Building a DIY LiFePO4 Battery Bank: What Comes After the Sizing Math

The sizing math tells you how big the bank must be; this is the companion piece about putting it together. A DIY LiFePO4 bank is genuinely within reach of a careful hobbyist — but lithium chemistry forgives nothing, so the build order matters. What follows is the checklist I use, in the order I do things.

1. The BMS is not optional equipment#

The battery management system protects against overcharge, deep discharge, overcurrent and temperature extremes, and it keeps the series cells balanced. Size its continuous current rating for your peak load: peak watts / pack voltage ≤ BMS continuous current. A 100 Ah cell pack with a 50 A BMS is a 50 A pack, no matter what the cells could deliver — the BMS is the bottleneck, and deliberately so. A lithium bank without a BMS is a chemistry experiment, not a power system.

2. Series inside the pack, parallel between packs#

A nominal 24 V LiFePO4 pack is 8 cells in series (8S) internally — the series connection lives inside each pack, supervised by its own BMS. When you need more capacity, add whole identical packs in parallel, never by building longer series strings across packs. Parallel packs scale both capacity and current; mismatched packs in parallel will fight each other through their BMSs.

3. Cells: grade and source matter#

Prismatic LiFePO4 cells (the EVE LF105 / LF280K class) are the standard building block. Buy grade-A cells from a reputable source and inspect on arrival: no bloating, no damaged terminals, and voltages within a few tens of millivolts of each other — they typically ship at storage charge, around 3.2–3.3 V per cell. A cell that arrives at 2.5 V while its siblings sit at 3.3 V is a return, not a project.

4. Top-balance before the first assembly#

Before wiring cells in series, connect them all in parallel and charge the group slowly to 3.65 V per cell, then let them rest. This top-balancing aligns every cell to the same full state of charge, so the BMS starts from a level field instead of fighting an imbalance from day one. Skip it and the weakest cell hits the top first on every charge, throttling the whole pack.

5. Busbars, torque, fuses#

• Clean every contact surface (terminals and busbars) before assembly — a thin oxide layer becomes a hot spot at tens of amps.
• Torque the terminal bolts to the cell manufacturer's spec; too loose arcs, too tight strips the threads.
• Fuse every parallel string individually, as close to the pack as possible.
• Size cables for both current and voltage drop — at 24 V, a 3% drop is only 0.77 V, and long thin runs eat it fast.
• Keep the bank ventilated and the cells mechanically restrained; prismatic cells swell slightly with cycling.

6. Commissioning: prove the capacity#

Before trusting the bank, run one full cycle: charge to the BMS cutoff, then discharge through a known, measured load while logging voltage and current. Integrate to watt-hours and compare against the nameplate. If a "2,560 Wh" 24 V 100 Ah pack delivers far less usable energy, you want to know that on the bench — not during the first cloudy week.

# rough capacity check: discharge through a known load,
# logging current (A) once per minute at 25.6 V nominal
discharged_ah = sum(current_samples_a) / 60
discharged_wh = discharged_ah * 25.6
print(f"Usable: {discharged_wh:.0f} Wh")

Build it like the datasheet is watching: torque specs, fuses, and one honest capacity test beat three forum threads of opinions.

These are planning-level practices, not a substitute for the cell and BMS datasheets or local electrical regulations. Anything grid-tied — and anything above extra-low voltage in some jurisdictions — deserves a qualified electrician's sign-off. Walk in with the sizing math done and a torqued, fused, tested bank, and that conversation goes much faster.