AMD's soldered dies are literally 300mm^2+. Intel's glued ones are half that.

So, why do they not glue Broadwell-E, then? And don't start on volume thing, Intel's server business is ridiculously big.

broadwell E is soldered because it is an expensive CPU derp derp and enthusiast. Also it is the same die as server just locked up....why make one soldered and one glued? Broadwell E is a server chip that lacks important features.

Also SL never fucked up the gluing...... They wouldnt be able to sell chips if they did.



your right i thought it hd an IHS.
http://www.pcgameshardware.de/screenshots/970x546/2013/08/Aufmacher_small-pcgh.JPG

10 degrees with a huge risk of just breaking your $400 plus cpu.. no thanks will be the answer for almost everyone. Solder cracking is a non issue, but if you think it is then why is AMD so good at it and yet Intel has all these issues and worries. Give me a break they did it to save cash and make more profit, it's what anyone as a CEO would do. Hell Chrysler deleted a schrader valve to save .03 cents on every vehicle, just to give you a little insight on cost cutting for profit. CPU's dont die from overheating hardly anymore due to thermal throttling and onboard temp sensors. Back in the day we used to jut apply our paste straight to the die no IHS and the best way to keep your chip cool. AMD has little choice since they need maximum speed out of their chip and that is why the FX is soldered on, otherwise I am sure they would paste it and pocket the money.

not really a big risk at all.....people do it for 50 bucks and warranty their work......

CPUs also almost never die these days because they can run at 100 C...back in the day some had 60C or less maximums.......

 

[Shintai]

CPUs also almost never die these days because they can run at 100 C...back in the day some had 60C or less maximums.......

No they didn't, people just mixed up tjmax and tcase.

 

[SomeGuy133]

http://ark.intel.com/products/27449/Intel-Pentium-4-Processor-505-1M-Cache-2_66-GHz-533-MHz-FSB

do they not release tjmax back than? I find no record. I recall my P4 shutting down at 60-70C back in 2003-2006

 

[pxc]

http://ark.intel.com/products/27449/Intel-Pentium-4-Processor-505-1M-Cache-2_66-GHz-533-MHz-FSB

do they not release tjmax back than? I find no record. I recall my P4 shutting down at 60-70C back in 2003-2006

The 90nm Pentiums only had a primitive on-die thermal diode. The TCC activation temperature (when throttle kicks in, yo) is essentially the the same as the maximum Tcase temperature given in the link. Almost certainly the junction temperature of some interesting point on the die is higher than max Tcase, but it's not specified in older desktop processors (I checked the 90nm Pentium data sheet and 90nm Pentium thermal guide).

IIRC, Tjunction was something specified starting with the Pentium M and carried on in other mobile heritage processors like Core, Core 2 and sucessors.

 

[Shintai]

Yep, it was apple and oranges in measurement compare. You just think its higher today because you get core temperatures.

You can actually only measure tcase by etching down in the IHS and placing a thermocouple between it and the heatsink. 380 microns deep into the IHS or so.

 

[SomeGuy133]

Yep, it was apple and oranges in measurement compare. You just think its higher today because you get core temperatures.

You can actually only measure tcase by etching down in the IHS and placing a thermocouple between it and the heatsink. 380 microns deep into the IHS or so.

so what was the temp that was measured bacvk than? Tcase or core? They sure died at around 60-70C back then

 

[Shintai]

so what was the temp that was measured bacvk than? Tcase or core? They sure died at around 60-70C back then

None. They measured Tdiode or the sensor via the mobo. Its pretty much the same as current AMD CPUs. Zen seems to introduce better located and a lot more accurate sensors.

 

[SomeGuy133]

None. They measured Tdiode or the sensor via the mobo. Its pretty much the same as current AMD CPUs. Zen seems to introduce better located and a lot more accurate sensors.

Okay so Tdiode crapped at 60-70C back then. So how close is that to TjMAX? Internal core temp?

 

[Shintai]

Okay so Tdiode crapped at 60-70C back then. So how close is that to TjMAX? Internal core temp?

Roll a dice

It depends on the math used. For example with AMD I would guess 65-70C is around 100C or so and 80-85C where they shutdown would be around 130C. I dont think its any different from older parts from Intel. However exact numbers and relation that is a mathematical guess is something you have to ask the manufactors for. Motherboard makers also made temperature guesses from the measurement below the socket.

If you remember old graphics cards you would recall some at 130-135C too. Even the modern chipset got a 102C max.

I would say the temperatures have always been the same, its just the accuracy of measurements that is getting better. More sensors, better located sensors, calibrated sensors etc.

