Reviews by gar
Kill-A-Watt single socket meter
121223-1300 EST

I believe most of the AADL single socket energy meters are Kill-A-Watt model P4400. But at some prior time there was available at least one model made by a different manufacturer. This unit was similar in operation to the Kill-A-Watt. The P4400 can be very useful, but if you plan to buy one, then purchase the P4460 Kill-A-Watt EZ which is available for around $30.

Use of two short extension cords with the Kill-A-Watt will make measurements easier. These must be 3-prong cords with #14 wire. I suggest one cord with a straight plug 9 ft long for connection of the Kill-A-Watt to the wall outlet. This is hard to find in #14, but some are available. The straight plug is to ease insertion and removal from the wall outlet. The other cord should have a right angle plug for insertion into the Kill-A-Watt. These are easy to find. A 6 ft length is probably adequate. The total cost for these two cords is probably somewhat greater than $20, but worthwhile for the convenience.

The P4460 EZ compared to the P4400 has more complete and better specifications. The 4400 has a major problem in that a short power interruption will cause a loss of stored data (kWh and elapsed time). The P4460 does not always properly retain data on loss of AC power, but usually works correctly. Why is this stored data important? Suppose you are measuring a freezer or refrigerator. Probably this test should last 2 days to a week. If a power interruption occurred at one or more intermediate time points thru the test when using a P4400, then all data up to the last interruption time would be lost.

If your test will exceed 99 hours, then manually record the start date and time (hours and minutes), and also the end date and time. Convert time into decimal hours, and calculate total test time using decimal time in hours. This is necessary for long tests because Kill-A-Watt stores time in hours with 1 hour resolution after 99 hours and 59 minutes. Even using the Kill-A-Watt internal clock for time less than 100 hours it is necessary to convert its reading to decimal hours to perform calculations from kWh to average kW.

The P4400 is easier to use because any one measurement is selected by a single button press, except two measurements require two presses. The P4460 requires many button presses to select most measurements. But loss of data on power loss, and better resolution are more important criteria than button presses.

The published specifications are:


115 volts ---- operating voltage --- whatever that means.
125 volts ---- maximum voltage --- probably means you should not exceed this value.
15 amperes --- maximum current.
1875 volt-amperes --- incorrectly labeled as maximum power instead of maximum apparent power.


85-125 volts ---- Voltage RMS --- probably means you should not exceed 125 as a maximum value.
15 amperes --- Current RMS --- complains above 15 A.
1875 watts --- Active power --- more typically referred to by the single word "power", or when more clarity is needed by "real power".
1875 volt-amperes --- Apparent power --- because it is simply Vrms * Irms and in an AC circuit may not equal real power.
0 to 1.00 PF --- power factor range --- the definition of PF is PF = Watts / volt-amperes.
0 to 9999 kWh --- energy measurement range --- in kilowatt-hours.
0 to 9999 hours --- cumulative time measurement.

The meter accuracy is quite good for an instrument of this price level and capability. The 0.1 V resolution means you can use this meter in combination with a 1500 W heater to test the quality of the of the installed wiring in your home for bad installation of wiring to sockets. I did not list Kill-A-Watt's accuracy specifications because you need to know how they are defined and that information is not provided.

The next comments are some that I have experimentally determined, and some are deduced.

The 4400 quantizes by 0.1 V, 0.01 A, 1 W, 1 VA, 0.01 PF, 0.01 kWh up thru 99.99 kwh then by 1 kWh, 1 minute up to 99:59 hours and minutes then by 1 hour.

The 4460 quantizes by 0.1 V, 0.01 A, 0.1 W up to 100 W then by 1 W, 0.1 VA up to 100 VA then by 1 Va, 0.01 PF, 0.01 kWh up thru 99.99 kwh then by 1 kWh, 1 minute up to 99:59 hours and minutes then by 1 hour. That the 4460 (newer ones) can and do resolve 0.1 W does not mean that all 4460s can produce good readings from 0 to 2.0 W in 0.1 W increments. One unit I have is fairly good in this region, and another is not. Above about 2 W all four 4460s I have were fairly good.

