One engineer’s experience with switching to solar power at home
I’ve long toyed with the idea of getting solar panels. My house in Boulder, Col., is perfectly situated: It has a south-facing asphalt shingle roof, including the roof over the four-car garage, and being a relatively new house, it has few trees to block the sun. Boulder, located just east of the continental divide and about 50 kilometers from Denver, gets an amazing amount of sun: an average of 157 sunny days and 184 partly sunny days each year, according to the Colorado Climate Center.
But whenever I tried to crunch the numbers—the cost of the equipment, the cost of installation and maintenance, the eventual payback, rebates, and so on—solar never seemed financially viable. In part that’s because electricity rates in Colorado are dirt cheap. Until just recently, rates were a flat US $0.12 per kilowatt-hour. In California, by contrast, tiered rates start at 11.9 cents/kWh and can climb as high as 49.8 cents/kWh. I figured solar would at most save me about $1000 per year in electricity, so how could I justify installing a $40 000 system?
Then one day a few summers ago, I was eating at Beau Jo’s, a pizza place in Idaho Springs, Col. All the Beau Jo’s restaurants switched to solar and other renewables a while back. [For more on the restaurants’ green initiatives, see Beau Jo’s Web site.] I spotted a big rotating LCD screen that showed the following:
- Kilowatts being produced now
- Kilowatts produced today
- Cumulative kilowatt-hours produced
- Cumulative tons of CO2 saved
Tons of CO2 saved? Of course! The financial picture was important, but there was the bigger picture to be considered: the impact on the planet. This aspect of solar hadn’t really been in the forefront of my thoughts, but now it made sense. Switching to solar wasn’t just about dollars and cents.
Step 1: Research
I’m an engineer, and so for me the first step in going solar was research. I started with my home’s electricity usage. The house, built in 2000, has 3500 square feet of finished space (about 325 square meters), and my family’s average annual electricity usage is about 8000 to 8500 kWh. Of that, our central air conditioning is a big contributor. Our wonderful view of the Rocky Mountains is provided by large windows on the south and west sides of the house. The downside to all those windows is that the solar gain indoors is truly impressive during the summer, when daytime temperatures can easily rise above 32° C, necessitating the cooling.
Next I looked at the rebates that my electric utility, Xcel Energy, was offering customers who switched to solar. At the time, the rebates were pretty generous. During the summer of 2008, the utility was paying $4.50 per watt for panels installed. Given that the price of a panel was around $4.50/W, Xcel would basically be paying for the panels while the homeowner paid for the installation. Based on what I’d been hearing, though, it was clear that this generosity wasn’t going to last forever. Also working in my favor was a federal government tax credit that would let me write off 30 percent of the system’s price tag after other state and local rebates had been subtracted, up to a cap of $2000. Suddenly the price of solar was looking very reasonable.
Step 2: Determining the System Size and Getting Bids
I then had to decide how large a system to install. It’s difficult to take an abstract system rating of, say, 5 kW and translate that into real power production. The PVWatts Solar Calculator, created by the National Renewable Energy Laboratory, was a great help. You enter in a few critical variables for your location and the type of solar system you intend to install, and you get a reasonably accurate estimate of how much energy your panels will produce. [See tables, “Results from the PV Watts Solar Calculator.”]
The toughest part about using this tool was figuring out the derate factor. This is a number from 0 to 1 that takes into account the loss in energy your system will experience from things like inverter and transformer inefficiency, wiring, shading, age, and so on. I tried to convince myself that recent advances in solar technology would justify increasing the number—indicating less overall energy loss—but ultimately I left it at the default.
The only quibble I had with the calculator’s results was the cost of electricity. The number it gave—8.4 cents/kWh—was correct for pure power-company rates. But it did not take into account federal and local taxes. The effective rate I pay is closer to 12 cents/kWh.
Next was to assess my energy consumption. A quick call to my energy company gave me my monthly usage for the previous year. [See the table, “Household Electricity Usage in 2008.”]
When I plugged these numbers into the solar calculator, it recommended that my system size be in the 5.5 to 6 kW range to cover all my electricity. Not everyone goes this route, but I did after getting bids from several installers in the area. One bid came in at $25 200 for a 3-kW system; the other was $43 000 for a 5.88-kW system. After factoring in the rebates and tax credits available at the time, the prices were $12 550 and $15 014, respectively.
The upshot was that the bigger the system, the cheaper the electricity was going to be. For example, the price per watt of the 3-kW system after rebates and taxes would be around $4/W. The 5.88-kW system, by contrast, would end up costing around $2.55/W. So for an additional $2500 or so, I could cover all my electricity needs. I decided to go all out and cover 100 percent. To do the installation, I went with REC Solar, a well-established California company that was expanding into Colorado.
