CNET's power testing goes live in desktop reviews

Scroll down to the bottom of our last six all-in-one desktop reviews and you'll find CNET Labs' latest addition to the desktop testing regimen. With the help of the Environmental Protection Agency's EnergyStar program, some impressive-looking power meters, and not a little trial and error, we're happy to report that we've finally published the first batch of power consumption results.

Our power results come courtesty of Chroma ATE's EnergyStar-certified 66202 Digital Power Meter.

(Credit: Rich Brown/CNET)

We have a few goals in mind with our new power tests. We want to help you make more informed buying decisions, for one thing. You'll see an annual energy cost chart in each review that compares the yearly estimated cost to operate a variety of systems. We don't expect the actual dollar amount will influence most of you one way or another (we're only talking about a range from $15 to $30 in this first round), but an annual cost figure also distills the relative efficiency between systems down to straightforward terms. You can also refer to our newly official Juice Box, located above the cost chart, for the individual power ratings across a variety of usage states for each system.

Apple does exceptionally well on our annual energy cost chart.

(Credit: CNET)

Another goal with this testing is to help incent the vendors to take power efficiency seriously. Just like with performance, you can't judge efficiency by the size of a PC, or even the by dimensions of its screen in the case of an all-in-one. We're happy to report that each system we tested falls within the guidelines for EnergyStar's near-official 5.0 specification for computers. Even so, some products, Apple's iMacs in particular, differentiate themselves beyond EnergyStar. We want to be able to highlight those vendors and products that have truly innovated with power efficiency, and also point to those that could improve.

If you're curious about our methodology, you can read about how we power test desktops here. We have a few updates to make to that document, and we also don't spend a lot of time explaining our thinking behind how we test. We've decided to follow the established EnergyStar 5.0 procedures, but with a twist of our own.

Lenovo's IdeaCentre A600 was surprisingly efficient given its large display.

(Credit: CNET)

EnergyStar is useful because it provides vendors with an efficiency baseline, but it doesn't yet include a typical computing workload test. Given the many and varied uses for the computers it certifies, the lack of an official EnergyStar load test isn't surprising. Fortunately, CNET has a mostly consumer-focused audience, which gives us a bit more freedom to focus on a specific workload. In addition to the standard round of Off, Sleep, and Idle EnergyStar testing, we've also added a load test that's basically an extended run of our Multimedia Multitasking performance benchmark.

After we compile the test results, we use our own variation of the EnergyStar equation to get the raw annual kilowatt hours (kWh). From there we simply multiply by the national average cost per kWh, currently $0.1135, to come up with our annual energy cost. We gain some efficiency of our own in this process, in that by gathering the requisite EnergyStar test data before adding our own load tests, we can also run just the basic Off, Sleep, and Idle results through the nonmodified EnergyStar formula to gauge whether a system is indeed EnergyStar 5.0 compliant.

You'll find our current power results in each of the following all-in-one PC reviews, as well as in all desktop reviews moving forward:

• Apple iMac 20-inch
• Apple iMac 24-inch
• Dell Studio One 19
• Lenovo IdeaCentre A600
• Sony Vaio JS250J
• Sony Vaio LV250B

(Credit: CNET)

You'll also see us add the CNET power rating meter, at left, and currently appearing at the top of more recent TV reviews, to the desktop reviews once we expand our power data pool. We'll be following up with power results on our laptop and monitor reviews shortly. And while our power testing doesn't tell a product's entire environmental story, we hope it will give you a new way to think about choosing your next computer.

Additional reading:
CNET Energy Efficiency Guide

[NewsReader_]

The iMac is $400 more than the Lenovo.

Congratulations Rich, it a PC. Go Green and save some green :-)

[kevinm6]

Yeah congrats on getting a computer with the old 2.0 Dual Pentium processor and intel graphics. iMac has 2.66 Core 2 Duo and NVIDIA 9400m. So expect rich to pay another $799 in a year.

[pithenumber]

@kevin
the processor is a Pentium DC [Core based] not a Pentium D [P4 based]
so its not old

but still, the iMac is faster

[rhbrown]

Lenovo has a great showing here, no doubt, and you also a good point about price. The 20-inch iMac is Apple's least expensive all-in-one, but at $1,200 it leaves ample room for more affordable competition.

I'd only point out that while the 20-inch iMac has a smaller screen than the 22-inch Lenovo, the iMac is significantly faster. Windows/Mac stuff aside (a large request, I suppose) there's a clear big picture trade-off between the two. Larger screen/lower price for the Lenovo vs. faster performance/slight efficiency edge for the iMac. I'm not sure there's an outright winner between the two. Depends on your needs.

