Power Consumption

They say (IDK who, really), that leaving your PC on all day every day for the entire year (24/365) would consume more energy (and cause more CO2) than a car. My FJ Cruiser (not the most efficient vehicle in the world) costs about $9 a day to commute to and from work. So, that's $2340 in fuel every year. A rough calc. I did not remove vacation days for not commuting, but I also did not add in weekend car usage or car usage over the vacation. It averages out - and is 12480 miles - not to far from the typical 12k/year average estimate.

So, me, who recently had as many as 8 desktops (6 winXP, 1 osx, 1 Linux) and 2 laptops (1 osX, 1 winXP) running, that should be (assuming laptops take 1/2 power usage) $21,060 in electric bills. I'm pretty sure I would have noticed that. And, over a decade ago when I used to host lan parties at my prior house, I had 50 PCs running for over 12 hours one day. Old PCs, the power consuming CRTs and graphic intense applications. I took the Kwh reading before, and after. I don't have the numbers any more, but the calc at the time was that it cost well under $10 for the power for the party.

But, that was then. Now, PCs have bigger power supplies - 500-700 watts on average versus 200-250 typical back then. And, cost per Kwh has also increased. So I decided to collect some data and do a little math. How much are my PCs costing me in dollars per year. I'd check the power bill for Kwh prices, and I have a spiffy Greenlee AC ammeter. How hard could it be?

Hard enough. First issue was the bill. Tiered pricing. Add to that, tiered and non tiered surcharges. Add to that, the prices change in the tiers each month. So, I figured I'd just use one month's pricing and go with that (inaccurate, but should still provide useful data). OK, so, for that month, the cheapest price per Kwh was 3.9 cents. The largest was 9.6 cents. The cheapest 'bonus' charge was .3 cents, while the most expensive was 3.4 cents. I crunched some numbers in my head (OK, I didn't crunch any, I just lightly bruised some). I came up with 8 cents per Kwh as a working average.

Next issue. My Greenlee meter was producing numbers that I thought were too low. A lot too low. A 40 watt bulb was coming up at .17 amps. A 500 watt halogen light was coming up at 2.33 to 2.85 amps. How does one know how accurate a meter really is?

Final issue: How to determine the actual power consumption. When is a device hibernating, when is it on, but not using a lot of CPU? When is it using more power for more intense apps? I chose not to answer that completely, and just chose to take readings in the different modes. hibernate/off, up/idle, and running 3 browsers, each with a different youTube video running.

I did not measure every device. Here are some of the amp figures:

MacBook: hibernate = 0.01; on = 0.15; loaded = 0.3

clock radio: 0.01

Dell XPS dual core: off = 0.04; on = 0.69; loaded = .9
Acer 24" LCD: off = 0.02; on = 0.27

NAS Storage device: (off switch did not do anything) on = 0.08;
accessing data on it = 0.1.

8 port gigabit switch: 0.01

1500 watt heater: low = 4.15; high = 6.9.

Knowing the dubious nature of the numbers thus far, and wondering what number to use for volts (110? 115? 120?), I settled for an average of 115 for yet another dubious number. 720 hours per month (used a 30 day month).

Dubious results (usage for a month):

Macbook: Hibernate = 6.6 cents; on = 99 cents; loaded = $1.98.

Dell XPS: off = 26.5 cents; on = $4.57; loaded = $5.96.
LCD Monitor: off = 13 cents; on $1.78.

Not a lot. But I figure those are low ball numbers (due mostly to the low ball readings).

But, they add up. 8 PCs running 24/7/52 (with monitors on all the time) = $617.

I now have 2 PCs and 1 MacBook (mostly hibernating) up all the time, with the occasional 1 or 2 more running as needed.

Still need to figure out what the deal is with the Greenlee readings...