Step 3: Site Survey
I signed a contract with the installer and put down a deposit of $1000. Shortly afterward, they did a site survey. That entailed measuring the area of the house’s south-facing roof surfaces and checking the attic space for ease of installation. They also went up to the roof and took pictures of the skyline, using a camera with a fish-eye lens. The photos gave them a good understanding of the amount of shading on the roof throughout the year [see photo and graph, “Partly Shady”]. In a perfect world, we would see zero shading throughout the year. Based on the assessment, though, the house was going to have shading problems in November, December, and January on some parts of the roof.
Next, the company prepared a schematic showing where they intended to place the panels. I found the placement of several panels over the garage to be suboptimal. Specifically, the garage had shading from the second story next to it that the contractor had not accounted for. In my research, one of the weird things I discovered about current solar technology is that the panels are all linked together in series. If one panel is in the shade and starts to underperform, they all underperform to the same degree. So making sure all the panels are unshaded improves the performance of the whole system.
I took pictures of the problem area and showed them to the contractor. Based on their initial calculations, the installation would produce about 6750 kWh/year. When we removed three of the shaded garage panels as well as two other panels from another surface and relocated them to a spot with no shading, we increased the expected annual production to 8263 kWh [see diagrams, “Before and After”].
Step 4: Installation
On a cold morning in December 2008, a crew of four guys showed up with a large enclosed trailer containing all the hardware, tools, and panels. The system I’d selected included 28 Kyocera KD210GX-LP 210-W photovoltaic panels and an SMA Sunny Boy 6000-W inverter, to convert the panels’ DC output into house-friendly AC.
Up On the Roof: A crew from REC Solar installs solar panels on the roof of the author’s garage. Photo: Steven Johnson
They started by drilling holes in the roof to run conduit for the wires [see photo, “Up on the Roof”]. They ran most of the wires through the attic, and the few pipes that were external were well hidden behind gutters and down spouts. Overall the install was very clean and tidy, with very few exposed electrical conduits. The electrician on the crew went to work putting in the 90-kilogram inverter in the garage and wiring it into the main fuse box [see photo, “Solar Hardware”]. The entire installation took about four days. Throughout the process, I kept the crew supplied with hot coffee and tea, and I bought them lunch each day to maximize their time for the installation.
After the work was finished, the city inspector came out and checked the system. No issues were discovered. I also had to fill out a lot of paperwork for the power company so that it would allow me to turn the system on. Among other things, I had to complete a nine-page contract that stated I was responsible for maintaining the system and carrying increased liability insurance and that any changes to the system would have to be reported to and approved by Xcel. My first day of solar production was 25 February 2009 [see photo, “Mission Accomplished”].
Step 5: Assessing the Performance
By February of this year, I’d had a full year of solar production. At the end of each month, I calculate the month’s power production, the net change on the meter (+ or -), and household electricity consumption. If my power production exceeds the household’s electrical needs, the excess power is returned to the grid to help power my neighbors’ houses, and I get credit for that excess. My net meter is basically a flow meter: During the day, if the system is producing more electricity than the house needs, the meter rolls backward. At night, when there’s no solar production, the house draws electricity from the grid, and the meter slowly runs forward. Effectively, the net meter allows me to treat the grid like a huge infinite battery. (Xcel Energy also recently installed 20 000 smart meters in Boulder as part of its SmartGridCity project, which will let customers closely monitor and manage their energy usage. But my house was not eligible. For more on SmartGridCity, see “How Smart Can You and Your Local Electricity Grid Get?” IEEE Spectrum, June 2009.)
Starting in March 2009, I saw dramatic changes in my utility bill, which includes both electricity and gas. The power company also began tacking on a $7.50 per month connection charge for the privilege of having the solar panels connected to the grid. That’s still a whole lot cheaper than having to put in a bank of batteries. In the table, "Savings from Solar," 0the green numbers show how solar production has reduced the total bill: In 2008, I paid $2013 for my utilities. In 2009, I paid just $1043. And from March 2009 through this past February, the first full year of solar production, I paid $899.
The system has generally outperformed the contractor’s estimate. In 2009, it generated 8524 kWh of power, about 3 percent higher than the 8263 kWh promised by the contractor. Over the year, the monthly reading on my net meter has varied from –874, meaning that I banked 874 kWh that month for later use, to +46, meaning that I pulled 46 kWh from the grid. In short, I’m covering 100 percent of my electricity usage. To be honest, I thought it would be a bit higher. I don’t know if I’ve become hypersensitive to the weather, but I swear it was cloudier than usual in 2009.
As it became clear that I was producing more electricity than I was consuming, my next task was to figure out how to use the excess electricity. I bought several space heaters to use in rooms that we spent the most time in. I figured we could turn down the thermostat and heat the rooms with electricity instead of gas.
