Power went out at 10:47 on a Tuesday night. I remember because I was on a call with a supplier, and the laptop dropped off the network. My wife came into the office holding a flashlight and said the fridge wasn't cooling anymore.
That's not a disaster. But when it happened twice in one month, I decided to do something about it.
I'm a procurement manager for a mid-sized manufacturing company. I've spent six years tracking invoices, vendor quotes, and hidden fees. So I didn't start by shopping for a battery. I started with a load calculation.
Why I started with loads, not batteries
The first mistake people make is picking the battery first. I almost did that. A friend said, just get 10kWh and you'll be fine. That's how you overpay.
I went through the house and listed what had to work during an outage:
- Fridge and small freezer
- Internet router and a laptop
- Well pump, on for about 15 minutes per hour
- Four LED lights
- Phone chargers
That gave me a rough overnight load of 3 to 4 kWh. With some margin for cloudy days, I decided on a 7.1kWh usable battery bank. Bigger isn't better. It's just more money sitting idle in a cabinet.
The quote that made me question turnkey
I called one company for an all-in-one installed system. The quote was $13,400. The line items said: battery system $9,800, installation $2,100, other $1,500.
That 'other' line to me is a red flag. What is other? When I asked, it was permit, extra cabling, and configuration. I get that, but I don't like 'other'.
Let me rephrase: I don't think turnkey systems are bad. I think that particular quote was bad for someone who already knows how to compare costs. I knew I could build the same idea around a Pylontech battery US3000 for less.
Building the Pylontech 48V system
I chose two Pylontech battery US3000 modules. Each is a 48V LFP battery with a built-in BMS. Two modules gave me about 7.1kWh of usable capacity. If one ever fails, the other still carries the loads. That's a nice failure mode, and a big reason I prefer modular stacks over a single sealed unit.
I also looked at the Pylontech Force H2 battery. That's the high-voltage route, and I think it's excellent for a whole-house design with a high-voltage hybrid inverter. But my setup was a 48V system, not high voltage. The Force H2 would have meant a different inverter and a different panel. For a small critical-loads setup, the US3000 stack was simpler and cost less.
The first lesson was compatibility. I assumed that if a battery is 48V, every 48V inverter can work with it. Didn't verify. Turned out one of the first inverters I ordered wasn't on the manufacturer's compatibility list. I had to return it and buy the compatible model. That cost me about a week and $75 in return shipping, though I might be misremembering the exact fee.
The hardware prices I tracked at the time were roughly:
- Two US3000 modules: $1,800
- 5kW hybrid inverter: $1,100
- DC breakers, cables, busbars, and mounting: $350
- Permit and electrical inspection: $120
- Licensed electrician to connect to the grid: $800
Total: around $4,170. The electrician's cost is not something to skip. Grid connection has to be done right. At least, that's been my experience with a 48V setup and a pretty particular local inspector.
The 10 amp solar charge controller mistake
I almost added a small solar panel with a 10 amp solar charge controller to keep the battery topped up. I had one on an old RV battery and assumed the same logic would work here. Wrong assumption.
This is a 48V LFP system with its own BMS, and the hybrid inverter is in charge of charging. A 10 amp solar charge controller is not designed for that job. If I had wired it in, I probably would have confused the inverter and maybe damaged the controller. In the end, I used the 10 amp solar charge controller for a separate 12V shed battery, where it was actually useful.
Monitoring: don't overbuild it
I wanted monitoring because I don't like guessing. Someone on a forum recommended a professional broadcast control and monitoring system. It sounded exactly like the kind of thing a procurement guy should want: alarms, remote access, logs.
Then I looked at the total cost. The broadcast control and monitoring system needed a gateway, a subscription, and more network setup than my home needed. The inverter has a built-in monitor and an app. That app is a control and monitoring system, and it's enough.
I have mixed feelings about that decision. Part of me still wants the professional dashboard. Another part knows it would have added complexity for no real gain. I can see state of charge, grid status, and alerts on my phone. That was the goal.
The final math and what I'd tell someone else
The surprise wasn't the hardware cost. It was how much time I spent checking compatibility. I want to say the whole process took eight weeks, but don't quote me on that. Installation itself was a weekend. The rest was reading datasheets and cross-referencing compatibility lists.
The turnkey quote was $13,400 for about 10kWh. That's roughly $1,340 per usable kWh. My Pylontech system was about $4,170 for 7.1kWh, or around $587 per usable kWh. Is the turnkey option overpriced? Not necessarily. It had a longer warranty and a bigger physical capacity. But for my needs, the modular route made more sense.
If you're searching how to build a home battery backup system, start with your loads, then the battery, then the inverter compatibility list. Don't start with the battery price. The price isn't the price. What I mean is, the cheapest quote is often the one with the most hidden cost.
The next time I do this, I'll also model the BMS communication path before ordering anything. Every battery and inverter pair needs to speak the same protocol. Inverter compatibility lists are there for a reason. I know that now.
Would I do it again? Yes. But I'd also pay more attention to cost per usable kWh instead of total system price. A battery backup is a long-term asset. I'd rather spend time in a spreadsheet now than pay the difference in hidden costs later.
