Some Numbers on Solar Power

8 October 2008

I have a home construction project looming in my future, and I decided that would be a good opportunity to upgrade my environmental footprint (and reduce my energy bills) by adding a solar electrical generation system to the mix. So I did some research on what it would cost me and what I could expect to gain from it. My summary: we aren’t there yet. Here are the numbers:

The performance of photovoltaic panels (ie. solar cells) is specified based on an assumed insolation (the amount of energy hitting the panel in the form of sunlight) of 1000 Watts/square meter, which is approximately the amount you would get at the surface of the earth at noon at the equator (in other words, under perfect conditions). If you extend that out over a 24 hour day of constant sunlight the insolence would be 24 kiloWatt-hours/square-meter/day.

To estimate actual performance you need to know what to expect for insolation where you are. A little research online found estimates for average daily insolation at various places in the United States derived from a NASA data set at:

Since my small town in southeastern New Hampshire was not among the locations listed in their tables I estimated what I might expect by averaging the numbers for Boston, MA (1 hour to the south and on the coast as I am) and Manchester, NH (45 minute east and inland) to get an average daily insolation at my house of approximately 3.6 kiloWatt-hours/square meter/day. That means I can expect to average about 15% (3.6 / 24) of the rated performance from any solar panel I use.

I next looked for a solar panel to use as my test case. I found a decent one at, a Sharp ND-216U1F, rated for 216W with a surface area of about 1.2 square meters (a bit more than 18% efficiency, which is pretty decent for the current technology). Applying the derating factor determined above, I could expect to average about 32 Watts out of that panel; over the course of a year it would produce about ( 24 hours/day x 365 days/year x 32 Watts ) = 280 kiloWatt-hours/year of energy.

At the website that panel is sold for $1088.

I next went to my local electric company (Public Service of New Hampshire) to get their electricity pricing ( At the moment residential electricity, including all taxes and tariffs and surcharges, is about 14.6 cents/kiloWatt-hour. Hence to pay for one year’s worth of solar panel output bought instead from the electric company would cost about $41 at current prices.

Now we can do the comparison: how many years would it take, saving $41/year, to pay for the $1088 purchase price of the solar panel?

Answer: ~26 years

This, of course, ignores the costs of panel intallation and the power converters/grid ties required to make a solar system work, and it ignores any maintenance costs on the system and effects of solar cell aging that reduce efficiency. Hence this estimate is likely low — it would probably take longer to break even. On the other hand, I didn’t try to predict future energy prices. If they rise faster than inflation the break-even time would get shorter.

Based on this I’ve abandoned my plans to install solar panels as part of my project. Maybe when the next generation is ready….

Update, 03/17/2018:  Since this was published, prices on solar panels and associated equipment have come down considerably.  Moreover, there are now new business models that allow you to get a solar system installed without the large up-front purchase/installation costs, and ‘net-metering’ arrangements that allow you to use the public grid as a battery, which make the cost/benefit tradeoffs much more attractive.  Thus, I now have a system installed that — averaged over the course of a year — produces nearly as much power as I consume.  That caveat — about averaging performance over a year — is worth its own, separate, discussion and I will write a new article about it once I get the time.

Update, 08/05/2018: I have now written about my experience with the new solar array here.

(C) Copyright 2008, Augustus P. Lowell

Leave a Reply