XFX TIM replacement = voids warranty?

Discussion in 'AMD Flavor' started by blade52x, Apr 22, 2011.

  1. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    So I've got a small problem with my 6850 crossfire setup... and that's temperature on one card in particular. Below is a snapshot of the temperatures from both cards after running the Heaven DX11 benchmark for 10 minutes or so.

    [​IMG]

    As you can see, one card maxes at 89C, while the other at 99C. Now you may think is an airflow issue. Below is a picture of my setup.

    [​IMG]

    The card that is reaching 99C is not the bottom card, but in fact the one at the top. Yes, there is a sound card sandwiched in between the two cards, but with the way these coolers are designed, its making little difference whether the sound card is there or not. I have an exhaust helper fan (what you see in front of the cards) to help exhaust the hot air in the area, which actually does the job pretty well. In comparison, a single card would reach 75C. I could have maybe understood if the bottom card was the one hitting 99C... but it's perfectly fine all cramped down there.

    So, I was lead to believe the issue was with the thermal compound of the card. However, when I took the card out and was getting ready to remove the heatsink, I saw stickers over the screws that said "void if removed". What the hell? Has it always been like this for XFX? If that's the case, and I cannot do this without voiding warranty, I don't really care. I'll just leave it as is and wait for the card to break and send it in for a replacement because I don't see it lasting very long at that load temperature. But maybe it will just fine... who knows.
     
  2. wra18th

    wra18th [H]ardness Supreme

    Messages:
    7,760
    Joined:
    Nov 11, 2009
    In a setup like yours the top card will always be a little hotter than the bottom one. The bottom card is breathing cool bottom of the case air. The top one is sucking up the heated air from the bottom card. Try getting some cross ventilation across the cards. It will definitely be an improvement.
     
  3. Mr Mean

    Mr Mean [H]ard|Gawd

    Messages:
    1,357
    Joined:
    Oct 2, 2004
    If there is a printed warning on the card don't mess with it, unless you want to void your warranty. Mine is already void as soon as I opened my HIS 5970 to replace its TIM. Already replaced the HIS BIOS with a Sapphire OC BIOS. Playing it safe is the correct step. But watching those temps would bother me to no end :)
     
    Last edited: Apr 22, 2011
  4. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    That I'm trying to figure out right now. As you can see I'm a little low on space. :p
     
  5. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    Ugh... Metro 2033. I definitely need to figure something out here. It's not even May yet so there may be worse to come.

    [​IMG]
     
  6. Gulkor

    Gulkor [H]ard|Gawd

    Messages:
    1,241
    Joined:
    Jan 26, 2007
  7. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    Thanks for in the information. Very helpful. However, I managed to get temperatures within check by moving my ssd (basically ghetto mounted above the second hard drive cage directly across from the cpu cooler and then mounting a 120mm fan onto the hard drive cage (good thing I didn't lose those clips).

    So now there is cool air being blown over the cards directly from the front intake, and it seems to helping out. Now the top card maxes at around 92-95C while the lower card at around 84-86C. Still not great, but really there's not much else I can do at this point.
     
  8. XacTactX

    XacTactX [H]ardness Supreme

    Messages:
    4,103
    Joined:
    Dec 13, 2010
    Where are you located. I know for a fact that XFX allows for removing the heatsink in order to install aftermarket cooling, which by virtue of its nature includes re-applying TIM for those in America, and possibly for North America, though for NA I'm not sure. The 'Void if removed' stickers are there because their policy doesn't allow European customers to remove the heatsink, and they are not sourced for America or Europe, they're all from the same bin. It doesn't apply to America.

    http://xfxforce.com/en-us/Help/Support/WarrantyInformation.aspx?ispreview=true

    If you decide to RMA the card for any reason they have to receive it in the stock configuration that it was in when it left the factory. Otherwise, go crazy.
     
    Last edited: Apr 23, 2011
  9. NKD

    NKD [H]ardness Supreme

    Messages:
    7,391
    Joined:
    Aug 26, 2007
    if its an airflow issue, which it seems like in your case, I dont think removing the TIM would help a whole lot. I really dont think replacing the TIM really helps a whole lot. I have done it countless times and rarely saw any difference. May be 1 or 2c. In your case going from 99 to 97, which really wont matter.
     
  10. gjs278

    gjs278 Gawd

    Messages:
    711
    Joined:
    Feb 22, 2011
    considering xfx lets you replace your cooler altogether without voiding the warranty, I doubt tim would matter to them. I also doubt it would make a difference though. even if you got slightly cooler with the tim, your vrms will still be lot as all hell.
     
