This article follows a single source video — DavidPoz's Build Your Own Solar Battery and Save Thousands, watched more than 1.4 million times. The premise is simple: a pre-built 48V server-rack style lithium battery costs about $300 per kilowatt-hour, and David wants to see whether a DIY build from new cells can land at roughly half that. Inside the crate he opens are 280Ah LiFePO4 prismatic cells — brand-new, zero-cycle cells that had been sitting in storage for about two years. He unboxes them, reconfigures the 20S pack into a 16S 48V battery plus a leftover 4S 12V pack, wires in a 300A BMS, runs three capacity tests, and installs the finished pack in a server rack cabinet. The whole build is documented below, every number taken from the video.
The interesting thing about this build is what it deliberately is not. It is not about reusing degraded EV packs or salvaging 18650s — it uses brand-new, zero-cycle prismatic cells. It is also honest about difficulty: David calls removing cells from the tray one of the more dangerous things he has done on the channel, and he recommends an easier route whenever possible. If you are weighing a DIY battery against a ready-made server rack unit, this video gives you the real numbers to make that call. The same voltage, capacity and BMS decisions apply whether you build the pack yourself or order a factory-assembled one from Dajiu Energy.
David has owned both kinds of batteries, and he is unusually clear about the trade-off. A pre-built server rack battery is ready to go out of the box, already tested, and it works well — but it costs around $300 per kilowatt-hour. A DIY build can land near $150 per kilowatt-hour once you count the cells, the BMS and the hardware. The catch is that DIY usually means buying used cells that already lived a life in a car, and used cells can be degraded. What changes the math in this video is the source: these are new cells, retired from a module assembly but with no cycles on them, so you get the DIY price without the used-cell risk.
Time is the other cost. David spent about two weeks at two to three hours a day on the first module — around 40 hours for both planned modules combined — and much of that time was filming, moving lighting and figuring out the process as he went. If you value your hours at hobby pace, the build is a great project; if you need the storage up this week, the ready-made rack is the honest answer. That is exactly the decision we help customers make every day: whether to buy a turnkey battery or spec a custom pack for their voltage and budget.
The crate contains two modules, each holding a 10S cell group — ten cells of 3.2V nominal, 280 amp-hours each, wired in series. The whole crate is a 20S pack: about 64V nominal at the terminals, which is too high for most 48V equipment. That is why the pack has to be rewired before it becomes useful. The modules sit on a steel base with a fan that blows air between the cells, and there is a plastic cover with vents. The cells are clamped in place with steel banding and glued to the tray — an arrangement that matters later, because the clamp pressure is meant to stop the cells from expanding as they cycle, not to compress them hard like a vise.
There is no true BMS on this assembly. Instead it has a BMU — battery management unit — that collects data and reports to a central computer that would have overseen many modules in an industrial installation. So part of the DIY job is adding your own BMS, which any lithium pack needs to protect the cells. When David checks the pack voltage, the two 10S groups each read 3.292V per cell, and the whole crate sits at 65.88V. The individual cells are the same chemistry we use in our 48V and 51.2V home storage batteries — prismatic LiFePO4 cells with real BMS protection and datasheet-verified capacity.
The most common arrangement for this chemistry is 16 cells in series — 16S — which gives 51.2V nominal and charges perfectly with standard off-the-shelf 48V components. Since the crate has 20 cells, four have to be electrically disconnected. David's plan: cut the jumper between the two 10S groups so the pack becomes a 16S 48V battery plus a separate 4S 12V battery still sitting in the same tray. Cutting the bus bar rather than removing the cells means the two batteries stay glued in place — which he later emphasizes was the safer and smarter route.
