Residential Power Monitoring
Author: Mark Anderson
Living with soaring fuel prices, a difficult economy, and the emerging Smart Grid, I decided I would start to monitor and manage my power consumption when I rehabbed my 1912 house. The monitoring has been running for several months now, so did it make me a more eco-friendly citizen and did it save me money?
There’s been lots of talk of the Smart Grid for several years and many utility companies have rolled out Smart Meters and accompanying utility plans. The core concept of the Smart Grid is that utility companies monitor power usage and, in return for cheaper electricity, you will allow the utility company to turn on and off certain circuits and/or appliances during peak demand periods. This has attractions for many. Some do it for the financial savings, while others do it for environmental reasons. Would I partake in such a scheme? In a word: “No”. Now that doesn’t mean I have money to burn, nor does it mean that I don’t care about the environment; although, I do have many friends whose philosophy is “if we’re all sailing on the Titanic, we may as well go first class”. I have a small house and I’m already fairly responsible about the energy I use. Now if I had a 6,000 square foot home, with central air, hot-tub, pool, etc., I may be inclined increase the temperature in certain areas of the house, or turn off the heat to the hot-tub/pool during peak demand periods in return for savings. But would I let the utility company shut-off the air when it’s 100-degrees and humid? No way.
My main reasons for monitoring my power usage are curiosity and cost saving. It’s clear that if you can’t measure it, you can’t manage it. So the first thing I needed to do was figure out how much power I was using and which devices were the real hogs. Some are obvious: air conditioning, pool heaters, etc., but does turning down the AC 1-degree really result in 3% saving (or whatever the oft-quoted number is)?
Most utility companies will do some basic analysis for you. Here in Chicago, ComEd sends me a bill with a nice bar graph showing my consumption by month for the last 12 months. Unfortunately, this has little use in determining if changes in your lifestyle are having a positive effect, (unless they are significant changes, such as insulating the house). For example, turning down the AC in August and then comparing the bill with July’s provides little useful data if August was five degrees hotter than July.
In my opinion, the only real way to do this, is to look at consumption relative to activity. For example, if I turn the AC on, how many more kilowatt-hours am I consuming. There are several ways to approach this. You can buy devices that monitor the main feed and note the difference. The problem with this approach is that everything in the house will affect the power consumption at the main feed. For example, if the fridge’s compressor turns on, you are now measuring the consumption change caused by the fridge and the AC. So this approach has limited usefulness.
The next option is to use something like a Kill-A-Watt device. This is a monitoring device that plugs into a receptacle and then the appliance to be measured plugs into the Kill-A-Watt. While this can be used for measuring 110v appliances, it is of no use in measuring the consumption of 240v appliances, such as wall ovens, washing machines, central air compressors, etc. Further, having to repeat the measurement process, and manually track the cost of each appliance is both time-consuming and tedious. I do have a Kill-A-Watt and it was useful in getting an idea of what my window AC unit was costing per hour, but for managing the entire residential usage, it’s of little use. I do have some other plug-in devices that monitor consumption and I’ll discuss these later.
For me, the only logical approach was to find something that could measure individual circuits. I rewired my entire house earlier this year, and according to code, all window AC units, fridges, microwaves, dishwashers, etc., needed to be on dedicated circuits. This meant that I knew where everything was connected at the main panel and, more importantly, all the power hungry devices were on dedicated circuits. I also had other dedicated circuits were for specific functions (e.g. TV’s and HiFi are all on one circuit). All I had to do was find a device that would measure individual circuits and report the results back to some software for analytics.
As luck would have it, as I commenced the search for a solution, Google PowerMeter launched. It looked like I’d solved the analytics piece of the puzzle. All I needed now was the hardware. After researching several devices, I settled on Brultech’s ECM-1240. So why the Brultech?
- It allowed me to monitor the main feed, plus 6 additional loads (each load can be several circuits, so I could, for example, combine all the 1st floor lighting circuits on a single input).
- I could connect multiple ECM-1240’s. I now have two, so can monitor 14 loads
- There are several current transformers (CT’s) to choose from, (more on this later)
- It can interface via Ethernet, Wifi or RS232
- My usage data could be forwarded to both Google and my home automation system (HomeSeer).
- Monitors voltage
I ordered a kit comprising one ECM-1240, a selection of CT’s, an Ethernet gateway and software to collect and forward data. Before continuing with an explanation of how it’s all hooked up, it’s worth explaining what a CT is.
