Why don't OEMs use better thermal paste?

Discussion in 'Hardware Components and Aftermarket Upgrades' started by Brad331, Jan 30, 2019.

  1. Brad331

    Brad331 Notebook Enthusiast

    Reputations:
    27
    Messages:
    44
    Likes Received:
    32
    Trophy Points:
    26
    Suppose I were a laptop manufacturer, I could just make the exact same laptop as some other laptop (same specs, heatsink capability etc.), and just switch to using better thermal paste, and I would instantly get up to about 10°C lower temperatures with very little increased cost per unit and zero R&D cost.

    I don't see how a "they don't care" explanation is valid. OEMs spend time and money engineering heatsinks, fans, etc, sometimes even use fancy stuff like vapor chambers and pompous gimmicks like GORE. So while they're at it, why don't they also use better thermal paste for an easy performance gain, better user experience, and a competitive edge? It seems like a no-brainer to me.
     
    Last edited: Jan 30, 2019
  2. Khenglish

    Khenglish Notebook Deity

    Reputations:
    592
    Messages:
    837
    Likes Received:
    678
    Trophy Points:
    106
    A lot of it is paste longevity. Most aftermarket thermal pastes start drying out badly after a few months. Sure the aftermarket paste performs a few degrees better at first, but a year later the OEM goop is a few degrees better.

    With that said, some aftermarket pastes do last a long time. I'm pretty sure GC extreme does while being at the top tier of nonmetal thermal pastes.
     
    Vasudev, Ashtrix, Papusan and 3 others like this.
  3. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,020
    Messages:
    11,515
    Likes Received:
    9,070
    Trophy Points:
    931
    this. also, consider cost. we are not talking about a syringe or two of thermal paste, but basically bucket loads. would be quite the significant cost increase.
     
  4. Richard Zheng

    Richard Zheng Notebook Evangelist

    Reputations:
    31
    Messages:
    335
    Likes Received:
    154
    Trophy Points:
    56
    Because it costs a LOT more for them to use, but would be of little gain
     
  5. saturnotaku

    saturnotaku Notebook Nobel Laureate

    Reputations:
    3,997
    Messages:
    7,375
    Likes Received:
    2,901
    Trophy Points:
    331
    Yeah, that's not how that works.
     
    Aroc likes this.
  6. Brad331

    Brad331 Notebook Enthusiast

    Reputations:
    27
    Messages:
    44
    Likes Received:
    32
    Trophy Points:
    26
    I'm not sure I get you. 10 grams of GC-Extreme is 35USD on Amazon. That's good for more than 40 applications, so less than a dollar per application. OEMs can order in bulk and get better prices too. That doesn't look like much on, let's say, a $2000 computer. Using better thermal paste would allow them to make the laptops faster, thinner, cooler, or quieter — whichever they're optimizing for, allowing them to be more competitive. Assuming $1 for thermal paste on $2000 computers, they only need to sell 0.05% more computers to make up the difference. I think the competitive advantage would guarantee those extra sales and profit.
     
    Last edited: Jan 31, 2019
    MiSJAH likes this.
  7. Raidriar

    Raidriar ლ(ಠ益ಠლ)

    Reputations:
    1,466
    Messages:
    5,367
    Likes Received:
    3,799
    Trophy Points:
    431
    conspiracy theory time: they use garbage thermal paste so that it dries out in little over a year so that CPUs get nice and hot, thermal throttle, making the computers "slow" -> people buy a new computer or pay to "fix" the existing one.

    Either way, the OEM makes money!
     
    Biotic, Starlight5, Ashtrix and 3 others like this.
  8. pitz

    pitz Notebook Deity

    Reputations:
    56
    Messages:
    1,033
    Likes Received:
    71
    Trophy Points:
    66
    The thermal interface is rarely the root cause of overheating issues. Transferring heat from the CPU to the heatsink isn't the issue in a laptop cooling system, actually convecting it away is.

    A better thermal interface might buy, at best, a few additional seconds of better thermal performance. Might even cause greater severity of fan operations, ie: the higher R heatsink interface may very well provide for overall less intrusive fan operation.

    I suspect when most consumers replace heatsink paste, they also give the heatsink/fan a good cleaning, which probably explains most of the efficacy of the procedure, not replacing the thermal paste.

    As far as the cost of paste, even the high-end formulations, trivial. I've had equal results with a $5 jumbo syringe from China (including shipping) as I've had with the expensive Arctic Silver, etc. at 10X the price. In fact, only a trivial amount of paste is required, basically enough to fill in the slight irregularities in flatness between the surfaces.
     
    Aroc likes this.
  9. Larry Q

    Larry Q Newbie

    Reputations:
    9
    Messages:
    9
    Likes Received:
    6
    Trophy Points:
    6
    DFM, or design for manufacturing is an important consideration too. A lot of the times the thermal paste is pre-dispensed automatically onto the heat sync assembly so that the assembly worker simply places the assembly and tightens the screws.

    For the paste to not run during storage or handling, it is usually rather thick, and it's almost like a very soft wax and not a liquid paste. Think butter. I have seen a picture of a bunch of heat sinks with a dotted pattern of pre-applied paste.

    It is honestly probably better this way; when paste is applied manually usually they put way too much. Keep in mind that the assembly line worker probably has about 3 seconds in total, maybe 5 to get the heat sync on there.
     
    Aroc and Brad331 like this.
  10. Brad331

    Brad331 Notebook Enthusiast

    Reputations:
    27
    Messages:
    44
    Likes Received:
    32
    Trophy Points:
    26
    There is no "root cause"; there is a multitude of factors that influence cooling capacity. Bad contact between CPU and heatsink? Bad cooling. Bad fans? Bad cooling. Not enough heatpipes? Also bad cooling. I'm not calling the thermal paste the one thing that decides all; I'm saying it's one link in the chain that can be easily improved with a good yield-effort ratio (since it requires 0 effort in engineering — just make a decision and reduce profit per unit by less than $1).

    The data suggests supports that better thermal paste increases overall cooling capacity, including during continuous load. By the way, notice how the GPU tests, which are in the 60-70°C range, show more benefit from better thermal paste than the CPU tests, which are in the 30-40°C range. That suggests the advantage of better thermal paste is greater in hotter or more thermally constrained systems like graphics cards and laptops. This observation is supported by thermodynamics (specifically Q=mcΔT): in graphics cards and laptops, the heatsink's thermal mass is smaller compared to the heat source. So, the heatsink gets saturated more easily and its temperature rises more, and heat transfer between the chip and heatsink decreases. This is when thermal paste becomes more critical in maximizing the heat transferred to the heatsink.

    In the controlled tests I've linked above, better thermal paste simply did better, not as a byproduct of cleaning.

    Better thermal paste does help. Not as much as some other things, but it does help, and it does so easily, without having to spend more R&D efforts in redesigning the heatsinks and fans. (Of course, doing that would be nice too. I'm saying do the paste, not do the paste and forget everything else.)
     
    Last edited: Jan 31, 2019
    Starlight5 likes this.
Loading...

Share This Page