1. 28
  1.  

  2. 5

    Heat causes more problems than just that your phone gets hot in your pocket. It causes degradation in the transistors and in the connections between them. This can affect both performance and reliability.

    Hello Nexus 5X?

    Actually it’s interesting that heat is supposedly the cause of the Nexus graveyard, while desktop CPUs can be run at 100 °C for ages just fine, and only overvoltage is known to actually degrade them…

    1. 2

      When I was collecting info on hardware design, I found that the big companies do thermal modeling to identify where the chips might overheat. Then, they adjust the design to reduce that risk. (wild guess) Here’s an example for 3D chips. It’s possible there’s less of that in the smartphone SoC’s for some reason.

      1. 2

        Perhaps because there are many of them, so less time and attention are lavished upon each?

        1. 2

          There is generally less attention paid to the quality of consumer electronics chips than automotive or military grade chips. Consumer stuff can be expected to fail more often, but cars and missiles tend to have more serious consequences when they fail…

          1. 1

            Automotive and aviation being reliable makes sense. I was thinking about consumer mobile SoCs vs say PC and laptop chipsets. Both consumer electronic crap, but I hear about one going wrong more often than the other.

            1. 1

              I’d bet it’s the combo of consumer-grade electronics, market pressure to cram more power into a smaller package, and little opportunity for cooling, compared to even a very thin laptop.

              1. 1

                I suspect that the fan in my laptop helps it avoid heat-related breakdown.

            2. 1

              That’s a decent idea.

          2. 2

            Those are two very different sets of chips with very different die sizes, very different layouts on die, very different mating to the board (LGA-style vs BGA-style), very different QA screening in fab and in assembly, and very different manufacturing processes/technologies.

            1. 3

              I would kill for an AMA with an engineer who worked on the 5X. I have tortured mine for thousands of hours and found it perfectly stable in spite of the countless stories of hardware failure.

            2. 1

              I see thermal numbers about 40 to 80 C, but I can imagine 100 C