  11. Gulkor

    Gulkor [H]ard|Gawd

    Messages:
    1,241
    Joined:
    Jan 26, 2007
    The link i posted was from march 1st 2011 and is for United States and Canada.
     
  12. tikiman2012

    tikiman2012 [H]ard|Gawd

    Messages:
    1,228
    Joined:
    Dec 16, 2009
    I replaced the TIM on my XFX HD6950. I got online with one of there techs who verified for me that American customers can remove the stickers and cooler without voiding their warranty. I used Ceramique. Once it cured, I had lower idle temps and lower operating temps. The TIM that came on mine had a fold in it. I'm glad I replaced it. It might have reduced it's life cycle.
     
  13. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Yeah replacing stock TIM on Nvidia and ATI/AMD if done properly by a skilled person usually always yields a 3-7*C drop depening on the card, obviously the higher end cards yielding better drops. I can get IC Diamond samples to people for freewhich is the best GPU TIM IMO if you post your results.
     
  14. WhiteFireDragon

    WhiteFireDragon Limp Gawd

    Messages:
    495
    Joined:
    Jul 16, 2009
    so free IC Diamond if i provide a before and after screenshot of temps showing the improvements? i'm down for that :D. any details to this offer?
     
  15. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Just PM me your name, email, and address usually takes a week or two. Then please post your before and after results and all is good.
     
  16. Spare-Flair

    Spare-Flair [H]ardness Supreme

    Messages:
    7,471
    Joined:
    Apr 4, 2003
    When I got an 8800GTX from XFX and sent in a ticket to the RMA department because it was running hot, the tech TOLD ME to reapply TIM myself.
     
  17. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Thats common from all GPU Vendors isn't the first person who has told me that, I replace the TIM on all my GPUs day 1 it's totally worth it and to be frank they need it but I don't recommend this for the casual gamer who doesn't do this technical level of work on their PC if you damage the GPU no free replacement or warranty. However most all GPU vendors allow TIM replacement and it does not void your warranty if done properly.
     
  18. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    I don't think it's an airflow issue. Case open versus case closed isn't major in difference. The idle are nearly identical, and the load temperatures are slightly warmer with the case closed.

    [​IMG]

    It's good to know I will not void warranty if I decide to replace the TIM. Now I need to figure out how exactly to remove these 6850 heatsinks. :eek:
     
  19. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    If you want to try IC Diamond TIM which is the best for GPUs IMO I can get you a free sample if you PM me your name, address, and email, for the sample its requested you post before and after results.
     
  20. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    Definitely! Sending a PM your way now!
     
  21. DanNeely

    DanNeely 2[H]4U

    Messages:
    3,467
    Joined:
    Aug 26, 2005
    IC diamond will scratch up whatever it's applied to. Will XFX, etc have a waranty problem with that?
     
  22. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Thats absolutely not true, I have put ICD on all my many many GPUs even ones sent for RMA and never ever once a single scratch or warranty denial. You are spreading mis-information.
    Also since you know so much about it wheres your GPU you were denied warranty on due to using ICD. Thats what I thought you have never even tried ICD on your GPUs to even have spoken on this topic.
     
    Last edited: Apr 24, 2011
  23. DanNeely

    DanNeely 2[H]4U

    Messages:
    3,467
    Joined:
    Aug 26, 2005
    I've never used it because I've seen many poeple complain about the damage it does and post pictures of their scratched up hardware after trying it out.

    Diamond is the hardest material known to man. Anything it's rubbed against (eg your chip/heatsink when spreading it, will be scratched. Diamond dust is commonly used as an industrial abrasive. Trying to pretend it doesn't result in abrasion does nothing except to destroy your credibility.

    http://skinneelabs.com/indigo-xtreme/4/http://skinneelabs.com/indigo-xtreme/4/

    " It [IC Diamond]'s also really good at polishing–as you can see from my pictures, the IHS of my CPU no longer has Intel’s markings"

    Skinnee did 5 test runs and the markings on his CPU were completely abraded away. He may have done additional mounts beyond the 5 he logged and graphed; but if so he didn't indicate the fact. He's mentioned the abrasion elsewhere as well; citing it as a reason not to include the product in additional tests due to concerns about the impact of the scratches on the comparability of tests done afterwards with those done before.

    http://www.overclockerstech.com/ic-diamond-thermal-compound-review/3/

    overclockers tech observed the same abrasions, but estimated it would take dozens of applications to completely abrade the markings off the chip. They did not provide any pictures to go with their review, leaving us with no way to evaluate their estimate. The degree of damage Skinnee saw makes me think they're being overly optimistic..

