9 ms·

I don't have enough background, but I recently read that some distributions could be a fire hazard due to the lack of proper temperature regulation: https://xn

I don't have enough background, but I recently read that some distributions could be a fire hazard due to the lack of proper temperature regulation: https://xnux.eu/log/#017 https://xnux.eu/log/#017 https://xnux.eu/log/#018 https://xnux.eu/log/#018 If this was the case, is that still a concern and which distributions/systems address this problem?

josteink 6y ago
I believe all those concerns were put to rest once the modularised design of the phone was taken into consideration. https://news.ycombinator.com/item?id=24596248 https://news.ycombinator.com/item?id=24596248
- Causality1 6y ago
  Not at all. The specced hardware charge controller has no thermal circuitry and only disables the battery at a 6 amp current draw.
  - DoingIsLearning 6y ago
    I'm not a power electronics guy but can a charging circuit realistically be a fire hazard without an initial overcurrent? It seems like good redundancy to have thermal cut-off but theoretically a fire would have to be preceded by a massive overcurrent anyway, no?
    - Causality1 6y ago
      A chip can only detect current that's running through it or a sensor. A breached or shorted cell has a massive overcurrent between two parts of the cell and hence can only be detected by heat
      - josteink 6y ago
        > A breached or shorted cell has a massive overcurrent between two parts of the cell and hence can only be detected by heat But that's not an error which can be caused by badly written software (i.e. like a Linux kernel). So there's theoreticly no increased risk with running different "home-brew" OS builds on this phone due to software-related aspects of those builds.
      - megous 6y ago
        Shorted cell is game over regardless of software, most probably.
    - megous 6y ago
      Heat can accumulate. You just need to operate near the safety cutoff limit and over the battery spec for long enough. Anyway, there are layers of protection. More the better.
- megous 6y ago
  "All those concerns" is a bit too much optimistic. There are several concerns, and while there were some ideas put forward, and some patches applied to non-mainline Linux trees, all of those are untested. Values for the NTC were taken from some random datasheet on the internet in the hopes most 10kOhm NTCs have a similar R-T characteristic (they don't), and that the random datasheet I picked matches NTC that's in the battery. Just to give you a perspective on the level of quality of existing solution. It's better than before, though.
megous 6y ago
It's probably not a fire hazard any more than other regular smartphones. Battery usually only gets somewhat warm when charged at "fast charging" current. My kernel has some patches that enable battery thermal regulation in the power management chip. Battery gets too warm/cold, charging stops.
- 0rjf83 6y ago
  Would you be willing to share these patches? I'm interested in this.
  - megous 6y ago
    https://megous.com/git/linux/log/?h=axp-5.10 https://megous.com/git/linux/log/?h=axp-5.10 - regulator: axp20x: Enable over-temperature protection and 16s reset function - power: supply: axp20x_battery: Setup thermal regulation (experimental) EDIT: I'm the author of the quoted articles on xnux.eu, just to be clear.

josteink 6y ago

I believe all those concerns were put to rest once the modularised design of the phone was taken into consideration. https://news.ycombinator.com/item?id=24596248 https://news.ycombinator.com/item?id=24596248

Causality1 6y ago
Not at all. The specced hardware charge controller has no thermal circuitry and only disables the battery at a 6 amp current draw.
- DoingIsLearning 6y ago
  I'm not a power electronics guy but can a charging circuit realistically be a fire hazard without an initial overcurrent? It seems like good redundancy to have thermal cut-off but theoretically a fire would have to be preceded by a massive overcurrent anyway, no?
  - Causality1 6y ago
    A chip can only detect current that's running through it or a sensor. A breached or shorted cell has a massive overcurrent between two parts of the cell and hence can only be detected by heat
    - josteink 6y ago
      > A breached or shorted cell has a massive overcurrent between two parts of the cell and hence can only be detected by heat But that's not an error which can be caused by badly written software (i.e. like a Linux kernel). So there's theoreticly no increased risk with running different "home-brew" OS builds on this phone due to software-related aspects of those builds.
    - megous 6y ago
      Shorted cell is game over regardless of software, most probably.
  - megous 6y ago
    Heat can accumulate. You just need to operate near the safety cutoff limit and over the battery spec for long enough. Anyway, there are layers of protection. More the better.
megous 6y ago
"All those concerns" is a bit too much optimistic. There are several concerns, and while there were some ideas put forward, and some patches applied to non-mainline Linux trees, all of those are untested. Values for the NTC were taken from some random datasheet on the internet in the hopes most 10kOhm NTCs have a similar R-T characteristic (they don't), and that the random datasheet I picked matches NTC that's in the battery. Just to give you a perspective on the level of quality of existing solution. It's better than before, though.

megous 6y ago

It's probably not a fire hazard any more than other regular smartphones. Battery usually only gets somewhat warm when charged at "fast charging" current. My kernel has some patches that enable battery thermal regulation in the power management chip. Battery gets too warm/cold, charging stops.

0rjf83 6y ago
Would you be willing to share these patches? I'm interested in this.
- megous 6y ago
  https://megous.com/git/linux/log/?h=axp-5.10 https://megous.com/git/linux/log/?h=axp-5.10 - regulator: axp20x: Enable over-temperature protection and 16s reset function - power: supply: axp20x_battery: Setup thermal regulation (experimental) EDIT: I'm the author of the quoted articles on xnux.eu, just to be clear.