It has been stated several times in this forum that the energy consumption of computer manufacturing is about 3-4 times greater than the energy consumption at the end-users home/workplace. My initial statement was that I hade a hard time believing that that would be the case for an always on (24/365) computer.
I have now hade a little time to review some of the source material for the above assumptions. (Thanks to MikeC
,also thank you for this great site)Abstract
Full Paper
There are two things in this otherwise good paper I would like to comment on and to have your thoughts on.
1. The data used in this paper are from the late nineties and basically all data and assumption is about 10 years old, the computer industry have changed a great deal the last 10 years and there is no data to indicate in which way the energy consumption have moved. All you can say really is that the energy consumption is somewhere in the same order as in the paper (maybe in the range 2000MJ-10000MJ, display and use phase not counted)
2. The assumption for the user case is really simplistic. The author assumes 3 hour use each day and the rest of the time the computer is unplugged. During use the assumed computer and display would consume 137W (this is based on a paper from 2002). For an always on pc that would mean 8 times longer life time use and while the display would use less since it would go to sleep mode for most of the duration it still dramatically changes the energy balance. if you instead have 90W 24/365 the balance is in the range of 60% of energy consumption during the use phase. (close to 8500MJ)
This doesn’t change the general conclusion that reusing old computers is generally a good idea but for an always on PC energy efficiency does make a big difference. And if you factor in economy it makes a big impact in your wallet if you don't go for an energy efficient solution. Good economy is also key for a sustainable future.