[sid] lenovo thinkpad x200s discharge battery while on hibernate

[Dot Deb]

I'm running debian/sid on a x200s.

Some time ago (don't know exactly when) it started this strange
behaviour of using battery power while on hibernate with AC
disconnected.

Today I made a check: 4% in 3 hr.

I don't know what to do to understand.
Any idea is very welcome.

augh


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Bob Proulx]

Dot Deb wrote:
> Some time ago (don't know exactly when) it started this strange
> behaviour of using battery power while on hibernate with AC
> disconnected.
>
> Today I made a check: 4% in 3 hr.

I suspect that your battery has developed a slow discharge due to
internal shorting.  This is one of the common failure modes of
rechargable batteries.

If it really is hibernated then it should be virtually zero power
(only enough to run the power button).  You can remove the battery.
To test this theory I would either hibernate or shutdown and remove
the battery then test the power level again after an overnight.  If
the power is the same as before then your battery is probably okay and
the fault is elsewhere.  If the power is drained with the battery
removed then you can be fairly certain that the problem is isolated to
the battery.

Hopefully that is it because then you have an easy, if moderatly
expensive, fix and then you are back solidly running again.  It that
isn't it then I have no more ideas.  Good luck!

Bob

[celejar]

On Sat, 24 Oct 2009 16:01:35 -0600
Bob Proulx <bob@...> wrote:

...

> If it really is hibernated then it should be virtually zero power
> (only enough to run the power button).  You can remove the battery.

By 'hibernate', I assume that we're talking about suspend-to-disk.
What do you mean by 'enough to run the power button'?  When suspended to
disk, isn't the system completely off?

Celejar
--
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Bob Proulx]

Celejar wrote:
> By 'hibernate', I assume that we're talking about suspend-to-disk.

I am assuming that too.  Hibernate is suspend-to-disk and the machine
is then turned off.  Standby/Suspend would be suspend-to-ram and would
need continuous power to keep the ram alive.

> What do you mean by 'enough to run the power button'?  When suspended to
> disk, isn't the system completely off?

Most of the system is off.  But the power button is electronic.  It
takes just a tiny little bit of power to sense the button and then
activate the power supply.  It is almost a zero power state.  Power
isn't needed for the computer since nothing is running but power is
needed for the switch.

In the old days the power switch was a simple physical power switch.
The switch went "clunk" and power was either connected or
disconnected.  The machine couldn't turn itself off but could only say
it was now safe to remove the power.  But these days almost all
computers use a smart electronically driven power supply that can turn
itself off and on.  Or allows the machine to wake up automatically
upon a set time.

Also machines might be configured for a Wake-On-LAN magic packet in
which case the networking would also be active.  You may have seen
Wake On LAN available in the BIOS settings.  (I don't know if the x200
in particular has WOL.)  The LAN would of course draw significantly
more power, for a laptop.  But for a desktop the amount of power to
drive the LAN card is small.  A lot of larger sites configure machines
for wake on LAN so that they can wake them up in the middle of the
night for backup and for updates and then shut them down again when
that is done.

Hmm...  Thinking about this more makes me think you should probably
make sure that Wake-On-LAN isn't enabled in the BIOS.  Since that
would need to keep the LAN powered up while off or hibernated.

Bob

[celejar]

On Sat, 24 Oct 2009 21:40:05 -0600
Bob Proulx <bob@...> wrote:

...

> Most of the system is off.  But the power button is electronic.  It
> takes just a tiny little bit of power to sense the button and then
> activate the power supply.  It is almost a zero power state.  Power
> isn't needed for the computer since nothing is running but power is
> needed for the switch.
>
> In the old days the power switch was a simple physical power switch.
> The switch went "clunk" and power was either connected or
> disconnected.  The machine couldn't turn itself off but could only say
> it was now safe to remove the power.  But these days almost all
> computers use a smart electronically driven power supply that can turn
> itself off and on.  Or allows the machine to wake up automatically
> upon a set time.

Thanks for the detailed explanation.

Celejar
--
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Dot Deb]

On Sat, Oct 24, 2009 at 11:01 PM, Bob Proulx <bob@...> wrote:
> To test this theory I would either hibernate or shutdown and remove
> the battery then test the power level again after an overnight.  If
> the power is the same as before then your battery is probably okay and
> the fault is elsewhere.

I just woke up, re-inserted the battery and switched the computer on:
the charge is the same as when I switched it off 7hr ago.

I found the the problem is actually elsewhere: in the meanwhile I got
the solution
from another list. For details see:
http://www.nabble.com/Battery-discharge-during-hibernation-td22417171.html

The problem is the suspend-to-disk method: in my laptop (lenovo
thinkpad) the "platform" method do both suspend-to-ram and then suspend
to disk. This is expected to make the resume process more rapid using
the image written to ram, while preserving those written to swap in case
of complete battery consumption. By the way, this doesn't work for me:
only resume from swap works making this method completely unuseful.

augh


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Bob Proulx]

Dot Deb wrote:
> I just woke up, re-inserted the battery and switched the computer on:
> the charge is the same as when I switched it off 7hr ago.