The Kill-A-Watt instruments do a fairly good job measuring power factor. Whereas a TED system does not (The Energy Detective).

Side note. Never buy a device that is wired to your main panel that claims to save energy by correcting power factor. As a residential customer these are a scam relative to saving you any money on your electric bill.

You need to understand the relationship between power and energy. Power is the derivative of energy with respect to time. And of course energy is the integral of power with respect to time. From the field of mechanics work is energy, and power is the rate of doing work. A 100 W (power) bulb on for 10 hours is 1 kWh (work or energy). What you buy from the power company is energy plus some fees and taxes. Presently averaged over the last year my cost has been about $0.16 / kWh. For normal service your cost per kWh won't be much different than mine.

In the 1930s thru the 1950s electricity in southeastern Michigan was about $0.025 / kWh. From the 30s to the 50s the value of the dollar dropped by about 50%. Thus, the cost of electricity dropped in terms of a constant dollar. The reason was that power plant efficiency increased by going to higher pressure and hotter steam as result of new materials.

If the power of a load is varying with time (could be gradual or on-off), then to determine or measure the energy used over some time period you measure the average power over a short time (short enough that the instantaneous power does not change much during the short time), and multiply by the short time. These measurements and calculations are done for contiguous time elements over some long time period of interest. Then all of these contiguous energy elements are added together to obtain the energy for the time period of interest. The average power over that long time period is the total energy of the period divided by the length of the long time period.

The 4400 and 4460 Kill-A-Watts do this averaging and accumulating process for you using a short time increment of about 1 second. In power mode a new short time average power is measured and displayed every 1 second. In kWh mode the accumulated energy measurement since the last RESET is displayed, and the display is updated each second.

It is interesting to measure the power consumed by a cellphone charger when plugged in, and no cellphone connected. With modern chargers you won't produce any significant reading on a library 4400 Kill-A-Watt. This power level today is probably a fractional watt. Politicians have made a big issue about having you constantly plug and unplug these chargers. This is nonsense and not worth your effort. Suppose this residual power consumption was as high as 1/2 W. Then the wasted energy consumption for one year would be 8766 * 0.5 / 1000 = 4.4 kWh and at $0.16 / kWh the cost per year is $0.70 . Is this saving worth your time, and the wear and tear on the outlet and charger plug? You can save this 4.4 kWh of energy by not leaving a 100 W bulb on for for an unneeded 44 hours. It is far more important to turn off lights and other major power loads went not needed, than to mess with cellphone chargers. The big energy users are where the emphasis should be.

Besides lights some major energy loads are refrigeration equipment (refrigerators, freezers, air-conditioners, and dehumidifiers), furnace blowers, entertainment centers, electric water heaters (change to gas). There are other big power loads, but many are not continuous loads. The less total time that these loads are on the less you need to be concerned.

What can you do with the Kill-A-Watt? Measure loads that do not exceed 15 amperes, and plug into a 120 V wall outlet. You can not measure your furnace blower, yet it might consume about 800 kWh per year or about $128 presently. You won't be able to measure a permanently wired in outdoor flood light, but assume it is on 50% of the year, then the energy consumed is 0.15 * 4383 = 658 kWh or $105 / year.

Most refrigerators or freezers continuously work with an ON cycle of about 20 to 30 minutes and an OFF period of about 30 minutes. Most are plug-in and probably under 500 W when on, except for when auto-defrost powers on.

A freezer is easier to measure because there are fewer door openings and loadings of warm items. If you randomly start a measurement and randomly stop a measurement then you potentially have an error of about 30 minutes in your time measurement based on the above assumption. Ideally you want to start the time and energy measurement from a defined point in the ON-OFF cycle, and stop at the same point N cycles later. Not easy to do. If you want 1% accuracy, then with a random start and stop time you need a test period of 100 * 30 minutes. This is 3000 minutes or 50 hours or 2.08 days. If it is frost free, then probably a longer time is needed. You start an energy measurement and the internal time measurement by operation of the RESET function (button). The RESET zeroes these registers.

