I have made repeated references in past posts to the modest off-grid photovoltaic (PV) system I built to cover a large share of our—again modest—electricity usage. By popular demand, I’ll take you on a tour of the system: it’s history, its composition, and adaptation to my house.
In 2007, I acquired a single, second-hand solar panel—intent on doing something useful with it. Confronted with a variety of options, and eager to explore multiple paths, I purchased a second panel and proceeded to set up a dual system: two stand-alone off-grid PV systems mounted side by side. It was really cool. I was able to power my television console and living room lights off of the two systems, while experimenting with different components and learning to live (part of) my life on natural power. I wrote a comprehensive article about how to size and design such a system, which may be worth reading first. Since that initial success, I have incrementally expanded my system so that I now get more than half of my electrical power from eight panels sitting in the sun. This is their story.
I have enough to say about my solar setup (and PV systems in general) that I must break this topic into multiple posts. In this, the first, I will describe the components, functions, and evolution of the system. In a future post, I will present system performance data and an assessment of efficiency of the various components. Perhaps even later I can explore the impacts of panel orientation, tracking, horizon obstructions, and geographic location.
I was teaching a class at UCSD on Energy and the Environment (for the second time), and was about to get to the part where I described solar power and photovoltaics when I happened upon an Earth Day demonstration on campus. Some guy had a truck decked out in solar panels, and was thumping out youth-magnet music on large speakers. I chatted with him for a bit about practical issues of PV panels and systems, which led to his giving me a sweet deal on a spare panel he had. I figured I owed it to myself and to my students to transcend the theoretical and learn more about the practical side of an apparently important player in our energy future.
To develop an idea of how to put a PV system together, I turned to the Solar Living Sourcebook (12th edition, in my case). This book—a mixture of tutorial and catalog—allowed me puzzle out the components I would need, conveniently interspersed with parts selection and prices. The book also offered an appendix on the National Electrical Code (NEC) suggested practices for photovoltaic installations, compiled by J. Wiles and also available here.
Even then, I was surprised at how difficult it was to find a definitive, comprehensive wiring diagram for an off-grid system. Picking up tidbits from a variety of sources (including helpful diagrams from Outback, and referring to the NEC standards), I was able to cobble together a code-compliant schematic, and proceeded to buy parts and build up my system.
Because the endeavor had a large educational focus for me, I was willing to: purchase a second panel and build two independent systems; try out multiple charge controllers; and spend money on monitoring/data collection. So my expenses are not representative of an installation where saving money on electricity is the primary focus. A small scale (few-panel) off-grid system is bound to come out less favorably than a large grid-tie system from a financial point of view. The latter type can certainly accomplish a financially-motivated quick payback, but my entry was more of the hobby variety.
Parts List for Initial System
The following list isn’t quite at the level of detail you would need to replicate my initial system without a bit of your own design/thought. But it certainly fleshes out the principal components.
System 1: Running Television, DVD/VCR, stereo
- Kyocera 130 W polycrystalline panel, 16% efficient
- Xantrex C-35 charge controller
- 400 W modified sine wave inverter (cheap!)
- Trojan T-1275 golf-cart battery (12-volt, 150 amp-hour)
- Class-T 110 Amp fuse & holder
System 2: Running Two CFL Torchier Lights
- Unisolar 64 W multi-junction thin-film PV panel, 8% efficient
- Phocos CML-20 charge controller
- 400 W modified sine wave inverter
- Trojan T-1275 golf-cart battery (12-volt, 150 amp-hour)
- Class-T 110 Amp fuse
- Current shunts for later use with system monitor
- Extension cords for delivering power indoors
- Lots of #6, #8, #10 stranded wire in red, white, and green
- Quality crimper, crimp rings, heat shrink
- Conduit, feedthroughs, terminals, ground clamp, etc.
