CB3-431-C5FM: AP13J3K and AP13J4K battery troubles

grjer
grjer Member Posts: 3 New User
edited November 2023 in 2019 Archives
Primary question:  Do I just have bad luck with batteries, or is my motherboard failing?

My refurbished CB3-431-C5FM came with an AP13J4K battery that only had 3 hours of life.  Reported battery health was high for the two hours, then rapidly dropped toward 0%.  Replaced it with an AP13J3K (vendors claim the 11.25v AP13J3K is appropriate to replace the 11.40v AP13J4K), this gave about 8 hours of useful life but battery health declined more or less linearly from 100% to 45%, at which point the battery was depleted and the CB3 just died with no warning.  Needless to say, I depleted the AP13J3K a good number of times, and now it's stuck reporting 1% and won't charge, dies immediately when unplugged.  Tried a hard reset, disconnecting battery, letting it sit charging for a day, etc.  So now back on my AP13J4K with 3 hours of life.
Poking around the net, I see lots of differing advice:
  Either battery should work, at least on a C720P:  https://community.acer.com/en/discussion/448684/c720p-battery-replacement
  Acer recommends a rather long winded new battery calibration procedure:  https://us.answers.acer.com/app/answers/detail/a_id/27667
  This guy recommends powering up without a battery first:  https://forums.hak5.org/topic/37227-chromebook-not-charging-stuck-at-1-heres-the-fix/
  Try a hard reset if doesn't charge, full discharges can brick the battery:  https://support.google.com/chromebook/thread/583023?hl=en
  It's a common problem across chromebooks:  https://www.reddit.com/r/chromeos/comments/a4cdm5/help_battery_wont_charge/

I have one more month to return this beast, I like it except for the battery issue.
I suppose I could send the chromebook back and get a new one, not so sure I want to take potluck on another refurbished one.
Is this just a battery issue, or is this possibly a problem with the charging circuit on the motherboard?
Is there an issue with using an AP13J3K battery on this CB3-431-C5FM, perhaps due to the 11.25v vs 11.40v nominal voltage?
Is there any way to un-brick my AP13J3K battery so it will take a charge, perhaps forcing 100ma into a couple of the wires on it for a couple hours?
Do some of the batteries available from after-market manufacturers simply not work properly?
Shouldn't a decent design prevent the laptop from damaging the battery with a full discharge?

I've replaced many laptop batteries, have never had this much trouble.

Curiously, under Crouton, the system reports that the AP13J3K battery is cycling between charging and discharging about once per second.

The AP13J3K battery I bought was this, no longer offered by Amazon for some reason, perhaps a bad clone?:
ZTHY AP13J3K AP13J4K Laptop Battery Replacement for Acer Chromebook 11.6" C720 C720P C740 C720-2848 14 CB3-431 CB3-431-C5EX Series Notebook 3ICP5/65/8  Sold by: ZTHY TECH


Answers

  • grjer
    grjer Member Posts: 3 New User
    Assuming it was toast, I took apart the AP13J3K battery that refused to charge.
    Learned a fair bit, notes included below.

    Summary: 
    The battery assembly reports battery status over a two wire SMB bus to the laptop, this includes things like percent remaining and hours remaining.  So any weird behavior in the battery status that a laptop reports is probably due to trouble inside the battery.
    The battery assembly includes NFET's to disconnect itself from the laptop when an internal fault is detected.  This is likely the case when it's stuck at 1% and refuses to charge.  
    I think my laptop if fine.

    Also this, a bit more esoteric:  The controller inside the battery assembly checks status roughly once every second, which explains why I was seeing a status message under Crouton that cycled between charging and discharging once every second.  With the battery on the bench, looking across the P- and P+ terminals of the battery with a voltmeter I could see 12.75v for one second, then 0v the next second.  Grounding the reset pin of the controller chip or the PRES (present) pin of the connector did not cure this behavior.  At 12.75v, each of my three cells were fully charged at 4.25v, so there was some other fault condition causing me grief.  I may probe it through the SMB bus using an Arduino Nano to see if I can learn why it fails.

    Here's my notes from the surgery:

    Opened on side opposite all the print, pealing off the film from over cells, then prying off the plastic cover at the side over the electronics PCB.
    PCB has silkscreen legends near the wires to the connector:       black+black:P-   blue:SMBD   green:SMBC   yellow:T   orange:PRES   red+red:P+
    When it's working, should see the battery voltage across P- and P+ terminals, but perhaps only when PRES is grounded (to indicate that a laptop is present).  The T pin provides a voltage proportional to the battery's internal temperature.  The two pins for the SMB bus allow the laptop to communicate with the controller chip inside the battery assembly.
    Battery is wired as 3 lithium cells in series with sense wires into the two intermediate nodes.  PCB glued down, no parts on the back but there are traces.
    Has one SOIC44 controller, schematic for a suggested design is on page 15 of the datasheet:   http://www.ti.com/lit/ds/symlink/bq20z95.pdf
    The Acer AP13J3K mostly follows this suggested design with a few of the optional features missing. 
    Has two power NFET's in SOIC8 packages, marking of "B06 N03":  http://ftp01.cystekec.com/MTB06N03Q8.pdf  (pins 1,2,3=Src, 4=Gate, 5,6,7,8=Drain)
    Positive cell lead goes to green 7A SM1206 polyswitch, then to source of one NFET, drain to drain, source of second NFET to P+ out to connector to computer, as per the datasheet's suggested design on page 15.  This allows shutting down current flow in either direction, charge or discharge.
    Two SOT23-3 packages with marking of "A6", PCB silk says D1 and D2 (shown as D7 and D1 respectively on p15 of bq20z95.pdf)
    Thermocouple on two magnet wires, says "NTC" in PCB silk where the magnet wires are soldered into the circuit.
    Large 0.010 ohm resistor marked "R010" in series with negative lead from the 3 cells so controller chip can monitor the current.
    One green surface mount 1206 device marked "7" (a 7A polyswitch?),    Twenty-three 0805 resistors, and twenty 0805 capacitors. 
    That's all there is to it.

    If you wish to dig into this further, here's some interesting webpages:
      Arduino code to play with the SMB bus of a laptop battery:  https://linuxehacking.ovh/2014/03/30/recharging-and-reusing-acer-laptop/
      https://batteryuniversity.com/learn/article/how_to_repair_a_laptop_battery

    If you choose to disassemble your battery, do be careful.  Shorting across one of the cells could conceivably cause it to explode.  There's a lot of energy packed in there, and a few nasty chemicals as well.

  • JackE
    JackE ACE Posts: 44,478 Trailblazer
    >>>So now back on my AP13J4K with 3 hours of life.>>>

    If the old one still works OK except for the 3-hr life, I'd have to guess the new battery was junk. Sanyo Part #KT.00303.014 & Sony part #KT.00304.001

    Jack E/NJ

    Jack E/NJ