I’ve been running a solar microgrid on my coffee farm for the last 7years. We started with a few golf cart batteries and 4 panels, these days we’re powering 4 houses, 7 cabins, water extraction, treatment, and RO processing, campus-wide fiber network and switches, path lighting, security systems, and a small server rack.
We’re running 6 inverters on our primary system in a three phase configuration, 35kw of panels and 160kwh of lithium iron batteries. About to add an additional 20kw of panels and a test bank of LiTo cells.
Our panels are a distributed set of rooftop mounted panels on various buildings, which also serves to shade the rooftops reducing cooling loads.
We still have to run a generator to supplement charging on dark overcast days, but it’s typically about 100 hours a year. Hooping to get that running on biomass eventually.
It’s strange to me that people in rural areas pay for electricity. It makes no economic sense, at least here in the Caribbean.
I have a solar system in my house in london. 5kw, 13kwhr battery. I am self sufficient from end of march to october.
I recently got a second hand electric car. I bought an EV plug (total fucking ripoff. its a fucking plug with a contactor, RCD and a CAN interface. no way is that worth fucking £600)
It has some basic control to allow me to charge from excess solar. What is not easy to do is charge at night without draining the house battery. Its fine for me, because I have Home Assistant, with enough fiddling I can get all the systems to talk to each other to play ball. (to add to the complication, I'm on a variable rate tariff, so price can be negative or £1 a kwhr)
I would really love a "house power API" that would allow a "controller" to locally control the power behavior of all the things in a house. Because at the moment, a "normal" person wouldn't be able to charge their car and have house batteries and have solar, and optimise for cost.
It's kind of telling/ironic that the author gripes (whether rightfully or wrongly) about PG&E wanting more return on investment due to their risk of doing business in California, but when the author decides to build their own power system to not deal with PG&E, literally almost every paragraph related to their planning/setup contains a mention of some different California/city regulation that the author had to meet or else it's illegal for them to build their own power system. To the point where they had to pay a professional to help them meet all the regulations.
I did build my own solar system too. In Switzerland.
Took me 1-2 month planning and then 3 month building it alone nearly each day. Sept 2023 til Xmas 2023. Got all the hardware from a PV dealer friend on his purchase price level. Even 24 panels I have put myself alone onto the roof. With two persons it was a bit better.
I've got: 420w x 71 Trina solar panels and two SolarEdge inverters. SE10K Hybrid and a SE17k. Also a 24kWh BYD LFP battery.
All prices without state funding:
Offers from local installers for 56*410W Panels without battery were around 65k CHF.
I've paid now 44k CHF including every kind of cost associated with building it.
> I found three companies and gave them my PG&E usage for the past year (about 16,000 kwh) and got three quotes ranging from ~45 — 55k.
Wow these rates are crazy. A 10kW setup costs you maybe €10.000 all-in here in the Netherlands.
What's going on with these rates? Do they already include the ridiculous tarrifs?
A new battery setup for a 20kWh LFP battery + 10 kW inverter + installation is €7000 now.
And dropping, fast.
Assuming batteries and PV come from China, someone in California is making a lot of money or the government is straining the process with bureaucracy costing $30.000 per setup.
Wow, what a fantastic write-up—thanks for sharing this! I’m a San Jose homeowner (and PG&E sufferer) with a homelab that pulls over 1 kW, and I’ve been down the DIY solar rabbit hole for the past two weeks. Based on my research, I’m planning a roughly 9 kW Signature Solar setup:
18 U server rack (~$500)
— total hardware ~$14 760
My big hang-up has been the rooftop work, permitting and inspections—almost no one I call will touch a true DIY system. If anyone here in the Bay Area has recommendations for installers or back-of-house permit-whisperers who’ll partner on a non-Tesla/Sunrun job, I’d love to hear how you made it happen. Thanks again for the inspiring guide!
Great to see DIY early adopters getting great savings here. I think the bigger trend here is lower cost, commoditization, and it eventually becoming a no-brainer for people that have the opportunity/space to be running their own micro grids for cost reasons. The cost of what you need here is still quite high. But making things easier to plug together helps. And of course component cost is coming down.