    various forum threads with people complaining about the scratches.
    http://forums.anandtech.com/showthread.php?t=2073442
    http://www.overclock.net/air-cooling/998711-oh-my-freaking-god-ic-diamond.html
    http://www.evga.com/forums/tm.aspx?m=308916&mpage=1&print=true
    http://www.evga.com/forums/tm.aspx?&m=695748&mpage=1
    http://www.overclockers.com/forums/showthread.php?t=649643
     
  24. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    So your spreading rumors since you never used it like I said. And this thread is on GPUs not CPUs so you have nothing. You look silly. And it's your credibilty that is being destroyed not mine I speak from first hand experience.

    GPUs markings are sealed much better than CPUs. I have never had any scratching or loss of markings on any GPU period nor has anyone every reported it on GPUs period.

    On CPUs if you use Artic Clean, or Isospropyl Alcohol and let it set a little it comes off scratch free. If your not to smart and wipe it off dry you may and I reapeat may get micro scratching.

    To get anything worse you have to be rubbing/polishing with force and yes if you do that you can lap your CPU and/or heatsink with it. Many other TIMs will do this as well.

    Get educated and try it or stop posting mis-information you know nothing about first hand is all I am saying. PM me your name, address, and email and try a free sample on your GPU. Then take it off its free and see for yourself and report out here your findings.

    Go to this link on my YouTube channel and you'll see a guy compared MX-4 to ICD (CPU since you brought it up) and ICD beat it hands down under load killed it: http://www.youtube.com/watch?v=tSq_xEx6vrM

    One last major point I make public I have a relationship with the owner of ICD and give out free TIM for posted results to anyone. You not having any skin in the game, never using ICD, a question comes to my mind why you are even posting on this thread? Stealth marketing (bashing other products too not just marketing for another) is illegal and can be fined. I am just saying no acusations just education.
     
    Last edited: Apr 24, 2011
  25. kniah

    kniah 2[H]4U

    Messages:
    2,323
    Joined:
    Oct 19, 2004
    OP- have you tried swapping the position of the two cards, i.e. move the top card to the bottom and vice versa? This can help you determine whether it's an airflow problem or whether it has something to do with your top card.
     
  26. nissanztt90

    nissanztt90 2[H]4U

    Messages:
    3,858
    Joined:
    Apr 13, 2007
    Diamond has the highest thermal conductivity known to man. Thats why i use it. Cry about it being abrassive.
     
  27. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Here is a FAQ on the scratching issue by the owner of Innovation Cooling:

    Some background on the subject - 100 years of standard Industry practice on abrasives is well documented and well understood by those in the business.
    Most thermal compounds are liquid sandpaper or lapping compounds by definition.

    Light gray compounds are usually/mostly made up of aluminum oxide. (Check MSDS for AS5,Ceramique, ShinEtsu, Dow etc. hard metal oxides are more common than you think)

    Aluminum oxide is what they make sandpaper out of.....MOHS Hardness scale 9

    Diamond on the MOHS hardness scale: 10

    Copper MOHS: 3

    To be abrasive you just have to be harder than the material to be cut.

    Most thermal compounds have particle sizes in the range of 600 - 800 same as your 600 - 800 sandpaper

    http://www.unc.edu/~rowlett/units/scales/grit.html

    If you are ambitious, you can calculate cutting speeds and material removal between diamond -- the hardest, (cubic boron nitride is actually harder than diamond) and the second hardest -- aluminum oxide, you will find an insignificant difference.

    For the insignificant distance traveled, equivalent pressures, You will not see any difference between the two under normal use.

    Abrasives have to move to work.


    For all compounds- if you are a habitual abuser who re-installs his sink 20 times a week or constantly rotates the sink under pressure then switch to something less abrasive like a zinc oxide compound .

    I would like to note tha IC Diamond is classified/called a Flour in the business. The snatches portrayed in the picture below are probably 50 times the size of the Diamond particles in IC Diamond, below is an image of a IHS lapped with ICD - See any comparable scratches?

    http://i291.photobucket.com/albums/ll300/ICD7/ICD7finalpolish.jpg

    below an image posted on overclock net claiming ICD damage

    http://img192.imageshack.us/img192/1655/1025315.jpg



    In three years we have not had one person email IC with the complaint. I have done 20 giveaways with public forum tests which are posted on our website and given away thousands of tubes and not had any complaints. I offer you the option of reading through over 700 hundred of end user tests (the latest at OCUK) to discover where this is a systemic problem as you claim.