He then removes the cells for a cleaner single battery, and this is where the build gets genuinely risky. The multi-colored sense wires must be cut one at a time and pulled from the plug — cutting them together would short the wires to each other and burn them out. He covers the exposed ends with liquid electrical tape as a temporary seal, and he puts a plastic shim under the bus bar to keep metal shavings from falling into the cells. Removing the steel banding did not spring the cells open — a good sign, because it means the pack was not under damaging tension. The lesson for anyone attempting this: keep the four cells in the tray and just cut the bus bar if you can. It is safer, and it leaves you a useful 12V pack for cabin lighting or a water pump.
A LiFePO4 pack needs restraint — not compression, but enough force to stop the cells from expanding as they charge and age. The factory did it with steel banding top and bottom. David replaces it with a banding kit from Amazon: a strap tool, clips and the banding itself. The goal is to match the original tension — he says it took about 12 separate tries to get the feel right, ratcheting until the handle tension felt the same as the factory band. He uses a leftover aluminum heat sink with a thermally conductive mat between the cells and the backing plate, and a plastic isolator between cells.
This is a detail that separates a battery that lasts from one that fails early. Over time, prismatic cells swell slightly; unrestrained, that swelling can stress the terminal welds and shorten life. Threaded rod — the classic DIY approach David used on his Chevy Volt pack — works, but the banding kit is cleaner and non-conductive. If you buy a ready-made battery instead, the manufacturer has already made this decision for you, which is why every Dajiu Energy pack is built with proper cell compression in a structural enclosure. It is one of the invisible details that determines whether a battery sees 2,000 cycles or dies early.
The BMS in this build is a 300A continuous contactor-based unit from Battery Hookup. David chose it because these cells can both charge and discharge at 280A continuously — powerful cells need a powerful BMS, even if it is overkill for a home install. The BMS reads the voltage of every single cell through two sense-lead harnesses, and it can shut the whole pack down if any cell goes over or under voltage. It also has passive balancing: resistors burn off excess current from high cells until all cells match.
Wiring the harness is where most DIY battery failures happen, and the video is methodical about it. Each sense lead goes to a specific cell junction, following the BMS wiring diagram rather than guesswork. The leads carry almost no current — a tenth of an amp at most during balancing — so small cold-solder joints are acceptable here, unlike power connections. Before plugging the harness into the BMS, David probes every pin and reads 3.2V, then 6.4V, then 9.6V up the line — each check confirming the wiring steps correctly and no wires are crossed. This test is non-negotiable: a crossed sense wire can confuse the BMS into under- or over-protecting the pack. It is the same verification our engineers run on every pack that leaves our production line.
With the BMS connected and Bluetooth app showing all 16 cells balanced, David runs his first capacity test: a 240V space heater drawing over 5,000 watts through a grow-watt inverter, with a Victron shunt measuring kilowatt-hours. Before closing the breaker he uses a small resistor to pre-charge the inverter's capacitors — a surge of dead capacitors would otherwise cause a huge current spike on connection. The first test delivers 14.2kWh and 279Ah, just shy of the pack's 14.3kWh / 280Ah rating, and the voltage cut off cleanly when the BMS opened the contactor.
The second test runs a 0.2C discharge — about 2.8kW — through a chargeverter into a second battery bank, and it delivers 14.3kWh / 280Ah exactly. The third test, back on the space heater, lands at 14.4kWh / 282Ah, above rating. David's explanation: the first cycle woke cells up after two years in storage, and the next two cycles met and exceeded the rating. The takeaway for anyone buying or building 280Ah cells: the first cycle can under-read slightly, so capacity-testing a pack more than once is the only way to trust its real number. That is why our home storage batteries ship with actual tested capacity rather than theoretical cell math.
The finished pack slides into an old server rack cabinet. David notches the rear rail to clear the battery, trims the original front panel by 5/8 of an inch on each side to make it fit, and re-drills the holes. He replaces the module's proprietary connectors with generic insulated bushing posts — a long copper stud with a 3/8-inch terminal that handles several hundred amps — so the battery connects to standard hardware. The front panel's disconnect switch survives because it is genuinely useful: flip it and the pack is dead for service.
