[2009-03-26] Power use of white vs black screens in LCDs and CRTs
Some time ago I read about www.blackle.com, which purports to be an energy-conscious version of the Google home page, using a black background instead of a white one on the argument that computer monitors require less energy to display black pixels than to display white ones. Considering how many computer screens across the world are displaying the Google homepage at any given moment, even a small savings per screen could amount to a lot of conserved energy in the aggregate.
An interesting idea, to be sure, but it didn't quite jive with my understanding of how monitors work. A CRT, certainly, requires energy to light up the phosphors on the inside of its screen, and displaying white requires maximal illumination of all three phosphors (red, green, and blue) at each pixel. CRTs are emissive displays, and stimulating that emission consumes energy. LCDs, on the other hand, are absorptive displays. That is, an LCD display includes a backlight which is always on. The backlight is white, so rendering a white pixel requires no additional energy input. To render a black pixel, however, all three color channel subpixels have to be "opaqued" by applying an electrical potential to the liquid crystal phase. Put simply, CRTs spend more energy to make white, while LCDs spend more energy to make black. Or at least, that's how it seemed like it should be to me. And since almost everyone these days uses LCDs instead of CRTs, www.blackle.com may actually be doing more harm than good.
So I did an easy experiment. Using a Kill-o-Watt meter, I recorded the power consumption of an LCD monitor and a CRT monitor over two periods of 24 hours. During the first period, each monitor displayed an all-black screen, and during the second period each monitor displayed an all-white screen. The LCD in question was a 15" Planar systems model 996-0547-00 that was new in 2004. The CRT was a 17" eView e17 monitor. Its exact date of manufacture is unknown. I used my laptop to generate the input signals. The Kill-o-Watt meter gave total power consumption over each 24 hour period in kilowatt-hours (KWH). Dividing by the exact length of the recording period in each case gave the average power consumption rate of each monitor and each color in watts (W). Normalizing these rates for screen area gave final values in watts per square inch (W/in2).
The results are as follows. The CRT monitor consumed 0.48 W/in2 to display black pixels, and 0.65 W/in2 to display white pixels. Thus, as expected, the CRT requires more (+34%) power to display a white background than a black one. The LCD monitor consumed 0.190 W/in2 to display black pixels and 0.183 W/in2 to display white pixels. Here, again as expected, the color preference is reversed: the LCD requires less (-5%) power to display a white background than a black one. The average power usage of the CRT was 0.57 W/in2, and that of the LCD was 0.187 W/in2. The LCD used 68% less power that the CRT on average.
last modified 2009-03-26