Kill-a-watt is a nifty, pretty reasonably priced outlet meter.
Plug the Kill-a-watt into a standard 110V US outlet, then plug the device (like a TV, air-conditioner, computer, etc) you want to measure into the Kill-a-watt. The KAW will measure voltage, amps, watts, Hz, kWhr, and elapsed time (since plugged in). The unit requires no batteries (it's plugged into the wall after all) and does a pretty good job. I bought two so I could measure intermediary loads, and of course so I could take one apart and not have to worry about destroying it.
The KAW cracked open
Remember: Power (Watts) = I (current in amps) * V (volts)
To give you an idea of the world I play in, 5V at 1A is a huge amount of power (5 watts)! Most of my projects use under 5V@200mA (1W).
Also realize the cost of a kilowatthour (the amount of 1000 watts used over an hour) varies widely between states and countries. Checkout the average kWh price across the US. For Boulder, CO we're looking at about $0.10 per kilowatthour. Sorry Hawaii ($0.33)!
Interesting finds:
My UPS uses 5 watts. When a KAW is plugged into wall, and a second KAW is simultaneously plugged into UPS, we can see how much power is actually delivered to the load (my computer) and how much is used up by the constant charging of the UPS battery. Subtracting, we see that the UPS is drawing 5Watts to trickle charge the internal battery (416mA@12VDC is pretty large).
My 19" LCD screen uses 50W, and I have two screens (100W). Thank goodness we have something called 'Power Save Mode' built into most modern monitors. My monitors automatically shut themselves down after a few minutes of non-use.
My computer (a modest 2.4GHz AMD 3700+ with dual HD), uses 270Watts. Not too bad.
The new SparkFun coke machine uses nearly 500W of power, 12kWh per day. It's brand new and the thing never turns off! You'd think it would either reach a cold temperature (my 15 year old fridge turns off!) or power cycles at night (who needs a tasty beverage at 4am?). Ahh well.
My Tivo! Oh Tivo. How you insist on recording Sabado Gigante even when I swear to you, I don't speak Spanish. TiVo is on all the time, there is no power down mode, and thankfully, it only uses about 30Watts. Not too shabby.
My trusty soldering iron uses about 60Watts. It's a bit of a hog. I usually remember to turn it off. Actually, I religiously remember to turn it off. A burned down house is no fun.
Now with everything off (computer, bumping stereo, iron, monitors) my desk uses 24Watts. What is using it? The WiFi router and cable modem. This is something I could turn off. I really don't need to have my computer or internet stuff on while I'm out of the house. How much does it cost me? 365 days * 24 hours/day * 0.024kW/hr * $0.1049/kWh = $21 / year.
Not horrible, but wouldn't it be handy if I had a way to turn that and all the other random junk plugged into my outlets when I left each day?
Enough talk - more pictures!
There's a basic LM2902 quad opamp in the lower-right corner.
The display board
No photo trickery - the main square IC had the label completely and cleanly etched off. Looks like an MSP? Looks like a connected I2C EEPROM from Atmel? The crystal is 4.194304MHz. This seems to be a common freq? But why?
In action. The KAW is a bit hard to read when plugged into the wall behind furniture so an extension cord helps.
There's a basic LM2902 quad opamp in the lower-right corner.
The display board
No photo trickery - the main square IC had the label completely and cleanly etched off. Looks like an MSP? Looks like a connected I2C EEPROM from Atmel? The crystal is 4.194304MHz. This seems to be a common freq? But why?
In action. The KAW is a bit hard to read when plugged into the wall behind furniture so an extension cord helps.
I've been using a similar product called "Watts Up" (I guess punny names are obligatory for power meters).
AFAIK, both of these devices measure true power by measuring instantaneous voltage and current throughout the cycle (rather than by measuring average voltage and current and adjusting for a phase difference). More accurate especially in the case of non-sinusoidal signals, which would be common in front of triac lamp dimmers, non-PFC switching supplies, and so on.