Trojan T-105 batteries (6-volt 225 amp-hours) are more frequently seen in PV systems than are T-1275 units. I opted for the T-1275 because I favored a single 12-volt unit for convenience. Later, in researching battery details for the nation-sized battery post, I learned from a Trojan engineer that the T-1275 and T-105 cells use exactly the same lead plates/grids. So they should have the same cycle performance—just packaged differently and with differing capacity. Incidentally, a battery’s storage capacity in kilowatt-hours can be obtained simply by multiplying voltage and amp-hour capacity (then dividing by 1000). So the T-1275 battery, at 12 V and 150 Ah, comes to 1.8 kWh, for instance.
Assembling the parts into a working system was not terribly difficult. It really comes down to lots of stripping and crimping large wires. Heat shrink (especially the kind that oozes sealant/goop as it melts/shrinks) is useful to protect the crimp joints from corrosion.
At the time, I was renting a condominium, and could not make arbitrary alterations to the place. Because I was simply running extension cords inside, I did not need to mess with the condo’s electrical system—and I found a way to get the extension cords inside without drilling any new holes. I was even able to follow existing holes in kitchen cabinet partitions made for supplying the refrigerator with water. I only needed to drill through one wall, with the landlord’s permission, to get into the living room. The panels sat on the carport, and as such (not being attached to a dwelling), I did not have to provision the system with a ground-fault protection device (GFPD). The rest of the electronics and batteries occupied a space outside in a protected alcove of our patio, safe from rain.
Big Lesson: Energy is Precious
Having such a small system, I had to be vigilant about energy use in the living room. The energy I was using had become very personal. I felt it was my energy in a way that I had not remotely felt before. I paid more attention to the weather, and to the forecast (boring as this tends to be in Southern California). Cloudy periods meant we should ration our television watching. A sunny afternoon when the batteries had reached full charge meant “free” energy that otherwise would go unused. Break out the movie!
The kind of energy awareness that accompanies personal on-site energy production—even if representing a small fraction of total use—turns out to have tremendous leverage. That’s because increased awareness and the resulting behavioral shifts transfer to all sectors. You’ll never look at energy the same way. Energy becomes personal; precious. Once you’ve experienced horror at realizing you’ve left the solar-powered lights on while out of the room (unnecessarily draining batteries and making the system’s job that much harder the next day), you’re unlikely to ever do it again, and that much less likely to perpetrate the same crime on any lights anywhere. Similarly, I have found that energy monitoring (as with a TED system) is another effective way to personalize energy use.
Buoyed by the proof-of concept, I settled into fleshing out the dual system a bit, doing things “right.” I added circuit breakers and a monitoring system.
I also tried several different charge controllers, learning the pros and cons of each type. The Phocos charge controller—despite being very affordable—had no equalization mode, so did not seem like a good long-term solution for keeping the battery happy. At first, I added a low-cost maximum power point tracking (MPPT) charge controller for the TV system, moving the Xantrex to the lighting system. A MPPT unit typically recovers 30% more energy than a simple charge controller by performing a DC-to-DC conversion at high efficiency so that the panel can be operated at its optimal voltage—while the battery is fed a reconfigured voltage compatible with its state of charge. The MPPT was useful and good, although it had trouble in lower-light situations, and also had a maximum power input capability of 250 W. With an eye on expansion, I upgraded to a serious charge controller: the Outback MX60 MPPT (affectionately called “the muppet” in our household), capable of 60 A of output current.
The added components to the (still dual) system were:
- 15 Amp DC circuit breakers for the 130 W PV panel and charge controller
- 8 Amp DC circuit breakers for the 64 W PV panel and charge controller
- 30 Amp DC circuit breakers for the inverters of both systems
- Outback “Combiner Box” and extra bus bar to house breakers and shunts
- Pentametric system monitor: measuring three currents and two voltages
- Indoor display (LCD) for Pentametric
- Outback MX60 MPPT charge controller (massive overkill for single PV panel)
It is useful to get breakers that are bi-directional (not uni-polar) so that the orientation in the breaker box is not determined by the direction of current—also providing fault tolerance for improper installation of the uni-polar type (there is no wrong way with the bi-directional sort).
