Re: Accuracy of the power meter?

2007-05-09 Thread hendrik
On Tue, May 08, 2007 at 08:16:40AM -0400, Theodore Tso wrote:
 On Tue, May 08, 2007 at 01:39:35PM +0200, Ralph Angenendt wrote:
  Lucky you. When I'm connected to my AP at home (The white fonera AP),
  I'll get 1.5 days max, even if I don't use the device and just omweather
  does periodic updates.
 
 Omeweather turned out to be it, it looks like.  I need to do some more
 tests to conclusively prove it, but it looks like I had set omweather
 (version 0.17.2) to never update, and my guess that set the interval
 time to 0, and that seems to have triggered some continue loop.

Sounds like a bug.  Report it.

-- hendrik
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Kalle Valo
ext Mikhail Sobolev [EMAIL PROTECTED] writes:

 One possibility is if you use RD mode or some other [non-default]
 options changed with flasher application.  In this case power management
 would not be able to deliver the promised time.

RD-mode itself shouldn't affect power consumption, but for example
serial-console RD flag will.

-- 
Kalle Valo

___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Kalle Valo
Theodore Ts'o [EMAIL PROTECTED] writes:

 I've been noticing that I don't get anywhere near the promised 10 days
 idle battery life on my N800.  With bluetooth and the wireless turned
 off (offline mode), and with the display off and keys/touchscreen
 locked, I'm getting at best 8-10 hours of idle time before the battery
 is completely drained.  

8-10 hours of standby time implies that the MPU cannot sleep.

 Has anyone seen anything like this or any thoughts about why I might be
 seeing this?

Usually there are two cases, either serial-console RD flag is enabled
or USB cable is connected. Both block MPU from sleeping (or going into
retention or whatever it is called). If it's not that, I would guess
that it's some badly behaving application.

-- 
Kalle Valo

___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Frantisek Dufka
Theodore Ts'o wrote:
 The second is that I'm running off a 2gig flash card as my root
 filesystem instead of the internal flash, per the instructions given in
 the Maemo wiki.  That's a distinct possibility, although I don't see how
 the flash card could be pulling so much current when the system is an
 idle state.

I'm running off a SD card too and with bluetooth on and wi-fi connected 
to AP it lasts couple of days.

 
 Has anyone seen anything like this or any thoughts about why I might be
 seeing this?


Apart from something you installed (some home screen or status bar 
applet?) there is this metalayer-crawler that can get stuck when 
indexing card, there is also getty respawning bug 
https://maemo.org/bugzilla/show_bug.cgi?id=1222 but still it is probably 
something else as the first one probably drains battery faster and the 
getty bug slower. Nothing suspicious in kernel log?

Frantisek
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Ralph Angenendt
Frantisek Dufka wrote:
 Theodore Ts'o wrote:
  The second is that I'm running off a 2gig flash card as my root
  filesystem instead of the internal flash, per the instructions given in
  the Maemo wiki.  That's a distinct possibility, although I don't see how
  the flash card could be pulling so much current when the system is an
  idle state.
 
 I'm running off a SD card too and with bluetooth on and wi-fi connected 
 to AP it lasts couple of days.

Lucky you. When I'm connected to my AP at home (The white fonera AP),
I'll get 1.5 days max, even if I don't use the device and just omweather
does periodic updates.

Ralph


pgp9vS37AbcSE.pgp
Description: PGP signature
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Frantisek Dufka
Ralph Angenendt wrote:
 I'm running off a SD card too and with bluetooth on and wi-fi connected 
 to AP it lasts couple of days.
 
 Lucky you. When I'm connected to my AP at home (The white fonera AP),
 I'll get 1.5 days max, even if I don't use the device and just omweather
 does periodic updates.
 