For example, you can buy kits on amazon for powering your shed or boat and it's essentially a smaller version of what you would put on your house. No electricians needed. No permits required. Here in Germany you can buy balcony solar kits in the supermarket. They only deliver a few hundred watts of power but it's plug and play. And you can get a nice little subsidy to do that. Some of these kits only cost a couple of hundred euro.
I could see that eventually adding a microgrid to a building is not going to break the bank. Car batteries are much larger than what goes in a house and kwh prices are trending well below 100$/kwh now. Meaning it should not cost tens of thousands to get a couple of tens of kwh to store energy. Inverters shouldn't break the bank either. The going rate for solar panels is around 200$.
Mostly current prices for home setups are much higher than the component cost mainly due to regulations, labor cost, certifications, etc. If you go off grid, you can just DIY and you end up much closer to the component cost. But of course long term both component cost and other cost are coming down. With the exception of labor cost probably. Though the skills needed will become more common and you might be able to do a lot of work yourself.
Having done some work around net zero policy I am increasingly convinced that this is the way forward, and indeed that this will be the way things are done normally maybe 10-20 years from now. The concept is called "distributed generation" and in UK each distribution network keeps an "embedded capacity register" which is basically all the distributed energy resources that are connected to the grid at distribution level (i.e. in the local area). Over here the national grid largely operates at or over capacity, which is a very serious problem for the immediate future, especially as more and more power is drained by compute-heavy infrastructure (data centers and such). Distributed generation is an attractive solution for households regardless of which angle you look at it from.
It’s a true shame that the public utility commission is so corrupt that we lose the obvious economies of scale from grid-scale installations of same in favor of expensive rent seeking by a state-sanctioned monopoly.
I also have 60kWh of batteries in my kitchen, but for the average person who doesn’t want to deal with this stuff, having to admin part of the power grid is a tragic waste.
If the PUC and power company weren’t bastards, this could all be in a giant field somewhere staffed by a tiny fraction of the people who have to waste their lives dealing with it in their garages. So many unnecessary struts, so much caulk and EPO switches, so many inverters.
I installed a similar EG4 inverter and battery system and had a similar experience. The big problem I ran into was that EG4 inverters are NOT compatible with my state's power grid (Hawaii) but because I had so much power, it turned out I never needed to connect to the power grid but I am loosing out on selling power. (Check with your local power company before buying!).
Another problem I had was that a week after installing, one of the TIGO units burned up and started a fire that burned a hole in a solar panel. I only noticed the problem when I saw that one of the solar arrays wasn't putting out power. Replacing the defective unit solved the problem.
Also, I had assumed I would need a generator to power through consecutive days of dark clouds but I instead opted for Ford's Pro Charger Station which has a feature that allows you to power the house; no need for a generator. However, its been over a year and not once have I needed this.
My big takeaway is this: having energy abundance is the good life when you have all electric appliances. My EV (a Ford Lightning Truck), hot tub, AC, water heaters etc have been running over a year with no problems and zero costs after the initial investment.
I feel like they overpaid maybe because they got direct to consumer rates for the gear. If you would have went through a full on solar installer that solar system would have come out to less than $15k, throw in having to get the subpanel and a reroof you would maybe be looking at $30k all in. (Not including the batteries, but by the time he got to the batteries I feel like his budget was way overboard even going the non-microinverter route).
Make sure you are buying and not leasing from the company, have that all rolled into a single loan and then you claim the tax credits to help pay for the reroof.
To add to this, they take care of getting the certified roofers, the city permits for both the roof and solar and handle the PTO for you, which from what you called out is even more costs.
This is brilliant. Energy self-sufficiency often associated with off-grid eccentricity,looks engineerable for city homes. Goes to show that the more "personal" our energy solutions become, the more they reflect institutional failure. Wonder how this scales and if it does what are the political implications?
Funnily enough back home along the equator, having a solar setup still is a social signal of luxury!
This disabused me of the notion of doing it myself, lol. This is super impressive. Hope to one day be confident enough in electrical work. That and plumbing are the areas of home improvement that I still just would rather sub out.
But, this is such a cool article. The quotes on solar installs for some reason seem a lot higher than what you'd expect, given all the charts showing the absolute cratering of solar energy prices – but doesn't seem to work out that way when it comes to residential installations.
