TL;DR: A supplier’s IEC 62619 certificate tells you almost nothing without the test report behind it — COA field completeness and incoming inspection thresholds are where qualification actually happens.
TL;DR: In our review of 31 supplier qualification packages over the past 18 months, fewer than 9 included a COA with all seven mandatory data fields populated — the rest required at least one follow-up round before we’d approve them for AVL entry.
COA Field Requirements: The Seven Fields That Determine Whether a Certificate Is Usable #
Most buyers treat the Certificate of Analysis as a checkbox. Either the supplier has one or they don’t. That framing misses the point entirely.
A COA for cells destined for an IEC 62619-compliant battery pack isn’t useful unless it contains traceable, lot-specific data. Generic COAs — the kind that list “Typical: 280Ah” without referencing the actual shipped lot — are cosmetically compliant and operationally worthless. We’ve rejected COAs from two Dongguan-area cell suppliers in 2024 specifically because capacity data was populated with spec-sheet values rather than measured outgoing QC results.
The seven fields we require in our COA intake form (internally designated QF-22 Cell Receipt) are:
- Lot number with traceability to production batch date
- Measured capacity at 0.2C discharge (25°C ± 2°C), not nominal
- Internal resistance at 1kHz AC impedance, per-cell average and standard deviation
- Self-discharge rate over 28-day OCV monitoring (expressed as % capacity loss)
- Cell weight, measured, not stated — deviations >1.2% from nominal flag counterfeit risk
- OCV at shipment, with acceptable range declared per the supplier’s SOC-OCV curve
- Test equipment calibration certificate reference number and validity date
Fields 4 and 7 are the most frequently missing. Self-discharge data requires a 28-day hold, which most factories skip on fast-turn orders. Calibration traceability is skipped because it exposes gaps in the supplier’s own metrology chain. If a factory can’t show you a calibration reference, their other numbers are suspect regardless of how clean the COA looks.
Under IEC 62619:2022 Clause 5.2, secondary cells used in industrial applications must meet defined safety requirements — the COA is the primary documentary evidence that cell-level requirements were verified before pack assembly. A COA missing measured data isn’t compliant with the spirit of that clause. Some auditors will pass it; we won’t.
Supplier Qualification Communication: What You Ask and What the Answers Reveal #
Ask for the full IEC 62619 test report — not the certificate, the report — within 48 hours of initial contact. The response time and completeness tells you as much as the data itself.
A supplier with genuine third-party certification has the report on file and sends it within a business day. A supplier operating on a shared or borrowed certificate stalls, sends a certificate image instead of the report, or claims “the report is confidential.” None of those responses are acceptable. UN38.3 Section 38.3.5 requires that transport test documentation be available to competent authorities — a supplier who won’t share it with a qualified buyer has a structural compliance problem, not a confidentiality policy.
When the report does arrive, check three things before reading any data:
First, verify the test house is accredited. The IEC 62619 test scope should appear on the lab’s ILAC-MRA or equivalent accreditation scope document. We’ve encountered reports from Chinese labs that hold CNAS accreditation for mechanical tests but not electrochemical safety tests — and the reports don’t disclose this distinction.
Second, check the cell configuration tested. The report must match your specific cell chemistry, form factor, and capacity. A 280Ah LFP prismatic cell report does not cover a 306Ah cell from the same factory, even if the physical dimensions are similar. Configuration drift is one of the most common compliance gaps we document.
Third, look at the overcharge test results from IEC 62619:2022 Clause 7.3.3. Specifically: did the cell vent, swell, or rupture? A pass result means no fire, no explosion — not necessarily no swelling. Buyers spec-ing tight enclosures need to know exactly what “pass” looked like in practice.
Request the incoming inspection sampling plan the supplier uses for outgoing QC. If they can’t produce a written AQL sampling procedure, they’re doing 100% visual at best. That’s not a quality system.