How often omweather updates data? Did you try with omweather disabled? 
Not sure what is default value now but you may try to change PSM timeout
http://www.gossamer-threads.com/lists/maemo/users/18057#18057
I am now at 100ms and did not notice any disadvantages (Asus wl500gx 
with Oleg's firmware http://oleg.wl500g.info/). Also I have both N770 
and N800 so one of them sits idle while I am using the other one. Mostly 
the idle one is N800. N770 has hardcover so I take it everywhere and 
then just continue where I left even when at home.

Frantisek
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Theodore Tso
On Tue, May 08, 2007 at 01:39:35PM +0200, Ralph Angenendt wrote:
 Lucky you. When I'm connected to my AP at home (The white fonera AP),
 I'll get 1.5 days max, even if I don't use the device and just omweather
 does periodic updates.

Omeweather turned out to be it, it looks like.  I need to do some more
tests to conclusively prove it, but it looks like I had set omweather
(version 0.17.2) to never update, and my guess that set the interval
time to 0, and that seems to have triggered some continue loop.
Setting the interval time to 4 hours seems to have extended my battery
life.

 - Ted
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Michael Wiktowy
On 5/8/07, Ralph Angenendt [EMAIL PROTECTED] wrote:
 Eero Tamminen wrote:
   Omeweather turned out to be it, it looks like.  I need to do some more
   tests to conclusively prove it, but it looks like I had set omweather
   (version 0.17.2) to never update, and my guess that set the interval
   time to 0, and that seems to have triggered some continue loop.
   Setting the interval time to 4 hours seems to have extended my battery
   life.
 
  AFAIK it can also make desktop take all the memory in the device within
  day or two and make some of the other applets not to work.

 That seems to have been fixed in one of the last updates to omweather. I
 haven't seen massive memory consumption from omweather during the last
 weeks.

Good news ... maybe I will try using it again as it was causing
spontaneous rebooting on my N800 when I had the applet enabled. As
soon as I turned it off, my N800 has been rock solid. I think that I
had it never to update too so that quirk that Theodore pointed out is
illuminating. Hopefully that gets fixed too.

/Mike
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-08 Thread Ralph Angenendt
Frantisek Dufka wrote:
 How often omweather updates data? Did you try with omweather disabled? 

It is disabled, as I just found out - I needed the Yellow Notes home
applet and had no place for omweather anymore :)

 Not sure what is default value now but you may try to change PSM timeout
 http://www.gossamer-threads.com/lists/maemo/users/18057#18057
 I am now at 100ms and did not notice any disadvantages (Asus wl500gx 
 with Oleg's firmware http://oleg.wl500g.info/). 

Thanks, will try that. Completely forgot about it.

Cheers,

Ralph


pgpbDAB78fXDA.pgp
Description: PGP signature
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Accuracy of the power meter?

2007-05-07 Thread Theodore Ts'o

I've been noticing that I don't get anywhere near the promised 10 days
idle battery life on my N800.  With bluetooth and the wireless turned
off (offline mode), and with the display off and keys/touchscreen
locked, I'm getting at best 8-10 hours of idle time before the battery
is completely drained.  

I have two theories why at this point.  The first is that the battery
has gone south, even though it I've only had the N800 for about a month
or os.  The second is that I'm running off a 2gig flash card as my root
filesystem instead of the internal flash, per the instructions given in
the Maemo wiki.  That's a distinct possibility, although I don't see how
the flash card could be pulling so much current when the system is an
idle state.

Has anyone seen anything like this or any thoughts about why I might be
seeing this?

Thanks,

- Ted

___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-07 Thread Mike Lococo
 I've been noticing that I don't get anywhere near the promised 10 days
 idle battery life on my N800.  With bluetooth and the wireless turned
 off (offline mode), and with the display off and keys/touchscreen
 locked, I'm getting at best 8-10 hours of idle time before the battery
 is completely drained.  

That's definitely weird.  I don't get 10 days between charges, but then 
my N800 is never idle for 10 days at a time.  I commonly run 3 days 
between charges, though, and I'm a fairly active user.

