Wow this is scary. LiPo are very dangerous and even NMC which are stabler are very error prone when assembled in an ebike pack: the cells need to be high quality (LG, Panasonic), the assembly need to be very careful (so many ways for the interconnects to become a liability and cause a fire) the BMS needs to be quality (or it'll overheat and cause a fire). It's hard to make it right when careful, and even decently sized companies mess it up.
Slapping together some LiPo? It's just a matter of when, not if.
Otherwise I think these things are super cool, but in the case of lithium ebike batteries, they look so innocent and easy to build, and are so dangerous. I'm speaking from being humbled myself after building two NMC packs with cell-fuses, a spot welder, being careful... I'm dead scared of my packs.
They aren't slapping LiPos together. They aren't even using LiPos. LiPo are usually single small cells pouches in phones and gizmos. These are Li-Ion 18650 cells like Samsung 50e's being nickle spot welded into ~60V arrangements, with every 4.2V "p-pack" getting a wire to the Battery Management System circuit board. All charge and discharge does thru the approval of the BMS and gets cut off if any cell shows anything abnormal (short, too much resistance (voltage dipping), undervolted, overvolted). The photos look just as professionally made as another other factory. Unwrap literally any random e-bike's battery and you'll find identical construction. A few rare brands also pot the battery in thermal epoxy to make them even more safe. Yes fire and liability are scary for battery pack builders but this guys shop doesn't appear to be doing anything bad.
You may think so, and I encourage you to watch the video I shared. I held many of the same conclusions/approximations until learning of Grin's experience.
The video you linked is problems that can happen and things that can go wrong. It seems more like something the battery makers should watch instead of proof that batteries are inherently unsafe.
The person who linked it was saying it is evidence of batteries being unsafe because it goes over things that can go wrong.
I was saying it's educational for battery makers and the video isn't about batteries being unsafe, the guy in the video makes his own batteries and is talking about what to avoid.
I am building similar battery packs for long range uavs. We have many different configurations. Like 12s24p. 48v with total of 252 4.2v cells. We put them in 3d printed holders and spotweld them. Takes about 30 minutes to make one battery.
Bms systems are available widely. Its rare to have dangerous issues. Even with chinese clones.
Most of the accidents happening when nickel strips are cut and accidentally falling on open battery packs.
I agree that it's fun and can often be safe if built correctly for the right application. But I think UAVs don't experience the same beating and vibrations that ebike batteries do, which I think is what the bigger problem is with ebike batteries in particular. Unless carefully built, things like balance wires can shafe and cause a short, the spot-welded tabs can break off or the connection resistance increase, leading to hot spot. The BMS can have hot spots (FET based ones in particular) that lead to failure or electrical fires over time. Humidity mixed with various electrical potentials on nearby PCB traces causes corrosion and eventually failure of the BMS, etc. And it's not all the BMSes that are built the same, they may all have the same set of features on the happy path, but they're not all created equal when their PCB starts failing and fail-safes are required.
Basically in a "tough" environment like ebikes, I think the aging of the pack is much accelerated, which leads to lower mean-time between incidents. Combine that with the profile of the average buyer (wants the cheapest, longest range battery they can get) and it's an even higher probability.
Don't get me wrong, I absolutely love custom DIY packs. It's just way more dangerous than meets the eye, especially in dirty and high vibration environments.
Having to pedal without power from battery is extremely rare on e-bikes. They have range of 100km+ and few people bike that much between charges.
LFP:s are less popular for DYI because you need a much smarter BMS to deal with the flat voltage curve when charging. A cheap, easily available open source LFP BMS should solve that.
For SoC many ebike BMS just use voltage curve to give a rough approximation. But with LFP you need a shunt and a coulomb counter, because the voltage curve is essentially flat until the very last Wh.
Slapping together some LiPo? It's just a matter of when, not if.
Recently published very relevant talk by Grin: https://youtu.be/j92Gt4VviSQ?si=0TV_HqZBEBuVDqdx
Otherwise I think these things are super cool, but in the case of lithium ebike batteries, they look so innocent and easy to build, and are so dangerous. I'm speaking from being humbled myself after building two NMC packs with cell-fuses, a spot welder, being careful... I'm dead scared of my packs.