Power Consumption

Taking in an interest in the environment (as all people should), I decided to monitor the power consumption of as many devices in my house as I could be bothered with. I’m more interested in devices that are on 24-7, but I couldn’t resist doing more for a bit of fun and curiosity.

Rather than buying a power meter, and in the spirit of eco-friendliness, I borrowed one from a friend (cheers, Ian) and set to work.

Musical Equipment

I love to play guitar and own 8 different guitars (pride of place being my Les Paul Custom). I usually play through a Zoom digital effects pedal with headphones plugged in, but every so often I hook the pedal up to a little VOX amp and let rip. The following are values for my normal (loud) volumes set. For the pedal, it made no real difference what I set the volume to - which you’d expect for a pair of headphones. Neither are left on and I only use them when I’m playing (which probably isn’t more than an hour or two every other day).

It sounds silly, but I expected the 22W amp to draw more (ie. Close to 20W) for the loud volume I tend to use it at, but it didn’t come close. It’s frankly incredible that the small pedal (which I power with batteries on the road) draws more power than the practice amplifier.

Entertainment Equipment

Our entertainment requirements are fairly modest - although they represent probably the largest constant power draw in the house.

I have two Hauppage MVP boxes and two televisions; a large, older tube telly downstairs (which I’m loathe to change since it works fine), and a smaller LCD television in our bedroom. In addition, the downstairs TV is output through a Sony amplifier with a pair of 100W Mordaunt-Short floor-standing speakers. Downstairs also has a SKY TV digital box (the old type before Sky+, so there’s no hard disk in it or fancy pants recording) and a cheapo region-free DVD player I bought in ASDA years ago.

Other than those main things, there are a few other miscellaneous things like an automatic SCART switcher and an ethernet IP camera (a tiny Linux machine based on the WinBond ARM chips) I can watch on my phone for security. There’s also a Wifi access point for our laptops and a wired hub with a link to the main server room upstairs.

A while ago, I did at least set up a kill switch for all the downstairs entertainment equipment, so it’s all turned off at night.

More than a few things of interest here:

• I have the oldest revision of the MVP box upstairs and the newest revision downstairs. The newest revision actually draws 3 watts more power than the older one, even though the hardware is almost exactly the same. Somehow Hauppage have made them 33 percent less efficient over the years. I’m also slightly disappointed as 10W seems a little high for an effectively diskless, fanless, slow computer with a simple MPEG2 decoder card in it. There is no discernible difference in consumption between standby and on and playing a video or music stream with the MVP, the draw is constant.

• The draw for the CRT television is astonishing. Almost 100W for the telly, which we probably leave on for nearly 12 hours a day some days. This has to change as it’s probably the overall largest power consumer in the house by some margin. By comparison, the LCD television is far better, although the power draw was less than constant - the meter stayed at 5W for the most part, but every 5 seconds or so, it would spike and go up to 40W for a fraction of a second. Not really being sure what this means, I marked the consumption to be a mid point between the two (although I was lazy and didn’t factor in time, or it should have been lower than the midpoint, but overestimating can’t hurt)

• That the 150W amplifier and 100W speakers only drew 30W at normal volume. If I change the TV, I may well focus carefully on speaker requirements and dispense with the separate amp altogether, however it’s handy for listening to streaming music through the MVP while playing MarioKart for example.

• The SKY box seemed to alternately draw 23W or 25W depending on what channel you were watching. I find this completely inexplicable, other than something to do with signal multiplexing, as it would switch between these consumption levels every time the channel number incremented - and it would stay constantly at that level while that channel was watched. Bizarre.

• 20W on standby for the SKY box (which I don’t turn off, what with it being a pain in the arse and screwing up the TV guide) seems a little excessive.

• I would have thought that the power requirements for the wireless access point (15W) would be far greater than the wired ethernet hub (14W), yet there’s just 1 watt of difference. Either that’s an efficient access point, or a very inefficient hub.