Fusing gets a professional-grade treatment. The BMS shipped with a 400A fuse with no brand name and no spec sheet, so David refuses to use it — an unverifiable fuse is a gamble in a 48V high-current path. Instead he installs a name-brand fuse from Battery Hookup's pulled-from-module stock, which costs a few dollars retail for what would be $100+ new. The distinction matters: the fuse is the last line of defense between a cell fault and a fire, and "400A, brand unknown" is not a spec. When you buy from a manufacturer, the fuse and BMS ratings should be printed on the datasheet — as they are on every Dajiu Energy battery.
Here is the math David gives at the end. The crate with two modules, plus BMS, fuse, terminals and hardware, lands at roughly $5,000 for 28.8kWh of capacity across both modules — about $173 per kilowatt-hour. That already beats the $300/kWh server rack units behind him. A discount code he arranged with Battery Hookup drops the price further, to about $150 per kilowatt-hour — roughly half the price of the pre-made racks. For the two-crate scenario — four modules, five batteries totaling almost 72kWh — the per-kWh figure comes to about $120 all-in after the discount, versus 72kWh of premium rack storage that would cost more than twice as much.
He is also clear about the caveats: about 40 hours of work, the risk of working on live cells, and the leftover cells only make sense if you actually use them. And the math assumes you buy new-old-stock cells with zero cycles, not degraded used EV cells. For a cabin with modest loads, a single 14.4kWh battery plus the free 12V pack is plenty. For a serious home system, five 48V 16S batteries in a pallet rack give 72kWh and 500A of BMS capacity. Those are the two sweet spots he recommends — and both are configurations we build every day as custom home storage systems.
David's honest answer: build it yourself if you have the time, the tools and the confidence to work safely with live lithium cells — and especially if you have a small-load application like a cabin where one 14.4kWh pack plus a free 12V battery covers everything. Buy two crates and build five 48V packs if you have the upfront funds and want roughly 72kWh at the lowest possible price per kilowatt-hour. But if removing and rewiring cells feels beyond your comfort zone, keep the cells in the tray and only cut the bus bars — or skip the build entirely and buy a tested rack unit.
What this video proves is that the gap between DIY and factory batteries is mostly labor and confidence, not magic. The cells, the BMS, the compression and the fusing are all standard engineering. That is also why the factory route is so competitive when you count your time: a ready-to-install 48V battery with tested capacity, proper compression and a documented BMS costs more per kWh than DIY, but it arrives working and it carries a warranty. For OEM and ODM projects especially, buying packs built to your spec — voltage, capacity, BMS and enclosure — is almost always faster and safer than a DIY build at scale.
And if you're not hunting for a consumer brand but need the battery itself — a custom LiFePO4 pack at a specific voltage and capacity, an OEM or ODM project, or wholesale supply for your own product line — that's what we do. Dajiu Energy is a Chinese manufacturer building LiFePO4 batteries for RVs, homes, golf carts and portable applications since 2017. Our product range covers LiFePO4 drop-in replacement batteries, EV and golf cart batteries, wall-mounted home storage, portable power stations, commercial energy storage and battery swap cabinets — explore the full lineup in our product center, or browse more buying guides. Tell us your voltage, capacity and size, and we'll engineer the pack to hit it.
These are the most useful of the video's top comments, with our practical answers. The creator's replies in the thread went to lower-liked comments, so the top threads here carry our answers instead. Thanks-only comments were left out on purpose; these are the ones with real technical or experience value.
Source video: DavidPoz — "Build Your Own Solar Battery and Save Thousands $$$" (youtube.com/watch?v=D2eXgiakD8U). Comment excerpts are quoted verbatim from the video's top comments; where the creator replied, his words are quoted, otherwise answers are Dajiu Energy's own. Build figures are transcribed from the video.
Tell us your voltage, capacity, size and application. We will get back to you with a professional quote and technical spec as soon as possible.