We just signed up for a solar install.
O.M.G.
One horrifying power bill after another, and the looming lowering of the Solar Energy Tax Credit, and it was time to do something drastic.
Some Basics You Need To Know
Cost per Watt
It's common to compare systems in terms of cost per watt. That's the total cost you pay for the materials and install, divided by the watts your system is rated for. If you put 30 panels on your roof, and each panel produces 290 watts (max output), that's a 8,700 watt system (30 panels * 290 w/panel).
If that system costs you $28,275 out of pocket (i.e. before incentives, rebates, or tax credits), then your system cost you $28,275 / 8,700, or $3.25 per watt. Using this, you can create a rough comparison of vendors and their offerings. There are finer details to consider, but if you have two vendors offering roughly the same products, and you want the lower price (often, but not always, a good choice), you can reduce them to price per watts and go from there.
Propaganda - I mean, Sales Pitches...
Here are some things you'll hear and read:
- Over 30 years, you'll pay X dollars in utility power, and it's only getting more expensive.
- These panels offer the highest efficiency/wattage on the market.
- These inverters/optimizers are 99% efficient.
- Solar payments will be lower than your existing power bill.
- Warranties!!!!
First, yes, power is getting more expensive every year. They usually also tell you how much you'll save by going solar. This is a gimme. The numbers are legit, but let's be real, this is an argument to go solar, not to go with a particular company.
The second is not untrue, but higher efficiency and higher wattage panels usually come at higher cost. They will take up less room on your roof for the same kWs, by the numbers. That's great if you don't have a lot of space, or have high-enough needs that you need to squeeze every watt you can out of every square-foot of roof. But will you see a return on that investment? For example, a 21kW system will require 65 panels at 325W per panel, and 55 panels at 385W. At $209 and $319 (prices of a couple of comparable panels I pulled from a solar component reseller), this yields $13,585 and $17,545, respectively, for the same total kW of generation. You need to crunch the numbers and not get blown away a mere data-point.
The third is another data-point, and this one makes me mad because it's misleading when presented the way it is. Optimizers and inverters condition or change the power (respectively), and to do this, they need electronics. No electronics are 100% efficient. So, touting this is just a comparison between one brand of optimizer with another, or one brand of microinverter with another. For our calculations, the efficiency they report translates to how much power you're going to lose to your optimizer/inverter. They consume power to operate.
Fourth, generally another gimme, but be careful. Companies often show up and look at your bill, then quote you a system that is $20-$50 less per month. But the price per watt might be through the roof (read on for examples, and no pun intended)! What seems like a great deal, relative to your current energy costs, might not be a great deal compared with other offerings. Even the same tech installed by a different company can yield savings.
Finally...Warranties. OK. These appear to be important. But like the other points, they need to be considered in the overall picture. Most solar tech seems to be quite robust, so offering a 25 year warranty on something that has an extremely low probability of failing is, well, almost pointless. Not entirely, but...almost. But everyone is doing it, and if it signs up more customers, so be it. On the other hand, at least you have that level of assurance that your new system will reach its maturity (and then some) before it starts having costly problems. If you opt for a more complicated system (i.e. microinverter or optimizer), those warranties will probably come in very handy.
Our Usage Needs
A 12-month moving average of our kW usage showed a recent usage trend of around 2300 kW / month. I set our target for 2400 kW / month of generation.
We have about 5.2 sun-hours (yearly low) in my location, and luckily no trees or buildings to shade our roof. We could expect to get a decent amount of input. If our per-day goal is 80 kW, then we'd need a system at least 15 kW in size. That's the layman's sizing; actual sizing is more complicated, and needs to take into consideration the fact that you won't generate that much power from your array every day, or every month (seasons, sun location in the sky, etc).
Every system we were quoted went above 17kW. Where I'm at, any system about 11.2kW is called a "type 2" system, and requires more gravy to get the train moving.
DIY Solar Options
If you are willing to forego better warranties, and you can either find someone to install it, or are willing to drill a few hundred holes in your own roof and reap that harvest during the next big rain, then you can buy kit systems online for $1.50 / W. I'd certainly seriously consider it for a ground-based system, but not a roof system. That said, the law may also not be on your side (and it wasn't on mine).
We need a certified installer to make the power company happy, otherwise no grid-tie for us. Why grid-tie? Batteries are too expensive, the ROI is just not there; and grid-tie is really the only option here. So, we had to find a vendor. We asked around for buy-and-install pricing, and found exactly one willing to do it, but the warranties offered on whole systems were just too appealing. Plus, there is a ton of work and permitting that has to be done with a solar install.