For a more decisive match-up, check out the Dell Studio One 19 vs. the 20-inch iMac. The iMac is faster, has a larger screen, costs only $175 more than the $1,024 Dell config we reviewed, and uses a lot less power. The Dell's only advantages are its touch screen (a thin lifesaver at best), the fact that you can configure a less expensive model, and its Blu-ray option if you want to pay an extra $150 (although if you want an all-in-one with Blu-ray at that price I'd point you to the Vaio JS250J). At these prices, and with these configs, it seems to me like a pretty clear win for Apple.

[heulenwolf]

Good job Rich and CNET. Comprehensive, reasonable power efficiency ratings have been missing from many review sites for a long time. The only thing I don't understand is why you aren't publishing the equation you use, as in, "We use our own variation of the EnergyStar equation to get the raw annual kilowatt hours (kWh)." Doing so could take your rating from being just a number to being a standard. It also helps make your results verifiable and comparable with 3rd party results. When I see a different power efficiency rating from a competing review site and all I know is that the test equations are different but not what they are, I'm left with more confusion and less information than if both sites hadn't bothered to report power efficiency ratings. Data without context is indistinguishable from noise. Data with "trust us" context requires independent verification. Data in the context of published acquisition methods and standards helps make it trustworthy and quickly becomes information.

[rhbrown]

I'd actually planned to add the equation to our Desktop Power Efficiency Guide today (http://reviews.cnet.com/green-tech/desktop-power-efficiency/). I don't mind putting it up here now, though.

Here's EnergyStar's equation for annual kWh for desktops, in handy Excel format:

(8760/1000)*((Off watts*.55)+(Sleep watts*.05)+(Idle watts*.4))

You can find that formula on page 26 of EnergyStar's 5.0 Computer Specification PDF: http://www.energystar.gov/ia/partners/prod_development/revisions/downloads/computer/Version5.0_Computer_Spec.pdf

8760 represents the number of hours in a year. We divide that by 1000 to get to the annual kilowatt hours once we multiply the result by the wattage.

We use the above formula to figure out whether a system is indeed EnergyStar 5.0 compliant, according to what EnergyStar calls its TEC (typical energy consumption) rating. Each kind of system has a different TEC rating, which further varies depending on the amount of memory, whether it has a discrete graphics card, etc.

Here's our formula, factoring in our custom made load test:
(8760/1000)*((Off watts*.55)+(Sleep watts*.05)+(Idle watts*.35)+(Load watts*.05))

As you can see, we took a bite out of EnergyStar's annual Idle time estimate and added in our load test. We then multiply that figure by the national average cost for kWh ($0.1135), to get our annual cost figure for each system.

For our load test, a variation of our Multimedia Multitasking test, described here (http://reviews.cnet.com/4520-6603_7-5020816-1.html), we don't really think the average consumer will spend 1.2 hours per day (or 438 hours a year) simultaneously converting an HD movie for an iPod while also converting a bunch of MP3s to ACC files. But given the relatively low dollar amounts involved, we felt that modeling the actual energy costs was less important than showing the differences between systems with a reasonably taxing workload that will run on most computers, regardless of whether they use Windows or OS X. The dollar figure is mostly there to provide a rough guide, and you'll likely never see us factor that amount into the price comparison between systems.

Both our load test and the weight we've given it are of course debatable. And the difficulty in coming up with a universal load test in part explains why EnergyStar doesn't have one of its own yet, although we know that's one of EnergyStar's goals. We also plan to come up with a 3D load test for gaming systems, and we can adjust both the test and our formula as necessary if we decide we need to.

Who cares ? I need a computer, I buy one and plug it in, it takes electricity to run it, they send me a bill, I pay it. What's the problem ? So I can save a dollar a month if I buy the energy efficient. My truck gets 7.5 mpg. I will not live long enough to buy a twenty five grand truck and see the savings in gas pay for the new truck.

Climate warming - a joke. Windmills and solar are not going to power my computer anytime soon.

[gadgetplanet]

Its people like you who are just ignorant and can't except anything that makes your life a little less convenient.Global warming is a joke - you are one intelligent ape aren't you.Solar panels and windmills won't power your computer because its not a priority for trailer parks."i want computer,i by computer,pay with money,i pay bill,whats the problem".If you can't see past that than your the problem,seriously where do you guys come from.

[bensagenius]

Whenever I see "green," "go green," "carbon footprint," or "going green," I completely ignore the article or advertisement which features those words. It's all generally just pandering nonsense.

[Dango517]

$12.00 or $30.00 a year is roughly a $15.00 difference. So, big deal!

This might be how many might respond to your analysis. To them I'd say .................. it all adds up! An extra $10.00 here, $50.00 there and before you know it your spending hundreds and thousands more dollars then others on energy. You owe it to yourself to save that hard earned money by going "green".

What I'd really like for you to check is refrigerates. These 24/7/365 ice boxes can consume loads of power and are always on. Big saving can be had by replacing a 15 year old fridge.

Hmmm, what about pilot lights on water heaters and furnaces? Are there any possible saving there or more importantly any alternatives to pilot light?

Good first effort by Cnet.