After the system stabilized and was happily powering my living room, I wrote an article for Physics Today on how to build and set up a small-scale off-grid PV system. It would be something of a waste for this post to rehash that work, so I strongly recommend you look at that article to fill in important gaps that I gloss over here, if you have not already (I’ll wait, in fact). It is there that you will find a more complete description of the roles that the various components play, how to size the system, and many other practical tips. In a sense, this post serves more as a detailed system composition and evolution and an update to the original article.
As satisfying as it was to watch movies and entertain guests on the modest system, the house was begging for more. Anything that I could plug into an extension cord was fair game. So I took the plunge and bought seven more 130 W panels, upgraded to a 3500 W Outback VFX3524 inverter (24 volt), and also purchased additional communication and indoor display units for the Outback devices (now inverter and charge controller). The 24-volt inverter demanded that I put my two 12-volt batteries in series, so at this point I abandoned my dual system and consolidated into one. The 64 W panel took a break from the sun.
- Outback VFX3524 3500 Watt, 24 volt inverter
- Outback “Mate” for indoor display and access to advanced inverter settings
- Outback Hub to link charge controller and inverter with the “Mate”
The MPPT charge controller allowed complete freedom as to how the panels were configured. So in February 2008, I switched over to the single system, using two 130 W panels in series. After a mere four days, I added a third panel in series (the MPPT makes it that easy). Three months later, I had four panels running in series. At this point, I had an open-circuit voltage (maximum voltage that panels reach when no current is delivered) in the neighborhood of 80 volts. My circuit breakers were rated for 80 V, so I became shy about simply extending series combinations beyond this—even though during proper operation the voltage drop across the breaker is trivially small. The point of a breaker is to offer protection if something shorts out or goes wrong.
After some extensive tests of panels hooked up in parallel under various states of partial shading—for which I built my own IV curve tracer—I concluded that there was no penalty in configuring parallel/series combinations.
So by May 2009, I was up to six panels in two parallel chains of three panels. That’s about as much as I could conveniently accommodate on the carport roof, so I reined in my ambitions for the moment, even though two panels still sat inside waiting to be used.
At this stage, I was powering a refrigerator that averaged 75 W (50 W in winter, 100 W in summer); the entertainment system, and the living room lights. The extension cords were almost entirely concealed, but further expansion would have required unsightly runs.
The more sophisticated inverter can be configured to sense a low battery charge state—at a user-selectable voltage threshold—switching to utility power input to give the batteries a break. It can also use utility input to recharge the batteries, but I consider this to be cheating, and have disabled this service. I want my batteries to be 100% solar, for whatever reason. My inverter does not export energy back to the grid: it’s a one-way utility connection. But that limited utility connection saves the batteries from deep depletion during poor weather periods. And I don’t have to be vigilant about the battery state-of-charge with the ever-watchful smart inverter on duty.
The Next Big Move (to the present)
In late 2009, we stepped back into the housing market after a crash-hiatus. This meant I could configure my house any way I wanted, up to spousal approval. Okay. Eight panels this time. More power. More stuff connected. But extension cords running through the house was a non-starter. So I set about running standard household electrical cable through the house (attic/walls) to dedicated PV power outlets (colored gray) throughout the house in strategic locations. I put a breaker box next to the PV installation so the dedicated PV circuits would have over-current protection. At present, I have five outlets throughout the house running on PV, plus the direct-wired attic fan. The items powered by the PV system change a bit from time to time (no longer run TiVo; changed fridge; changed television). At present, we run:
- Refrigerator (40 W average)
- attic fan (on thermostat; can easily switch to utility as needed)
- LCD TV (20–45 W when on, depending on Eco mode)
- Entertainment cabinet (stereo, DVD/VCR, Roku)
- Cable modem and wireless router
- “Normal” locations of two laptop computers
- Printer (hog when off, at 9 W!)
- TED LCD display
- garage plug for electronics projects, charging cordless tool batteries, etc.
