6700K delid before and after results

Discussion in 'Overclocking & Cooling' started by Eldata, Oct 16, 2015.

  1. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
  2. n=1

    n=1 2[H]4U

    Messages:
    2,359
    Joined:
    Sep 2, 2014
    1.424V Vcore damn you're brave.

    And I lol'd @ 181W CPU package power, that's like Ivy-E level power draw. Congrats on the 4.8 OC though
     
  3. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    "1.424V Vcore damn you're brave."

    Silicon Lottery is selling binned 4.8GHz@1.424V 6700K's for $540.00 a pop and they're currently sold out (I paid $320.00 for mine from MC);

    http://siliconlottery.com/collections/lga-1151

    In terms of Vcore, this is Skylake, not Haswell, IB or SB.

    "And I lol'd @ 181W CPU package power, that's like Ivy-E level power draw."

    Well, I did indicate it was Prime95 28.7, one of the "hotter" stress testers around. Here is a four hour run of a much less aggressive stress tester. CPU package power max is only 117W;

    http://tinyurl.com/oqy7q3z
     
  4. n=1

    n=1 2[H]4U

    Messages:
    2,359
    Joined:
    Sep 2, 2014
    Well SB I'd have no issues running 1.45V for 24/7, and 1.4V for IB (I ran my old 4930K at 1.4V for an entire year without issue). Haswell's kind of an oddball since the FIVR just pump out too much heat to really push Vcore hard. I know Skylake isn't Haswell but 1.4V+ on 14nm? I just don't know.

    Btw the P95 comment wasn't meant to be a slight, I was genuinely impressed at how much power the quad core was sucking down.
     
  5. Dayaks

    Dayaks [H]ardness Supreme

    Messages:
    6,307
    Joined:
    Feb 22, 2012
    Apparently Asus and Intel suggest a max of 1.45V. He's pretty close to that. Personally I thought the voltage max would have dropped due to the die shrink but it seems to be hovering around that point for a few years now.

    Nice job with the delid. That's a pretty massive temp drop you got there.
     
  6. Nenu

    Nenu PM Kyle for the TItle You Pick.

    Messages:
    18,011
    Joined:
    Apr 28, 2007
    1.424V is fine if temps are in check, as long as 1.424V is the measured voltage, not what was set in the BIOS.
    Asus stated that for a normal chip, 1.43V is ok on air if not hard benching and 1.45V is ok for water cooling.
    Delidding has a similar effect to using water on a normal chip, those temps are great!

    Ah, Dayaks got there first with the Asus quote :)
     
  7. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
  8. new2019

    new2019 [H]ard|Gawd

    Messages:
    1,398
    Joined:
    Sep 29, 2012
    Did you try OCCT ? My OC was stable in Prime95 but not in OCCT.
    Would this bring same effect on e5 1620 @ 4.5GHz ?
     
  9. Nenu

    Nenu PM Kyle for the TItle You Pick.

    Messages:
    18,011
    Joined:
    Apr 28, 2007
    Op, can I ask about the delid.
    What was the procedure you used?
    Was there any problem adjusting the heatsink height lower?
    Did you take any pictures while you did it?

    I fancy having a go :)
     
  10. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    No, I didn't try OCCT. The die may be soldered to the IHS if E5-1620 is a Sandy Bridge Xeon CPU.
     
  11. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    "What was the procedure you used?"

    Razor blade to separate IHS from PCB and IHS resealed when done like the original.

    "Was there any problem adjusting the heatsink height lower?"

    Not certain whether you're referring to a bare die setup or you're referring to the IHS making contact with the PCB before making contact with the die. Whichever, bare die was not used and IHS made contact with die before making contact with PCB.

    "Did you take any pictures while you did it?"

    Yes, but there are so many videos and pictures already out there. Search You Tube, also, here's a link to a bare die setup (Ivy Bridge), just use like 1/20 of the amount of CLU shown and use some painter's tape to mask around the die when applying;

    http://forums.anandtech.com/showthread.php?t=2285595
    http://tinyurl.com/m258vjf
     
  12. Chas

    Chas [H]ardness Supreme

    Messages:
    6,883
    Joined:
    Oct 31, 2005
    There's usually a height differential between the original uncut IHS and the replaced IHS after you cut it off then re-place it with your thermal compound of choice. Not a HUGE one. But enough that some thermal solutions won't make good contact with the IHS plate afterwards.
     
  13. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    I've read accounts where the rim of the IHS made contact with the PCB and had to be sanded down but that wasn't the case here. Made certain of that. Wondering whether all traces of the old sealant were removed from the PCB and IHS in such cases. Almost clean as a whistle here;

    http://tinyurl.com/ock3hr2
     
  14. Suprfire

    Suprfire 2[H]4U

    Messages:
    2,092
    Joined:
    Sep 1, 2008
  15. Copyright

    Copyright [H]ardness Supreme

    Messages:
    5,938
    Joined:
    Nov 17, 2000
    Wow.. amazing results.. I got like a 16C drop when I did my 4770K but I also have my system built to run silent. This makes me want to delid my 4790k...
     
  16. datap1mp

    datap1mp Limp Gawd

    Messages:
    138
    Joined:
    Feb 23, 2006
    What TIM did you use?
     
  17. Millerboy3

    Millerboy3 Limp Gawd

    Messages:
    154
    Joined:
    Jun 5, 2015
    I'm also curious to the thermal paste you used
     
  18. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    TIMs used:
    1. CLU between die and IHS (IHS resealed to PCB)
    2. Noctua NT-H1 between IHS and waterblock

    BTW, here's an eight hour prime95 28.7 run @4.8GHz;

    http://tinyurl.com/qcqzjx4
     
  19. DaedalusHelios

    DaedalusHelios 2[H]4U

    Messages:
    2,182
    Joined:
    May 19, 2008
    Nice job. The delid temp drop is the impressive thing. That and the professional clean way you did it. Thank you for sharing with us. :)
     
  20. Eldata

    Eldata Limp Gawd

    Messages:
    367
    Joined:
    May 10, 2011
    Thanks! No problem.
     
  21. WarLust

    WarLust Limp Gawd

    Messages:
    329
    Joined:
    Oct 20, 2004
    No delid (thank god!), 4.8 on 1.35. Can get 5.0, but over 1.4.
     
  22. Millerboy3

    Millerboy3 Limp Gawd

    Messages:
    154
    Joined:
    Jun 5, 2015

    Thanks I did the same with my 2x g3258 and a 4690k with CLP
     
  23. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,477
    Joined:
    Apr 12, 2015
    Sucks intel could have just soldered it for a mere dollar and gave better results