Thanks for your help.
I have an 820-3536 macbook and I need to know what info (or how smart) the battery packs are. Obviously the SMC talks to it constantly and to the point of if I inject power on the batt line with no batt connected it won't turn on, it knows there's nothing to talk to so ignores it.
Anyway, the pack discharged to 1V/cell due to a faulty charging FET and sat there for some weeks. Could the battery chip have logged this and locked out the pack for safety so it can never be used again?
The output FET on the packs PCB is never turned on. If I jumper this so that B+ is present at MB it will turn on but shuts down during boot.
Runs fine on charger and reports battery level accurately. All current sensing circuitry appears to be ok. Because all rails perform when on charger I'm struggling to find a reason other than the battery has condemned itself and telling system to not use it.
Cheers
Could the battery chip have logged this and locked out the pack for safety so it can never be used again?
Definitely a possibility. Some laptop battery packs will do that.
Also, when you leave lithium cells at 1V per cell for weeks they get ruined to the point where charging brings them to 4.2V almost instantly but as soon as you try and discharge them they fall back to flat. So this could be happening to your pack.
It may just have no capacity left.
Should have mentioned I manually charged them to 4.2 which took a couple hours or so at 1A and they took an external load ok, pretty sure the cells are healthy enough to run the unit.
Definitely a possibility. Some laptop battery packs will do that.
Not some but almost all of them.
there is 'one wire' link between battery and laptop, EC wont let you turn on battery if it doesnt like what its saying
Should have mentioned I manually charged them to 4.2 which took a couple hours or so at 1A and they took an external load ok, pretty sure the cells are healthy enough to run the unit.
yep, that sounds like bat. controller deciding its dead and permanently bricking itself
sounds like you really want to fix this battery
maybe buy another used one with very little capacity, as long as its working it should be able to calibrate itself during full charge/discharge cycle with new cells.
What is the battery controller IC? It's probably one of the standard ones from TI or Maxim.
If you disconnect the cells from the controller, most controllers will lock out . Many have volatile SRAM bits that erase when power is lost. I have worked on batteries that have a volatile ID code in SRAM. No power erases the ID rendering the battery useless.
Battery replacement can be done if you maintain the correct voltages at the controller.
Aurora
Should have mentioned I manually charged them to 4.2 which took a couple hours or so at 1A and they took an external load ok, pretty sure the cells are healthy enough to run the unit.
Its good practice to bring cells back up to ~3v or so at a low current (50-100mA) before charging them at 1A. Then again, its also bad practice to use cells that have spent time below 1.5V so...
I'm not one of those people who like sowing fear about every lithium battery someone asks about, but these sound like ones to be cautious about.
As others have said, the pack has almost certainly perm-failed itself. There is software out there used by battery pack refurbishers that can reverse this condition, but it doesn't work for all packs by all manufacturers, and generally isn't priced for one-offs.
You might be able to hack it yourself. The packs just talk over SMBus using the Smart Battery Standard protocol. My sig links to a half-assed arduino project to read out pack data. The reprogramming commands are often proprietary to the chipset or firmware, and generally locked by a passcode. With enough determination, you might be able to round up the necessary information from public sources, though it would probably be easier if you can read and write in one or more of Chinese, Vietnamese and/or Russian.
Thanks guys. On close analysis with a logic analyzer I find the initialization bit has been cleared. It appears factory set info was lost when the cells died out. Also the Full & Stop Charging, and the Depleted Stop Discharging bits are set which is entirely contradictory and as a result the pack will never be usable without re-initialization of the chip
At least I learned heaps about SMBus
Among the info it probably lost was calibration of the coulomb-counter's current shunt, and information about the cells themselves.
I think Apple has generally used TI's bq battery management chips, probably with customized firmware. However, there is a fair amount of documentation available publicly about the firmware for the battery management dev board(s) that TI sells.
Whether you want to go to all that trouble at all is up to you, but I wouldn't go to all that trouble and use the existing cells after they sat below 1.5v for a while.