The Cost of Running Your PC
by Christoph Katzer on November 14, 2008 3:00 AM EST- Posted in
- Cases/Cooling/PSUs
Using a Higher Efficiency PSU to Reduce Costs
These days manufacturers are all promoting their high efficiency power supplies, and we have organizations and certifications like 80 Plus encouraging even small boosts in efficiency. Not surprisingly, plenty of users have been sucked in by the marketing and are now convinced that they need to purchase a new power supply in order to save money each year. Does it really make that much of a difference? The answer as usual depends on how you use your system. The previous page provided a baseline measurement, but now let's look at how much money you can save if you go out and purchase a new 80 Plus Bronze or Silver certified power supply as an upgrade to a slightly older ~80% efficiency PSU.
Our sample power supplies on the previous page are all relatively high-end choices for the specific market. Many (most) systems don't have power supplies anywhere near that nice, relatively speaking. So what happens when we switch to an older ATX 1.3 PSU -- something that would have been more or less state-of-the-art three years ago? Will a newer power supply really help you save the planet? Will it at least reduce your power costs and save you money? Let's find out, this time looking at power costs over the course of a full year: 24 hours a day, seven days a week.
For reference, we looked at some PSU efficiency results stashed away in our files and estimated ATX1.3 PSU efficiency at 75% idle and 78% load. That represents a high-end ATX1.3 PSU, and in some cases the discussion is hypothetical as it wouldn't be possible to find an older PSU with the necessary output rating. (That applies specifically to the high-end system.)
| System 1 24/7 Yearly Costs ATX1.3 versus ATX2.2 |
||||||
| Outlet Power ATX v2.2 |
Outlet Power ATX v1.3 |
Wattage Difference |
Savings NC | Savings CA | Savings GER | |
| Idle | 110 | 120 | 10 | $6.57 | $11.21 | €19.27 ($25.05) |
| Load | 167 | 179 | 12 | $7.88 | $13.46 | €23.13 ($30.06) |
| System 2 24/7 Yearly Costs ATX1.3 versus ATX2.2 |
||||||
| Outlet Power ATX v2.2 |
Outlet Power ATX v1.3 |
Wattage Difference |
Savings NC | Savings CA | Savings GER | |
| Idle | 190 | 213 | 23 | $15.11 | $25.79 | €44.33 ($57.62) |
| Load | 412 | 449 | 37 | $24.31 | $41.49 | €71.31 ($92.70) |
| System 3 24/7 Yearly Costs ATX1.3 versus ATX2.2 |
||||||
| Outlet Power ATX v2.2 |
Outlet Power ATX v1.3 |
Wattage Difference |
Savings NC | Savings CA | Savings GER | |
| Idle | 369 | 413 | 44 | $28.91 | $49.34 | €84.80 ($110.24) |
| Load | 663 | 705 | 42 | $27.59 | $47.09 | €80.94 ($105.23) |
Now we can see exactly how much money you might save during the course of a year by purchasing a new high efficiency power supply. Obviously, the more power your computer uses, the better your monetary savings. Looking at these tables, you might begin to think there's actually a point in upgrading power supplies -- and there is, provided you're running your computer a large portion of the time.
What happens if we change our usage model to something more realistic for most families? Instead of looking at 24/7 usage, let's change it to three hours of use per day on average, with two hours at idle and one hour at load.
| Yearly Power Savings for 3 Hrs/Day | |||
| Savings NC | Savings CA | Savings GER | |
| System 1 | $0.88 | $1.50 | €2.57 ($3.34) |
| System 2 | $2.27 | $3.88 | €6.66 ($8.66) |
| System 3 | $3.56 | $6.07 | €10.44 ($13.57) |
The need to upgrade power supplies suddenly doesn't seem as dire once we switch to a more realistic usage model. Particularly on low-end systems that only use 100W of power give or take, even an extremely inefficient PSU probably doesn't matter too much if the system isn't on more than a few hours per day. Even with power costs that are up to three times higher in some parts of Europe compared to areas in the US, the savings don't make sense.
If you happen to be the type of user that leaves your system on all the time, certainly you can save a fair amount of money by purchasing a better power supply. An easier solution would simply be to turn off your computer when it's not in use, unless you have a really good reason to leave it running overnight. Similarly, if your current PSU happens to fail, it might be worthwhile to spend a little bit more money to get a higher efficiency, better quality power supply. If you figure on a moderate amount of use and a five-year lifespan, you might want to spend as much as $50-$100 extra. Otherwise, there's very little incentive to go out and spend $150 on a top quality power supply just so you can save $10-$15 per year (or less).