A refrigerator will require a longer time because of the many door openings and closings. Most refrigerators are frost free which introduces a high power resistive load at random times. A large variation in food loading and unloading causes additional variations that need to be averaged over a longer time than a freezer..

With cycling loads you want to measure energy used for some known adequately long time period. Then divide the energy (kWh) by the measurement time (decimal hours) to obtain an average kW. Next multiply the average kW by 8766 to get an estimate of the energy (kWh) used per year. What is a decimal hour? It is hours, minutes, and seconds converted to a single hour value. 31 hours, 20 minutes, and 15 seconds is 31.3375 hours.

If you have some device that is continuously on (an AC powered LED clock for example at possibly 1.3 W), then just measure its power consumption and multiply by 8766 hours to get the energy per year. In this example 0.0013 * 8766 = 11.4 kWh / year or about $0.16 * 11.4 = $1.82 per year.

Entertainment centers present a problem with how to minimize residual loads. Many systems probably have devices that requiring some sort of programming or a clock. Turning power off when the system is not needed may cause loss of stored data. Thus, you may need to provide power to some components all the time. Note: the front panel switch on some equipment may not eliminate a residual power load on the wall outlet.

Your computers and entertainment centers may be areas where the Kill-A-Watt can help you reduce energy use. Measurement of refrigerators and freezers may help you decide whether it is worth while to buy new equipment. In all cases make sure that purchase of new equipment will result in performance of the basic function of the equipment that you expect.

If you determine the energy use of your present refrigerator and/or freezer, and then compare this with the advertized value of a new one, then it is necessary to know how the new equipment was tested. For example I run my freezers at an average temperature of 0 F, and at an ambient temperature around the freezer that ranges from possibly a high of 80 F in the summer to 20 F in the winter. The temperature difference between the ambient around the outside and the interior of the freezer is the primary determining factor of the average cooling power.

To make a comparison between your present equipment and an advertized energy use of a new model may be difficult. Only after you replace your present equipment with something new and run your own experiment will you have a good comparison. Suppose on average you could save 100 watts, then this is 877 kWh per year or about $140 / year at $0.16 / kWh.

See my book "Electrical Energy Measurement, Conservation, & Methods to reduce Your Electric Bill", AADL call number 696 Ro.

On page 37 are two plots. is of an old freezer. Note, the duty cycle is less than 50%, and this may be a design criteria to provide faster freezing for a load of room temperature food. is of a fairly recent side-by-side refrigerator-freezer. Here the duty cycle is greater than 50% which reduces the size of the motor and compressor. Not as much need to provide quick cooling as with a freezer only unit.

Both have about the same average power consumption, but different duty cycles. The freezer while running is about 300 W and the refrigerator about 200 W. Note, you can not judge energy consumption over a period, such as a year, by just looking at power consumption at one instant. More information needs to be known about the load and how it varies with time.

1-26-2013 Some data from my freezers. Both are in the garage and I will estimate an ambient temperature of about 20 F over the last two days.

The Amana, in the range of 40 years old, used 3.61 kWh in the last 61.83 hours. Thus, its average power was 58 W and its duty cycle was around 20%. In the summer its average power is more than double. A possible error of 2% to 3% in this power value is likely because the measurement was not synchronized with the freezer cycling period.

My other freezer is somewhat newer and slightly smaller, but it used 4.49 kWh in the same time period. Thus, its average power was about 73 W. Both interior temperatures were about 0 F.

8-Sensor Energy Monitor
121210-2101 EST

The information provided by AADL with any of the energy monitors is relatively meager, and there is no suggested chart for recording measurements. There is no discussion on what you should measure, how measurements should be made, and the significance of any measurements. The relationship of power, energy, voltage, current, volt-amperes, and power factor are not mentioned. To help in some of these areas you can look at portions of my book "Electrical Energy Measurement, Conservation, & Methods to Reduce Your Electric Bill", AADL call number 696 Ro.

