4 ms·

The main risk factor is cells, you have to source them from reputable manufacturers. After having monitored a few battery fires, we went on to design a casing w

The main risk factor is cells, you have to source them from reputable manufacturers. After having monitored a few battery fires, we went on to design a casing with multiple features to contain fires, you can check it here: https://www.youtube.com/watch?v=v0NXXfCA2CY https://www.youtube.com/watch?v=v0NXXfCA2CY If you're interested, you can order them here https://infinite-battery.com https://infinite-battery.com

gertlex 1y ago
How much is this mitigated by well-behaved charging circuitry? I.e. my understanding is most devices are not like the 3s lipos I put in my hobby robots, but instead have integrated charging circuitry that you just give an appropriate voltage to... and that proper charging behavior avoids a lot of the dangerous scenarios with lipos? I ask because, loosely, the # of battery fires seems like a function in part of: (1) use of good charging logic; (2) cell manufacturing quality; (3) # of cells in the wild. While the growth of 3 probably dominates the improvements to 1 and 2, I'm guessing the number of battery fires has grown but not "exponentially".
- joecool1029 1y ago
  > I.e. my understanding is most devices are not like the 3s lipos I put in my hobby robots, but instead have integrated charging circuitry that you just give an appropriate voltage to... and that proper charging behavior avoids a lot of the dangerous scenarios with lipos? I'm not sure your understanding is correct. There are assembled packs with a BMS on them, laptop batteries this is usually the case. The cells themselves can be lipol, prismatic, or cylindrical (like 18650's). The cells almost never have active BMS built-in. It's always external (either on a board on the pack or on the device itself charging it). I really think the battery fires are mainly qc issues and running bad qc cells beyond limits (either fast charging or discharging). If capacity is staggered a ton between cells then even charge balancing the cells isn't going to do much good. Pouch cell fires probably more related to physical damage due to expansion in places they aren't designed well to expand (so it pushes the jellyroll down and causes a short/thermal runaway).
  - gertlex 1y ago
    Thanks for the reply! I was not thinking/claiming individual cells had a BMS on them, but was referring to the "pack" level protections. Your mentioning of specific failure modes helps me, specifically breaking up the failures into multiple categories.
Doohickey-d 1y ago
This seems like a cool design! One question though: it seems like the contact between the PCB side panels and the cells relies on flexing the (FR4?) PCB material of the small per-cell "fingers" that are surrounded by cutouts, on the side panels. I wonder if once it's in this "flexed state", it will eventually fatigue, or settle into the bent state, after some time (years?), and no longer press on the cell ends strongly enough? A bit like if you were to leave a rubber band in stretched state for a while, it won't fully go back to the diameter it was before.

gertlex 1y ago

How much is this mitigated by well-behaved charging circuitry? I.e. my understanding is most devices are not like the 3s lipos I put in my hobby robots, but instead have integrated charging circuitry that you just give an appropriate voltage to... and that proper charging behavior avoids a lot of the dangerous scenarios with lipos? I ask because, loosely, the # of battery fires seems like a function in part of: (1) use of good charging logic; (2) cell manufacturing quality; (3) # of cells in the wild. While the growth of 3 probably dominates the improvements to 1 and 2, I'm guessing the number of battery fires has grown but not "exponentially".

joecool1029 1y ago
> I.e. my understanding is most devices are not like the 3s lipos I put in my hobby robots, but instead have integrated charging circuitry that you just give an appropriate voltage to... and that proper charging behavior avoids a lot of the dangerous scenarios with lipos? I'm not sure your understanding is correct. There are assembled packs with a BMS on them, laptop batteries this is usually the case. The cells themselves can be lipol, prismatic, or cylindrical (like 18650's). The cells almost never have active BMS built-in. It's always external (either on a board on the pack or on the device itself charging it). I really think the battery fires are mainly qc issues and running bad qc cells beyond limits (either fast charging or discharging). If capacity is staggered a ton between cells then even charge balancing the cells isn't going to do much good. Pouch cell fires probably more related to physical damage due to expansion in places they aren't designed well to expand (so it pushes the jellyroll down and causes a short/thermal runaway).
- gertlex 1y ago
  Thanks for the reply! I was not thinking/claiming individual cells had a BMS on them, but was referring to the "pack" level protections. Your mentioning of specific failure modes helps me, specifically breaking up the failures into multiple categories.

Doohickey-d 1y ago

This seems like a cool design! One question though: it seems like the contact between the PCB side panels and the cells relies on flexing the (FR4?) PCB material of the small per-cell "fingers" that are surrounded by cutouts, on the side panels. I wonder if once it's in this "flexed state", it will eventually fatigue, or settle into the bent state, after some time (years?), and no longer press on the cell ends strongly enough? A bit like if you were to leave a rubber band in stretched state for a while, it won't fully go back to the diameter it was before.