For an AC system, Power = Volts x Amps x Power Factor
Your description defines VoltAmps (which in a DC system = Watts)
Power Factor is just how closly the Voltage and Current waveforms are in sync. If you place an inductor or capacitor on an AC line, you'll have current and voltage, but no power used (the power factor would be a lagging or leading 1)
The crystal is 4.194304MHz. This seems to be a common freq? But why?
Divide by 2^22 and you get 1Hz which is a pretty good frequency to use for anything that involves time, like say measuring energy consumption!
If you take a look at the LM2902, you'll see that a current monitor circuit can be achieved.
Well, May I know what is the type of LCD being used by the Kill A Watt? Thanks
may be you can sell it? i will buy!
If you really want to get into this, it's not hard to build such a meter with a microcontroller. Lots of projects out there like this. You've got the controller's ground on AC neutral, a low-side current shunt resistor and a voltage divider for the voltage. Just need 2 ADC inputs reading several times per cycle. You'll need to pull some tricks for powering it because no common 5v reg has a 200V absolute max input. Like I say, there's projects out there for this.
Be aware that with the controller's ground at AC neutral it is UNSAFE to use these signals to interface with anything outside the box. You will need optical isolation.
When I pulled mine apart many years ago, they were not yet polishing off the chip so the part number is visible (not that it helped me find it back then). Pictures are here:
http://cafeelectric.com/killawatt/index.php
-Otmar
"The new SparkFun coke machine uses nearly 500W of power, 12kWh per day. It's brand new and the thing never turns off!"
You can get PIR devices to lower the energy usage of the coke machine. info
I have a Kill-A-Watt and ran some very similar experiments with various equipment. A real killer for the electric bill is the stuff that draws current 24 x 7 whether you use it or not. It convinced me to give away my hot tub ;-). Plenty of computer equipment falls into the always-on category. My rule of thumb is that for every 15-watt persistent load, it adds a dollar to your monthly electric bill.
As a related experiment, I switched to CFs lately and computed that reduces a typical electric bill by about 7%. I also took into account the effect of the lost heat from incandescent bulbs that requires my furnace to make up for in the winter and the CF's reduced heat load on the air conditioner in the summer.
I've got a Kill-A-Watt as well (well, only one) and in my research and experience, the units are more accurate than their watt displays imply: if you let it run for, say, 100 hours, then read the KWH and divide by hours, the wattage is actually more accurately accumulated than the units-place watts.
I'd be curious where to put a supercapacitor or battery so I could keep readings after unplugging the darn thing -- doing a week-long average of my refrigerator sucked when I accidentally unplugged the Kill-A-Watt, losing all data!
Second, a handy back-of-hand calculation I use is that at about 11 cents/KWH, running a device 24 hours a day for a whole year costs $1 per watt. In a specific example, a 100-watt lightbulb will cost you $100 to run it all year.
Finally, I did the math on gas versus electric heat in my blog ... the results are convenient as well: the break-even point between gas and electric heat is where 1 therm of heat costs 25 times the cost of a kilowatt of electricity. So if electricity is $0.10/KWH, then gas would have to cost more than $2.50/therm for electric heat to be cheaper.
Enjoy!
---Jason Olshefsky
is it just me or do power outlets look like a suprised-shocked emoticon :)
=(]
I have a quick question. if I'm seeing this correctly the i2c connection between the unidentified micro and the Atmel EEPROM has no pullup resistors. whats going on there?
The micro could have internal pull-up resistors.
Take the circuit board and place it in the freezer for 20 minutes or so and when u pull it out look for the outline of the etched off brand on the IC chip.
I really wish there was a way to hack a serial port or something into this thing. It would be nice to be able to hook up a data logger to it and see how my power consumption changes over time.
Maybe looking at the bill every month?