Cost-Performance Trade-offs in IEC 62619 Certification Scope #
Certification costs for a complete IEC 62619 test campaign at a recognized Chinese lab (SGS, TÜV Rheinland China, Bureau Veritas) currently run between $8,400 and $14,700 USD per cell configuration, depending on scope and lab queue time. That range is based on quotes we collected in Q1 2025 for 280Ah LFP prismatic cells across three labs.
This creates a real cost pressure for smaller pack manufacturers. A factory producing 12 SKUs across three cell chemistries faces potential certification spend exceeding $120,000 before any pack-level testing. Many Shenzhen-area pack houses respond to this by certifying one representative configuration and applying it to adjacent SKUs. That’s where the compliance gap lives.
The counterargument for accepting a narrower certification scope: for a buyer sourcing a single, stable SKU at volume, a factory with deep certification coverage on exactly your configuration is more valuable than one with broad but shallow coverage across many variants. We’d rather see a supplier who certified 280Ah LFP at three temperature extremes and two aging states than one who certified eight cell types superficially. Scope depth matters more than breadth when your product doesn’t change.
Where the cheaper option is actually correct: if you’re sourcing for a low-cycle application (backup power, less than 200 cycles per year), Grade-B cells from a certified lot with verified COA data can be technically appropriate. The performance penalty at low cycle count is minimal, and the cost difference is real — roughly 18-22% ex-works compared to Grade-A from the same factory, based on our 2024 procurement data from two Dongguan suppliers. The certification question doesn’t change; the cell grade question does.
Incoming Inspection Protocol: Pass/Fail Thresholds and Sample Sizing #
This is where IEC 62619 compliance either holds in the field or quietly degrades over time.
Most buyers run incoming inspection on battery packs as a visual and functional check: power it on, read the SOC, ship it. That protocol catches maybe 30% of the failure modes we track in our Category C incident log (cells underperforming spec, BMS misconfiguration, thermal protection gaps). The other 70% require measurement.
The incoming inspection protocol we recommend for industrial battery packs claiming IEC 62619 compliance covers five test layers:
Capacity verification at 0.5C discharge rate, 25°C ± 3°C, full charge to cutoff. Accept threshold: ≥97% of COA-stated capacity. Sample size: minimum 3 units per 50-unit lot, or 5% of lot, whichever is greater. A result below 94% on any single unit triggers 100% lot hold pending supplier response.
Internal resistance delta across the cell string. Measure each cell’s AC impedance at 1kHz and calculate the standard deviation across the pack. Accept threshold: σ ≤ 4.2 mΩ for a 16S LFP configuration. High variance indicates mixed-grade cells or a pack assembled without sorting — both are supplier process failures, not just QC escapes.
BMS protection threshold verification. Trigger the over-voltage protection by charging a single cell past the OVP setpoint using a bench supply. Confirm the BMS disconnects within 200ms. This test is absent from most buyer incoming protocols and present in fewer than 40% of the supplier qualification packages we’ve reviewed. IEEE 1625-2008 Section 6.4 specifies cell-level protection requirements for lithium-based batteries — use it as your baseline for setpoint validation.
Thermal sensor continuity. Confirm all thermistors are wired, reading, and within ±2°C of ambient at rest. One of the clearest field failure patterns we document involves packs where a thermistor connector was seated but not crimped — it reads correctly at ambient but drifts under thermal load, defeating the over-temperature protection entirely.
Label and marking compliance. IEC 62619:2022 Clause 9 requires specific marking content. Verify the label matches the certified configuration, not a generic factory template. Mismatched labels on certified products are a regulatory liability for the importer, not just the manufacturer.
The table below shows the thresholds we apply, with disposition logic for common failure patterns.
| Test Parameter | Accept Threshold | Reject Trigger | Disposition on Fail |
|---|---|---|---|
| Capacity at 0.5C | ≥97% of COA value | <94% any unit | 100% lot hold |
| Cell IR standard deviation (16S) | σ ≤ 4.2 mΩ | σ > 6.0 mΩ | Return + root cause required |
| BMS OVP response time | ≤200ms | >350ms or no trip | Immediate reject |
| Thermistor reading delta | ±2°C vs ambient | >5°C or open circuit | 100% pack inspection |
| Label configuration match | Exact match to cert | Any mismatch | Hold pending cert review |
One question we’re still tracking: whether incoming IR variance thresholds should be tightened for packs destined for high-ambient deployments (>40°C operating environment). Our current dataset only covers packs tested at standard conditions — we expect to have field return data from two Southeast Asian integrator deployments by Q3 2025 that will let us recalibrate.