Well, at least you know it isn't the battery.

> The problem is the suspend-to-disk method: in my laptop (lenovo
> thinkpad) the "platform" method do both suspend-to-ram and then suspend
> to disk. This is expected to make the resume process more rapid using
> the image written to ram, while preserving those written to swap in case
> of complete battery consumption.

I am curious.  On my ThinkPad T42 when suspend-to-ram is active there
is a little "moon" led lit to indicate that power is being used to
keep the ram alive.  When illuminated it shows the machine as
sleeping, drawing power, and not hibernated.  Does your not have such
an indicator?

> By the way, this doesn't work for me: only resume from swap works
> making this method completely unuseful.

Unfortunately it is not uncommon on laptops to find that one method or
the other is not functional.  Or worse that neither works.

If you haven't yet then browsing the wonderful thinkwiki.org site may
turn up some hints as to your problem.

  http://www.thinkwiki.org/wiki/Problems_with_ACPI_suspend-to-ram

Bob

[celejar]

On Sun, 25 Oct 2009 12:05:26 -0600
Bob Proulx <bob@...> wrote:

...

> Unfortunately it is not uncommon on laptops to find that one method or
> the other is not functional.  Or worse that neither works.

You know a lot more about this stuff than I do, but I'm curious - I
thought that s2disk, unlike s2ram, really doesn't involve any special
low-level, device-specific quirks, and should work pretty much
anywhere.  I have never gotten s2ram to work on my Acer 3690, but
s2disk has almost always pretty much worked out of the box (except for
one brief period with a bleeding edge kernel, which may have been
caused by a problem with a buggy ide driver).  Are there indeed common
cases where s2disk doesn't work?

Celejar
--
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Bob Proulx]

Celejar wrote:
> I thought that s2disk, unlike s2ram, really doesn't involve any
> special low-level, device-specific quirks, and should work pretty
> much anywhere.

Unfortunately no.  For example even after working the problem very
hard for a significant time I never got the stock in-kernel swsusp to
function on my HP nc4000 or my older machine before it.  From my
perspective the default in kernel suspend only works on some
architectures and often not the one I have in my hand.  It has worked
on my ThinkPads though.  Suspend2 by Nigel Cunningham has worked
perfectly on a range of systems.  It isn't in the kernel because the
in kernel suspend is maintained by someone who has consistently
blocked it.  This has been the source of long running controversy.  If
you search the web email archives for suspend2 vs swsusp you will find
thousands of postings and arguments about it.

> I have never gotten s2ram to work on my Acer 3690, but s2disk has
> almost always pretty much worked out of the box (except for one
> brief period with a bleeding edge kernel, which may have been caused
> by a problem with a buggy ide driver).  Are there indeed common
> cases where s2disk doesn't work?

Personally I think you have been lucky!

Bob

[celejar]

On Sun, 25 Oct 2009 13:17:55 -0600
Bob Proulx <bob@...> wrote:

> Celejar wrote:
> > I thought that s2disk, unlike s2ram, really doesn't involve any
> > special low-level, device-specific quirks, and should work pretty
> > much anywhere.
>
> Unfortunately no.  For example even after working the problem very

...

> Personally I think you have been lucky!

Oh, well, I guess this is real YMMV territory ...

Celejar
--
mailmin.sourceforge.net - remote access via secure (OpenPGP) email
ssuds.sourceforge.net - A Simple Sudoku Solver and Generator


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Dot Deb]

On Sun, Oct 25, 2009 at 7:05 PM, Bob Proulx <bob@...> wrote:
> I am curious.  On my ThinkPad T42 when suspend-to-ram is active there
> is a little "moon" led lit to indicate that power is being used to
> keep the ram alive.  When illuminated it shows the machine as
> sleeping, drawing power, and not hibernated.  Does your not have such
> an indicator?

The little moon you refer to is present only when suspend-to-ram.
As far as I understand, on my laptop model (don't know if this apply
to any thinkpad)
the "platform" method to suspend-to-disk, performs two actions: ram
image + an actual suspend-to-disk.
In this case no "little moon" appears on the horizon!

> If you haven't yet then browsing the wonderful thinkwiki.org site may
> turn up some hints as to your problem.
>
>  http://www.thinkwiki.org/wiki/Problems_with_ACPI_suspend-to-ram

I know thinkwiki.org and I agree with you it's wonderful.

Alberto


--
To UNSUBSCRIBE, email to debian-laptop-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...