The main configuration change in the new house—besides eight panels arranged as two parallel strings of four—is the requirement for a GFPD in the circuit. Because they sit atop our dwelling, the panel frames must be grounded, and a special breaker set between the ground and neutral buses that will also kill the power to/from the panels if current begins to flow to ground (e.g., if a short at the panel connects the positive terminal/lead to ground, thus potentially creating a fire-starting arc).
I also added a Lantronix UDS1100 terminal server to form an interface between the serial communication spoken by the Pentametric monitor and the ethernet protocol of the internets. As a result, I can query my system externally no matter where I am (once the home’s router is properly configured). I can also automate retrieval of the data logs twice a day (or more if I wanted) to guarantee that I do not lose any data (monitoring unit stores 1.3 days’-worth at 5 minute intervals). This way, I can disappear from internet access for days on end without losing knowledge of what my precious energy system is up to. My wife wants cat-cams to check up on our cats while we’re away. But I already have in place a way to check up on the PV system and on utility usage via TED—prompting my wife to question which I love more: our cats or my energy devices. I’m smart enough to change the subject.
As far as my blogging duties are concerned, I still owe you an efficiency analysis of my PV system. But for me personally, I’m pretty happy with my current PV setup. I have gained valuable experience through the process of setting up the various stages of the system. I have a system that can move with me wherever I go. The PV system has helped me develop a keener awareness of wasteful energy practices. I don’t have to worry about loss of refrigeration during power outages. The door is open to expansion if I need it: I can always throw more panels on the roof or add batteries for greater capacity. I can add circuits to my house to support more devices.
But mainly, having learned first-hand what it means to build, operate, and maintain a PV system has been hugely rewarding. I’m pretty content with the current setup, and have no burning drive to grow further. After all, we can’t expect growth forever. At some point, it’s nice to sit back and enjoy the steady-state.
Warning: Do Not Try this at Home (Apparently)
A few readers have informed me that the 2011 NEC standards on PV installations have taken the DIY out of solar installations. So doing what I did would now be against code, since I am not an authorized installer. Even John Wiles, who wrote much of the NEC code is not authorized to install a system, and another individual who trains installers to take the test is not himself eligible to take the test, and could not today install the 7 kW system that he previously installed at his home. So here I thought I was doing people a favor by providing information on how I did it myself. Turns out you can’t. Bummer.
What (if anything) holds your panels down in the event of high winds?
And are there any issues with local zoning? I have the “privilege” (sigh) of living in the one and only town in Massachusetts that has special regulations governing the placement of solar energy systems (some guy ruined it for everyone by siting his smack dab in his front yard on a busy street; it was bright orange, almost objectively ugly. So now there’s rules about placement on roof, side of house, back yard, etc.)
I performed calculations to indicate that my panel setup was going to remain grounded for winds up to about 55 mph (about 90 km/h). Unheard of in San Diego. But indeed a freak wind has the potential of making me very sad.
I am not aware of zoning restrictions in my current residential location, but in the condo I had to seek approval from the board. My white-painted wooden frames were attractive enough, and solar power is seen in a positive-enough light that I was never hassled.
55mph is frequently exceeded in some other parts of the country. We had a nice little storm full of microbursts last week that tore down some good-sized trees.
And 55mph is not even a hurricane.
One reason to avoid the use of glass plates in solar hot water systems is that if one breaks in a windstorm, the others can be lifted off (I have seen the results of this). You can design around this, or you can just use something tough like polycarbonate (my father has taken to using translucent fiberglass panels; they’re cheap, and good enough).
You mention a rain-catchment system… is this something fairly common in SoCal, or another of your projects? I am interested in learning more about the system, how it works and what you use it for.
It’s not unheard of in SoCal, but not very common, either. It would certainly classify as “another of my projects.” I have a single “tote” (275 gallon polyethylene cube surrounded by galvanized tube frame; comes to about 1000 liters, or a cubic meter) collecting water from about 1000 square feet (approx 100 sq. m). A half-inch of rain (bit over 1 cm) is enough to fill it up. Our rain is seasonal, and although we only get 10 inches (25 cm) per year, I am storage-limited rather than roof limited—by a large factor.
