Mac Pro 2009 upgrade running hot

Discussion in 'Mac Pro' started by JD575, Jan 25, 2015.

  1. JD575 macrumors member

    Joined:
    Nov 12, 2014
    #1
    Hello,

    I have an issue with my upgrade to a x5677. I had it de-lidded and I have attempted to install several times, but it is running hot: CPU A 150 Heat sink 98, CPU B 163 Heat sink 87. It seems to me like they are not making proper contact.

    I used the spread method and the bead method. Nothing seems to work. I installed each one three times so far and all with the same readings.

    I tighten by hand which maybe preventing proper contact?
     
  2. h9826790 macrumors 604

    h9826790

    Joined:
    Apr 3, 2014
    Location:
    Hong Kong
    #2
    First of all, I assume you talking about Fahrenheit. If it's idle, yes, the CPU is too hot and the heat sink is too cold.

    If that's under load, the CPU temperature is normal and your ambient temperature must very low. However, IMO, the temperature difference between the CPU and the heat sink is too much, that means the heat is not effectively transfer from the CPU to the heat sink.
     
  3. JD575 thread starter macrumors member

    Joined:
    Nov 12, 2014
    #3
    I agree. What could be some of the causes? I would like to figure out a fix.
     
  4. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #4
    Hello JD575 ,

    I assume you have a DP 2009 ? And your upgrade processors had the integrated heat spreaders removed ? It's very difficult to remove the soldered spreaders without damaging the chips , so that is the likely reason why you're having trouble .

    Any red lights on the CPU tray ?

    Are your thermals in C or F ? And are your temps at idle or load ? Need some info , here .

    I've done many of these upgrades and the first ones I did were done just like you – completely manually and by “eye” or “feel” . Tricky . But, I'll walk you through as best I can long distance .

    Most important ! Your Mac was never designed to run with a pair of these Big Boys . They consume 30 percent more juice than the factory likes and you definitely need to bump up your case fans to 1600-1700 RPMs to deal with the accompanying heat increase . Get yourself a copy of SMCfancontrol now for the long term health of your Mac .

    Intel likes these chips to run no hotter than 78 to 80 C .

    Just to be on the safe side, what's the sSpec ? Did you purchase engineering samples by mistake ?
     
  5. JD575 thread starter macrumors member

    Joined:
    Nov 12, 2014
    #5
    Sorry I forgot to post that it is in Fahrenheit at idle.

    Machines I will check out the red light. Yes, its a 2009 DP. I had the de-lidding done by third party who is very experience. In fact I am working with him on this issue which may require me returning the CPU along with the tray. I thought I would pick the MacRumors brains to see if anyone had a same situation.

    Appreciate the advice, I am aware that these chips use more juice but others have been successful in getting them work at temperatures around 110-120F.

    I use iStat and all other temps are normal except the CPUs. I have the Boost fans running at 1600 and 850 for Intake and Exhaust.
     
  6. h9826790 macrumors 604

    h9826790

    Joined:
    Apr 3, 2014
    Location:
    Hong Kong
    #6
    Improper thermal paste application is the biggest suspect here.

    Or may be heat sink too loose.
     
  7. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #7
    I just fired up a re-certified 12 Core @ 3.46 GHz X5690 Mac Pro 4,1 > 5,1 I have in house – a very similar Mac to yours . Her idle temps after 5 minutes with factory cooling ( I disabled SMCFancontrol to simulate your situation ) are as follow : CPU Cores between 70 -102 F and the CPU Heatsinks at 84 – 95 F . They basically match each other . My temps are cool and well below the danger zone . That zone begins at 173 F . Again, these are idle temps .

    Your idle temps are not dangerous yet, but are extremely elevated . I would not even think to put your Mac at load until you lower those idle thermals . There is also a serious mismatch between your CPU and Heatsink temps . Your booster fan rotations are also accelerated as mine are at 1100 RPM . Your 850 RPM for air intake and exhaust are normal .

    I'm surprised you haven't already gotten an overtemp warning light on your Tray . You're getting close to the point where your Mac's safeties kick in and prevent a boot . It's a scary situation but not necessarily dangerous when she red lights .

    Likely causes for your situation : I concur with our Hong Kong friend : your Heatsinks are either not installed properly (not lowered enough or are uneven) . Or you are using excessive amounts of thermal paste and the paste is acting like an insulator and has lost it's thermal conductive properties .

    Let me know the answers to my other questions and we'll proceed further tomorrow . I would like some pictures . Do you have a camera with a macro function ? I am very interested in close up pictures of your CPU socket standoffs with the heatsinks as they are currently installed . I'm looking for something in particular .
     
  8. Macsonic macrumors 65816

    Macsonic

    Joined:
    Sep 6, 2009
    Location:
    Earth
    #8
    Hi JD575. I also think the comments of h9826790 may also be correct. About the thermal paste, if it is too abundant, it may "block" the heat transfer to the heatsink.
     
  9. box185, Jan 25, 2015
    Last edited: Jan 27, 2015

    box185 macrumors member

    box185

    Joined:
    Sep 12, 2013
    #9
    IHS Removal and JD575's Thermal Issue

    I am the third party that de-lidded JD575's processors. I am working with him to resolve this issue. I appreciate that the MacRumors community is offering him help as well - a great resource.

    The de-lidding process that I use is very safe. The CPU's were never exposed to the extreme temperatures that Intel used when assembling the processors. The only difficulties my customers have had has been related to not tightening the heatsink bolts enough - causing boot failure. JD575's machine does boot.