 I have two theories why at this point.  The first is that the battery
 has gone south, even though it I've only had the N800 for about a month
 or os.  The second is that I'm running off a 2gig flash card as my root
 filesystem instead of the internal flash, per the instructions given in
 the Maemo wiki.

I don't do this.  If you don't get a more helpful response, I would run 
with a stock fs and no applications installed for a day (no fun, but an 
important troubleshooting step if you're not sure where the problem 
lies).  If you still have low standby life in offline/screenlocked mode 
then you definitely have hardware issues.  If not, then you can start 
introducing variables in a controlled way until you pin down the rogue 
config/app that's killing your battery life.  Please report back so 
others can learn as well.

Thanks,
Mike
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-07 Thread Mikhail Sobolev
On Mon, May 07, 2007 at 06:56:44PM -0400, Theodore Ts'o wrote:
 I've been noticing that I don't get anywhere near the promised 10 days
 idle battery life on my N800.  With bluetooth and the wireless turned
 off (offline mode), and with the display off and keys/touchscreen
 locked, I'm getting at best 8-10 hours of idle time before the battery
 is completely drained.  
One possibility is if you use RD mode or some other [non-default]
options changed with flasher application.  In this case power management
would not be able to deliver the promised time.

--
Misha


signature.asc
Description: Digital signature
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users


Re: Accuracy of the power meter?

2007-05-07 Thread Igor Stoppa
Hi,
On Mon, 2007-05-07 at 23:25 -0400, ext Mike Lococo wrote:
  I've been noticing that I don't get anywhere near the promised 10 days
  idle battery life on my N800.  With bluetooth and the wireless turned
  off (offline mode), and with the display off and keys/touchscreen
  locked, I'm getting at best 8-10 hours of idle time before the battery
  is completely drained.  
 
 That's definitely weird.  I don't get 10 days between charges, but then 
 my N800 is never idle for 10 days at a time.  I commonly run 3 days 
 between charges, though, and I'm a fairly active user.

Yes, apparently it's somehow problematic to convey the difference between
idle and almost idle.

To give a rough idea, the power used by the device while being running
can be _at_least_ 10 times higher, with peaks in the order of a couple
of hundred times.

So it's very hard if possible at all to define what is your expected
runtime apart from few selected use cases that are using for marketing
the device and that we refer to when measuring the power performance.

Of course our use-cases are tailored to cover most of the common uses,
but it's easy for a power user to come up with some combinatioin of
software/use that doesn't allow for proper power saving.

  I have two theories why at this point.  The first is that the battery
  has gone south, even though it I've only had the N800 for about a month
  or os.

I have seen it happening somehow only to 3 years old batteries from the
early 770 era. But that's only my personal experience.

 The second is that I'm running off a 2gig flash card as my root
  filesystem instead of the internal flash, per the instructions given in
  the Maemo wiki.
 
 I don't do this.  If you don't get a more helpful response, I would run 
 with a stock fs and no applications installed for a day (no fun, but an 
 important troubleshooting step if you're not sure where the problem 
 lies). 

MMC / SD shouldn't be an issue. Actually the throughput could/should be
higher because jffs2 does dynamic compression/decompression while the SD
doesn't. Of course there could be a bug in the driver, but history has
usually shown that it's some application/service/daemon that is run too
often or that does busy loop or similar nasty things.

 If you still have low standby life in offline/screenlocked mode 
 then you definitely have hardware issues.  If not, then you can start 
 introducing variables in a controlled way until you pin down the rogue 
 config/app that's killing your battery life.  Please report back so 
 others can learn as well.

Yes, that's the only certain way. My advice is to start removing first
those apps that are likely to introduce periodic activity.
 
-- 
Cheers, Igor

Igor Stoppa [EMAIL PROTECTED]
(Nokia Multimedia - CP - OSSO / Helsinki, Finland)
___
maemo-users mailing list
maemo-users@maemo.org
https://maemo.org/mailman/listinfo/maemo-users