If you can do it, go do it. If you are old and tired, and have multiple kids driving you crazy and a need to get stuff done, better to let a good installer take it on.
Finding a Vendor
After a fateful visit from a door-to-door solar salesman, I started deep-digging into vendors.
I found a list of the best local installers, who had hundreds of ratings and reviews each, plus certifications or accreditations from NABCEP (that's important): The North American Board of Certified Energy Practitioners. I contacted four of them. At least one was based in California. The others were local of various size.
Spoiler Alert: a local company, and NOT the door-to-door guy, came out on top. A good company also shouldn't hound you or pressure you. Watch out for gimmicks like "improving your A/C system to cut down on your solar bill" - these just equate to easy money for them, and significantly higher $/W for you from your solar system. And no, the guy who offered me that didn't have a crystal ball, but he should have for making such sweeping predictions about how great my A/C would be after that $5,200 treatment.
My takeaway: If you have already optimized your energy savings in your house, then size your system to cover your needs. Don't under-size a system in anticipation of how much power you can maybe save after fixing your home's energy leaks. That's just putting the cart before the horse.
The System Quotes
All sorts of different systems were offered: string-inverter, optimizer, micro-inverter, different kWs, even one A/C treatment. Names hidden, we received the following quotes (cost per watt):
- A: $2.82 / W
- B: $3.11 / W
- C: $3.40 / W
- D: $4.02 / W
Lower is better for the wallet, but only if the other features match your needs and expectations. We were offered systems of sizes:
- A: 22.4 kW (string-inverter)
- B: 19.24 kW (optimizer)
- C: 22.11 kW (micro-inverter)
- D: 17.985 kW (micro-inverter)
All $/W values are raw (before tax credits).
The Technology
There's a lot of information, and misinformation, on the Internet. I'll try to sum up the realities here.
Inverter Tech
You have three major choices. All inverters need to find the Maximum Power Point (MPP) of the panels - this the point in the panel IV curve that gives maximum Power output (where Power is Current (I) times Voltage (V)). It's the "knee" in the IV diagrams you'll see for panels, where it stops going sideways and starts plummeting like a rock.
The MPP varies with sun, shade, soiling, etc. There are different algorithms for finding it and keeping it optimal. A less-intelligent inverter can get stuck in a local maxima. Smarter ones will occasionally sweep to ensure the global maxima is actually the one it's on.
The String Inverter
Large inverters on the side of your house or garage, ideally near the meter. No electronics on your roof, just panels. SMA makes one of the best, and claims to compete with the newer tech.
Overall, it's a simple, durable system with the fewest parts, and time-proven technology. On the downside, it might not offer the overall energy generation of newer tech. One reason for this might be it doesn't optimize as frequently as it could or should; another might be that individual panel losses add up more in a system like this. However, take all this with a grain of salt: the research on all these claims is extremely lacking.
Optimizers
Basically DC-to-DC power conditioners. They send ideal voltage to a central inverter, modulating the amperage accordingly all over the array. The idea is to make the inverter work as efficiently as possible, by insuring it always has the right voltage. An under-performing panel will be adjusted at the optimizer, along with all the related higher-performing panels, such that the total voltage is what's expected.
There are costs for this: every panel needs an optimizer. The optimizers need power to do their job, and they are not 100% efficient at the transformation. The sales sheets will tout 98% efficiency like it's a great thing (and compared to less efficiency, it is). But that's 2% loss from every panel during the conversion process. What that means in terms of actual power production, I do not know. It could be 2% voltage loss, or 2% amperage loss, or even 2% overall power loss.
If it's a power loss, a 70 panel system at 320 W per panel would lose 462W due to the conversion loss.
Also, because every panel needs one, they are up there on the hot, hot roof, under a hot, hot panel. So, there is a reliability concern. You also still have a central inverter, but it might be smaller than the string inverter unit.
On the upside, if they do indeed perform better in intermittent shade scenarios, the losses during high-sun might be made up in gains during less optimal solar conditions. You also get per-panel statistics, which is nice for diagnosing panel issues...or, I guess, bad optimizers.
Micro Inverters
These are like optimizers, but instead of doing DC-to-DC, they do DC-to-AC right at the panel. No central AC inverter is required. I'm not sure if they require any special junction equipment, but if you wanted ultimate efficiency with maximum shade tolerance, this is probably the way to go.
