[Olpc-sysadmin] Fwd: [support-gang] The great battery experiment

[Henry Edward Hardy]

FYI:

---------- Forwarded message ----------
From: Richard A. Smith <richard at laptop.org>
Date: Wed, Oct 15, 2008 at 3:56 AM
Subject: [support-gang] The great battery experiment
To: OLPC Developer's List <devel at laptop.org>, sugar <sugar at laptop.org>,
Support Gangsters <support-gang at laptop.org>, OLPC Community Support <
community-support at lists.laptop.org>


As I reported in my report that went to community news I'm looking for
power use and battery life information.

Here's a modified version of what I sent in for community news:

========
Richard would like to call for help from the G1G1, developer, and
testing community.  Gathering large amount of data to establish a good
power baseline is a very time consuming task.  Repeated power runs tests
take  around 6 hours to run.  Richard would like to request that all the
developers, G1G1 users, and testers that are active do a power testing
run and send the resulting log file to Richard.

If a large amount of people run 1 or 2 logs then Richard will have a
wealth of data to process.

The steps:  (Please follow these steps as close as possible so the data
is apples to apples)

1) Download the latest olpc-pwr-log:

 wget http://dev.laptop.org/~rsmith/olpc-pwr-log<http://dev.laptop.org/%7Ersmith/olpc-pwr-log>

2) Copy this into your XO where you can run it from a VT.

VT in this case means Virtual Terminal.  A VT is the text console that
you get to by pressing the Ctrl-Alt-F1 key.  F1 in the XO case is the
mesh key which is to the right of the magnifying glass.  For more
information on the VT:

http://wiki.laptop.org/go/Virtual_terminal

3) Fully charge the battery of your XO and leave the external power
plugged in.

3a) Please remove _all_ external devices connected to the XO.  This
includes SD cards, USB keyboards, USB disks, etc.  External devices will
 skew the power readings and create situations where the CPU will not
go idle.
I'm not really concerned with what your network is doing.  Thats one of
the things I want to see is how much variability there is in the
readings.  So association to an access point or running mesh is fine.
However, please don't disable the WLAN with the extreme power magement
seting.

If you wish to record a brief note on what you had the network doing you
can do so by adding a message after './olpc-pwr-log' It will be copied
into the log file as a note.  See 5)

4) Switch to a VT and cd to where the olpc-pwr-log script is.  Running
from a VT is very important.  The output on the VT console will prevent
the system from going into idle-suspend if you have that enabled.

5) ./olpc-pwr-log [message]

[message] is any descriptive text you want to provide about this log.
Useful info would be if you were connected to an AP or in the mesh.  Its
not required though.

6) The moment you see the first line of log output (ie right after it
says it found a battery and a series of numbers displays) pull the
external power.  The status should go from 'Full' to  'Discharging' on
the next print which should be in about 15 seconds.

7) Let the system run untouched until it dies.  Do not switch back into
sugar.  If you go back into sugar the DPMS power saving will kick in in
20 minutes and turn off the dcon.

After a while of inactivity the console screen blanker will kick in and
the text on the screen will go a bit screwy. Don't worry its a known
bug.  Just ignore it.  If however your screen goes dark and gray (No
backlight and looks just like the powered is off) then DPMS screen saver
has kicked in and turned off your DCON.  Please review the earlier steps
and try to figure out why DPMS activated.  If you have to you can
disable DPMS with by using xset:

 "xset -dpms"

8) Mail the logs or any questions you have to richard at laptop.org or find
me on IRC in the devel room.

--
Richard Smith  <richard at laptop.org>
One Laptop Per Child
_______________________________________________
support-gang mailing list
support-gang at lists.laptop.org
http://lists.laptop.org/listinfo/support-gang



-- 
The best material model of a cat is another, or preferably the same, cat.
--Norbert Wiener, Philosophy of Science, 1945
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/olpc-sysadmin/attachments/20081015/7720ee67/attachment.htm