I'm in Texas and designed our own solar+battery setup in '21. I had an installer handle that bit and the interconnect though. I found the "secret" was the to leave room for expansion.
In Texas, residential grid-tied solar can be up to 20kWh so I chose cells that got me to 19.96kWh
Then I added an 29kWh battery stacks (18kWh total) but each stack could go to 18kWh (36kWh total). Then I also used a pair of Generac inverters which lets me get two more stacks for 72kWh total.
My goal was resiliency, not sufficiency, so I use it grid-tied so I charge during the day and use at night. When there are storms coming, I flip it to "priority backup" which will prioritize charging.. even using the grid to do it if there's not enough sun.
Anyway, it's been in production since January '22 and my annual* electric bill is ~$1100. Even better, when we have an outage (occasional as we live in a heavily wooded area) if our stack is full, we have ~5 days of battery backup for the fridge, a handful of appliances, and our Starlink.. therefore we have safe food storage, safe food prep, and non-terrestrial communications.
Of course, this cost ~$80k to get there in 2021 (aka before crazy inflation)
This is fantastic. Looking at the absolutely massive cost differential between DIY and full-service solar installers, the DIY option looks prettypretty tempting. My main concern was 1) actually getting through the local permitting process, and 2) what a potential purchaser would think of a DIY system when I go to sell the house.
Seeing that somebody has done it is very inspiring, and if I didn't see a high chance of moving in the next 5 years I'd be on it tomorrow.
The author says that the resale value to PGE is 25% of the buying price.
Is there some way to get a better rate by selling CPU time instead, when the sun shines and batteries are full? Especially if you have your own server rack.
I have computing needs that can wait for 12/18 hours. Would it make any sense to have those processed by distributed solar server farms between 11am and 5pm? They just take data in and send results back, using their own "free" marginal kwh.
It would reduce the need the connect to the power grid, but still benefit from selling solar power indirectly.
Cloud providers may just be so much better at doing this already?
I've interested in solar and batteries but unfortunately the return would be too long for us if we hired a contractor and I'm not confident in my own skills to pull it off correctly.
Mostly I want an easy whole home battery just for some arbitrage. Our rate plan has a 4 hour peak period of $0.27/kwh vs $0.14/kwh off peak. Of course the peak time is when we actually want to use a lot of energy running our AC and other appliances.
Unfortunately I haven't found any DIY stand alone battery systems that offer the simple scheduling I need since most seem to be only backup systems.
Even with tariffs batteries are very cheap. Into the 4 year mark my cost per kwh is down below $0.02 because I end up using excess power for all sorts of things because it is "free".
Pretty cool. I'd love to get into solar DIY but it seems pretty daunting. I moved in to a place recently with an (apparently very old) 7kW rooftop solar system without a battery, and I called the original installer to talk about what my options are if I wanted to expand it, either adding more watts or adding a battery. The overall take-away from talking to the installer was that it wasn't worth it because my existing system uses some different, older technology, and that making any substantial improvements to it would change my "NEM" whatever that means. The way I understand it, it would move me into some other, more expensive regulatory regime. I don't know--my head was spinning after the conversation, and I was basically convinced to just ignore the system I have and not touch it.
I have a minor interest in this subject as well, and after some thinking I realized that my personal solar interests are best served by electro-mechanical harvesting of solar energy, first and foremost .. since a full-blown EV solar rig would require the purchase of components I cannot myself personally make very easily.
This summer I'm building a solar oven to cook bread and veggies with .. and if this works well, I'll build a solar death ray to play with while I wait for lunch.
It seems to me that this is a potential route for the popularization of off-grid/local-energy-harvesting movements to gain more traction. Sure, its nice to have a whole roof full of PV panels and a battery bank to sip juice from now and then, but this still requires a heavy investment in foreign-originated parts and materials.
A solar oven/solar death ray, however, is a lot more feasible to produce locally.
Okay I finally did this earlier this year too and it took quite a bit less than that. Started with impulse buy of invert at an online auction end of January. Installed last panels end of April. 3 months. Quite a bit of it was waiting for third parties or just me being lazy and not ordering parts I needed. I've wanted them for years beforehand so I did some research.