Facebook
Twitter
RSS
59 Comments
View All Comments
Kyanzes - Friday, November 14, 2008 - link
I've kind of anticipated a calculator but still a nice read.JarredWalton - Saturday, November 15, 2008 - link
Isn't that what the spreadsheet is?vandaliser - Friday, November 14, 2008 - link
All you had to do is buy a Watt Meter which is kind of like a surge protector (but with a digital reader) where you connects your PC's power plug to the meter, then the meter to the main. (just go to ebay search Energy Meter and you will know what I'm on about)Take the reading in watts, divide it by 1000 to gives you the number of kwph. Finally, multiply it by the cost of one kwph on your electricity bill and numbers of hours you want to run it for.
I'm not sure about their expected cost of running, but it actually surprises many people that their PC uses a lot less power then what they expects.
Griswold - Friday, November 14, 2008 - link
Just that truly el-cheapo equipment will give you horribly wrong readings (cos-phi anyone?). Not saying a "watt-meter" must be expensive to give you acurate readings for home use, but there is way too much junk on the shelves out there.Souka - Friday, November 14, 2008 - link
Buy a Kill-a-watt meter of eBay.... I did years ago, still using it today.It'll show real time Amps, Volts, Watt load, KWhr used, and time.
http://energyseeds.com/2007/10/11/go-solar-and-kil...">http://energyseeds.com/2007/10/11/go-solar-and-kil...
I just pulled it out for a co-worker to try at her home. :)
DeepThought86 - Friday, November 14, 2008 - link
Given how little power even beefy systems consume, why is it that Anandtech continually reviews rediculously overpowered PSUs? What % of the market is made up of those 600W-1000W monsters? How about comprehensive reviews of the 300-500W marketanartik - Tuesday, November 18, 2008 - link
I would have to say that is a common misconception... There are reasons to buy more power than you "need". I bought "extra" for future upgrades and headroom. The problem with the calculator is most people plug in and come to the conclusion they need some fixed amount of power. All power supplies degrade in output over time with the cheaper ones faster (or use misleading claims as to output in the first place). If the calc says you need 400 and you buy 400 you’re in for trouble as the output deteriorates even quicker from running it at full capacity. The more you strain the PS the hotter its going to run and the louder its fans get. Plus you decide to run out and buy the latest power sucking hardware and voila you need a new power supply.I have a 4.3ghz E8500/X48 (SB w/bay,2 sticks DDR2, 2x drives,1 dvd burner and 3x120, 3x140) system and according to the calc I only need 462 with my current OC'd 8800 GTX. My old 550 Antec couldn't hang, screeched harmonics and was replaced with a Corsair HX1000. If I did a worse case upgrade... OC'd Q9550, more HD's, bluray burner and either a single 4870 X2 or possibly two and that power jumps to the range of 650-850 on paper. Factor in overages for peaks, efficiency, deterioration, percent utilization and it ranges from in the ball park to pushing it.
nilepez - Wednesday, November 19, 2008 - link
With all due respect, sites have been pushing large PSUs for years. As I posted earlier, people were trying to convince me I need 600w 3 or 4 years ago, when I built an Athlon 64 Rig with an X800XL: a rig that couldn't not possibly have used 300w, even when overclocked, from the wall, much less from the PSU.As for the idea of what you'll need down the road, by the time you need more PSU (esp due age), you could just buy a new quieter, more efficient PSU, with more bells and whistles of equal or higher quality with the money you saved.
Besides, in 20 years of computing, I've never had a PSU die. The worst thing that happened was a fan died. Bought a new fan and it worked like a champ, and that was some POS PSU that came with my Inwin Case (I think I still use that PSU, 10 years later!).
buying a quality PSU makes sense. Buying 750w+ PSUs only makes sense for someone running Tri or quad SLI, which means almost nobody. I've seen developers at work return 600W PSUs, because they feared that they'd need more to run to 8800GTs.
Those 2 cards pull at most 160w...add in a Core2 CPU, and you're looking a rig that is unlikely to pull 300W while playing far cry with Super Pi (just in case there's an idle cycle) running in the background.
It's almost all marketing hype.
Griswold - Friday, November 14, 2008 - link
I've been saying this since Chris' first (excellent!) review here at AT. I really wish he would push those insane power monsters with extra bling off his workbench and start reviewing those PSUs the majority actually buys.Christoph Katzer - Friday, November 14, 2008 - link
Next one up will be most probably the Thermaltake TR2 QFan series with 300, 350, 400, and 450W. Everyone cheer up! ;)