A 55 gallon barrel atop a steep bank provides 10 psi (70 kPa) of head pressure, replenished by a DC pump connected to the spare PV panel (with a custom regulator circuit in between). A float switch in the barrel lets the pump/circuit know when to stop. We use the system to water our (expanding) garden. It’s still a third full, with our last significant rain being in April. As our garden expands, so might the catchment storage…
“As our garden expands, so might the catchment storage…”
I’m sure it will 😉
We have about the same amount of capacity for our garden but it’s all gone within two weeks of sunny weather. I’m looking to add another 600 gallons or so…
I’m building my own house at the moment and included a 2200 gallon cistern as part of the design. Ideally, considering rainfall data, the amount of catchement area and usage estimates, I should have have something closer to 10000 gallons of storage – but the cost of the tanks was a limiting factor.
It’s hard to have enough water storage capacity…
P.S. that looks like a really nice kayak hanging on the wall…
.. and sorry I’m so off topic.
I’ve been putting a lot of thought into your posts here, especially the one where you show that converting to renewable energy requires diverting energy from current needs to invest in the machinery to produce renewable energy. That there’s no way around it : fossil fuels are super easy, mainly because they come out of the ground as both an energy source and an extremely good STORAGE medium. So as we run out, life will get harder.
It doesn’t mean the apocalypse or rioting in the streets. But with some back of the envelope calculations, I noticed some disturbing numbers. Check this for yourself : if you go to sunelec.com, you’ll see the cheapest panel prices available. They have been offering the lower grade thin film panels at $1/watt for a while now.
Let’s be rationally optimistic and assume that this isn’t the limit. The reason why photovoltaic cells have a big future is because they are something that is very easy to optimize, and each improvement grants an immediate advantage to the developer. Every cell is the same as every other cell in a panel, and every panel is identical to every other panel. The junctions themselves are not particularly complicated, even though the manufacturing process is.
Contrast this to other power generation methods. Nuclear reactors use tens of thousands of unique parts specific to a particular model of reactor. Same with gas turbines and the accompanying production well infrastructure to supply the turbines with natural gas via pipeline. (you should mentally imagine the whole infrastructure to get the gas to the turbine as part of the power generator, not just the piece that goes in the power plant)
So if you want to optimize for cost and performance, what is easier? Making a single cell and a circuitry package more efficient and cheaper, or trying to improve a vastly more complex infrastructure chain?
Obviously that’s a rhetorical question, and this is why solar cells have gotten consistently cheaper year over year.
Anyways, if you assume $0.25 a watt delivered, and assume that the needed inverters bring it up to $0.50 a watt, you still get some nasty numbers. You should do this math yourself. Assume high temperature electrolysis of water to hydrogen, then a high efficiency Sabatier process to methane. The resulting natural gas is pretty [darned] expensive (though fully compatible with today’s entire infrastructure chain)
The voltage rating of a breaker (or fuse or any hard contact) is related to the ability of the device to extinguish the arc that occurs when the contact opens with current flowing through it. DC is more problematic in this regard since an AC waveform goes through zero as part of the cycle. This becomes more of an issue with an inductive load since the collapsing field provides additional voltage which the open contact must withstand.
Exactly: DC is scarier without that self-extinguishing opportunity 120 times per second (100 per second in Europe). If we had DC homes, we’d see more fires. Recreational vehicles (RVs) burn up all the time. And I had a truck burn up once. Those are just 12-volt systems!