For actual labor - it's about half days to install roof racks (I have shingle roof so quite a bit of time on angle grinder). Another half day to put panels on (requires 1 or more helpers), run a cable thru roof space. I've installed 12 panels on 2 facia.
My hack was hiring electrician to install inverter so I can export to grid (I'm in New Zealand).
We’ve recently had some crazy storms and power outages and my Ecoflow units kicked in and worked perfectly. I have a River 2 which I use as a UPS for my Mac Mini + Monitor at my desk and a River 2 Pro as a UPS for all my Ubiquiti networking, Protect cameras, and severs in my racks. Then I have a Delta 2 I can hookup to the two Rivers to add additional charge. Just pulled the trigger on a 3600watt gas generator off Amazon as well incase of extended outages. While not whole home backup by any means it’s a solid setup for electronics and server gear.
Wow. That's ten years of electrical bills where I live in Texas, and you also have to deal with the increased water leakage risks, any maintenance issues, storm problems, etc.
I missed the "power" when I skimmed the headline and clicked the link. I was slightly dissapointed when I finally realised that the OP is in fact not building his own solar system.
With power issues in Nigeria (basically no power from my perspective), I have been building my home solar system. Currently at 9kw solar panels, 10kva inverter with 30kwh Chinese LFP. Spent about $7,000 so far. Looking to eventually get to 60kwh and at least 20kw in panels because of air conditioning needs.
Waiting for government, my kids currently in primary school would probably have graduated from college before they fix the power situation (they are currently fixated on building roads to nowhere).
This solar installation is quite large and quite expensive.
There is a whole cottage industry of DIY 48V 15kWh batteries based on LFP prismatic cells (16x) and special battery case resulting a price of around 1500 Euro for 15 kWh.
A DIY setup is quite doable, Deye (EG4) or Victron make suitable inverters.
The continuous 1 kW power draw I find Ludicrous, probably especially as a European. I would realy rethink what is absolutely necessary. Huge data storage was my hobby but the storage server is only turned on when required, saves a lot of power.
We also setup our own off-grid solar system, almost 4 years ago now. The cost was much lower due to lower import fees and lack of regulations where I live, it was altogether somewhere between $3-4k total, for 16 panels, mounting rack, inverter (with backup inverter), mptt, 12 acid-lead batteries, etc.. We saved so much money, and fewer headaches, since state electricity only came around 5-6h a day and people relied on scammy-mafia generator providers that ask for insane prices.
The technical work is lovely. I would have liked to have seen (though I respect that it might be personal or PII) a working out of the accounting, specifically about the amortized cost of batteries vs just eating PG&E costs and treating them as the battery.
Saving this for when my parents moves - they're on NEM 1 and will freak at the rates for a new solar install. Right now the excess power goes to an EV and then the grid since the power import/export rates are almost 1:1 but after the move these in home batteries are going to be crucial.
Suspect it would have been cheaper to invest money in upgrading the homelab and aircon to more efficient. Not downgrade, but rather moving to newer lith nodes etc. Tackling the usage before bigger battery system often makes sense
Am interested in the DC-AC—DC- Battery-AC vs DC-DC-Battery-AC efficiency argument. As in there must be a DC-DC conversion happening to charge batteries appropriately.
Guy claims he's tired of paying for electricity because of the "ludacris" (sic) costs so he's going to "DIY". Yet, first thing he does is pay a guy to do it.
Building my own solar power system
(medium.com)433 points by JKCalhoun 18 May 2025 | 340 comments
Comments
We’re running 6 inverters on our primary system in a three phase configuration, 35kw of panels and 160kwh of lithium iron batteries. About to add an additional 20kw of panels and a test bank of LiTo cells.
Our panels are a distributed set of rooftop mounted panels on various buildings, which also serves to shade the rooftops reducing cooling loads.
We still have to run a generator to supplement charging on dark overcast days, but it’s typically about 100 hours a year. Hooping to get that running on biomass eventually.
It’s strange to me that people in rural areas pay for electricity. It makes no economic sense, at least here in the Caribbean.
