The Skylake i5-6600K/i7-6700K delid thread

lehtv

Elite Member
Dec 8, 2010
11,897
74
91
As with Haswell and Ivy Bridge CPUs, the Skylake IHS has poor quality thermal paste under it, and replacing it with higher quality paste can have a drastic effect on temperatures. However, given that the Skylake PCB is much thinner than before, there seems to be a little more risk associated with delidding. The vice+hammer method in particular is strongly advised against, while the razor method should be a little safer than before given the smaller die and fewer critical components on the PCB.

I figured it'd be a good idea to collect delid results in order to have more info available on different methods and what to expect as a result. Also, for fun! :p

Answer template:

Processor _______________ i5-6600K or i7-6700K
Cooler __________________ e.g. 212 EVO
Delid method ____________ e.g. razor, vice, vice+hot air gun (to soften the glue)
Thermal paste under IHS __ e.g. Coollaboratory Liquid Pro, Liquid Ultra, NT-H1 etc
Thermal paste on top _____ e.g. Coollaboratory Liquid Pro, Liquid Ultra, NT-H1 etc
°C before and after _______ post all core temperatures, along with the GHz and VCore used!
Other comments _________ e.g.. "easier than I expected, here's why...", "definitely worth it, couldn't break 4.6ghz before" and so on

Please use Prime95 27.7 SmallFFT test for 10 minutes and record temperatures before and after delid (I recommend HWinfo64 for monitoring). Let us know if you used a different stress test.

All the better if you have photos to share! :biggrin:
 
Last edited:

lehtv

Elite Member
Dec 8, 2010
11,897
74
91
Let me begin...

Processor _______________ i5-6600K
Cooler __________________ Scythe Ninja 4
Delid method ____________ vice+hot air gun (to soften the glue); see photo shoot: http://imgur.com/a/OxqEG
Thermal paste under IHS __ Coollaboratory Liquid Pro
Thermal paste on top _____ Arctic MX-4
°C before and after _______ 71/69/66/69°C before, 59/58/57/58°C after, 4.5GHz @ 1.344V
Other comments _________ It was tougher than delidding 3770K. The glue seemed tougher than before, possibly because the 3770K had been in use for years. I had to be careful applying the CLP on to the tiny die, actually applied way too much of it at first. I didn't actually have to delid but did it anyway because... well, no good reason, really. Fun?

See also:



It seems the decrease in temperatures grows linearly with heat output, if measured in degrees centigrade. The percentual benefit of delidding, however, seems to stay roughly the same. There might be a logarithmic effect where lower heat output shows less % benefit.
 
Last edited:

readers

Member
Oct 29, 2013
93
0
0
Nice, but...


Coollaboratory Liquid Pro, Liquid Ultra...+ delid....


And you use 212 Evo?
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
Nice job!

Not much difference at stock. But with a solid OC the difference gets quite large.
 

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,250
136
But.. did you improve your OC headroom?

Probably not. Doesn't look like he was thermally limited to begin with according to his charts.

Nice results non the less.

I just can't get passed the warranty is void part of the delid process.
 

lehtv

Elite Member
Dec 8, 2010
11,897
74
91
But.. did you improve your OC headroom?

No, because I wasn't thermally limited before I delidded. 4.7GHz requires over 1.45V on my chip which I'm not comfortable with.

If my chip ran hotter (due to being a 6700K or having a weaker cooler) then I might have hit thermal limits before I hit voltage limits, and delidding would've helped me get another 100 or 200 MHz out of the chip.

I just can't get passed the warranty is void part of the delid process.

Meh... it's not anything will happen to the processor if you don't delid it.

OIC, Ninja 4 is a great cooler.

Yup :)

I don't see anything wrong with cooling a delidded chip with a 212 EVO though. In fact, there's more benefit to be had compared to using a higher end cooler
 
Last edited: