Can you save the planet by switching to a "black" Google?

Over the past few weeks I've come across dozens of websites promoting the idea that switching over to using a version of Google that had a black background as opposed to a white one would save me money/electricity/the planet^* (delete as appropriate). 

The alternative search site most often linked to is one called Blackle, a website that offers visitors the ability to use Google via an interface consisting of a black background and grey font.  I've had dozens of emails from people suggesting that I switch to this search site, dozens from people wanting my opinion on it and dozens more from people wanting to know if switching to it can help them save the planet.

I decided to find out for myself what the energy savings, if any, are likely to be and if switching all my searches to Blackle really would help me save the planet.

Next -->

The Test

Over at the PC Doc HQ we came up with two very different approaches to testing this energy savings. 

• Go large scale The first way was to go large scale.  Hook up half a dozen monitors together, have them display mostly black for a fixed period of time (a week say) and then switch them over to white and see which use the most power.
• Go small scale The alternative method was to go small scale.  Take one LCD panel and one CRT panel and do some testing over a few hours and see what the results suggested. 

In the end we decided to go for small scale testing.  Not only would we get quicker results but it really didn't feel like chucking kilowatts of power at trying to find out what was the best method for saving electricity. And anyway, if there's any substance to this black background hypothesis, it should be easy to spot.

So here's what we decided to use for the test:

• One generic 19 inch LCD panel.  This was purchased within the last 12 months.
• One generic 15 inch CRT unit.  This has been hanging around the PC Doc HQ for some time now.
• One power meter capable of measuring, amongst other things, power usage in Watts.

For the test we would measure the power consumption of both the flat panel and the CRT when displaying a variety of different images.

The first thing that we needed was a baseline figure for power usage for both the LCD and the CRT.  The LCD panel was fitted to a PC running Windows Vista with the default desktop while the CRT was on a system running Ubuntu, again with the default desktop. 

Baseline power usage:

• LCD - 28 Watts
• CRT - 57 Watts

So, right off the bat we find that the small 15 inch CRT is consuming twice as much power as the much larger LCD panel.  But what variation does changing the background color have on these baseline figures.

First off, we tested the LCD panel.  We set it up to display first a completely white screen and then a completely black screen (by doing nothing more sophisticated than changing the background desktop color and hiding the icons).  Here are the results:

• LCD displaying black screen - 28 Watts
• LCD displaying white screen - 28 Watts

As you can see there's no detectable change in power consumption.

Next we repeated the test on the CRT.  This time we saw a pretty dramatic variation:

• CRT displaying black screen - 49 Watts
• CRT displaying white screen - 60 Watts

So running a CRT with the screen totally black uses 14% less power than the baseline measurement, while a completely white screen uses 5% more power. 

Since no one in their right mind is going to run a screen 100% black or white, we decided that some more real world testing was required.

Since both systems were running at 1024x768 screen resolution we fired up a web browser (Internet Explorer on the Windows Vista system and Firefox on Ubuntu, both systems switched back to default themes) and set it to be an 800x600 pixel window and into each loaded in turn Google and Blackle.  Here are the results:

LCD Panel

• Google - 28 Watts
• Blackle- 28 Watts

CRT Monitor (Ubuntu)

• Standard desktop - 56 Watts
• Google - 56 Watts
• Blackle- 53 Watts

Does the operating make any difference?  I switched the CRT to Windows XP system running at 1024x768 and duplicated the test.  Here are the results from that test:

CRT Monitor (Windows XP)

• Standard desktop - 57 Watts
• Google - 60 Watts
• Blackle- 58 Watts

Next -->

Conclusions

OK, so what's the bottom line here?  Well, first off, if you're running an LCD panel then none of this means anything to you.  LCD panels seem to draw the same power no matter what (dimming the display did cause a drop in power consumption but only when dimmed beyond the point of being readable).

If you're running a CRT and are interested in doing your bit to save the planet then one of the best things that you could do is bin the CRT and replace it with an LCD panel.  Not only will you get a better screen, it'll consume half the power.  However, if you don't want to change your monitor, then darkening your screen does have an effect.  Running a CRT with the screen totally black uses 14% less power than the baseline measurement, while a completely white screen uses 5% more power but these numbers are nowhere near as dramatic when you consider a standard user desktop.  Here the difference between using Blackle compared to the baseline is small and we certainly couldn't get power consumption anywhere near as low as 49 Watts (which was the power consumption of the CRT displaying a black image).

We also tested to see whether dimming the CRT had any significant effect.  We found that the difference between running the CRT at full brightness and being barely readable was only 1 Watt.  Measurable but hardly significant.

The bottom line is that if you want to save the planet, there are far better ways to do it that shifting your searches to Blackle. 

Personal Note:  Going through this experiment has made me glad of the fact that over the past few years that we've eliminated almost all CRTs from the PC Doc HQ and shifted over to LCD panels.  We've also taken other steps such as:

• Changed all light bulbs for energy saving ones
• Implemented power saving features on most systems
• Made it easier to switch off the power to devices on standby by hooking them up to a single isolation point

Thoughts?

<< Home >>