Asus ROG Zephyrus GX501 - Dissasembly and Repaste?

Discussion in 'Asus' started by pitha1337, Sep 5, 2017.

  1. pitha1337

    pitha1337 Notebook Geek

    Reputations:
    2
    Messages:
    79
    Likes Received:
    40
    Trophy Points:
    26
    Hi,

    I want to repaste my GX501 with Liquid Metal because I want to get the best out of this beatiful laptop.

    I have done that for all new Alienware models and I wonder how I can do it with my new ASUS.

    A few questions regarding the repaste:

    1. Is it possible to use liquid metal in the ASUS GX501?
    2. Is there any dissasembly guide especially for the heatsink?
    3. What would be good CPU temperatures in OCCT with and without liquid metal?
    Would be nice to get some help because I do not want to harm the system.
     
  2. ThatOldGuy

    ThatOldGuy Notebook Deity

    Reputations:
    642
    Messages:
    1,710
    Likes Received:
    1,191
    Trophy Points:
    181
    Repaste of the GX501 voids the warranty, as you have to remove the "warranty void if removed stickers" to remove the heatsink.

    That being said, It definitely could use some liquid metal. I would expect mid 70's to 80's with repaste... depending on a lot of factors
     
  3. pitha1337

    pitha1337 Notebook Geek

    Reputations:
    2
    Messages:
    79
    Likes Received:
    40
    Trophy Points:
    26
    Hm, voiding the warranty is a problem for me. Yesterday I ran two OCCT 5 minutes session. One run without undervolting and One run with -100mv undervolt.

    Seems to me that I already got a very good pasted unit. What do you think?
     

    Attached Files:

  4. ThatOldGuy

    ThatOldGuy Notebook Deity

    Reputations:
    642
    Messages:
    1,710
    Likes Received:
    1,191
    Trophy Points:
    181
    5 min is a very short run time, that data is suspect. OCCT isn't a best representation of Gaming/ normal use anyway. Do a Firestrike stress test (20 min).

    The 11C difference between core 0 and core 3 suggests there might be a seating issue. I don't know why that stabilized with the undervolt (but like I said before 5 min is not enough time for good data)

    With proper seating and good paste job, there could be even more improvement.

    Call/email Asus. Ask if repasting voids warranty. If you get it in writing (that it doesn't) from any company rep, that is legally binding and they will have to support any warranty issue after repaste.
     
    hmscott likes this.
  5. pitha1337

    pitha1337 Notebook Geek

    Reputations:
    2
    Messages:
    79
    Likes Received:
    40
    Trophy Points:
    26
    The core differential seems to be a "miss meassurement". In Fire Strike stress test I don't have this high differentials. Nevertheless ASUS replied that repasting will void the warranty.

    "Thank you for contacting the online ASUS tech support.

    I am sorry to hear you are having problems with your GX501VIK notebook.

    Unfortunately replacing the CPU paste would void the warranty. Could you please tell me how high the CPU temperature is and when it is at its highest (while playing or during normal use)?"
     
  6. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,631
    Messages:
    13,543
    Likes Received:
    15,684
    Trophy Points:
    931
    If you have 4x threads of equal compute utilization, like prime95 or other CPU benchmark tool, tune it for 4 or 8 threads with 100% CPU utilization - I use prime95 small FFT - don't run a simultaneous GPU load test that's too much for a thin laptop to handle, just stick with testing the CPU core temp differentials first.

    If you have a misaligned heatplate, not getting even pressure across the CPU, you can get Cores 0/1 with lower/higher temps than Cores 2/4 (or their equivalents in hyperthreaded cores), basically half the "cores" will run hotter than the other half.

    3c-5c core differential is "normal", from 6c-10c it's "acceptable" if undervolting keeps the hottest cores from entering "thermal throttling" range of 93c. If half the cores run hot enough to thermal throttling, then all the cores are throttled, so you don't want that to happen.

    Before testing all this, undervolt first. It's a waste of time to do this under the overvolted BIOS CPU voltage setting.

    Asus used to say re-pasting was ok, at least in the US, where are you located?
     
    Darkhan likes this.
  7. pitha1337

    pitha1337 Notebook Geek

    Reputations:
    2
    Messages:
    79
    Likes Received:
    40
    Trophy Points:
    26
    Hi,
    my location is Germany. Attached you can see my results with Prime95. What do you think? CPU is undervolted with -125mv.
     

    Attached Files:

  8. IKAS V

    IKAS V Notebook Prophet

    Reputations:
    985
    Messages:
    5,647
    Likes Received:
    298
    Trophy Points:
    251
    Temps look fine no need to repaste.
     
  9. pitha1337

    pitha1337 Notebook Geek

    Reputations:
    2
    Messages:
    79
    Likes Received:
    40
    Trophy Points:
    26
    hmscott likes this.
  10. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,631
    Messages:
    13,543
    Likes Received:
    15,684
    Trophy Points:
    931
    You only ran the hwinfo64 for 1 minute and 27 seconds, so that wasn't long enough to get good numbers, the short measurement time and the CPU temps are way too low to show anything useful.

    Try running Prim95 small FFT on 8 threads for at least 5-15 minutes, start the hwinfo64 monitoring before you start the Prime95 run.

    -125mV is good, but is that your first setting? Are you stable with that value at idle and for long gaming sessions? Did you try a larger undervolt? Try increasing -5mV at a time until you crash under 100% CPU load and then back off +10mV, if you haven't already :)
     
    ThatOldGuy likes this.
Loading...

Share This Page