A current transformer (CT) is a device for measurement of electrical current. They produce a reduced current accurately proportional to the current in the circuit being measured. This allows high current loads to be measured by sensitive electronic equipment. The CT, like any transformer, has a primary winding, a magnetic core, and a secondary winding. As alternating current flows through the primary windings, it produces a magnetic field in the core, which then induces a smaller, but exactly proportional current in the secondary winding circuit. So, for example, I could be measuring a 15A current and producing a 15mA current (1/1000th of the measured current). The easiest way to think of it is as a step down transformer for current rather than voltage.
There are two main types of CT’s: split core (left) and solid core (right).
The split core is generally used when the load cannot be disconnected and the solid core (donut style) requires the load to be disconnected and passed through the CT. The two types are shown above and Brultech have a good description of the different types and their usage.
Connecting the ECM-1240’s isn’t too difficult. It should really be carried out by a licensed electrician, but is simple enough to install for a competent DIYer with basic electrical knowledge. The main feed is measured by two split CT’s, these are hinged so they can be clamped around the main feed without removing the wires (definitely not a good idea). The photo below shows the installation of the split CT’s on my electrical panel.
Individual circuits are monitored by disconnecting the wire from the circuit breaker (after turning it off!) and threading it through the micro-CT.
Finally, the CT’s were connected to the ECM-1240 and via RS232 to the Ethernet gateway, which would send measurements to the Brultech software. Below, you can see my current installation.
The Brultech software was then configured to send data to Google PowerMeter and my home automation system. If the data only needs to be forwarded to a single location, there’s a good chance that the Ethernet gateway can forward it there directly without additional PC software. In my case, I wanted to forward the data to several destinations, so I had to use Brultech’s intermediate EngineG software. I am currently forwarding data to the following: HomeSeer (with UltraECM plug-in), Pachube.com and Check-it.ca (more on these later). I am also beginning to test SmartEnergyGroups.com
After a few hours of wiring, I was getting real-time power consumption data. I have since purchased a second ECM-1240 to monitor up to 7 additional loads, as my “Other” power was still 40% of the total and, damn it, I wanted to know where that 40% was going!
Every week, I received an e-mail from Google with my energy usage, and I could go online whenever I wanted and view my real-time usage.
The charts above show 1 week’s daily usage and 6 weeks weekly usage. These were a revelation: I could see how much I was using at different times of the day (on the main feed and 5 other circuits) , as well as my “Always On” usage. Once I started receiving these, I started trying to figure out the trends and their root causes. I set about trying to reduce my Always On consumption (more on this later), and slowly but surely got my energy consumption down.
So how did I do it? First of all I targeted phantom loads (devices that draw power when in standby mode), which were a lot of my Always On consumption. I have a 50” and a 42” plasma. These are well known for their vampire power. Together, mine draw about 80W and they are both off for over 20 hours a day. At 10 cents/Kwh, this represents a waste of about $5/Month Now, I could unplug them from the wall, but I’m too lazy for that, and if I want to turn on the TV in the morning I don’t want to get out of bed. I had to find something that was smart enough to know when I’d gone to sleep.
There are plenty of smart energy strips out there that can turn off several devices when a “control device” is turned off. The problem with these is selecting a control device, whose power cannot be killed. These work great for computers, where the monitor, printer, etc. are turned off when the PC is shutdown, but in my bedroom, the only device I want turned off besides the TV is the DVD player. Given that I have other sources besides the DVD, the DVD cannot be the control device, so the only candidate is the TV. Clearly, this wouldn’t work: the TV is the device whose power I want killed.
I finally invested in a couple of Smart Energy Switches from Aeon Labs. These are Kill-A-Watts on steroids. Not only do they give me real-time consumption to my home automation system, I also control them from my iPhone, etc. So here’s how I set them up. When the TV is on, it’s consuming in excess of 300w, so I needed to determine when I’d turned if off for the evening and then kill the power. For the living room, this was easy: I have an event on my home automation system called “bedtime”. This turns on the stairway lights for 5 mins, turns off the 1st floor lights after a 2 minute delay and kills the power to the TV. I have an alarm function on my home automation system, so it powers the TV back on when the alarm goes off. I also have an event called “Work”, which I run when I go down to my basement office. This kills all the other lights in the house and kills the power to the TV again.
For the bedroom, my HomeSeer system looks for a power drop from the TV to below 60w after 10pm. This is a reasonable indication that the TV is off for the night. When this drop is detected, the TV (and a few other devices are powered off until the alarm goes off in the morning). On the rare occasion that my wife goes to bed and turns off the TV after 10pm, and I come to bed later and turn it back on, I just have to push a button on my Logitech Harmony, whose IR code HomeSeer recognizes, and power is restored. If I then turn off the TV later, HomeSeer detects the drop to less than 60w and kills the power again.