To date, I’ve done zero maintenance to the system. I haven’t even climbed up on the roof yet to wash the dirt off the panels. Colorado has had some impressive hailstorms since the system was installed, but the panels came through without a scratch.
I got a nice surprise from the federal government. As part of the big bank bailout of 2008, the government continued the 30 percent tax credit on solar photovoltaic installations and also eliminated the cap of $2000. So the cost of my system, which started at $43 000 list and was reduced to $17 000 after Xcel’s power rebates, was cut by another $5100 with the federal tax credit. The final system price: $11 900.
The solar system may pay off further when my wife and I eventually sell the house. If you believe the Web site of the U.S. Department of Housing and Urban Development, any home improvement that saves $1000 a year in energy bills will increase the value of the home by $20 000. So with a $12 000 investment, I’ve raised the value of my house $20 000 and I get free electricity. Even if it only increases the house value by half that much, the system will pay for itself in 2 to 3 years.
In short, I can’t think of any other single change I could make to my house that would cut my utility bill in half and provide me with free electricity.
Mission Accomplished: The back of the author’s house with all three sets of panels installed. Photo: Steven Johnson
The Dark Side of Solar
Obviously, I believe adding solar panels can be a smart idea. But I’d be irresponsible if I didn’t mention solar’s dark side. Namely, components wear out and need replacing. The panels have about a 25-year life, and they become slightly less efficient over time, about 1 percent a year. The inverter has an expected life of 10 to 15 years. To replace an inverter currently costs around $3000 to $5000.
The other worry is if the roof needs to be replaced. To get at the roof shingles, all the panels would need to be removed and then reinstalled. I really hope I’m out of the house before that happens.
Changes Since I Installed My System
In the last two years, Boulder has seen a few changes in terms of solar power. In 2008 the city passed a ballot measure that allowed it to issue up to $40 million in special bonds; the money raised would be used to make loans to commercial and residential green projects. For residences, you would receive a low-interest loan—with rates between 5.2 and 6.68 percent—to cover the cost of the system, and you’d repay the loan through your property taxes. If you sold the house before the loan was paid off, the remaining principal and interest would be assessed for the current and future owners of the property.
The ClimateSmart loan program began in May 2009 and in its first year financed 280 projects for a total of $10 million. Despite its popularity, the residential portion of the program was recently put on hold due to objections from the Federal Housing Finance Agency, the Federal National Mortgage Association (otherwise known as Fannie Mae), and the Federal Home Loan Mortgage Corporation (Freddie Mac). According to a Boulder County press release announcing the decision, the federal policy "prohibits new financing, secured by a Fannie Mae or Freddie Mac mortgage, on properties on which there exists a property assessment for any energy efficiency/renewable energy improvements." Basically, the federal government doesn’t like its mortgages going to properties that have liens on them, even if the lien is for a renewable energy system. The county says it is now exploring other ways to finance such projects.
Another significant change is that Xcel Energy has reduced its solar rewards program from $4.50/W to $2/W, to reflect the significant drop in the price of solar panels as well as the improvement in their efficiency. The price per panel has fallen to $2 to 2.50/W, compared to $4.50/W when I purchased my system just two years ago. Current panels are also about 10 percent more efficient than the ones in my system. So even though the rebates have been cut, the ultimate end cost to the homeowner is basically the same.
Xcel’s electricity rates have changed, too. This year it began using a tiered pricing scheme during the summer when demand is highest. Previously, residential customers paid a flat rate no matter how much electricity they used. Now in June, July, August, and September, the rate is 4.6 cents/kWh for the first 500 kWh each month (or 9.7 cents/kWh with taxes and other fees factored in); then it jumps to 9 cents/kWh (or 14.3 cents/kWh with taxes).
To see how the new tiered rates would have affected my utility bill if I didn’t have solar, I applied the tiered prices to last summer’s usage. I would have seen my electricity costs rise substantially, by 33 percent and 26 percent in July and August, respectively. Needless to say, I would have been really upset to see my summer rates jump so dramatically. But with my solar system, I don’t care; I’m now virtually immune to any changes in electricity rates.
Having had more than a year to wrap my mind around the whole solar power addition, I’m really pleased with the results. Saving $1000 a year in energy bills is very significant, and my family hasn’t had to make any major changes in our lifestyle. If anything, I’m feeling a bit less guilty for running the air conditioner during the summer. When I drive up to my house and see my solar panels, it makes me smile. Sure, it was a major investment up front, but I think the payoff is clear and practical.
And for the record, as of this July I have saved 14 658 pounds of CO2 (about 6650 kilograms). That’s equivalent to planting 36 trees. Hopefully, it was enough to put a smile on at least one polar bear’s face.
About the Author
Steven Johnson is a performance scientist at Oracle Corp., in Broomfield, Col., where he studies all forms of computer-related performance problems. He still can’t explain to his family what a performance scientist does for a living.