    I have done this test and you can do it yourself - Take a piece of oxidized copper and some ICD, AS5, MX or Shin Etsu and a piece of cloth and polish them separately for 20 seconds then compare them they all polish more or less equally and under a microscope I see no difference in polish finishes.
     
  28. Makeroflostsouls

    Makeroflostsouls 2[H]4U

    Messages:
    2,520
    Joined:
    Jun 4, 2008
    I just put a water block on my 6970 yesterday. Been using the card for a few months now. The TIM that came on the card so do die I had to use my finger nail to scratch it off wasn't to happy about that. It was like dried sand. I can't say how much my temps droped as its on water and doesn't get over 36c now at full load @1005 clock.
     
  29. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Try Artic Clean next time it works magic on crusted TIM removal.
     
  30. Makeroflostsouls

    Makeroflostsouls 2[H]4U

    Messages:
    2,520
    Joined:
    Jun 4, 2008
    Yea the Isopropyl alchohol just wasn't cutting it.
     
  31. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    Yeah I don't use Isopropyl anymore for TIM it doesn't cut into the dense TIMs at all it just sits on top,
     
  32. tikiman2012

    tikiman2012 [H]ard|Gawd

    Messages:
    1,228
    Joined:
    Dec 16, 2009
    I use 3M wax and grease remover or PPG adhesive remover from the local body shop supply store. Seams to work the best.
     
    Last edited: Apr 25, 2011
  33. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    I'll wait for the ICD sample before doing this. I'm not exactly sure how scratches could ever be an issue, as the surfaces between the chip and heatsink aren't exactly rubbing against each other non-stop

    No but I have tested them separately, and the warmer of the two was usually about 3-5C warmer under load (separation increases a little bit in crossfire).
     
  34. tangoseal

    tangoseal [H]ardness Supreme

    Messages:
    6,999
    Joined:
    Dec 18, 2010
    I like how you mounted that suction fan next to your two GPUs. Did you have to cut your case?
     
  35. Gorilla

    Gorilla [H]ardness Supreme

    Messages:
    5,902
    Joined:
    Sep 14, 2007
    I would think that he must have had to cut it. I'm pretty sure that the p182 just has a couple of water cooling grommets there.
     
  36. tikiman2012

    tikiman2012 [H]ard|Gawd

    Messages:
    1,228
    Joined:
    Dec 16, 2009
    So here are my results with Ceramique. The stock results were 10c higher on idle and about the same on load because I never let it get higher. I shut down and replaced the TIM. These results are with the stock fan profile. 10 minutes idle and 10 minutes in Crysis. Max load was 89c with stock fan profile and Ceramique.

    http://img824.imageshack.us/img824/53/gpuz1.gif

    Idle
    http://img88.imageshack.us/img88/3273/gpuzidle.gif

    Load
    http://img703.imageshack.us/img703/330/gpuzload.gif


    This is also with unlocked shaders and overclock on the core only.
     
  37. Falcon_CMH

    Falcon_CMH Gawd

    Messages:
    524
    Joined:
    Feb 28, 2009
    This is a perfect example of why I always replace my TIM on GPUs. I buy high end GPUs and I see 10*C- results or better every time. But its not for evryone once again I repeat casual users that haven't done this before be wary you can damage your GPU and lose RMA/warranty. But those that know how or are ready to learn/take the risk done properly yields great results. Factory TIM quality is usually bad and the application even worse.
     
  38. tikiman2012

    tikiman2012 [H]ard|Gawd

    Messages:
    1,228
    Joined:
    Dec 16, 2009
    And by worse, in my case, the TIM was folded over itself where it made contact with the GPU. The cooler itself had oil marks from a memory thermal pad on the GPU contact surface. There was actually an oil image of the memory thermal pad on it! If I hadn't done anything with it, it would have been an RMA card very shortly I think.

    I replace the TIM on every GPU that I put in a system. And the results always speak for themselves. Never had to RMA a single card. My brother in law never wanted to mess with it. He has RMA'd one of his cards 3 times now.
     
  39. blade52x

    blade52x 2[H]4U

    Messages:
    2,937
    Joined:
    Sep 9, 2006
    Yeah. The water cooling holes (which I will never use) made it easy to cut. But it looks a bit ghetto with duct tape covering the openings around it. But it works. ;)
     
  40. Cowboy420

    Cowboy420 Gawd

    Messages:
    1,014
    Joined:
    Jul 19, 2005
    Opinions vary. I've NEVER replaced the TIM on any card and have never had an RMA.