#15
Post
by xiphmont » Tue Jan 19, 2016 3:46 pm
Weighing in, as I am capable of some battery hardware and firmware hacking (I own some of the earlier mentioned 'specialized equipment' for hacking on smart batteries).
LiIon batteries, hardware and software, are super-paranoid about any out-of-spec conditions. Any overvoltage, undervoltage, or unexpected behavior will cause the firmware controller to permanently mark the battery 'failed' as well as intentionally blow hardware fuses that isolate the cells. This is partially because damaged cells will not only occasionally overheat/vent (ie, explode) when charging, but also when discharged. Candlepower forums has some nasty pictures of people who accidentally overcharged cells, and the cells didn't explode until they were used. So yeah, anything goes wrong, the battery quietl commits suicide and renders itself as inert as possible.
You can open up the battery and try to replace cells, but then the controller will do the same thing--- see something weird going on and electronically blow all the fuses. And those fuses are *really* annoying to find. I have to get them from random sellers in China, since they're purposely not available to the public.
The eight-cell X6X batteries use two banks of cells that are controlled and balanced separately (One ~2000mAH bank of rectangular cells, one ~2400mAH bank of can cells). If one bank fails, you'll suddenly see the reported capacity drop by half or more. And if one bank failed, the second is usually not far behind. Deep-discharging a battery on its last legs is often the straw that breaks the camel's back--- poof go the fuses.
I'll also add--- having been inside many of the cheap Chinese eBay batteries, I'd never buy one on purpose myself. The 'recycled' cells they use are often in horrific condition, and the assembly quality is anywhere from 'meh' to 'horrifying'. The cheaper the battery, the more likely it works purely by accident. If you're re-celling a battery properly, at least as a hobbyist buying retail, the new cells alone should cost most of what a new battery does anyway.
if you really want to know how and why a better failed, send it to me and I can try to dump/parse the nvflash state from the controller. The standard smart battery protocols don't really have a good way of reporting failures or reasons for failures, you need to get into vendor-specific queries, and the vendors don't publish any of the how-to.
Homepage
Forum
Drivers
HMM
MTM
Feed
