How to Evaluate a LiFePO4 Factory's Cell Sorting & Matching Capability

📅 Nov 9, 2026 📁 Category: Supplier Audit ⏱ 11 min read

"Grade A cells" is a label, not proof. For procurement managers, the real question is: can the factory measure, sort, match, and trace every cell before it is welded? This guide breaks cell-sorting capability into an auditable checklist you can put in an RFQ.

RYDBATT (Shenzhen Ryder Electronics, est. 2005) designs LiFePO4 systems with in-house BMS, pack engineering, and cell-to-pack QC. The evaluation framework below is the same one we expect buyers to use on us.

1. Why Cell Sorting Decides Pack Lifetime

In a series string, the pack is only as strong as its weakest cell. A 1% capacity gap does not stay at 1% - it widens every cycle. Poor sorting causes:

Early BMS low-voltage cutoffs and lost usable capacity
Uneven heat generation under load
Faster capacity fade in the lagging cell
More frequent balancing events and higher RMA risk

BMS balancing is a safeguard, not a cure. Matching must happen before welding.

2. The Three Stages: Screening → Sorting → Matching

① Screening (reject)

Mechanical defects, leaked electrolyte, damaged terminals
Missing or unreadable QR/laser traceability
Abnormal self-discharge or OCV drift after rest
Out-of-spec thickness, weight, or appearance

② Sorting (classify)

Capacity bins (e.g. ±1% within a build set)
Internal-resistance bands (AC-IR or DC-IR)
Resting OCV bands at controlled SOC
Same production lot / date code group

③ Matching (group)

Assign cells to the same module using multi-parameter windows
Lock the group ID before spot/laser welding
Record every cell serial against the final pack serial

3. Parameters a Factory Must Be Able to Show

Capacity: measured Ah at defined C-rate, temperature, cutoff voltage
Internal resistance: AC-IR (1 kHz) or DC-IR, with tolerance window
OCV: rest voltage after formation/aging, delta ≤ 5 mV across a string
Self-discharge: 24–72 h rest stability check
Batch: cell maker, model, lot code, production date
Grade definition: written, numeric, not just "A/B/C"

4. Incoming QC: What to Ask For

Per-cell test log: capacity, IR, OCV, serial number
Sampling plan (AQL) + 100% critical-parameter checks
Formation and aging cycle records
Rework/quarantine workflow for out-of-window cells
Calibration records for grader, IR meter, and capacity tester

A generic CoC without cell-level data is a red flag.

5. Weld-Before-Matching Controls

Once cells are welded, imbalance cannot be economically corrected. Audit these controls:

Matched group ID printed/scanning into the MES before welding
Spot/laser weld parameters locked: current, pressure, dwell, spot count
Pull-test criteria per joint type
First-article inspection before volume run

🔗 Reference: RYDBATT RYD-51100 51.2V 100Ah LiFePO4 Pack - 16S, cell-graded, CAN/RS485, IP67

6. Traceability Architecture

Procurement-grade traceability maps:

Cell serial/lot → module group ID
Module → pack serial number
Pack serial → BMS firmware version
Pack serial → outgoing test report
Retention period defined in the quality agreement

This is what makes warranty claims, recalls, and batch investigations possible.

7. Factory Audit: Questions That Expose Real Capability

Can you show the incoming cell-sorting station live, with labels readable?
What are your written capacity and IR windows for "matched"?
Do you mix lots within one pack? Under what approval?
Where is the matching data stored, and for how long?
What triggers quarantine vs re-grade?
Can we audit a random pack serial back to its cell records?

8. RFQ Language to Lock Sorting Standards

Requirement: - Cell grade: written definition, not marketing label - Capacity match: ±___% within module - IR match: ±___mΩ within module - OCV delta: ≤___mV after ___h rest - Same lot code per pack: YES / NO - Per-cell test report: required - Matching before weld: mandatory - Traceability: cell lot → pack serial - Quarantine process: documented - No silent cell substitution without re-approval

9. How RYDBATT Controls Cell Consistency

Incoming capacity, IR, and OCV grading on every batch
Multi-parameter matching before module welding
Formation + aging cycles before pack assembly
Per-serial outgoing test report and BMS config record
ISO-certified lab, UN38.3/MSDS, CE/UL docs per project
In-house BMS: balancing, protection, and communication verified

🔗 RYDBATT 51.2V 100Ah Smart-Communication LiFePO4 Module
🔗 RYDBATT 51.2V 100Ah Low-Speed EV / Fleet Platform

Send RYDBATT Your Cell-Sorting Audit Questionnaire →

FAQ

Q: Is "Grade A" enough to judge cell quality?

A: No. "Grade A" is a supplier-internal label. Procurement should require written grading criteria: capacity tolerance, IR window, OCV spread, and per-cell test data.

Q: What parameters must be matched before welding?

A: Actual discharge capacity, resting OCV at controlled SOC, internal resistance, and same production batch - measured and recorded before any cell is welded into a module.

Q: Can BMS balancing replace cell matching?

A: No. BMS balancing is an operational safeguard. It cannot fix fundamental capacity or IR mismatch once cells are welded. Matching must happen before assembly.

Q: How should traceability be structured?

A: Cell serial/lot code should map to module, pack serial number, BMS firmware version, and outgoing test report - retained for the warranty period.

Q: What is a red flag during a cell-sorting audit?

A: No per-cell data, only aggregate certificates, no written grading bins, mixing lots freely, or inability to show the incoming inspection station live.