HumaniNet Home Page Summary HumaniNet Partners and Supporters HumaniNet Leadership Operational Scenario Frequently Asked Questions
ICT Features
ICT Articles
HumaniNet Consortium
How You Can Help
Donate Now

How much power does your laptop need?
By Larry Bentley

Editor's note: we are pleased to present this article by Larry Bentley, a HumaniNet volunteer in Plano, Texas who is leading our new mobile power information sharing cluster. Larry is a senior electrical engineer with Verizon and has extensive experience in power and energy conservation. He has won an award from the U.S. Department of Energy and has worked in support of projects in India as well as hurricane responses in 2003-2005. He welcomes questions and comments at larry.bentley(at)engineer.com.

Note - this is not Larry in the picture, but one of our field partners in Africa, and his assistant!

So how much power does that laptop really draw, and what can you do to reduce the power draw for longer life on battery? Or if you are operating on portable power, (solar, wind, battery, etc.), what can you do to minimize power consumption from the limited supply?

Estimating laptop power consumption .

Well, you just look at the rated power draw printed on the laptop, right? Volts multiplied by amps gives watts. So, read the power draw label and you know what the laptop will draw in operation. Not quite! The manufacturers sizes the power supply to be able to handle the worst case in their tests with some safety reserve. The reserve may vary by manufacturer. Now worst case is the laptop doing the most power intense operation, WHILE recharging a fully discharged battery.

Some manufacturers may limit recharge during operation to a slower rate, lowering the needed power rating. Since many laptops can recharge the battery faster than the laptops consumes the power from the battery, often the power supply is at least twice as large as the actual power draw of the laptop. Secondly, even thought we think of solid state equipment as having a flat power draw, laptops power draw varies almost constantly.

Measuring Laptop Power Draws

I placed a digital amp meter in series with the DC side of the AC power supply for a couple of laptops and made the following observations. These won't be iron clad rules for EVERY laptop but should give you general guidance about how much power the laptop draws in various operations. These numbers are at best an average over a continuously varying range. Boot up reading are peak values observed during the boot cycle. I'd say no better than +/-10% accuracy, but the differences between operating states is where power can be saved.

First, screen size affects power, but with only two laptops measured and with different models at different CPU speeds and hard drives we cannot give exact proportions, but the larger the screen the larger the power draw is going to be, just from the backlight. Obviously hard drive usage and CD operation draw more power, less obviously, audio output, WiFi use, recharging the internal battery all increase power draw.

In my measurements, recharging a partially discharged battery with the computer OFF, still drew about 20% more power draw than the highest noted operational power draw. It was noted that the boot operation draws an almost equal power draw to the operation of a diagnostic software package. This makes sense, the boot process tests memory, loads memory, runs hard drive and CD ROM, just as a diagnostic load program does.

The lesson from this is: if the machine can be kept in standby, very little power is drawn, but by timing the boot process you can estimate how many minutes of standby power will equal to the power draw of a whole boot process. Leaving it in standby for three days to avoid the power draw of a reboot isn't going to save battery power. Another option, if your OS and laptop BIOS support it, is hibernation. The data is written to the hard drive and loads back when the machine is powered up. The power draw during this appears about the same as a boot process, although it is much shorter.

For all these power readings, normal desktop screen, with no deliberate activity was used as a baseline for operational power draw. The Delta Watts column shows how much extra power is used (or saved if -) for that condition. Normal typing into a program didn't affect power draw very much, (lost in normal variances) similarly loading a webpage or changing screen display info aren't major power impacts.

The first laptop tested was a Panasonic Toughbook CF-27 300Mhz, 12” screen, 192 meg memory, 20 GB hard drive, CD-RW

 

 

 

 

STATE

Amps

Watts

Delta Watts.

Recharge

1.79

28.96

 

Charged

0.01

0.16

 

Below readings with Full Charged Battery

 

 

 

Boot w/Floppy

1.4

22.65

11.81

Boot w/CD

1.43

23.14

12.30

Steady screen

0.67

10.84

0.00

Screen OFF

0.42

6.80

-4.04

Dim Screen

0.54

8.74

-2.10

Bright Screen

0.89

14.40

3.56

HD Copying

1.07

17.31

6.47

CD Audio play

0.95

15.37

4.53

CD Audio no speaker

0.88

14.24

3.40

Run Diagnostic -HD

1.37

22.17

11.33

Diagnostic -CPU

1.13

18.28

7.44

WiFi Search

1.22

19.74

8.90

USB Stick

0.7

11.33

0.49

Read floppy

1.22

19.74

8.90

Standby

0.01

0.16

-10.68

Play MP3 speakers

0.75

12.14

1.30

Power Draw Rated

3.35

60

 

Here we see the recharge is about 29 watts and the most power consumption we see for the laptop operation is about 23 watts. The total of 52 watts matches pretty well with the 60 watt rated power draw printed on the laptop.