Miscellaneous Entertainment Equipment

These things don’t really fit in the other categories I’ve laid out. My radio/alarm clock - which seems to be suitably low powered. But it doesn’t do much except be an LCD clock with a radio that plays for 5 seconds each day until I punch it. There’s also a radio/CD player in the kitchen, which is rarely used, but left on standby. There aren’t really any surprises here.

Games Consoles

And what would a house be without a bit of fun? I’m probably a bit behind the times console-wise, but I couldn’t stand to give my money to Microsoft or Sony for an Xbox or PS3 (I’m aware other Sony gear is on here, but it was either a gift, or bought before Sony became the contents of bielzebub’s trousers), so I have 2 PS2s (largely for playing the Guitar Hero games) - a newer slimline model a year or so old (because my original PS2 broke and I needed to play Guitar Hero right then) and an original revision that I’ve repaired the optical drive on (after I’d run out and bought a new PS2 and satisfied my Guitar Hero fix). I also have a Nintendo Wii with a USB battery charger for the Wiimotes.

• The Wii has 2 modes, either normal standby (what it goes to when you switch it on from cold) and Wii Connect24, which it goes to when you turn it off. In this state, it remains connected to the net via the wireless router so it can retrieve messages to your Wii email (this is how online game notifications are received).

• What’s interesting here is that the USB charger only made a difference when the Wii was actually on and playing a game - the draw didn’t change for Connect24 with the charger in or out. I also metered it with wiimotes charging and the charger draw was the same (3W if the Wii was on) irrespective of whether it was charging anything or not.

• I also tried to load the Wii by playing games, connecting online to complicated Mario Kart tracks and racing other people. No matter what I did, the power draw remained constant at 25W or 28W depending on whether the USB charger was plugged in and completely irrespective of what the Wii was actually doing.

• I tried disconnecting the charger at the charger end as well as from the USB socket. If the USB cable was plugged into the Wii, but not the charger, that 3W extra power draw still remained until I pulled the USB cable from the Wii, suggesting that the Wii’s USB controller only powers down when nothing is physically attached to it, rather than when nothing is drawing power from it - this makes sense, since the USB controller should try to power external devices.

• Power consumption for the newer PS2 was very erratic, jumping to 31W when it was accessing the optical drive and busy, but falling right down to just 1W at times when it was idle (but typically around 5W). If you’re using it to play games though, it’s unlikely to be idle…

• The two revisions of the PS2 are like night and day - the original PS2 never dipped below 30W and went upto 46W under load and when using the optical drive. The only place where the original PS2 scored over it’s newer brother was standby, which drew nothing at all.

Laptops

My wife and I both have laptops that we use for general computing and tossing-it-off-on-the-internet activities. Both are on sporadically through the day for a few hours, but tend to be idle most of the time.

Mine is a recently purchased HP 530 (I wanted a T61, but couldn’t justify the expense), and the wife has my aging (but wonderful) Thinkpad T23.

• I find it unbelievable that the HP laptop draws 10W when it is off and without the battery connected, compared to 1W with it connected and charged. I can only assume that the charge controller wastes 10W looking for a battery. I’m actually horrified.

• The on value for both laptops represents them both after startup (when they’re relatively idle). I was too lazy to load them with an identical CPU-intensive activity as I just want a rough idea of the draw in different scenarios (and yes, I missed on with the battery full, but I’d already drained them and couldn’t be bothered to wait for them to charge to try again).

• The Thinkpad performance is excellent - no draw when off with a full battery or the battery disconnected. Also, it seems to draw the same amount when on, whether or not the battery is charging. I would expect the Thinkpad to be a lot less though, as it has a 1Ghz P3 CPU, compared to the 1.6Ghz Pentium M CPU in the HP. Yes, I’m aware of the Mhz myth, but the CPU is the biggest draw in any computer equipment and the fact that the HP CPU is more powerful suggests it will draw more (and cpufreq can step the Thinkpad down to 700Mhz when it is idle).