 

[MTDEW]

Obvious is obvious, dont get your panties in a wad
You didnt consider that my core flexes a little when hot, no magic.

The point is that Intel need to leave a gap with tolerance for problems, I gave a major reason.
It explains why they dont solder.

I detailed my experience so that others might not screw up their own kit after delidding.
When I first apply paste my load temps are low 50C at 4.7GHz on all cores.
After a few days of the core flexing under load, this changes and I have to reduce the overclock to 4.6GHz.

Right now, 2 of my 8 cores are 25C higher than the rest under Prime95.
That is the best I have managed to get it so far with only a very mild tightening of the heatsink.
Shrug.
Its still better than before the delid, my VCore is slightly lower with the same clock speed.

What heatsink are you using?
And did you re-attach the lid so that it stays centered when clamping down on it in the socket?
Or did you leave it loose and just placed it on top and centered it by eye when installing?

 

[Comixbooks]

I forget where I think the chips are made in Santa Clara or whatever but the packaging in done overseas I'm not sure if the heatsink is put on overseas or here in the USA or just packaged overseas. This has been a problem since Ivy Bridge I think they improved it with Haswell but maybe they slacked off again.

 

[thesmokingman]

I applied too much pressure which may have deformed the socket a little.
But that highlighted why the lid is used for stress relief.

Right because you applying too much force had nothing to do with it. Nor did coming up with a fantastical theory to explain it, cushion under the IHS to protect the die.... lmao.

more so if that was the case there would be massive air bubbles and it would get worse after each use.

True, I hadn't even touched on that aspect either.

 

[NamelessPFG]

Your question should be, why isn't Intel and AMD selling delidded CPUs to enthusiasts. If you think AMD uses solder in its newer CPUs you will be disappointed.

Funny thing is, both Intel and AMD used to do exactly that for FC-PGA(2) Socket 370 and Socket A, respectively. Bare die, no heatspreaders in the way!

Now count the horror stories of people accidentally destroying their expensive new CPUs via improper heatsink installation chipping, cracking and/or crushing the delicate die. They're not exactly uncommon.

 

[Shintai]

Funny thing is, both Intel and AMD used to do exactly that for FC-PGA(2) Socket 370 and Socket A, respectively. Bare die, no heatspreaders in the way!

Now count the horror stories of people accidentally destroying their expensive new CPUs via improper heatsink installation chipping, cracking and/or crushing the delicate die. They're not exactly uncommon.

Oh I know, I should have added something without warranty or like a 40$ replacement policy

 

[Palladium@SG]

Funny thing is, both Intel and AMD used to do exactly that for FC-PGA(2) Socket 370 and Socket A, respectively. Bare die, no heatspreaders in the way!

Now count the horror stories of people accidentally destroying their expensive new CPUs via improper heatsink installation chipping, cracking and/or crushing the delicate die. They're not exactly uncommon.

Every broken bare dies I heard are always Socket A. Pre-IHS Intel chips are built like tanks which is why they never had the 4 sponges on the corners.

 

[NamelessPFG]

Every broken bare dies I heard are always Socket A. Pre-IHS Intel chips are built like tanks which is why they never had the 4 sponges on the corners.

I have heard that it's pretty easy to kill a Coppermine or Tualatin chip if you're nuts enough to try mounting a Thermaltake Golden Orb on it, though.

For those of you not familiar with the HSF in question, it's a fairly popular one for the time made with PPGA-packaged CPUs in mind, since you have to twist-fit it for mounting (twisting it causes these clips to retract and grab the Socket 370 mounting tabs). I actually got an Abit BP6 board years ago with two of 'em plopped on, fins pre-filed to clear the capacitors near the sockets and all.

Using one of the more typical mounting systems for the period like a lever latch probably wouldn't risk harming the die, though.

 

[thesmokingman]

I have heard that it's pretty easy to kill a Coppermine or Tualatin chip if you're nuts enough to try mounting a Thermaltake Golden Orb on it, though.

For those of you not familiar with the HSF in question, it's a fairly popular one for the time made with PPGA-packaged CPUs in mind, since you have to twist-fit it for mounting (twisting it causes these clips to retract and grab the Socket 370 mounting tabs). I actually got an Abit BP6 board years ago with two of 'em plopped on, fins pre-filed to clear the capacitors near the sockets and all.

Using one of the more typical mounting systems for the period like a lever latch probably wouldn't risk harming the die, though.

Dude that is old school. Dual Celeries for the win! I used Swiftech MCX something the black sinks with pins instead of fins, so no fear during mounting.

 

[PliotronX]

It's genius: anyone wishing to perform decent overclocking needs to relid or delid, absolving Intel from any warranty coverage beyond relying on the honor system (no sir, I did not OC this).

