PowerBook energy usage

Kynan Shook kshook at cae.wisc.edu
Sun May 16 01:43:15 PDT 2004


Well, I just thought some people on this list might be interested in 
how much power they're using by doing certain things, especially for 
those extremely concerned about making their battery last as long as 
possible before they need to recharge; I borrowed a portable energy 
meter from the local library, and I figured out approximately how much 
power various features use on my 12 and 17" PowerBooks.  Now, I'm not a 
total geek - the energy meter was actually intended for use determining 
how much power all my roommate's networking equipment is using.  We're 
probably the only apartment in the world (and almost certainly the 
smallest and crappiest - this is student housing, after all) using a 
Cisco 5000 router instead of an Ethernet hub.  He actually has a 4-post 
rack with that and a number of other devices (a few routers, a few 
switches, some fiber optic cable running here and there...).

Anyway, if anybody else decides to do this for whatever reason, your 
results WILL be different.  Also, the power usage jumps around a LOT in 
general; the meter updates the display about every quarter second or 
so, and it would fluctuate by anywhere from a few tenths of a Watt to 5 
or more Watts, just as normal operation of the computer as the CPU 
turns itself on and off, and the disk cache gets flushed, and all that. 
  I did a fair amount of averaging in my head, and a little bit of 
guessing.  Configurations are as follows:
17"/1 GHz/60 GB/1 GB (55 Watt-hour battery)
12"/1 GHz/80 GB/768 MB/SuperDrive (47 Watt-hour battery)
Approximate power use for various things, and which of the computers 
the measurements apply to:

12":
Airport: 0.8 W
Bluetooth: around 0.3 W
Optical drive spinning at low speed (versus off): 3 W
Optical drive spinning at high speed (versus off): 4 W
Hard drive (idle, versus off): 0.8 W
Duplicating a large file: 2-4 W more than the hard drive at idle.
Duplicating many small files: 3-5 W more than the HD at idle (more CPU 
usage).
iTunes visuals: 7 W
Backlight dimmed (via brightness keys) versus display off (via Energy 
Saver): 1 W
Backlight on lowest setting versus backlight dimmed: 1 W
Backlight on highest setting versus dimmed: 4 W
Minimum power consumption: HD spinning but idle, backlight at minimum, 
Airport and Bluetooth off, minimal CPU and GPU usage: 12 W total for 
the whole computer.

17":
Airport: 0.8 W
Backlight dimmed versus display off: 1.5 W
Backlight dimmed to lowest brightness setting: 1.5 W
Backlight dimmed to highest brightness setting: 8 W
CPU at 100% versus ~3%, full speed (1 GHz): 11-12 W
CPU at 100%, full-speed versus reduced speed (1 GHz versus 667 MHz): 8 W
CPU at ~3%, full speed versus reduced speed: minimal difference (this 
doesn't surprise me too much; the CPU shuts off any parts of itself 
that aren't being used; changing the clock speed of something that's 
taking power isn't going to make it take any less power.  ;-))
Keyboard backlight: no noticeable difference
Minimum power consumption: HD spinning but idle, Airport and Bluetooth 
off, backlight on at lowest brightness, minimal CPU and GPU usage: 16 W 
for the whole computer.
Typical power consumption: HD reading and writing occasionally, Airport 
on, backlight at about half, moderate CPU usage, etc: 20-35 W.
Maximum power consumption: battery at fast charge (other sources of 
power drain, eg display brightness, just take power that would 
otherwise go to the battery, the total power usage doesn't change at 
all): 60 W

Both: the fan being on or off did not make a large difference in power 
usage.  Of course, if the fan is on, that probably means that the CPU 
has been under moderate to heavy use, and so it's an indicator of high 
power consumption.  That 11 extra Watts that your computer uses while 
very active would nearly halve your battery life.  And if you have the 
backlight on full, you would chop off a fair amount of time, too.

It should be noted that these measurements (especially the ones for the 
whole computer) are NOT the same as what they'd be while on battery 
power: the power adapter takes some power, and some things are affected 
by a power change that can't be simulated on AC (some circuitry isn't 
used on battery, the speakers and sound circuitry won't turn off while 
on AC, etc.).  The 16 W minimal usable power drain would likely be 
several Watts less while on battery power, but I don't have an easy way 
of testing that.

Anyway, hope somebody else finds this interesting; you can see how much 
faster your battery will drain at the highest brightness, certainly!



More information about the Titanium mailing list