You should check out this test on the accuracy of the kill-a-watt. It is not exactly a professional measuring tool, but it is still good for getting a rough estimate (ex: to see if you are using 5 or 500 watts).
http://hardforum.com/showpost.php?p=1032584323&postcount=68
Also, before the kill-a-watt became popular the only equivalent was the seasonic power angel. I have heard that they are both very similar, I should check mine to see if it has the same unlabeled ic as the kill-a-watt.
"I'm curious how well a KAW measures True Power. Has anyone measured (with a multimeter) voltage (V) and current (I) of various devices, multiplied the measured V and I, and compared the result to what a KAW indicates? If there is a large difference for a motor load, and a small difference for a lamp or toaster, then the KAW is reporting True Power."
The KAW measures true power. If you measure both current and voltage at the same time, and find the phase difference, it's pretty easy to calculate.
The KAW will give you Volts, Amps, frequency, power factor, watts, Voltamps and kilowatthours
"But how does it actually measure current? I've wondered whether those things have a hall effect sensor, or they just measure the voltage across a small (<1ohm) high power resistor."
The latter. The big loop of wire you see in the white box just to the left of the receptacle plug in the 3rd picture looks to be what they are using for a shunt.
Maybe they put a current transformer around that loop in the factory and measure a known current to calibrate the measurement?
Id like to know the UPS size, I have a UPS that used to run a phone system. And if it too only uses about 5w to charge, well WOOT!
I think Ill have to get a KAW next payday if Sparkfuns turns out to be smaller to find out.
My UPS says: APC Back-UPS ES 500 (Model: BE500R)
http://www.apc.com/resource/include/techspec_index.cfm?base_sku=BE500R
Probably not suitable for a phone system, but close.
I'm curious how well a KAW measures True Power. Has anyone measured (with a multimeter) voltage (V) and current (I) of various devices, multiplied the measured V and I, and compared the result to what a KAW indicates? If there is a large difference for a motor load, and a small difference for a lamp or toaster, then the KAW is reporting True Power.
But how does it actually measure current? I've wondered whether those things have a hall effect sensor, or they just measure the voltage across a small (<1ohm) high power resistor.
Better than a remote controlled switch is a "Smart Strip". They have one sensing plug for your computer or TV. When you turn off that device it turns off all of the other outlets (speakers, printer, router, etc).
http://bitsltd.net/ConsumerProducts/index.htm
http://www.homehardware.ca/Products/index/show/product/I3665742/name/strip_outl_smart_srg_7_plug
I was going to try to figure out how to build one but then I found out they already exist. Mine works great.
Give this a try for shutting everything off: http://catalog.belkin.com/IWCatProductPage.process?Product_Id=459516
A little review of it: http://www.twit.tv/dgw694
For those that don't want to click on the links... it's a multi-outlet power strip with a WIRELESS REMOTE CONTROLLED POWER SWITCH. The outlets are even spaced apart so you can plug in the transformers without blocking out other outlets. And it's only $49.
Switch your monitor to,
http://www.smartplanet.com/news/tech/10001590/hands-on-with-the-lg-w2252te-the-world-s-most-eco-friendly-monitor.htm
and CPU to D945GCLF2 (~40W full load)
Which TiVo do you have?
Dah - one of the old 40Hr units (series 2?) with a second 400GB drive dropped into it. Sorry I can't remember more accurately.
With my KAW I noticed my laptop drew 30-31 watts when on, but only 1 watt when in suspend, so I put it in suspend a lot more often. Switching power supplies are the trick.
I thought about using X10 devices with a simple uC turning things like my wireless router/cable modem, phone chargers, etc based on a time schedule. But then I wondered how much all of the X10 stuff would draw, I need to dig out some and measure them.
Analog Devices makes a series of easy to use power meter ICs as well. No need to dig into the math for figuring out all the power factors, etc.
so... how much power does one kill-a-watt draw? Was I the only one who had that thought when you said you had 2?
That was one of the plans. I'll see if I can do the numbers. I've got to switch them (A plugged into B and vice versa) to eliminate inaccuracies within each unit.
that was my first thought too