 

[Dan_D]

Thanks for all the informative comments. But back to my original question. Why would you cripple your top of the line consumer CPU which being unlocked and aimed at enthusiasts like us with Tim that's not up to the job. I'm convinced that 6 months on they will release a 5k kabylake with a good time in order to have something to gazump AMD new offering.

Intel hasn't crippled anything. The CPU still clocks higher than its predecessor, and does so more easily. Kaby Lake is an improvement even if it isn't the largest improvement Intel could have made. Intel could do better with the TIM, but Intel's a business first and foremost. It must maximize profits to keep shareholder confidence high. Its that simple.

 

[SomeGuy133]

Funny thing is, both Intel and AMD used to do exactly that for FC-PGA(2) Socket 370 and Socket A, respectively. Bare die, no heatspreaders in the way!

Now count the horror stories of people accidentally destroying their expensive new CPUs via improper heatsink installation chipping, cracking and/or crushing the delicate die. They're not exactly uncommon.

you ccan apply 60-80PSI and maybe even more to the die directly in nahalem and IB builds. My 920XM and 3920XM both had 60-80PSI on them from what i could tell. might have been more on certain spots. This was like 4 years ago so it is a little hard to remember.

Intel hasn't crippled anything. The CPU still clocks higher than its predecessor, and does so more easily. Kaby Lake is an improvement even if it isn't the largest improvement Intel could have made. Intel could do better with the TIM, but Intel's a business first and foremost. It must maximize profits to keep shareholder confidence high. Its that simple.

yes by pissing off customer base to save 25 vents per build instead of just charging an extra fucking dollar for decent TIM?

People wouldnt have an issue with it if they just used a better TIM. I would gladly pay 10 extra dollars to Intel inorder to not have to pay someone 50-75 to do it for me -_-

No matter how you spin it DAN Intel is creating a lot of waste for no fucking reason by cutting an unneeded corner


How much money does Intel pay you to spin that bullshit?Basic logic and economics calls bullshit.

 

[lolfail9001]

yes by pissing off customer base to save 25 vents per build instead of just charging an extra fucking dollar for decent TIM?

You are literally explaining why it is not economics that force usage of TIM. I mean, heck, every below 200mm^2 die Intel packages with IHS uses glue. And about every above uses solder. Also, it was explained thousand times already that it is not TIM at fault, TIM is irrelevant in fact, it is glue going rogue that causes temps.

 

[SomeGuy133]

You are literally explaining why it is not economics that force usage of TIM. I mean, heck, every below 200mm^2 die Intel packages with IHS uses glue. And about every above uses solder. Also, it was explained thousand times already that it is not TIM at fault, TIM is irrelevant in fact, it is glue going rogue that causes temps.

again i call bullshit on the glue because Silicon lottery reglues the shit with better TIMs and it works fine. Your glue argument is garbage.

and la-dy-da

http://www.overclock.net/t/1619534/is-it-the-tim-or-the-glue-that-is-the-issue#post_25739925

not the glue...Intel just doesn't use enough pressure and shitty TIM

 

[Shintai]

Oh the crusade continues

It can only be evil, no other reason can possibly exist and all facts presented is rejected! Hahaha

 

[lolfail9001]

not the glue...Intel just doesn't use enough pressure and shitty TIM

Obviously it is not the glue material, it is the gluing quality hence the gap, for pete's sake, stop being obtuse.

And yes, obviously Intel's TIM is not as good of conductor as LM, nothing usable as TIM is.

 

[SomeGuy133]

Obviously it is not the glue material, it is the gluing quality hence the gap, for pete's sake, stop being obtuse.

And yes, obviously Intel's TIM is not as good of conductor as LM, nothing usable as TIM is.

be clear....derp

back track all you want you said the glue was the issue which it isn't. Your wrong face it. Its the IHS application and TIM and the glue has nothing to do with it. derp derp

 

[lolfail9001]

be clear....derp

Reasonable enough.

back track all you want you said the glue was the issue which it isn't.

Oh, it is. It is not the glue material, and nobody ever claimed that. But glue application is just as freaking essential as any other property of glue, so you may as well consider it part of term "glue" in regards to the complex.

I'll give it that i was not clear enough, but my point was always constant.

 

[Nenu]

be clear....derp

back track all you want you said the glue was the issue which it isn't. Your wrong face it. Its the IHS application and TIM and the glue has nothing to do with it. derp derp

He is right.
The major issue is how high the lid is raised off the core, meaning more TIM is needed to fill the gap.
ie the lid is glued in place with a large gap filled with paste.