    That being said, I agree with many here who rightly understand that the issue is the junction between the processor and the heatsink - there is seemingly no other way to explain how the processors could be at 150-160 degrees F while the two heatsinks remain at 100 degrees F or so.

    One thought - is it even possible that the two heatsinks could be installed in the wrong position? I know they are different, but I don't know how different. Do they have different bolt patterns?
     
  10. box185 macrumors member

    box185

    Joined:
    Sep 12, 2013
    #10
    I agree - there appears to be a thermal barrier between each processor and it's respective heatsink. Regarding CPU socket standoffs, JD575 should not be using any additional standoffs as these are de-lidded CPU's. If standoffs are being used, that would explain the thermal barrier.
     
  11. Machines, Jan 26, 2015
    Last edited: Jan 26, 2015

    Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #11
    I don't know if the CPU A and CPU B heatsink fasteners are physically different enough to act like a “key” to prevent improper installation as I was always careful never to mix them up . They are different and are definitely not interchangeable . The TIM (Thermal Interface material) pads on the lower part of the heatsinks are of two different lengths . We should verify your client used the right heatsink positioning with some photos . It could explain this situation .
     
  12. Macshroomer macrumors 65816

    Macshroomer

    Joined:
    Dec 6, 2009
    #12
    A little off topic:

    I am running an upgrade hex ( 3.46ghz ) and have what I consider to be great CPU temps (F), 80-85 idle, 100-110 under load, SMC fan control used often.

    How can I tell the heat sink temp? Using iStat.
     
  13. JD575 thread starter macrumors member

    Joined:
    Nov 12, 2014
    #13
    CPU A Pics

    Hello Everyone,

    At the request of Machines, I attached a few photos of my CPU A side. I will do a subsequent post with CPU B.

    To answer a few questions: No red light has come and it boots fine. I use a thin layer of thermal paste (see pictures). At least I believe I do. Let me know if you see different.

    I use the crisscross method to tighten down the sinks.

    Thanks, JD
     

    Attached Files:

  14. JD575 thread starter macrumors member

    Joined:
    Nov 12, 2014
    #14
    Cpu b

    Here is my CPU B side.

    JD
     

    Attached Files:

  15. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #15
    I think I see the problem .

    You are using the proper heatsink for CPU A .

    But , those metal rotational cylinders on the standoffs must rest flush with the CPU Tray PCB . Two of them are at a funny angle (the ones closest to the heatsink's fan black plastic power / sensor connector . The other two look proper . No good . All of them must rest flush with the PCB or your heatsink will be unevenly installed . Knock the two until they rest proper . Reinstall that heatsink . Paste looks good . Examine the same standoff cylinders with CPU B and fix them , too , if needed .
     
  16. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #16

    Buy a copy of Bresink's Hardware Monitor .
     
  17. Macshroomer macrumors 65816

    Macshroomer

    Joined:
    Dec 6, 2009
    #17
    Thx!
     
  18. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #18
    I did not notice any physical issues with CPU B .

    Reinstall gently CPU B's heatsink .

    Once you have addressed the standoff issue with CPU A , gently reinstall that processor's heatsink .

    CPU A is the master socket . It must work properly or CPU B may experience issues as well , especially with fan rotationals .

    Boot her up and see if the thermals are improved .
     
  19. box185 macrumors member

    box185

    Joined:
    Sep 12, 2013
    #19
    Agreed - install CPU A and run your machine without CPU B installed. The fan will run fast, but you will be able to measure the temperature of both the CPU and the heatsink - maybe confirming that the orientation of those cylinders on CPU A were causing a problem.
     
  20. flowrider macrumors 601

    flowrider

    Joined:
    Nov 23, 2012
    #20
    Just a quick question for the OP are you looking at the temps for your CPUs at the Temperature Diode or Core 0 Relative to ProcHot?

    Lou
     
  21. JD575 thread starter macrumors member

    Joined:
    Nov 12, 2014
    #21
    Thanks everyone. I will reinstall CPU A only and let you know the results.

    Lou, its the Core 0 Relative to ProcHot. From your other posts, I know you installed these particular CPUs. Were they "matched?"

    JD
     
  22. flowrider macrumors 601

    flowrider

    Joined:
    Nov 23, 2012
    #22
    ^^^^You're looking at the wrong Temp measurement. Look at the Temperature Diode or TDiode measurement. I don't fully understand the meaning of the Relative measurement, but it is a measurement of some sort of difference and should be high. What are your TDiode and Heatsink Temps temps? Those are the important ones.

    And No my CPUs were not "matched" but they had the same markings, so the same CPUs.

    Lou
     
  23. Machines macrumors 6502

    Machines

    Joined:
    Jan 23, 2015
    Location:
    Fox River Valley , Illinois
    #23
    IF the OP runs his Mac with just CPU A socket installed with a processor and no processor in CPU B socket , his booster fan rotations will go to max RPM . What type of valid thermals will he get when his processor is artificially cooled ? Also , his CPU B socket is now unprotected and I noticed a lot of dust on his Tray . It's not likely he is in possession of a socket connector cap . He could contaminate socket CPU B with particulate matter .
     
  24. box185 macrumors member

    box185

    Joined:
    Sep 12, 2013
    #24
    There is big difference between the temperature of his CPU and the heatsink. So even if the booster fan does go to the maximum speed, he will only have an artificially cooled heatsink. The question about whether the thermal barrier still exists after adjusting the misaligned standoffs could be answered.
     
  25. flowrider macrumors 601

    flowrider

    Joined:
    Nov 23, 2012
    #25
    There is not - He is not measuring the temp at the TDiode!

    Lou
     

Share This Page