Dan and Liz's Solar Retrofit

I've always been interested in solar energy, so when we heard that there were good subsidies available for home solar power, we decided to go for it.

Group purchase being organized

But first, I'd like to mention a group named One Block Off The Grid, which is organizing a group purchase of solar power systems in Los Angeles. Visit their site if you want to see what their package is like.

Picking a contractor and type of panel

We looked around (gosolarcalifornia.org has a list of contractors) and got four bids. The one we went with was Absolutely Solar of Los Angeles.

Three of the bids specified BP Solar panels, which are ok, but the fourth specified the more expensive Mitsubishi Solar panels. Looking at the specs, I found two things that made Mitsubishi stand out: first, their power tolerances are tighter (+-3% instead of +- 9%), which is a mark of quality in my book. Second, they're lead free, which is attractive for someone trying to go green. (One of these days we've got to stop using poisonous heavy metals in everything we build, may as well be now.) So that's what we went with, and how we picked our contractor.

Technical Details

As it turned out, our duplex's large, flat roof was just right for a system that met 80-90% of our power needs.

We are installing 46 panels, each with a peak power rating of 175 watt, for a total of 8050 watts, made with polycrystalline cells, model Mitsubishi PV-UD175. The power is converted to AC by two Xantrex GT4.0 inverters (one for each unit of our duplex). The inverters are supposedly 95.5% efficient.

Cost

The total cost is $69K for our whole duplex before subsidies: for a net cost of $23K. Assuming a 20 year lifespan, that's a net cost of $1200/year, or $100/month, or $3/day. Our power bills are currently about $5/day for about 40 kilowatt-hours. If our unit produces 6 kilowatts for six hours / day, that's 36 KWH/day, which is in fact 90% of our power. (See the solar power basics articles at Green Living Tips and Backwoods Home Magazine for how I arrived at that guess. gosolarcalifornia.org looks like a good source of info, too.)

So, if those assumptions hold, we'll only be paying the power company about $0.50/day, and our total daily cost for power will go down from $5 to $3.50. (It can't be true, can it? Tell me where I pencilled that out wrong...)

Incidentally, if you're a do-it-yourselfer, you might be able to get a better deal by doing it yourself. We were too lazy to even think about this, but if you've got the energy, consider it.

Results - first month

Our system was turned on about February 1st. Peak power so far has been 3.0 KW (about 75% of the panels' 4.0 KW peak rating). During February, the upstairs and downstairs systems generated 371/424 and 433/480 KWH, or 88% and 90% of our power, respectively.

One set of panels produces about 14% less power than the other. This is probably a result of shade (the east side of our roof has a 2-3 foot high false pitched roof, and maybe only one system is in its shade) and fewer solar panels (we don't know which one has 24 and which one has 22). The difference is mostly during the first hour of the day, which makes me suspect shade is the bigger factor.

If the systems were reversed, we'd be at 371/480 and 433/424, or 77% and 102% respectively. Anything over 100% would be a waste because our power company doesn't pay for extra energy. I don't know if our installer was brilliant, or just lucky.

The PVWATTS calculator at nrel.gov lets you calculate estimated output over the year. It predicts a 4KW system will produce 441 KWH in February in Los Angeles (if tilted at 34 degrees). We almost hit that with one of our two systems.

Looking at the power indicator made me want to get to 100% solar power, so now I'm more careful about turning things off when we're not using them. We've set our computer to suspend itself when idle, which saves a lot of power.

Results - first six months

As of mid-August, the upstairs and downstairs units have generated a total of 3000 and 4200 KWH, and our net electrical energy usage is -300 and -1100 KHW, respectively. In other words, we're generating 24% more power than we use.

How much energy is 7.2MWH?

For comparison, it takes 2500 joules aka watt-seconds to boil a gram of hot water. 7.2MWH = 2.6GWS, or enough energy to boil 10 million grams = 10 thousand KG = ten tons of water. (Eight tons for our own use, and two tons pumped back into the grid.)

How much CO2 did we save?

Coal-fired power plants emit about one gram of CO2 per watt-hour. The LADWP says its plants emit 0.55 grams per watt-hour. 7.2 MWH * 0.55 g/WH = 4.0 thousand KG = 4 tons of CO2.

According to a 2005 UK report, the CO2 footprint of PV panels averaged over their lifetime is one-tenth that of LADWP power, ~0.06 grams per watt-hour. So our panels "generated" 0.4 tons of CO2, for a net savings of 3.6 tons.

Burning gasoline produces about 9KG of CO2 per gallon, so this saved the CO2 equivalent of 400 gallons of gas.

How much sun do our panels get?

The 8KW panels are producing 1.7 KW on average, 24 hours a day. That's about 4.8 hours of full rated sunshine per day during the first six months.

Pictures

Here are a few pictures of the system being installed.

Here's a much larger installation a few miles north of us, using the same solar panels (well, they use the 185 watt panels; I suspect Mitsubishi sorts cells by efficiency and sells the less-efficient ones a bit cheaper).

Where does the power go?