Of course, the cost here is that the micro-inverters are significantly more complicated than optimizers, and like optimizers you need one for each panel, at the panel itself. So, complicated electronics in a harsh environment. Again.
Like optimizers, you get per-panel info. You also get a pretty hefty price-tag increase.
But What the Heck is a "String"?!
If you have 70 panels, you're not going to run 70 individual wires down your wall into your inverter or meter box. You're going to group, or "string," them together, upwards of six at a time. Usually wired serially, to kick up the voltage while keeping the amps relatively low. Each string will run down to the inverter. Each inverter can take in so many strings.
For the SMA, a 70 panel system will require 3 SMA boxes (at least of the model we're getting - OOH SPOILER ALERT!): 6 panels per string, 4 strings per inverter, 3 inverters.
Panel Tech
People like to say that panels wired serially in a string will fail like Christmas tree lights. But that is not the case for most modern panels. For both power production and - more importantly - fire concerns, bypass diodes are pretty standard. Most panels seem to set up groups of cells, and wire them to a bypass diode, with multiple diodes in a panel. This means that if part of the panel is obstructed or fails, the other groups of cells can keep working. If the whole panel is obstructed or fails, the whole panel can be bypassed.
Panel efficiency is a rather stupid number: aside from telling you how many watts per square meter you can generate, it tells you really nothing else, and ceases to be applicable very quickly during most actual operating conditions. If you really wanted to compare apples-to-apples, and you had nothing else handy, panel efficiency will tell you the panel that will give you the most for the least space. Is that the whole truth, though? No.
There are two kinds of degradation you have to be aware of, at least to filter out the bull that will be handed to you by the sales guys: temperature, and standard lifetime panel degradation.
Panels perform worse when hot, and they give their stats on their spec sheets. I studied two: a 320W and a 370W panel. The 320W will put out 320W in optimal light conditions (read the spec sheet if you want to know what they are, and don't expect them to be numbers that match actual sun-obtained reality), at 25C. My roof can easily get to 70C in places during a nice hot day, which should drop the 320W panel's production to 280W.
Panels degrade over time as well. Industry standard warranty is 80% production at the 20 year mark, with the degradation being linear (meaning you should NOT see a lousy 80% production on year 3). Below I'll talk about how linear degradation isn't quite the selling point the salespeople wish it was.
An Estimate - PVWatts
The
PVWatts tool is pretty awesome: you can enter in your address, punch in some configuration details, and get a simulation on how your system will perform. I did this for the two top systems quoted to us. In order to get relevant results, I had to break the systems down into three estimates for the tool, since the tool doesn't handle multiple roof faces.
Here's my yearly production estimates, after adding my roof simulations together:
- A: 22.4kW => 31,101 kWh
- B: 19.2kW => 27,071 kWh
I can only hope that PVWatts actually accounted for panel losses due to heat, but I honestly am not sure how much it did. It's probably in the tool documentation. One thing I could do, however, is calculate system losses over time. Given each systems' panel specs, I calculated the amount of degradation I could expect at the 5, 10, 15, and 20 year marks.
At the 20 year mark, I had the following two numbers:
- A: 22.4kW => 26,125 kWh
- B: 19.2kW => 25,643 kWh
Even though the A panels degraded faster, the larger (yet cheaper per watt) system still produced more at the 20 year mark.
Extrapolating for Like-Kind Analysis
Is it worth it to beef up the B system and get both the better life and greater overall generation? Let's look.
If we resized the B system to match its sizing to the A system, it should produce 31,582 kWh per year on year 1. For that amount of array, the price tag would increase by $9,952. An already expensive system (by comparison to A) would go up in cost by 16%. This, to get a return of 481 kWh over the course of the year.
Other Considerations
There are other things you need to look into, when shopping solar installers. They are just as important as the tech.
Financing versus Cash Price
Some vendors will sell you a much better cash-price, though one potential downside of that is you might not be able to claim any closing costs or origination fees for the federal tax credit. Of course, that'll all be moot once the tax credit vanishes.
Financing Terms
Most vendors will sell 20 year financing. 25 years is probably not a good deal, and the "lower" payment might just be equal to or worse than a better 20 year system. The internal pricing of parts and labor is completely opaque until after you've signed (and even then you might not see it). Shopping around seems to be the only way to ensure that both your terms, and your total system costs, are optimal.