Sourcing Guidance for Buyers #
When evaluating Chinese suppliers for IEC 62619-compliant battery packs, the first document to request is the full third-party test report, not the certificate. A certificate with no backing report is unverifiable. A supplier who responds to this request with delay or deflection is almost certainly working from a shared or configuration-mismatched certificate — a compliance risk that transfers to you as importer once product crosses the border.
The qualification red flag specific to this category: COA data that matches datasheet nominal values exactly, across multiple lot numbers. Real measured data has variance. If every lot shows exactly 280.0Ah and exactly 0.25mΩ average IR, those numbers are being transcribed from spec sheets, not measured. One Shenzhen-area pack supplier we evaluated in 2023 had 14 consecutive COAs with identical cell-level values. Their QC process was a copy-paste operation.
For incoming inspection, prioritize the BMS OVP trigger test on at least 10% of units in the first three lots from any new supplier. Threshold: BMS must disconnect within 200ms of cell voltage exceeding the OVP setpoint by 50mV. Suppliers who object to this test — claiming it damages the BMS or voids warranty — have a product design problem worth understanding before you commit to volume. Also review BMS engineering fundamentals to align your internal spec requirements with what’s actually achievable from the Shenzhen BMS supply chain before you go to RFQ.
FAQ
What is the minimum COA content required for IEC 62619-compliant cell procurement?
At minimum, a usable COA must include lot-specific measured capacity (not nominal), internal resistance with standard deviation, self-discharge rate over a defined monitoring period, and a calibration traceability reference for the test equipment. Generic COAs listing only typical values against a spec sheet do not satisfy the traceability intent of IEC 62619 Clause 5.2 — and they give you no recourse if performance doesn’t match claims.
Can a factory certify one cell configuration and apply it to adjacent SKUs?
Technically, some tolerance exists for minor parameter variation within a tested configuration range, but this is narrower than most factories apply it. A 280Ah certification does not extend to a 306Ah cell from the same line. If your SKU differs in capacity, form factor, or chemistry from the tested configuration, you need a separate or supplemental test report. Ask the factory to show you exactly which cells are covered in Annex A of their test report — the answer tells you whether the scope is real or assumed.
How should I handle a supplier who won’t share the full IEC 62619 test report?
Remove them from your AVL consideration until they provide it. There is no legitimate reason to withhold a test report from a qualified buyer conducting procurement due diligence. UN38.3 requires transport test documentation to be available to competent authorities, and IEC-certified industrial batteries operate under the same disclosure expectation. Confidentiality claims around safety test reports are a red flag with no defensible basis.
Is Grade-B cell sourcing compatible with IEC 62619 compliance?
It depends on the application cycle profile. IEC 62619 defines safety requirements, not performance grades — a Grade-B cell can be part of a certified pack if the pack-level design and BMS configuration account for the cell’s actual performance envelope. For applications running fewer than 200 cycles per year, Grade-B cells from a verified lot with accurate COA data present acceptable risk. For daily-cycle industrial applications, the cycle life degradation curve on Grade-B cells will cause pack-level capacity drift that undermines BMS SOC calibration within 18-24 months.
What’s the right sample size for incoming capacity verification on large lots?
For lots of 50 units or fewer, test a minimum of 3 units. For lots above 50, apply 5% sampling with a floor of 5 units and a ceiling of 20 — beyond 20 units the incremental detection probability for systematic defects is marginal. Tighten to 10% for the first three lots from any new supplier regardless of lot size. Once a supplier has passed six consecutive lots without a threshold failure, 3% sampling is defensible for ongoing qualification.
Published by compactbess.com Technical Team | Request a sourcing consultation
