You can also get cheaper batteries and larger panels off Craigslist for much less. You can easily find much larger panels for around $100.00 - same kind you see on roofs. Batteries are more difficult to find cheap, but 6 volt golf cart batteries are readily available and low cost. Sometimes you can also find the 12V 35 AH wheelchair batteries (same as what the author used) cheaply too.

Point being, if you look around, solar electric generation at a small scale can be done even cheaper than purchasing off Amazon for some of the components (inverters and charge controllers are mostly exempt - unless you go to a hamfest).

Very good point! Hadn't thought of second-hand components. That'd help with the environmental footprint, too!

Unfortunately it isn't a 16% return because he can't sell that system for what he paid. If you buy a stock at $300 and sell it for $348, that's a 16% return. But for him, he probably would have trouble selling all of that for $100 (the battery will be worthless in about 2 years given how far it is being discharged and very few people buy these things on craigslist for anywhere near what they cost). If he can sell it in 6.5 for $100 he has now made $400 in 6.5 years, a 4.5% annual return.

Being able to sell the system is irrelevant as long as the system will last indefinitely (or say, 30 years, which approximates the same value anyway). Everything but the battery is solid state and should last indefinitely -- or certainly more than 8 years. I already excluded the battery as a consumable to calculate the 13% rate of return. But one thing you are right about is the other parts will fail eventually, dragging the return down as well.

Solid state parts can last decades though. The best way to calculate the rate of return would be to actually get accurate numbers for lifespans for all the parts and amortize them. It's probably greater than 4.5% and lower than 13%. Regardless, at near zero risk, even 5% is a great return these days, so it's totally worth doing.

Since you like the stock analogy -- a share bought at $100, that earns $10 a year, and that for whatever reason we confidently expect to keep earning $10 every year for decades, returns a 10% rate of return. The fact that the stock now sells for $50 makes no difference, as long as we believe we are right and the market is wrong about the likely future earnings.

Honestly I think even my number is a little optimistic because I doubt his battery lasts 2 years. Most lead acid batteries are only good for 100-200 100% cycles. I also doubt that the PWM controller or the inverter last 30 years. I've seen failure rates as high as 50% for cheap PWM controllers like that after 10 years. The inverter is likely similar. Sure you can buy stuff that will last 30 years but that will be much more expensive. 8 years seems like an appropriate amortization schedule which puts my $100 at 6.5 years as a little generous.

> as long as we believe we are right and the market is wrong about the likely future earnings

In a liquid market like the stock market, the odds of this being true is small. Recognizing a 10% return against a stock that's crashed 50% borders on, depending on if you're investing your money or others', delusion and fraud, respectively. There is no liquid secondary market for home-solar components, however, so approximating cash flows and amortization schedules works.

(If you've heard the terms "mark to market" and "mark to model", this is what they're talking about.)

Yup, it's very hard to make a good return in the competitive stock market, but easy to make a good return in the market you have exclusive monopoly access to: making investments that cut your own personal costs of living

> Solid state parts can last decades though.

But every year we seem to get better silicon. Better CPUs, GPUS, SSDs, etc. Same holds for photovoltaics.

To run with the investing analogy, it's like claiming you bought a bond with a 5 percent return, when six months later interest rates shift in the bond market. Now the rest of society is getting 7 percent returns, and you have to sell your to sell your bond at a discount to compensate. Sure you can hold the bond to maturity, but you're still missing out on that 2 percent return in the greater market.

To run further with that same analogy, what if you expected bond rates to keep increasing for 50 years? It would not seem wise to keep your savings in a checkings account for these same 50 years, to then get the highest rate.

Yet that's what most of us keep doing with solar, missing out on a good deal because it gets better every year.

Exactly! Thank you!

Assuming these are real, not nominal rates, who cares? I'd rather maybe miss out on the extra 2% than definitely miss out on the original 5%. Why be jealous?

You can also hedge your bets by gradually increasing your investment (maybe even "dollar cost averaging" or some other strategy)

For solar, that could mean getting a small system like the OP describes, trying it for a year or two, then expanding it a little more every year or two with the latest cheapest most effecient panels. This should work even better than dollar cost averaging does for securities, since you know it's unlikely the cost of panels will go up.

Those calculations also assume he'll use all of the energy it produces, every day, with no loss; it's a back-of-envelope best-case calculation, at best.

If we're getting into the nitty-gritty, then let's also include energy cost inflation into the equation. These numbers are for PG&E which is roughly equivalent for SF:

    2001: 0.10c kw/h
    2017: 0.28c kw/h (as per my recent bill)

That's an average of 10% increase in energy cost per year!

https://www.pge.com/tariffs/electric.shtml

Isn't 2001 during the Enron fucking with the West Coast for profit years? A quick review of wikipedia suggests that in 2001 prices may have been held artificially low. A quote by Gray Davis seem instructive:

> "Believe me, if I wanted to raise rates I could have solved this problem in 20 minutes."

2.8x increase over 16 years is a CAGR of 6.6%, not 10%.

(1.066 ^ 16 = 2.8)

Correction acknowledges!

Hahahaha that's an awesome point and not something I'd thought of.

But it seems that on your recent bill they're including the service fee in that – I'm only paying 14 I believe? Also, wasn't energy super expensive in 2001 due to deregulation? (I'd need to research this, not sure).

Given his current setup, you have a good point.

However, with a setup that includes say, his refrigerator, there'll be more than enough load to ensure no collected energy is ever "wasted"

Indeed, this is why the number of solar farms being planned is sky rocketing at the moment across many countries.

Another way to maximise output of your panel is to make sure it's in the best location, taking into account seasonal change of the sun depending on your location, avoid shading etc. The azimuth (direction) and tilt of the panel is make a big difference to your output.

Shameless plug, you can forecast your output using the (currently free beta) Solcast API.

https://solcast.com.au/api/

US forecasts update every ~30 minutes.

I played around with a solar savings calculator a while back out of curiosity, there is an excel spreadsheet on an Australian website I believe that's pretty thorough, but it's not the only one.

In short, I don't think it pays to sell electricity back to any utility (they pay back 1/2 of what they charge you) and a battery system probably won't be worth it unless you truly want to stay off the grid. Get a system that's enough for your maximum daytime usage, that's the fastest way to break even, and spend the least amount of money. For such a system, given my local electricity rates, the break even point was ~15 yrs. a decade ago, and it's 7-8 years now. If you can find cheaper panels/inverter the break even can be ~6 years, after which you'll only be left with a much smaller electricity bill. It even makes sense to add panels later if your usage increases.

The way the break even time has been shortening is very exciting, more people will be able to afford it; and if you can, it's a no-brainer to do it sooner rather than wait.

Thanks to put that in number. I was unsatisfied with the "not good enough" conclusion after reading it, but couldn't exactly say why.

CO2 footprint probably also is better than mentioned. Certainly, it would take 4+ year to pay back for it, but what would have been the footprint for those 4+ years using coal based electricity? It's not only "how does the system perform", it's also "how does it compare".

Well but 6.5 years sound s like a lot of the parts are getting replaced 1.x times which increases costs again.

Wow that's a really good point.