I recently got a second hand electric car. I bought an EV plug (total fucking ripoff. its a fucking plug with a contactor, RCD and a CAN interface. no way is that worth fucking £600)
It has some basic control to allow me to charge from excess solar. What is not easy to do is charge at night without draining the house battery. Its fine for me, because I have Home Assistant, with enough fiddling I can get all the systems to talk to each other to play ball. (to add to the complication, I'm on a variable rate tariff, so price can be negative or £1 a kwhr)
I would really love a "house power API" that would allow a "controller" to locally control the power behavior of all the things in a house. Because at the moment, a "normal" person wouldn't be able to charge their car and have house batteries and have solar, and optimise for cost.
Took me 1-2 month planning and then 3 month building it alone nearly each day. Sept 2023 til Xmas 2023. Got all the hardware from a PV dealer friend on his purchase price level. Even 24 panels I have put myself alone onto the roof. With two persons it was a bit better.
I've got: 420w x 71 Trina solar panels and two SolarEdge inverters. SE10K Hybrid and a SE17k. Also a 24kWh BYD LFP battery.
All prices without state funding: Offers from local installers for 56*410W Panels without battery were around 65k CHF.
I've paid now 44k CHF including every kind of cost associated with building it.
Should write a blog post about it :-)
Next project is a solar fence with 6kWp.
Wow these rates are crazy. A 10kW setup costs you maybe €10.000 all-in here in the Netherlands.
What's going on with these rates? Do they already include the ridiculous tarrifs?
A new battery setup for a 20kWh LFP battery + 10 kW inverter + installation is €7000 now.
And dropping, fast.
Assuming batteries and PV come from China, someone in California is making a lot of money or the government is straining the process with bureaucracy costing $30.000 per setup.
20× 455 W Canadian Solar panels (~$173 ea)
1× GridBoss MID V2 (~$2 400)
1× FlexBoss 21 (~$2 400)
4× Eco-Worthy 48 V 100 Ah LiFePO₄ batteries (~$1 500 ea)
18 U server rack (~$500) — total hardware ~$14 760
My big hang-up has been the rooftop work, permitting and inspections—almost no one I call will touch a true DIY system. If anyone here in the Bay Area has recommendations for installers or back-of-house permit-whisperers who’ll partner on a non-Tesla/Sunrun job, I’d love to hear how you made it happen. Thanks again for the inspiring guide!
For example, you can buy kits on amazon for powering your shed or boat and it's essentially a smaller version of what you would put on your house. No electricians needed. No permits required. Here in Germany you can buy balcony solar kits in the supermarket. They only deliver a few hundred watts of power but it's plug and play. And you can get a nice little subsidy to do that. Some of these kits only cost a couple of hundred euro.
I could see that eventually adding a microgrid to a building is not going to break the bank. Car batteries are much larger than what goes in a house and kwh prices are trending well below 100$/kwh now. Meaning it should not cost tens of thousands to get a couple of tens of kwh to store energy. Inverters shouldn't break the bank either. The going rate for solar panels is around 200$.
Mostly current prices for home setups are much higher than the component cost mainly due to regulations, labor cost, certifications, etc. If you go off grid, you can just DIY and you end up much closer to the component cost. But of course long term both component cost and other cost are coming down. With the exception of labor cost probably. Though the skills needed will become more common and you might be able to do a lot of work yourself.
I also have 60kWh of batteries in my kitchen, but for the average person who doesn’t want to deal with this stuff, having to admin part of the power grid is a tragic waste.
If the PUC and power company weren’t bastards, this could all be in a giant field somewhere staffed by a tiny fraction of the people who have to waste their lives dealing with it in their garages. So many unnecessary struts, so much caulk and EPO switches, so many inverters.
Another problem I had was that a week after installing, one of the TIGO units burned up and started a fire that burned a hole in a solar panel. I only noticed the problem when I saw that one of the solar arrays wasn't putting out power. Replacing the defective unit solved the problem.
Also, I had assumed I would need a generator to power through consecutive days of dark clouds but I instead opted for Ford's Pro Charger Station which has a feature that allows you to power the house; no need for a generator. However, its been over a year and not once have I needed this.
My big takeaway is this: having energy abundance is the good life when you have all electric appliances. My EV (a Ford Lightning Truck), hot tub, AC, water heaters etc have been running over a year with no problems and zero costs after the initial investment.
This is insane. And here I am shutting down nightly the drives in my synology to save 20W.