Server Room

My office. My favourite room in the house and the hub of computing activity. My main AMD64 Sempron workstation (Beefy), lives here, along with the Slug, the ADSL internet router, house ethernet hub, and my Buffalo Terastation which I use for serving music and films to the MVP boxes among other things. There’s also my 18" Fujitsu LCD monitor and a USB hard disk caddy with a 200GB disk in it (which I don’t really use, but I tested it anyway)

There is also a lot of other hardware in here (2 spare laptops, lots of other junk - it’s a hardware graveyard), but since I don’t use it, I’m not testing it.

Definitely some items of interest here:

• The Samsung router is considerably more power hungry than I thought it would be. Given the low power rating, I can only assume that its brick of a transformer is not particularly efficient.

• Beefy fared better than I thought, idling at a constant 69W (tee hee), and peaking at over 100W when I thrashed the CPU. The 7W draw for idling the ATX motherboard seems a bit much, but unfortunately is normal for motherboards.

• The slug was actually not quite as good as I was expecting. Sure, 7W is still fantastic, but I was expecting something on the order of 3W. I can only assume that powering the USB key it uses for a hard drive sucks a bit of extra juice, but I can live with that. It’s still 10 times more efficient than Beefy in an idle state (which I bought it to replace for 24/7 running). Also, that 7W never wavered, even when I loaded it with a busy task to do.

• The Terastation was better than I expected - not by a huge amount, but the manufacturer’s website had prepared me for a minumum of 40W and it came at 38W when idle. It did spike to 70W during startup, when I requested a large file transfer, the power consumption didn’t move from 38W.

• The caddy is particularly inefficient, as I understand it minimum power to spin an IDE disk is about 9W, so 17W for the caddy seems pretty poor.

• The 18" LCD monitor’s performance was amazing - just 2W and I foolishly tried making the screen busy to see if it made any difference (yes, I know LCD monitors are all about powering the backlight, but I was still in the CRT mindset). That is incredible. So incredible that I went back to test the LCD TV and found that it still exhibited the same behaviour, so I can only guess the speaker amp and digital receiver in it are the main culprit for its power consumption.

• The large 16-port ethernet hub in my office draws half the power of the crappy little 5-port Belkin hub I have downstairs. I’m not entirely sure how that works. It doesn’t have an external transformer though and has a normal 3-prong kettle lead, which I link up to my UPS for failover.

Kitchen

I couldn’t really be bothered to go round testing all the kitchen appliances, as they aren’t used that frequently. I don’t have an electric cooker. However, we do drink a lot of tea - the kettle must be boiled at least 10 times a day, so I tested that. And I just so happened to want to make some toast, so I tested that.

Now, this is quite interesting. The kettle was a fairly recent purchase because our previous fast-boil electric kettle died after 8 years service. This new one has some electronics in the base that detect the power being cut by the kettle (when the thermostat turns it off after it’s boiled). A small piezo buzzer issues a pathetic impression of a real kettle whine in this eventuality, so you know it’s boiled. I’m not complaining, this is certainly a very useful feature when you’re in the other room.

Problem is, it seems that the kettle now has a standby state as a result of the electronics and it draws 7W when the kettle isn’t doing anything! This is unforgivable waste!

As for the 3.2Kw draw during boiling, that’s just the way it is. It takes a massive amount of energy to heat water and we’d rather draw more and heat it faster since you’ll lose less energy to heat loss if you can boil it quicker. We generally drink between 7 and 10 cups of tea a day. However, in the spirit of experimentation, I emptied the kettle and boiled 1 litre of water from cold, which took 1 minute 55 seconds. Since I rarely boil quite that much water and only every other boil tends to be from cold, I’ll call it an equivalent of boiling 5 litres from cold per day (which is probably overestimating).

If I round the boil time to 2 minutes to make calculating easier, that’s 3.2Kw/H * 10 minutes / (60 / 10) = 5Kw/H per day. At 11p per Kw/H on average, that’s about 55p per day for tea. Added up over 31 days that’s a whopping £17.05 per month and approaching half of my electricity bill. I must have overestimated more than I thought here.