Warranties
A good vendor offers warranties. Our select vendor offered warranties from the manufacturer (so we're told) that go beyond the standard, due to their rating with the NABCEP.
Panel warranties come in two parts: performance, and the panel itself. The performance warranty covers that linear degradation I discussed above. The panel warranty should cover a panel outright failing, de-laminating, or suffering some other catastrophic pre-end-of-life fault.
The roof penetration warranty is your installer saying your roof won't leak for some period of time after the install is complete. Ideally, it's at least the life of the roof. One vendor put a cap on top of that (albeit it was 25 years). Our selected vendor said they'd restart the warranty following a reroof reinstall, and warranty for the life of the new roof.
The inverter needs a warranty as well, and it will often be different from the panels.
Finally, a workmanship warranty will cover a job-well-done (or maybe not so much). The longer the better: if something goes belly-up, the vendor shouldn't be charging you to fix their mistakes.
Reroof Costs
If you need a new roof, it is not advisable to let Handy Roofer Dude take your Expensive High Voltage Solar System off your roof. Let the solar installer do it. It's another cost, of course, but varies widely. Prices are sometimes listed per panel (like $50/panel or $100/panel for take-down and reinstall, plus storage fees). Prices are also sometimes given in flat rates. Most vendors couldn't guarantee what the panel removal and reinstallation costs in the future would be, but I settled for a "right now" cost estimate.
The Tax Credit
Get it while you can. Once you do, you have five (5) tax seasons in which to fully recover it. This means that if you can't claim enough taxes to get your full credit from the IRS, you can roll over the remainder to the next year. After five years, you're out of luck, but I imagine for most people this won't be a problem.
Secondly, my thinking is that it's stupid to spend a dollar to save a quarter. While it may be appealing to buy a larger system to get a larger tax credit, think about what you're doing here:
First, every solar loan operated with the expectation that you'd take that nice, fat tax credit and stuff it into your loan repayment by the 18 month mark, otherwise they'd re-amortize you at a higher rate. Ironically, at least with one lender if you pay more than that by the 18 month point, you can get re-amortized to a lower monthly payment.
Second, the overall cost of the system is something you have to pay, credit or no credit. Even after the tax credit, you're only getting a 26% discount (for 2020). A $60,000 system is cheaper than an $80,000 system, and it will STILL be cheaper after you take that 26% off the price tag. It seems to me the best way to overpay for your solar is to try to get your credit as large as possible. Just Don't Do It. Get a good system and let your actual power savings do the rest.
Additional Insurance
We have to obtain a $1 million general liability policy. That amounts to about $19/month with our insurance provider, and is called an "umbrella policy". Not every power company requires this, but ours does. Lovely. It's something to keep in mind, though, when you think about your total monthly cost.
Unused kW Credits
If I produce more power than I consume, my power company will roll those "credits" from month to month. The idea here is that I can generate more than I use in the cooler months, and use more than I generate in the hotter months, and still come out net-zero for overall power cost. At the end of the year, my company will supposedly "buy back" the unused credits, effectively zeroing me out to start afresh on the new year.
The buy-back isn't much - it's significantly lower than what they charge for their power. But meh. As for the production goals, I took a close look at the per-month report from PVWatts. After comparing each month of simulated generation to the previous matching billing cycles, I will probably pay for some power in January and February. The rest of the year I should be good. I should generate enough to get a bit of a refund back. In real numbers, I should go from spending $3900 / year on power to $133.
Remember, however, I still have a loan to deal with. That's the trade off. But, if the utility rates continue to increase, it will speed up my effective ROI, since that's more expensive power that I'm no longer buying. I trade a highly variable, increasing power bill for a relatively stable and fixed solar loan. The catch to this is generating enough to make that a reality.
In Conclusion
There are many ways to analyze the value and return of solar.
- In 10 years, I would have paid my power company enough to have bought my chosen system for cash (after tax credits).
- In about 16 years, I should see my ROI from the amount of savings I should get by going solar. (That's my break-even point)
- After the system is paid off, it will continue to generate power, even if we move and rent out the house. Offering a relatively fixed, low power bill to the renters will probably be very alluring.
That said, what remains to be seen is the actual system install and performance. I will write again once I have more to say about that. The install process should take only about 3 days, including a meter panel swap. We're going with the string-inverter system, for simplicity and because our shading is practically non-existent. It's hoped that however the panels are wired, they will perform optimally regardless and we should see at least what PVWatts estimated for production.