The 8-Outlet Wireless Energy Meter is a great disappointment. It is distributed by P3INTERNATIONAL, has the Kill-A-Watt trademark, and consists of two parts. One part of the system is remote sensors that are radio frequency linked to the second part, a display unit. The part numbers are P4220 (each sensor) and P4225 (display).

I expected the sensor to be a single outlet P4460 Kill-A-Watt EZ with the display and buttons removed, and the addition of the RF link. It is not close to comparable. It is substantially downgraded from the 4460. Resolution (smallest readable unit) is worse, and the response time is very long. The P4460 EZ is very good and useful at its price of about $30

The wireless system display emphasizes $ cost and carbon, and de-emphasizes the basic measurements. The wireless display is hard to read. The P4460 EZ has a much better display than the wireless P4225 display. The only advantage the wireless system has is that the sensor can be put behind whatever equipment is being monitored without the need for extension cords to bring the display to a readable position. Mapping multiple sensors to one readout has little value as configured. Because of the long time delay between reading updates you loose any quick sense of whether the system is working or not.

My philosophy is that all energy measurements and thinking should be done in kWh (kilowatt-hours) to avoid confusion and provide a stable consistent reference. To use $ or carbon is only going to lead to confusion. When you really need to know a $ or carbon value, then do the conversion at the current rates.

Illustration on my use of energy:
2006 was 14073 kWh @ $ 0.116 /kWh,
2011 was 14346 kWh @ $ 0.141/kWh,
2012 esti 11521 kWh @ $ 0.162 /kWh.
Comparing total cost of energy between these years is not useful in determining energy usage because the rate in $/kWh is not constant. To judge my energy change efforts I need to look at kWh, not $.

In 2012 the average cost has increased considerably, and I have reduced consumption.
For the 2012 year my consumption will be down almost 20% compared to 2011, but cost is up about 15% from 2011, and about 40% since 2006.

I believe you will find this wireless system non-intuitive, awkward, frustrating, and slow in use. Furthermore tech support at P3 seems to be lacking in knowledge on their own system.

With the wireless system from AADL the response time for a measurement with only one sensor active was about 2.75 minutes. I would have expected 1 to 2 seconds, which is the response time for the 4460 EZ. This wireless system response is so slow I can not practically test for minimum voltage or resolutions, and other factors.

In the auto scan mode, to scan 8 sensors, the total scan time is about 12 seconds. Scanning thru all 8 sensors, vs, just locking the display on one sensor, seems to make no difference in the data update rate. It is still about 2.75 minutes. Also note that in the auto scan mode all sensor addresses are scanned whether a sensor is present or not.

Just after address association the time to the first read may be about 30 seconds. Considerably better than the normal response time, but it is a one-shot.

"Reception Status", a small disk at the display top right, effectively tells you when a new sensor value will be displayed. When the disk transitions from solid black to white the data is updated to the present value from the selected sensor. This disk rotation is 2.75 seconds. Correction 2-2-2013 should read 2.75 minutes, not seconds.

From what is easy to perceive, the wireless system voltage measurement resolution is no better than 1 V compared to 0.1 V for the EZ unit. I can't, meaning won't because of the time delay, determine the minimum working voltage for the sensor, and it is not in the specification. The wireless upper spec is 125 V. The EZ spec is 85 to 125 V. Actual residential voltages may exceed 125 V some of the time in some areas today. In the early Edison days 110 V was nominal, today it is 120 V. The Kill-A-Watt upper specification limit is a stress limit, whereas, if provided, a low limit would be a functional limit. One EZ unit I have works (reads) from about 50 V to 145 V. This would be a functional range, but since the upper specification limit is 125 V the 145 V I applied to the EZ is probably an excessive stress level. Older P4400s tended to fail to operate at much higher low end voltages. I had one 4400 that was unusable because it quit at 107 V. More normally the 4400s were about 100 V.