I couldn’t believe the toaster. I must have the coolest toaster in Britain, drawing exactly 1337W while it cooks my bread! Since I was in testing mode while I made toast, I timed it. To cook 2 slices of toast in the toaster took 1 minute and 53 seconds. Since I cook toast probably once every couple of days, I wouldn’t worry too much about that one as I’d have to toast about 32 slices of bread to use 1Kw/h

Telephones

We have two landline phones in the house, an answer machine and our mobile phones. The main phone is an Amstrad/BT Easicom thing with an LCD screen that does caller ID and can remember your phone book for you. It runs from external power. In addition, we have a solid state answering machine and a cordless phone upstairs (which is never used, because I like to know who’s ringing and I never bothered putting any numbers in - that and I ignore the landline if I’m in my office anyway).

• The initial thing that leaps out at me is the Amstrad phone. It draws a constant 20W and is left on 24/7. 20 watts for a mono LCD screen and simple device that can remember phone numbers? It also has a failed experiment that BT did for a sort of cut-down web and can do email as well, but still - 20 watts for something with no moving parts? This phone has to go.

• Both mobile phone chargers didn’t draw any power when no phone was plugged into them. I’m happy with the performance of both - my Sony Ericsson phone draws a bit more than the Nokia, but it charges a lot quicker (about half an hour), and the battery lasts a week or so, so it gets charged a lot less.

• The answer machine performed very well - just 2W for a device that is always on and I’m happy with that.

• The cordless phone also fared a lot better than I expected at just 1W to run the base station and 7W while it was charging the phone.

Conclusions

There are a few things I could do to reduce my house’s overall power consumption:

1. Replace the monstrous Amstrad phone downstairs with the cordless phone from upstairs (which isn’t currently used). It’d be more convenient and would reduce the power consumption by more than half.

2. I need to replace the power-hungry CRT TV. I understand the argument that getting rid of a useful device is not eco-friendly, but at the same time it is a monster. Initially, I’ll try and reduce the TV’s usage (ie. Turn it off when we’re doing other things, don’t leave it on for the dog when we go out, etc).

3. Quite a few devices (eg: The Wii, the kitchen radio, etc) are left on standby for days and aren’t used. I’ll replace their plugs with switched plugs, or use switched multiplugs to make sure they are off when not needed.

4. I’ve been thinking about buying a 15W caravan solar panel, car battery and inverter for running some of the lower-powered devices. Free power would be nice! :-)

Update (13th January, 2009)

Since I originally wrote this article, I’ve replaced some of the devices in the house to reduce power consumption further:

• The Samsung router has been replaced (although it’s on standby if I ever need it again) with a Safecom ADSL router that only consumes half the power. With a saving of 10 watts, it will take 4,500 hours (or 187 days) of use to recoup the £5 cost of the router. After that, it’s saving all the way muwahahaha.

• The nasty 3Com wireless access point has been replaced with China’s finest - a new TP-Link router that is less than half the consumption at just 7W max. This also allows me to use WPA encryption again. I had to give up on the 3Com, not just for power consumption but because whatever it used for storing settings died and every night it would reset itself back to the factory defaults. Unfortunately, I still need a separate switch because a few things in the living room rely on broadcast packets and the new access point has to work as a separate DMZ.

• I finally replaced the CRT telly (a friend’s caught fire and he needed a new one, giving me the opportunity to go shopping). Unfortunately, that one has backfired a bit as at 122W it’s almost 30W more than the old CRT TV. The only saving grace is that we no longer leave the telly on when we’re not watching it and just have the sound on through the amp - the standby draw is at least quite a bit lower than the 10W of the previous TV at 3W. I don’t really care so much though as I now have a wonderful big-ass TV for watching sci-fi on :-)

• I had an answer for my previous question about the downstairs switch - it was horribly inefficient. It’s now been replaced with an equivalent 5-port switch (which is half the size) that uses just 2.5W - since this is a device that has to be always on, that’s a significant improvement over the previous 14W Belkin switch.