Breaking Eggs And Making Omelettes

Topics On Multimedia Technology and Reverse Engineering


Archives:

Monster Netbook Battery

April 23rd, 2010 by Multimedia Mike

I stubbornly refuse to give up my classic Asus Eee PC 701, one of the original netbooks. It’s 2.5 years old now but still serving me well. While these are supposed to be fairly disposable machines, I’m actually using this thing more and more these days (longer commute may have something to do with it). I decided to upgrade the battery from the included one (4400 mAh, rated for 2-2.5 hours). 7200 mAh batteries abounded for this Eee PC model but I decided to go crazy and buy the 10400 mAh battery.

And it’s huge. No one can keep a straight face when gazing upon this beast.



Naturally, I’m curious whether this battery is actually that much better. I searched to find if there are any established methodologies for testing battery life. It seems that the most established method is the most intuitive method, scientifically: Find a way to simulate typical usage and measure how long it takes before the machine dies from lack of battery charge.

Methodology
To that end, here is my testing method:

  • Charge a battery all the way (until the Easy Peasy/Ubuntu battery meter no longer appears)


  • Prevent the machine from sleeping or even blanking its screen; for this, I found the program xdotool which programmatically actuates various X windows actions, like mouse movement
  • On a different machine, run netcat in listening mode
  • Create a script that wakes up periodically and connects via netcat to the other machine and prints a timestamp
  • Unplug power cable and note the starting timestamp
  • Go to bed
  • The next morning, observe the final timestamp transmitted to the other computer

This may seem like a fairly non-stressful test. However, I think it simulates my normal working conditions in which I mostly use the netbook as a glorified portable word processor, typing things like this very blog post whilst riding a train.

This is the Bash script I came up with for testing:

while [ 1 ]
do
  date
  x=$RANDOM
  let "x%=800"
  y=$RANDOM
  let "y%=480"
  xdotool mousemove $x $y
  sleep 30
done | nc <dest-machine> <port-number>

Sit in a loop, print the current timestamp, move the mouse pointer to a random coordinate (800×480 screen), and sleep for 30 seconds.

Raw Data

4400 mAh battery:
1st run: 2h29m
2nd run: 2h31m

10400 mAh battery:
1st run: 5h22m
2nd run: 5h23m

Conclusion
So, yay, I’m glad to learn that the enormous new battery lasts proportionally longer than the original battery. This is especially useful since the battery indicator isn’t that helpful in estimating the charge of the larger battery. I suspect that either the software or the hardware simply isn’t aware that a longer battery life is possible.

Posted in Science Projects | 3 Comments »

3 Responses

  1. Bastien Says:

    Pretty much all the data about the battery life time will come from the battery itself. You can use gnome-power-manager’s stats program to check on the battery life time, and its level of fill over time.

  2. Fruit Says:

    Well, there is a way… but you need to be travelling often/regularly.

    Just take a note of time when you resume/turn the thing on in a train (for example), then do the usual (read stuff, write, should be either cpu-light, or same all the time if not) and then look at your watches when the machine decides to power down due to low battery (if it’s configured to!).

  3. whitetiger Says:

    http://xkcd.com/196/