• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

I7 8700K Running hot... or is it?

Caporegime
Joined
18 Oct 2002
Posts
39,314
Location
Ireland
I just don't get why Intel do this, I mean what does it take for them to apply decent paste?

They claim they use TIM because solder can crack after a while, its also much cheaper to use TIM. I think the last chips they used solder on was sandy bridge and broadwell, and i don't see masses of posts with people saying those chips have died.
 
Associate
Joined
23 Jun 2018
Posts
189
Location
Leicester
They claim they use TIM because solder can crack after a while, its also much cheaper to use TIM. I think the last chips they used solder on was sandy bridge and broadwell, and i don't see masses of posts with people saying those chips have died.
Yeah, I'll just go for delided chips in the future, would have been better for intel to use liquid metal and add a bit extra to the cost but then again OCUK do this, lesson learnt.
 
Associate
Joined
12 Jan 2017
Posts
406
Its pathetic, they just dollop this stuff on and hope for the best. You can see most of the core was barely touched at all, at best it was fractionally touching which obviously would do almost bugger all to transfer the heat to the ihs.

If intel are too cheap to use solder (despite their hurr durr bs of it cracking) then why not the liquid metal route? Or would that take too long for them to do properly?
no, the core was covered fine, problem was that the thickness and thermal conductivity level of the paste. You can see the paste of part where the core "barely touched" stick to the IHS instead, this is pretty common when you take them apart.
 
Caporegime
Joined
18 Oct 2002
Posts
39,314
Location
Ireland
no, the core was covered fine, problem was that the thickness and thermal conductivity level of the paste. You can see the paste of part where the core "barely touched" stick to the IHS instead, this is pretty common when you take them apart.

But you would still see a pretty decent imprint on the actual die, there's quite substantial parts of the die that are quite clearly totally clean of any type of thermal compound. If it was touching properly it would have made a mark at least. It's like they put this stuff on in small rectangular clumps and they're not all even so 1 end can be higher than the other which causes issues in regards to it touching the actual die.


That's my take on it anyway, if intel are so concerned with "cracks" in the solder (cough penny pinching) then surely liquid metal is a far better alternative (and presumably cheaper than the solder setup) than this crap they're using? Then again liquid metal just can't be globbed on as easily as TIM can so i doubt they would go that route.
 
Last edited:
Soldato
Joined
13 Jun 2009
Posts
6,847
They claim they use TIM because solder can crack after a while, its also much cheaper to use TIM. I think the last chips they used solder on was sandy bridge and broadwell, and i don't see masses of posts with people saying those chips have died.
Intel's main argument is that dies are smaller these days and that can cause issues with solder that didn't materialise with Sandy Bridge, etc. It's a fair enough argument (although AMD doesn't seem to agree since they continue to use solder), but it's still no excuse for terrible application of TIM.
 
Caporegime
Joined
18 Oct 2002
Posts
39,314
Location
Ireland
Intel's main argument is that dies are smaller these days and that can cause issues with solder that didn't materialise with Sandy Bridge, etc. It's a fair enough argument (although AMD doesn't seem to agree since they continue to use solder), but it's still no excuse for terrible application of TIM.

If that's intels stance then why do they insist on using TIM which seems to crank temps up by around 15c or so compared to liquid metal? Its pretty obvious the TIM isn't doing a good job . They can glob on the TIM easily while liquid metal requires careful application, just seems to be a quicker and cheaper method rather than them doing it right.
 
Soldato
Joined
28 Dec 2017
Posts
8,453
Location
Beds
I don't understand the people suggesting liquid metal in stock Intel CPUs. Their TIM situation sucks but liquid metal in millions of consumer products is a TERRIBLE idea. It's just too risky when it will go through postage, etc.

They need their chips to last a safe minimum of years and ideally a decade when taken care of, regardless of warranty. You wouldn't buy a product that literally was guaranteed to fail after two years and one day. Liquid metal simply isn't a proven ten-year solution. Crappy TIM is doing the job just well enough to let them keep selling the chips.
 
Caporegime
Joined
18 Oct 2002
Posts
39,314
Location
Ireland
I don't understand the people suggesting liquid metal in stock Intel CPUs. Their TIM situation sucks but liquid metal in millions of consumer products is a TERRIBLE idea. It's just too risky when it will go through postage, etc.

