Share This Article
Why Cell Selection Matters More Than Cell Brand
Ask any experienced pack builder what determines a finished battery’s real-world performance and the answer is rarely the brand printed on the cell. It is the consistency between cells. Two packs built from the same manufacturer’s cells can differ dramatically in usable capacity, cycle life, and safety — purely because of how carefully the cells were selected and matched.
The Weakest Cell Defines Your Pack
In a series string, current flows through every cell equally. When the weakest cell hits its discharge cutoff (typically 2.5–2.8V per cell), the BMS terminates discharge to protect it — even if every other cell still holds 40% charge. A 16S pack with one cell 5% below the rest therefore delivers roughly 5% less usable energy, regardless of how good the other fifteen are.
What No BMS Can Fix
A common misconception is that a “smart BMS with active balancing” can compensate for poorly matched cells. Balancing currents are measured in milliamps — a typical active balancer moves 1–2A between cells. If your pack has a 10% capacity spread on 100Ah cells, that is a 10Ah deficit — closing it at 1A takes 10 hours of balancing per cycle. Our cell balancing deep-dive explains why balancing exists to fix minor drift, not to rescue mismatched cells. Selection is the cure; balancing is the maintenance.
The 5 Parameters That Actually Matter
Cell datasheets list dozens of specifications. For pack assembly, five parameters decide outcomes, in this order of importance:
1. Capacity Consistency — The Primary Metric
Why it ranks first: Capacity mismatch translates directly into lost pack capacity. A 10% capacity difference means a 10% usable-energy loss in series — no balancing circuit can recover it.
Target: ±2–5% within the same pack. For 100Ah cells, every cell should test between 98 and 102Ah.
How to measure: Full charge to 3.65V, rest, then discharge at constant current (typically 0.2C or 0.5C per the cell’s datasheet) down to 2.5V, recording delivered amp-hours.
2. Self-Discharge Rate — The Most Underestimated
Why it ranks second: A pack can be perfectly top-balanced on day one and drift out of balance within weeks if cells self-discharge at different rates. Experienced builders identify uneven self-discharge — not capacity — as the primary reason packs lose balance over time.
Target: ≤3% per month is a typical Grade A specification. What matters for matching is the spread — cells in one pack should self-discharge at similar rates.
How to measure: Fully charge all cells, let them rest at identical ambient temperature for 15–30 days, and record each cell’s voltage drop. Discard outliers.
3. Internal Resistance Matching — The Heat and Voltage-Drop Driver
Why it ranks third: IR differences cause unequal heating and unequal voltage sag under load. However, a 10% IR difference matters far less than a 10% capacity difference.
Target: Relative ±5% within the same pack. Typical: 280Ah-class ~0.15–0.3mΩ; 100Ah-class ~0.3–0.5mΩ (AC 1kHz).
How to measure: AC 1kHz method with a dedicated IR tester. Measure all cells at the same SoC and temperature. Never mix AC and DC readings.
4. Rest Voltage — A Fast First Screen, Never a Verdict
LiFePO4’s flat discharge curve means a cell at 40% and one at 70% can both read 3.3V. Use ±0.05V spread as an initial screen, not a matching criterion.
5. Batch and Date Code — The Traceability Dimension
Cells from the same production batch share electrode lots and formation cycles — they age alike. Always request batch numbers and build packs from a single batch.
Step 1 — Incoming Inspection
- QR codes intact: Original factory QR codes should be present and scannable. Re-applied insulation covering the QR area signals re-graded cells. See our Grade A vs Grade B verification guide.
- Physical condition: No dents, bulging, or terminal corrosion. Reject swollen cells outright.
- Documentation: Demand per-cell capacity/IR distribution, batch codes, discharge curves, self-discharge data, sorting criteria, and test conditions.
Step 2 — Testing: Three Tiers
| Tier | Equipment | Best For |
|---|---|---|
| Basic | Multimeter + LiFePO4 charger | Small DIY builds |
| Intermediate | Above + AC IR tester (1kHz) | Installer batches |
| Professional | Above + capacity grading cabinet | Pack factories, OEM |
Step 3 — Sorting: Tolerance Bands
| Parameter | DIY | Commercial | Critical |
|---|---|---|---|
| Capacity spread | ±5% | ±3% | ±2% |
| IR spread | ±10% | ±5% | ±3% |
| Voltage spread | ±0.05V | ±0.03V | ±0.02V |
| Batch mixing | Acceptable | Single batch preferred | Single batch mandatory |
“Matched” vs “Batched”
Sellers advertise “matched cells,” but often mean only similar voltage — a single-point check. Ask: (1) At how many voltage points was consistency measured? (2) Can I see the test report? (3) Single-point or full-range matching?
Step 4 — Matching: Build Strings That Age Together
Series Strings
Group tested cells into capacity bands; draw each string from a single band. A 16S pack mixing 98Ah and 105Ah cells is effectively a 98Ah pack.
Parallel Groups: The IR Trap
Lower-IR cells hog more current, aging faster. For parallel groups, prioritize tight IR matching (±3%).
Step 5 — Assembly Constraints
Decide your joining process before ordering cells. Laser welding needs flat-tab cells; bolted busbars need threaded studs with confirmed torque specs. Terminal height consistency matters for flat busbar contact.
BMS Coordination
- Balancing current: Scale with cell size. 50mA passive is inadequate for 100Ah+ cells; specify 1–2A active balancing.
- Protection thresholds: From your cell’s datasheet — overvoltage ≤3.65V, undervoltage 2.5–2.8V, temperature cutoffs per operating envelope.
- Communication: CAN/RS485 for inverter integration; non-communicating BMS is fine for standalone packs. See our BMS selection guide.
Red Flags — When to Reject a Batch
- QR codes removed or covered
- Supplier refuses per-cell test data
- Capacity spread exceeds ±5% out of the box
- Mixed date codes in one “matched” shipment
- Price far below market for claimed Grade A
Frequently Asked Questions
Which matters more: capacity or internal resistance?
Capacity. But uneven self-discharge matters more than either for long-term balance.
How do I test self-discharge?
Fully charge, record voltages, store at identical temperature 15–30 days. Exclude fast-decay outliers.
Can I mix batches?
Not recommended. If you must, retest everything and tighten tolerances one level.
How can I tell if a cell is really Grade A?
Intact QR, factory test report, capacity recovery ≥95%, no re-applied insulation. See our verification guide.
What cell data should I request?
Per-cell capacity/IR distribution, batch codes, discharge curves, self-discharge data, sorting criteria, QR traceability, test conditions, IR method, defect rate.
LJY Energy supplies wholesale Grade A prismatic LiFePO4 cells with matching services. Browse our cell wholesale page or contact us.
Share This Article