The AC power supply was measured with a “Kill A Watt” device and showed about 41 VA. The much higher VA number would impact inverter operation. For those who understand AC power factor, the non-OEM power supply (older design) showed a power factor of 0.56. This is very low and again would point toward the option of using DC power directly from the low voltage system to power the laptop operations. Again, these measurements with two separate devices at two separate points (AC and DC draw) involve both AC to DC losses and random calibration errors between the devices. Newer design laptop designs have improved power factor readings and would impact an inverter system with lower VA draw.

Similar numbers with an IBM T-22 Thinkpad, 900Mhz, 384 meg memory, 20 GB hard drive, DVD/RW with 14.1" screen shows:

 

IBM T-22

14.1" screen

 

State

Amps

WATTS

Delta Watts.

Standby Full Charge

0.03

0.486

-10.854

Boot - Charged

2.15

34.83

23.49

Splash Screen

0.7

11.34

0

HD Copy

1.2

19.44

8.1

Typing

0.8

12.96

1.62

Bright Screen

0.9

14.58

3.24

Dim

0.6

9.72

-1.62

Play Audio CD

1.45

23.49

12.15

CD & HD active

2.2

35.64

24.3

Recharging Battery

1.68

27.2

 

Power Draw Rated

3.36

54

 

OK, from these we can see some items that provide power saving.

Dimming the screen to the lowest comfortable level saves power. I got both machines, with comfortable screen reading levels, with power draws 3-4 watts below full brightness. Since this a over 25% of normal power operational draw, that is a big factor in total power consumption Toggling the display screen OFF will also save power in idle use periods. The normal power management tools in the OS, that power down hard drive after X seconds, and screen after Y seconds, all will save power when it is limited. Placing the machine into standby, if you aren't going to use it for a while, saves more power. But with a boot cycle of about 90 seconds and realizing the amp draw readings taken during boot were PEAK values, we can see that for the IBM T-22 standby for about 100 minutes is going to draw more power than a reboot, and for the Panasonic Toughbook CF-27 it is about 200 minutes of standby versus a reboot. Again since boot power draw runs a wide range, the actual breakeven times will be lower than these calculations, since they are based on peak amp draw seen during bootup. In reality, it appears a reboot wins if you are going to be away much over an hour. One commenter suggested using hibernation.

Other impacts. WiFi -Turn if off if you aren't using it, it can draw another 3-8 watts, especially in searching mode. This can be done with a switch on some units, on other models, it requires a software access generally via “Control Panel”, or right clicking on the WiFi toolbar icon. Audio over a headphone will draw a little less power than using the speakers. Playing a CD appears to draw a little more power than playing MP-3s from hard drive, since CD motor also operates. Remember some operating systems spin up a CD ROM drive periodically to check for the presence of a CD. If yours does, see if it stops once you leave a CD in place. There may be ways to stop this operation, but you would have to check your PC user guide. The USB thumb drives draw very little power and really then only during read/write access, yes there is a very small amount of power to light the LED on the drive but that is below the measurement threshold. A USB powered white LED flexible keyboard was also tried but it drew so little power it was lost in the normal variations of power draw.

If you are using standard AC power supply, expect the combination laptop to consume AT LEAST 15% more than these DC measured numbers and at least 10% more if you are using a DC-DC converter, just due to losses in the conversion process. These numbers are about the best that can be obtained and some model’s power supplies may be even LESS efficient (uses more power). These numbers are on top of any losses to drive an AC inverter from the portable power or battery source. Remember all the manufacturer claims for DC-DC, AC-DC and AC inverter efficiency numbers are typically based on full rated load. At lower loads the losses form a higher percentage than at full load.

If the laptop’s internal battery no longer holds any useful charge, take it out. If it doesn't come up to normal voltage, the laptop’s internal power supply will keep drawing extra power to push into it, that can more than double the power draw on the AC adapter or even from a DC-DC adapter. That power winds up as heat, not useful computing power.