The next big challenge was the AC. I’ll write more fully about my ideas on temperature management in a future article but present the general concept here. I have 4 window AC units. What I was looking for was the ability to turn them off when I left the house, turn them back on 20 mins before I returned home, and turn on the bedroom one at will (say 10 mins before I went to bed). Again, the Smart Energy switches came to the rescue, but the real gem was integration with my security system via HomeSeer.
I installed a regular DSC alarm system with a connection to HomeSeer. (The most important criteria here was that alarm system should be fully functional without my home automation system.) Besides using the normal keypads, this allows me to arm and disarm the system from my iPhone, etc. It also sends messages on major events (arm, disarm, intrusion, etc.). What this allows me to do is to turn off the AC the minute I arm (away) the security system and then turn it on again 20 minutes before I return home from my iPhone. I did the same with the heating in the winter using a z-wave thermostat. When I leave home, the temperature is dropped to 60-degrees and 30 minutes before I return home, I remotely set it back to 70. In my opinion, this beats the heck out of any system that tries to learn schedules or uses motion sensors and waits for an hour or more of inactivity to determine if the house is unoccupied.
In addition to controlling heating, the arming of the security systems kills most of the lights and turns on my FakeTV in the guest bedroom at night so that the house looks somewhat occupied.
My next plan is to install motorized blinds, so that I can use passive heating and cooling. In the summer, it will automatically lower the blinds on the east side of the house in the morning to keep it cool. In the winter, it will do the opposite to take advantage of sunlight.
Demise of Google PowerMeter
In June, I received a blow from Google: they were discontinuing Google PowerMeter on Sept 16th. I immediately started hunting for an alternative, and after a few weeks I came across Pachube, which was fortunately supported by the Brultech software. Pachube offers much more than power monitoring. It can pretty much monitor anything that produces data temperature, barometric pressure, wind, water levels, radiation, etc.). This allowed me to get the real-time feeds and limited history, but there were no analytics. (Usage is available in XML or CSV format, so it can be analyzed, but that’s too much work for me.) Below is a screenshot of the consumption on my main feed.
For a few months I relied on the analytics provided by my HomeSeer plug-in for Brultech devices, but the charts required Flash, so I was unable to use my iOS devices to view them. Further, they could only display 1 day of data and 1 ECM-1240 at a time.
I added a 2nd ECM-1240 a few weeks ago and have been searching for a good online alternative to Google PowerMeter. I’m currently running a trial with Check-it.ca and preparing to try SmartEnergyGroups, who not only provide monitoring, but supply all the hardware required to monitor your equipment. So far, the analytics from Check-it.ca look pretty good and their support has been great. They also offer apps for iOS and Android.
Below are some screenshots from the Check-it.ca service.
Besides real-time and historical monitoring, their service allows you to setup notifications when devices go offline or when specific circuits exceed a given threshold. There is also a facility to control devices remotely if you use Insteon and have an ISY device. I have no need for this, as I can access everything remotely through HomeSeer. It also has a nice performance tracking feature, where you can enter improvements/modifications made on given dates and their associated it cost. From there it will track any changes in consumption and attribute those to the improvement. Their weekly scorecard (which can be automatically e-mailed) shows you performance by week as well as the projected energy bill for the month:
The cost for Check-it’s service is $99 per year.
Have I saved money?
Firstly, I’d like to qualify something: I’m probably not a typical case for measuring savings through power monitoring. Because I have a home automation system, a lot of normal energy wastage has been curtailed. When I go to bed, I run an event that turns off all lights. When I go to work or leave the house, another vent runs that adjusts the AC/heating and kills most of the lights. It would be an interesting exercise to turn off my home automation system for a week and see the difference.
So to answer the question: “not yet”. The Smart Energy Switches are about $60 each, so the two on the TV’s will take 24 months to pay for themselves. Our window AC units cost about 25c per hour to run. I reckon we save about an hour a day on unnecessary cooling (factoring in that it takes longer to cool a hotter house), so it would take about 240 days (2 years @ 4 months a year) of cooling to recoup the investment of the Smart Energy Switches used for thse. (I use a few of those in the winter for local electric heating, so the payback period will be slightly less.) On top of that, about $300 invested in Brultech equipment is going to extend the return on investment quite a bit; however, money saving isn’t the only factor to consider. There’s the effect on the environment and my inner-geek’s satisfaction. I still have a lot more analytics to examine and more optimization to carry out when I’ve finally settled on my Google PowerMeter alternative.
Am I a better eco-citizen?
Definitely. My carbon footprint has reduced and will continue to reduce as I further optimize things. I’m definitely smarter about the way I use energy and incorporate savings where I can in my normal home automation routines.
Mark Anderson is a full-time software consultant. In his spare time, he's and avid Home Theater and Home Automation enthusiast and occasional contributor.