They need their chips to last a safe minimum of years and ideally a decade when taken care of, regardless of warranty. You wouldn't buy a product that literally was guaranteed to fail after two years and one day. Liquid metal simply isn't a proven ten-year solution. Crappy TIM is doing the job just well enough to let them keep selling the chips.


You honestly think that there's not a way they could engineer the ihs so they could keep the liquid metal centered on the core? If they're not gonna to do it properly with solder then liquid metal is an alternative that doesn't have all these OTT intel dramatisations of cracks possibly appearing. The thermal material they're using simply isn't up to the job. It would be one thing if it was only a few c difference, but we're talking 15-20c difference in some cases.

They obviously wouldn't even consider liquid metal as it requires far more careful application than the crap they currently use which seems to be slung at the core from several feet away.
 
Last edited:
Soldato
Joined
28 Dec 2017
Posts
8,453
Location
Beds
I just think that their method of throwing crappy TIM from several feet away is producing chips that rarely fail, they just run hot. But the vast majority of Intel customers aren't even checking temperatures, and certainly not relying on low ones being available. Even lots of people who buy K-series CPUs end up never overclocking them.

Basically it's working for Intel to do a pretty bad job but a consistent one. Maybe it's enabled by their complacency due to market leadership, and will change. I'm still getting used to the fact Intel's chips are considered the burn risks and AMD are the ones staying cool, after the entire FX era CPUs and GPUs.
 
Soldato
Joined
13 Jun 2009
Posts
6,847
Liquid metal isn't required on a consumer product but certainly they could put more effort into using better TIM and applying it better.
 
Soldato
Joined
22 Nov 2009
Posts
13,252
Location
Under the hot sun.
Probably down to the crap TIM they use, when i delidded my 6700k last year the application was pathetic.

scmKmdX.jpg

Barely any real coverage on most of the core and a thick coverage on a few small parts of it.

My 8600K had less paste, you should feel better :p
And wasn't there some videos last year showing that some have good paste amount others bordering to almost nothing?
 
Soldato
Joined
22 Nov 2009
Posts
13,252
Location
Under the hot sun.
Intel's main argument is that dies are smaller these days and that can cause issues with solder that didn't materialise with Sandy Bridge, etc. It's a fair enough argument (although AMD doesn't seem to agree since they continue to use solder), but it's still no excuse for terrible application of TIM.

Yet they are using it on Skylake-X, and especially the 7900+ are pretty huge silicons. The 7980XE takes almost the whole package. One of the biggest silicon blocks the consumer systems ever had.

Intel are just money pinching. 4820K was soldered, a small chip, was run cool and was overclocking pretty high. It's sibling the 4770K was boiling water even at stock speeds. Same chip sizes. Didn't see the 4820K cracking.. :D
 
Soldato
Joined
13 Jun 2009
Posts
6,847
I completely agree they are penny pinching. Even if their argument has some merit, their actions don't add up and they are clearly just trying to save money for some reason.
 
Caporegime
Joined
18 Oct 2002
Posts
39,314
Location
Ireland
My 8600K had less paste, you should feel better :p
And wasn't there some videos last year showing that some have good paste amount others bordering to almost nothing?

Same situation with gpu's in regards to paste but at least those are relatively easy to replace without the oem being any the wiser.
 
Soldato
Joined
19 Feb 2007
Posts
14,343
Location
ArcCorp
Probably down to the crap TIM they use, when i delidded my 6700k last year the application was pathetic.
Barely any real coverage on most of the core and a thick coverage on a few small parts of it.

It's not actually the TIM they use that is the reason for the temps, The TIM is designed for very long term use, The reason for the bad temps is the amount of silicon they use to attach the IHS which leaves a sizeable gap between the die and the IHS.
 
Permabanned
Joined
1 Jun 2004
Posts
2,019
Location
London
They should solder the K series chips and charge us for it. As for the rest of them, what's the point, they're not overclocked. It's us that run them over specification and therefore benefit from additional cooling. Look at the standard heatsinks they provide, it all works as it should.
 
Back
Top Bottom