We have a very similar system at our remote cabin with the same Outback MPPT controller that you have and a VFX 2800 watt 12 volt inverter. I agree with you that the Outback wiring diagrams are extremely valuable for figuring out what to do. We use four 150 watt panels and about 7.5 kw of AGM battery capacity (when new) on a good clear day we get 2.5 to 2.7 kw of power into the batteries. Our system has to be windproof up to at least 70 miles per hour and be carefully lightning arrested because we get thunderstorms regularly. The reason I chose AGM batteries is that we aren’t there very often in the winter time and the temperature occasionally drops to -45 C and if there are a lot of cloudy cold days the batteries can freeze and AGM batteries are the only ones that can survive freezing without serious damage. Two years ago we left them at about 40% charge at Christmas time and when we came back in February after 2 months of cloudy and extremely cold temperatures they were frozen and at 0 volts they seem to work fine now with a little less capacity. The biggest by far energy hog in the solar power world is refrigeration. We have a little 3.2 cubic foot Energy Star rated refrigerator that will use up to 2.2 kw a day depending on the temperature and how often you open the door. Everything else seems pretty good as long as you unplug it to avoid phantom loads. Older CRT televisions seem to use about as much energy whether they are off or on. What I found is that a solar system really opens your eyes as to how precious electric power is and power in general. It really shocks me how much energy is wasted on phantom loads and unnecessary refrigeration. Have you looked into the costs involved in refrigeration, such as comparing the energy requirements of a chest type super insulated refrigerator compared to a conventional upright refrigerator. I think you will be quite surprised and it will make all the other losses and wasted energy seem quite minor.
I really enjoy reading your blog and enjoy making similar calculations myself (like 60 pounds of wood being equal to a barbecue tank of propane) and I think that gasoline should be worshipped as one of the most wonderful things on earth.
Keep up the god work.
Correction: you’re using kw (power) where you mean kWh (energy) (see useful relations page).
Your refrigerator surprises me. 2.2 kWh per day implies an operating average near 100 W. Our much larger (18 cubic foot) refrigerator averages 40 W, or about 1 kWh per day.
A note about CRT vs. LCD TVs: LCD TVs—whether fluorescent- or LED-illuminated—put out the same power regardless of the intensity of the scene. The LCD acts to block (absorb) light, but the background light doesn’t change (although some smarter TVs also lower overall illumination when no bright points appear in the scene, increasing dynamic range). CRTs, on the other hand, need to shoot fewer electrons at the screen for dark scenes, so that the power output can vary by more than a factor of five depending on scene brightness/darkness. Horror movies take far less energy than documentaries about the Arctic ice, when viewed on a CRT.
Thank you for the correction. It also occurred to me that they are 125 W and not 150 W panels.
It must just seem like the refrigerator uses that much power. The fridge is rated at 75 W so it would only run more or less continuously on days when it is over 30 °C outside or the kids are opening the door every few minutes. I have my eye on a chest type refrigerator with 4.25″ walls that is advertised to be able to run on a single 85 W panel and an RV battery. I really need to get an electricity monitor to be sure of exactly what is going on.
Is there an attachment mechanism for those panel frames to the garage roof? One of the factors holding up installation costs for solar PV is the panel framing and the labor-expertise required to punch water tight holes in the roof surface. I see the convenient flat mounting surface, but is there no concern for wind loads, ever?
See my reply to Joel; I live dangerously, with no attachment.
Your solar panel stands look homemade, do you by any chance have a drawing or maybe a closeup picture?
I just got a couple of panels to play around with and I still need to figure out mounting, and I’d like it to have adjustable angle. A wood frame looks like it would be both reasonable cheap and easy to build.
Ballast trays/pans – very common on flat-roof/zero-penetration set-ups… Handle more wind loading, easy to move or reconfigure.
You make a good point when you mention how much more personal energy use becomes when you are capturing it yourself. I think think this is one of the positive things we can look forward to in our future of energy decline and likely energy decentralization, that when more and more people are forced to generate their own electricity themselves they will gain a deeper appreciation for energy as well as shift their behaviors to become more efficient.