Also check you laptop’s INTERNAL BATTERY for voltage markings, if it shows under 12-13 volts, there is a very good chance you can run the laptop direct from the DC power source, just wire a suitable power plug to match the polarity, pull the internal battery out and power it up via the normal power socket. The laptops own internal power converter takes the rated voltage and converts it down to charge the battery and supply the needed voltages (typically 5 and 3.3 volts) to the various internal components. I have successfully run several laptops rated at up to 16 volts on 12-13 volts from standard lead acid car battery systems using that method.

By the way, a “12 volt” car battery isn’t exactly 12 volts. At full charge is should show 12.6 to 12.7 range and in a vehicle the alternator charges it to about 13.2-13.5 volts or so. So that cigarette lighter plug will see above 12 volts when the vehicle is running and 12.6 or so with it off and lower as the battery discharges. Also be aware some heavy vehicles ( buses and trucks, and some light aircraft) have 24 volt systems. So don’t fry your gear with those, check it first with a voltmeter.

If you have an AC power supply that fails, SAVE the laptop cord power connector end. If you have the option, try to buy a laptop that uses a standard available type power connector. Proprietary plugs are very hard to come by to make direct DC cords.

Here are some last minute AC power draws using the Kill A Watt metering device:

Dell Latitude D600 – 1.6GHz, 512 memory, 30 GB drive, 14.1” screen, in a mini dock station. AC Power supply is rated as - 19.6 volts 4.62 amps or 90 watts.

Normal display – 26 Watts Dimmer (usable) display – 21 watts,

In SUSPEND Mode 7 watts (this is probably battery charger power)

During wakeup from suspend mode power draw peaked at about 34 watts.

This modern design power supply showed a power factor of 0.98, and should be more representative of modern power factor corrected power supplies shipped with newer computers.

As an aside, a also tested a Motorola cell phone charger with the Kill A Watt device.

Charger is rated at 4.6 V @0.775 A, and it draws about 4 watts as indicated.

Hopefully these numbers will give you useful information in better use of your power limited systems for laptop and small DC charging devices. If we are Humaninet can provide additional help, please let us know.

Larry Bentley

----------------------------------------------------------------------------

Now after I did these experiments, I found a few websites with similar studies:

http://www.crhc.uiuc.edu/~mahesri/classes/project_report_cs497yyz.pdf

http://www.crhc.uiuc.edu/~mahesri/research/PACS_paper.pdf

These measured power draw on an IBM R40 laptop.

Apple fans here you go:

http://www.girr.org/mac_stuff/laptop_power.html

http://staff.science.uva.nl/~delaat/powertest.html

Using OS PowerManagement settings:

http://www.microsoft.com/atwork/stayconnected/battery.mspx

http://www.microsoft.com/windowsxp/using/setup/learnmore/russel_02march25.mspx

Other Power consumption notes. Most of these measure AC power to the adapter, rather than the direct DC reading I took, so they will show slightly higher power draw numbers due to losses in the AC-DC power supply conversion.

http://www.codinghorror.com/blog/archives/000562.html

http://cweiske.de/howto-laptop/ch13.html

Impact of WiFi cards and devices:

http://www3.intel.com/cd/ids/developer/asmo-na/eng/dc/mobile/333927.htm

http://www.fieldlines.com/story/2004/10/10/173020/87

http://www.thinkwiki.org/wiki/How_to_reduce_power_consumption

http://neasia.nikkeibp.com/archive_magazine/nea/200312/inst_278995.php

===============================================================

OLPC - http://wiki.laptop.org/go/Battery_and_power

* Power: 2-pin DC-input, 10 to 25 V, -23 to -10 V

* Line output: Standard 3.5mm 3-pin switched stereo audio jack

* Microphone: Standard 3.5mm 2-pin switched mono microphone jack; selectable sensor-input mode

* Expansion: 3 Type-A USB-2.0 connectors; SD Card slot

* Maximum power: 500 mA (total)

http://www.righthandeng.com/docs/rhe_pcpower.pdf

“Power: How much power are we talking about, exactly? A FAQ estimates milliamps, but that is not a power. How many watts, typical?”

“ Figures from the hardware specification page

Capacity, 22.8 Watt-hours; 6 volts; maximum power 500 mA; so 3 watts for 7+ hours.--Ed Cherlin 2006/09/14 21:42 GMT - 07:00 “



 


| Home | Summary | Partners and Supporters | Leadership | Operational Scenario | FAQ |
Designed by One Dog Designs