1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

8600K lapped and delidded; pics + results

Discussion in 'CPUs' started by 3t3P, 20 Apr 2020.

  1. 3t3P

    Mobster

    Joined: 20 Dec 2006

    Posts: 3,471

    Running folding at home (FAH) with both CPU and GPU maxed out was resulting in CPU temp peaks of 94-96 degrees celsius. No throttling but close to TJMax (100 degrees AFIK)

    Per my other thread doesn't seem worth it to upgrade on 1151 for gaming.

    So decided to lap and separately delid the processor, recording results as I went.

    One variable to account for is I changed the intake fans to NB E-loop RGB's.

    Testing was done by running Aida64 for 30 mins then recording temps.

    Then FAH when GPU and CPU workunits being crunched, again for ~30mins where possible.

    The patient is prepared and gowned in painters masking tape, initial scratches to show it's an 8600k if I ever sell it! The text quickly comes off.
    [​IMG]
    I had wanted to further record the text but it was already gone!
    [​IMG]
    [​IMG]
    This is shiny enough for me.
    [​IMG]

    [​IMG]

    Rockit delid tool bought off spockwiththeears on the MM:) He posted it out really quickly and gave me some insight into the delid process, so thanks.
    [​IMG]
    Into the tool.
    [​IMG]
    Digging the copper and black.
    [​IMG]

    Didn't make a gunshot sound just gave way with a bit less force than I thought.
    [​IMG]

    The face of God:
    [​IMG]

    Just before cleaning off all the Intel TIM I took a file and gently scratched out where the TIM rectangle was, so I would better be able to know where to put the liquid metal on the IHS side:
    [​IMG]

    Scraping away the silicone was easier than expected, use something plastic with a sharp edge (I used an uneeded three pin connector!). Then used the wire wool 'brillo' (included in the Coollaboratory kit) to scour the inside of the IHS for better liquid metal purchase:
    [​IMG]

    Coollaboratory ultra went on easily just take your time and it's true a little goes a long way. Probably used a 1.5mm (cubic) amount in total as had to top up slightly even with a small amount remaining in the supplied brushes:
    [​IMG]

    Used my nail varnish (ok the Gf's)...ok it was mine... to insulate the four contacts near the die as a precaution:
    [​IMG]

    Dries quickly. This pic was with it in liquid form still however:
    [​IMG]

    Paint the die in LM:
    [​IMG]

    Used this £2.75 tube of black high temp silicone glue. Now don't go crazy here, from research I did most failed re-lids are from too much silicone sealant or glue applied causing a gap between the Die and the IHS. As you can see I used the faintest amount only on the sides, enough to keep the IHS in place.

    [​IMG]
    [​IMG]
    Back in the tool for relidding, doing up the screws finger tight, bit of tape for protection. NB make sure you put the IHS back in the right orientation! Bottom right on the IHS is a small circular indent that aligns with a small gold dot on the bottom right of the PCB. This is less of a problem when you have text to orientate you!:
    [​IMG]

    Resists gravity, check!:
    [​IMG]

    My 'hybrid' paste application I invented. Cryonaut by thermalgrizzly (he so grizzly):
    [​IMG]

    Gratuitous system shot. Have put in some phanteks cables, RGB NB 140mm E-loop fans, RGB LED strips behind the motherboard and increased the intensity of the CPU white light.
    Please excuse the fan cable at the bottom, I have put in the hardrive cage fan again but didn't plug it in as needed to keep things as even as possible for testing.


    [​IMG]

    So did it work?

    YESSSSSS!

    Lapping on it's own did almost nothing, maybe I dropped 1-2 degrees but margin of error stuff.

    Delidding though, take a look at my spreadsheet, lost over 20 degrees at peak temps.

    The fans don't have to ramp up, the PC when folding was at full tilt and moving air and making not insigifnicant noise. Now max temps are in early seventies the fans cruise at 850-900rpm and it all stays cool and quiet.

    Really fun project, took me all day (not just the delidding but making a fan spacer, it's a long tediuous story involving mistakes..)

    [​IMG]
     
    Last edited: 20 Apr 2020
  2. FreeStream

    Capodecina

    Joined: 10 Apr 2004

    Posts: 13,314

    Standard result of de-lidding, mine dropped ~20C too.

    Although when I leave the machine folding I run my 8600K at 4ghz cores / 3.6ghz cache @ 930mv, doing ~80% of the work for ~40% the power (~55W).

    Normally have it at something like 4.9/4.6/1.35v.
     
  3. pugheaven

    Mobster

    Joined: 22 Apr 2008

    Posts: 2,819

    Location: Bryn Celyn Wales

    WOW that['s fasntastic mate, reall;y good read thanks for posting and very very sexy system BTW ;-)
     
  4. Matt26LFC

    Gangster

    Joined: 21 Feb 2008

    Posts: 213

    Location: Kent

    Nice work there!
     
  5. Rannoch

    Soldato

    Joined: 15 Apr 2012

    Posts: 5,457

    Location: Rannoch / Edinburgh

    That’s great. Must get a cpu like that in the future. Nice
     
  6. reflux

    Capodecina

    Joined: 15 Nov 2007

    Posts: 12,808

    Location: Enfield

    Looks good, glad it was worth all the effort m8! :D
     
  7. 3t3P

    Mobster

    Joined: 20 Dec 2006

    Posts: 3,471

    It was fun the PC ticks along so quietly now it was worth it for that as don't anticipate I'll get any higher clocks as a result of the headroom.

    Waiting for a wireless power draw socket next to see what effect voltages really have on total system power.

    Have a feeling I won't beat the +900W I got from a pair of volt modded 680s in SLI and a Q6600 on a suicide run overclock!

    Maybe 450-500W at the most? 2080S being main culprit:D
     
  8. Steedie

    Man of Honour

    Joined: 29 Jun 2004

    Posts: 20,279

    Location: Oxfordshire

    The lack of pics of the polished and spotless IHS made me cry :(

    Great job though
     
  9. 3t3P

    Mobster

    Joined: 20 Dec 2006

    Posts: 3,471


    Huh? What do you mean.
     
  10. tommoT.Striker

    Gangster

    Joined: 1 Jun 2019

    Posts: 185

    cool, glad it worked out for you.
     
  11. 3t3P

    Mobster

    Joined: 20 Dec 2006

    Posts: 3,471

    495-506 Watts from the wall when folding for what it's worth.
     
  12. EsaT

    Sgarrista

    Joined: 6 Jun 2008

    Posts: 8,617

    Location: Finland

    Looks like that heatspreader was pretty flat to start with.
    Some lapping pics show lots of wear on edges while center is only slightly scratched.



    Lack of mirror polish isn't any problem for cooling.
    In fact it might be bad unless going for flat enough surfaces to stick into each others level not needing anything in between:
    https://www.overclockers.com/why-heatsink-polishing-might-be-a-bad-idea/
     
  13. 3t3P

    Mobster

    Joined: 20 Dec 2006

    Posts: 3,471

    I held a steel rule edge to the IHS before lapping and it was surprisingly flat. Only in some some tiny places could I see between the rule and the IHS.

    As above polished is not desirable can be worse, hence me saying 'shiny enough', there was more to that statement than meets the eye.
     
    Last edited: 22 Apr 2020