Make sure you are buying and not leasing from the company, have that all rolled into a single loan and then you claim the tax credits to help pay for the reroof.
To add to this, they take care of getting the certified roofers, the city permits for both the roof and solar and handle the PTO for you, which from what you called out is even more costs.
Funnily enough back home along the equator, having a solar setup still is a social signal of luxury!
But, this is such a cool article. The quotes on solar installs for some reason seem a lot higher than what you'd expect, given all the charts showing the absolute cratering of solar energy prices – but doesn't seem to work out that way when it comes to residential installations.
In Texas, residential grid-tied solar can be up to 20kWh so I chose cells that got me to 19.96kWh
Then I added an 29kWh battery stacks (18kWh total) but each stack could go to 18kWh (36kWh total). Then I also used a pair of Generac inverters which lets me get two more stacks for 72kWh total.
My goal was resiliency, not sufficiency, so I use it grid-tied so I charge during the day and use at night. When there are storms coming, I flip it to "priority backup" which will prioritize charging.. even using the grid to do it if there's not enough sun.
Anyway, it's been in production since January '22 and my annual* electric bill is ~$1100. Even better, when we have an outage (occasional as we live in a heavily wooded area) if our stack is full, we have ~5 days of battery backup for the fridge, a handful of appliances, and our Starlink.. therefore we have safe food storage, safe food prep, and non-terrestrial communications.
Of course, this cost ~$80k to get there in 2021 (aka before crazy inflation)
Seeing that somebody has done it is very inspiring, and if I didn't see a high chance of moving in the next 5 years I'd be on it tomorrow.
Not the detract from the rest of the article, but - it's a company, what did you expect?
Is there some way to get a better rate by selling CPU time instead, when the sun shines and batteries are full? Especially if you have your own server rack.
I have computing needs that can wait for 12/18 hours. Would it make any sense to have those processed by distributed solar server farms between 11am and 5pm? They just take data in and send results back, using their own "free" marginal kwh.
It would reduce the need the connect to the power grid, but still benefit from selling solar power indirectly.
Cloud providers may just be so much better at doing this already?
Mostly I want an easy whole home battery just for some arbitrage. Our rate plan has a 4 hour peak period of $0.27/kwh vs $0.14/kwh off peak. Of course the peak time is when we actually want to use a lot of energy running our AC and other appliances.
Unfortunately I haven't found any DIY stand alone battery systems that offer the simple scheduling I need since most seem to be only backup systems.
https://news.ycombinator.com/item?id=40877337
Thanks to California zoning
This summer I'm building a solar oven to cook bread and veggies with .. and if this works well, I'll build a solar death ray to play with while I wait for lunch.
It seems to me that this is a potential route for the popularization of off-grid/local-energy-harvesting movements to gain more traction. Sure, its nice to have a whole roof full of PV panels and a battery bank to sip juice from now and then, but this still requires a heavy investment in foreign-originated parts and materials.
A solar oven/solar death ray, however, is a lot more feasible to produce locally.
For actual labor - it's about half days to install roof racks (I have shingle roof so quite a bit of time on angle grinder). Another half day to put panels on (requires 1 or more helpers), run a cable thru roof space. I've installed 12 panels on 2 facia.
My hack was hiring electrician to install inverter so I can export to grid (I'm in New Zealand).
Wow. That's ten years of electrical bills where I live in Texas, and you also have to deal with the increased water leakage risks, any maintenance issues, storm problems, etc.
Waiting for government, my kids currently in primary school would probably have graduated from college before they fix the power situation (they are currently fixated on building roads to nowhere).
There is a whole cottage industry of DIY 48V 15kWh batteries based on LFP prismatic cells (16x) and special battery case resulting a price of around 1500 Euro for 15 kWh.
A DIY setup is quite doable, Deye (EG4) or Victron make suitable inverters.
The continuous 1 kW power draw I find Ludicrous, probably especially as a European. I would realy rethink what is absolutely necessary. Huge data storage was my hobby but the storage server is only turned on when required, saves a lot of power.
I got dinged for this on the first inspection of my own self-installed solar+battery system.
A lot of signage companies don't know how to do the placard, but the company that I found that knows how was PVLabels:
https://www.pvlabels.com/04-645-map-placard/
And not a cosmic solar system.