Our idle power consumption seems to be about 500 watts. I have no idea where it all goes! Let's see, here are all the things that should be on in our downstairs unit:
What Model Power each#Total power
Refrigerator GE PSF26MGW 615 KWH/year = 70W average170
Cable box Mot. DCT2224 15 watts230
PVR Tivo series 240 280
Cordless phoneATT SL82408 5? 525
Burglar alarm ? 10 110
DVD D-RW2SU 3 26
Night Light incandescent 4 14
Ethernet hub Asante 10? 110
So our idle power should be 70 + 30 + 80 + 25 + 10 + 6 + 4 + 10 = 235 watts.

To check this, I used the Itron Centrol Watthour meters's one dot per watt-hour display and a stopwatch.

(To do this, turn off all lights in the house, turn off and unplug all computers in the house, throw all the circuit breakers to OFF (including the one for the solar feed), and then for each breaker, turn it on by itself and measure how many seconds between dots. (If you don't see any dots change in a minute, give up and move on to the next. You might want to group all of those together.) To convert those times into watts, divide them into 3600. So, if you counted ten seconds between dots, that's 360 watts.)

I had some trouble getting repeatable measurements, but here's my data:
Upstairs
breakersecondswatts
3 bedroom4678
1-2, 4-1013427
18 dining room5072
21 living room5072
sumn/a247
actual measurement of alln/a309
all plus mac8.4430
all plus mac and PC7.0514

Downstairs
breakersecondswatts
7,8 counter, bedroom14325
25 plugs5763
27 garage4678
28 living room plugs (including r3000 laptop)5269
sumn/a235
actual measurement of all18200
all plus kitchen lights8450

I omitted individual measurements of the fridges because they suck lots of watts briefly at startup, then nothing during idle; we probably have to trust the mfr who says they take 70 watts average. To avoid counting them in the actual whole-house measurement, you'd have to unplug them; I just waited a while until they probably finished running.

Note that the sum of the measurements for downstairs exactly matches the estimate I made earlier. This is just a lucky coincidence, though, since my total measurement is lower.

Doing accurate measurement by flipping breakers and counting dots is hard, plan to iterate a few times until you really understand the results. I have not quite finished iterating. A Kill-a-watt power meter would be a heck of a lot easier way to measure things that can be plugged and unplugged.

Orientation of Panels

Charles Landau's "Optimum Orientation of Solar Panels" recommends that, to maximize output during winter, panels should be tilted at (latitude * 0.9) + 29 degrees. For Los Angeles, that's 60 degrees (!). I plan to check our system to compare its tilt with his recommendation sometime soon. (Have to get the big ladder out of storage first...)

Greenhouse gas reduction

There are lots of greenhouse gas calculators out there, but the basic idea is to add up the CO2 emissions from all your utilities. Here's my guess at how to do that in Los Angeles:
total pounds of CO2 emitted = 1.4 * KWH electricity used + 12 * therms natural gas burned in home + 18 * number of thousand gallons of water used
So not only do you want to switch to solar, you also want to minimize your use of natural gas and water.

(sources: [1, 2.1 lbs CO2/KWH for coal, 1.3 lbs CO2/KWH for natural gas], [2, LADWP 45% coal, 33% natural gas], [3, 13 KHW/thousand gallons of water]

Carbon Footprint of Manufacturing Solar Panels

I've heard it asked several times: are solar cells worth it from a global warming perspective, i.e. do they save CO2 when you include the energy needed to manufacture them? Turns out the answer is "yes, they do save CO2". Wikipedia quotes research which estimates that similar photovoltaic panels have a carbon footprint of 25 to 32 grams of CO2 per KWH of power generated (mostly from the energy needed to purify the silicon), compared to 400 g/KWH for a gas-fired power plant.

Another measure is how long it takes before the panels have generated as much energy as it took to make and install them. Wikipedia says that recent studies estimate this at 1.5 to 3.5 years.

Mitsubishi has lots of environmental information online at global.mitsubishielectric.com/company/csr. They even mention at global.mitsubishielectric.com/company/csr/ecotopics/pv/manufacture that at their new solar panel factory, "PV systems will be installed on the roof of the building, reducing carbon dioxide emissions from PV cell production." However, I have not yet seen actual figures on the carbon footprint of the panels we are installing. It would be nice if Mitsubishi provided carbon emission labels on their solar panels so consumers could make informed choices more easily.

But what about solar hot water systems?

We use natural gas to heat our water... and in Los Angeles, of all places, one would expect to be able to use a solar water heater. Perhaps we missed a bet by not planning a photovoltaic + thermal solar energy system rather than just a photovoltaic one. See also:

Update Winter 2011

The downstairs (903 S. Sycamore Ave) power bill usage history (in kwh per two month chunks) is Average daily cost of electricity this month: $2.10.
Average daily usage: 16 kWh.
Our cost per net kWh is $0.0702 base plus taxes and surcharges = $0.14.

16 kWh is a metric f-ton of power use. Where does it all go, again?

Total estimated usage: 31 kWh/day. Minus estimated solar output of 13 kWh/day winter
Net estimated energy usage: 18 kWh/day (not too far off from actual of 16 kWh/day)

Hmm, why didn't we get a gas-heated hot tub, again?

Copyright 2008-2009, Dan Kegel