Kill-A-Watt wireless resolution is no better than 1 W, 1 VA, and 0.01 A, and may not be this good. The response time is too slow to run tests. The 4460 EZ resolves 0.1 W up to 100 W, then switches to 1 W. Similarly for volt-amperes. The EZ current resolution is 0.01 A.

The various Kill-A-Watt models are usable in a nominal 120 V circuit with plug-in loads up to 15 amperes. This means you can test entertainment centers, refrigerators and freezers, toasters, flat irons, small kitchen appliances, vacuum sweepers, electric mowers and other tools, etc.

You can not measure most air conditioners, installed electric water heaters, furnace blowers, wall ovens, range tops, and dryers. Also built-in lights can not be measured with any Kill-A-Watt unit. Some of these are your largest power users, but may not be your greatest energy user. We seldom use our wall oven, thus not a lot of energy is used, but power is high when on. Energy is the integral of power over time.

You can measure the motor on a gas clothes dryer, and a washing machine.

Unlikely you can measure a microwave because the RF radiation will probably jam the Kill-A-Watt communication channel.

None of the Kill-A-Watts provide data output to a computer. Data output is available with a TED System (The Energy Detective). But the TED System is not feasible for the AADL to provide on loan. Some data collected using a TED System are provided at the following page on my web site . This provides you an idea of how my power use varies. Most homes will drop off more at night than mine. Homes with no one, but the dog, home during the day may show a daytime dip.

If P3International had taken their P4460 EZ, removed the display and keys, added a good RF link, combined these with a communication unit that interfaced with a personal computer, and good software, then a desirable system could have been provided. I have to classify the present wireless Kill-A-Watt system as unacceptable. For individual measurement of plug-in 120 V loads I suggest using the Kill-A-Watt EZ P4460 with two #14 wire extension cords. One cord should have a straight plug and the other a right angle plug. The straight plug is used to plug into the wall and make removal easy.

The Kill-A-Watt instruction manual is too small, hard to read, and provides no information on different loads and how to interpret the data. The provided specifications are on page 23.

In the AADL Kill-A-Watt wireless system each sensor has a number, 1 - 8, on the front to identify the sensor address. This seems to be absolute in contrast to the description under flashing L in the manual, p 15. So AADL units may be different than over the counter systems. To associate (teach) a sensor with the display you must first have that sensor unplugged from the wall, get the display into the flashing L mode (press - key for about 3 seconds to enter flashing L mode), then early on within a limited time (flashing L disappears after about 20 seconds) plug the sensor into a hot (meaning on) wall outlet. When association has taken place the sensor address is displayed. About 10 seconds to associate. I do not know what happens to sensor association on loss of AC power. A test below seems to indicate the association is retained.

The sensors are not supposed to lose or change accumulated data (cumulative time and kWh) with loss or large fluctuations in AC voltage. This is true of the Kill-A-Watt EZ also. But with an EZ unit I have experienced scrambled data sometimes. It does appear that loss of power for a short time does not cause a loss of association of a sensor with the readout. But I did not see any indication on the display that power was lost, and the presently displayed values at power loss remained at their last value. Should have been a change of the display in some fashion to indicate the power loss.

For evaluation of a refrigerator or freezer a monitoring test needs to run for several days to a week. You do not want a small power interruption to cause your cumulative data to change. This problem exists in the standard Kill-A-Watt P4400.

When you checkout the 8-Outlet Wireless Energy Meter verify that everything is in the package that is on the list of items for that package.

121211-0931 EST

An additional point.

Short time loss of power to a sensor does not cause a loss of address registration information with the display. This I determined and mentioned above.

Last night I depowered the only active sensor. This morning, about 8 or 9 hours later, I found the sensor was not recognized by the display. Thus, the registration process had to be repeated. This meant removing power from the sensor, put the display in the flashing L mode, and reapplying power to the sensor.

I don't know and won't find out how long power can be removed from the sensor before it has to be re-registered.