This is something I’m experiencing myself with my Leaf EV. The other weekend I drove 200 km through the mountains and I was concerned about dying in the middle of nowhere (I bring an emergency generator with me that I haven’t had to use yet). But the intense analysis I do with my batteries’ state of charge and how this relates to driving style and range really makes you more aware. As you say, that 24 kW-hrs in my battery is mine to manage and use how I see fit. If I ever get to the stage where I generate it myself with solar panels, then I have closed the loop, except for the manufacture of the PV installation.
Just as adapting to a PV system becomes natural, living with a short range EV isn’t really that difficult once you adjust and now that Level 2 chargers are popping up all over the place. Soon standard EV’s will have a range of 200 km which is more than enough for 99.9% of uses with the emergence of public chargers.
The problem with our current economic system is that people have no connection with the things they’re buying. Most have no idea what’s involved in providing a gallon of gasoline or a cucumber at the local store. The only connection between producer and consumer today is a number — the dollar, mediated by that wonderful theoretical invention, the supply / demand chart. This is basically what people have been reduced to in today’s economic models — statistics feeding into supply / demand charts. This isn’t how humans evolved and I believe that we do not function well in this environment.
Tom, one concern I have always had in a power down situation is water. Like many people, I have a 220v 2ph submerged pump at about 100 ft. How hard and expensive do you think it would be to have a system that could run this pump?
If you can tolerate pumping only during the daytime (or even just sunny weather), then you can have a fairly modest setup without a battery. The only part I am not sure about is whether you can get a single inverter to do 220 V 2-phase, or if you need to gang two inverters together (which I know can work). The power demand of your pump will determine how much panel you need (oversize by factor of two to handle losses and non-optimal illumination).
I’ve chewed on this problem for some time…
My current thinking is that it just isn’t worth the effort/expense to have an on-demand off-grid pumping system.
Instead, I have decided to install a hand pump – I’ve seen pumps that claim to be able to pump from as deep as 350′.
Hand pumps are a simple and robust backup system that will get you by when the grid is down – or even if your PV system isn’t supplying energy.
Jon: another way to work on this is get a different pump. Grundfos has one (I think it costs about $1600, so maybe $1000 more than my last 220V pump). It can run on 12V DC, or 120V or 240V AC, or directly from a solar panel.
Someone else mentioned hand pumps, I looked at the Bison ones, they are tempting. Still ended up being $1600 for my depth well.
(PS – pumps also have a “soft start” feature, so they don’t require a big rush of current to get going. My well guy said it was included on most pumps now. My 4500 watt gas generator runs my current pump fine, while not “renewable energy”, it was about $400)
“So here I thought I was doing people a favor by providing information on how I did it myself. Turns out you can’t. Bummer.”
I’m all for safe practices, but codes should not be so restrictive – if an installation can pass inspection who cares who did it?
Funny how until only recently, residential PV installation was such a niche market that no one really seemed to care…
Is this just a case of attempting to force economic activity – by forcing people to pay other people for things they can possibly do for themselves?
Jon – As Lucas said, accommodating 220 in your off-grid or backup system adds a lot of complexity. Assuming the focus of your system is backup, I would suggest a cheap loud gas generator on the order of 5000W. As someone said, you can get this for $500. If you have an appropriate pressure tank (i.e. fairly large, 40 gallon drawdown), at the worst you would only have to run the generator for a few minutes several times a day. In the meantime your solar system could provide modest continuous quite power.
Tom – I second your observation about personalizing energy use. I built & installed a solar water system this spring, now I resent my teenage daughter’s long showers more than ever, because I can see how much it knocks my tank temperature down.
Last year I had a 3Kw grid tied system installed. 13 x 230 watt panels in series giving a total of 486 volt open circuit. I presume there is no advantage to using different parallel/series connection.
I have kept a daily log and currently over the year it has produced about 84% of the total possible power (due to cloudy days etc). Interestingly even on dull cloudy days it still often produces large amounts of power. On a recent day with a couple of hours of rain (and at no time during the day could one actually see the disc of the sun) and it still produced about 65% of the power produced on a totally clear sunny day a couple of days later. In fact, there were only 20 days during the entire year when the system produced less than 50% of the expected power for that date.