[lug] question about laptop battery

Nate Duehr nate at natetech.com
Mon Jan 18 13:15:52 MST 2010


Short version:

You've got it Ken, LIon is far more energy-dense than NiCd or NiMH, but 
has the downside that "cycles" of the battery slowly reduce its 
capacity.  It's about time to replace that one, looks like. Usually the 
OS can ask the battery how many cycles it's been through, but not sure 
how to do it on Linux.  The batteries have smart charging controllers 
that keep track of all this stuff.

Longer version:

SOMETIMES the LIon charge controller gets confused in cases where 
multiple partial charges have been done.  It's NOT like the old chemical 
"memory" issue some people claim happens in NiCD and NiMH.  (Which NASA 
took hundreds and hundreds of batteries, abused them, measured them, and 
debunked.  They've been using NiCD on spacecraft for decades.  They know 
how to charge/discharge them properly.  Many satellites have operated 
for YEARS beyond the battery's projected lifespan, so it prompted a 
study, and NiCD "memory" was bunk.  Go figure.  The CHARGERS of the day, 
and badly designed current draws of some devices,  just abused the crap 
out of them and they died.)

Anyway, back to LIon... Some machines (Lenovo has this in their power 
manager under Windows, for example) can signal the controller in the 
battery to "forget" all of their "how charged is the battery" data, then 
force a full charge/discharge/recharge cycle (takes many hours, and you 
want to let it finish).

Other manufacturers (Apple) have instructions to fully drain and 
recharge the battery to help the controller figure out what's going on, 
with at least 20 minutes at each "extreme" with no charge/discharge.  
(Sadly, this is a full "cycle" so their recommendation to handle the 
confusion in the controller in the battery also takes a tiny percentage 
of the battery's life away.  Oh well, small price to pay I guess.)

The best thing to do, see if your machine's manufacturer has a 
recommendation as to how their system should work.  Reason is, if a LIon 
is wearing out pre-maturely and you've followed all their 
recommendations, you can often twist their little arms into replacing it 
under warranty (or extended warranty if you have one).

If you remember all those LIon laptop "fires" a while back (really most 
were just overheating that burnt things outside the laptop), many of 
those were a manufacturing defect in the packs, but others were bad code 
in the battery's internal controllers.  LIon have a very narrow window 
of voltages/currents that they can be charged at without overheating the 
chemicals and creating a major problem/headache/loss of job for the poor 
EE who designed the charging circuit.  Most LIon EE work is done by 
CYA... they buy the "manufacturer recommended" charge control chip for a 
particular pack from a maker, and then design the computer's battery 
information bus around that chipset.  :-)

I tracked a huge discussion/debate between some brilliant EE's a couple 
of years ago about various battery issues.  The world really needs a new 
battery technology... but nothing looks all that great on the horizon, 
yet.  Fuel cells in a laptop aren't realistic, and we're kinda "stuck" 
in battery technology for many years now.  We have gotten better at 
*charging* those batteries, but often there's trade-offs between 
lifespan of the battery and reasonable charge times, etc.

Nate

On 1/18/2010 10:16 AM, Kenneth D Weinert wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> OK, so I took a look and I see the following two lines from
> /proc/acpi/battery/info:
>
> design capacity:         6000 mAh
> last full capacity:      2848 mAh
>
> and from state:
>
> charging state:          charged
>
>
> I presume this means I should start looking for a replacement battery,
> right?
>
> It's a LION, will taking it down to zero and then recharging make a
> difference?
>
> Thanks.
>
> Ken
>    




More information about the LUG mailing list