Why a PCM Needs Its Own Verification Plan
A battery protection module (PCM) is a small electronics assembly — protection IC, MOSFETs, sense resistor, capacitors and a PCB — that is asked to interrupt charge and discharge current before a cell is damaged. Buyers often treat it as an accessory to the cell, so the purchase order references the cell standard and assumes the module is covered. It usually is not.
The practical consequence is that two suppliers can both sell a "protected pack" while one module has a defined trip window, a documented release condition and an auditable component list, and the other has a copied schematic with no traceable specification. This guide sets out how to structure quality requirements for a PCM, which standards actually reach the module, and what evidence to request so the claim can be checked rather than believed.
First Define the Scope: What the Standard Covers and What It Does Not
The three standards most often quoted in PCM sourcing cover different things, and none of them is a component-level PCM specification. The scopes below are verifiable from the standards bodies' own publications:
- IEC 62133-2:2017 — Safety requirements for portable sealed secondary lithium cells, and for batteries made from them, for use in portable applications. First edition published February 2017; a consolidated version with Amendment 1 (2021) also exists. The object of the standard is the cell and the battery, not the protection PCB as a standalone article.
- IEC 62619 — Safety requirements for secondary lithium cells and batteries for use in industrial applications, including stationary applications such as telecom equipment, UPS, energy storage and emergency power, and motive applications such as forklifts, AGVs, rail and marine. Where an application-specific IEC standard sets conflicting test conditions, that standard takes precedence — IEC 62660 series for road vehicles is the cited example.
- UN 38.3 — Transport safety testing of cells, battery modules and battery packs under the UN Manual of Tests and Criteria. It is a transport test regime, not a product-safety qualification of a protection module.
Two conclusions follow. First, a PCM inherits requirements from the battery-level standard it is used in, so the requirement set depends on the end product (portable, industrial/stationary, motive). Second, the specific trip voltages, release voltages and delay times that a buyer actually needs will not be found in the standards text — they have to come from the cell specification and the protection IC datasheet.
Why UN 38.3 Cannot Substitute for PCM Verification
UN 38.3's overcharge test (T.7) is aimed at the battery: for rechargeable batteries the test current is twice the manufacturer's maximum continuous charging current, and the minimum test voltage is set relative to the battery's maximum charging voltage — for batteries whose maximum charging voltage does not exceed 18 V the lower of twice that voltage and 22 V applies, while above 18 V it is 1.2 times the maximum charging voltage. The test runs at ambient temperature for 24 hours, and the pass criterion is no disassembly and no fire during the test and within seven days after it.
Those numbers describe a battery-level abuse condition. They say nothing about whether a particular PCM trips at the right cell voltage, releases cleanly, or survives repeated cycling. Public UN 38.3 documentation also recognises that where a battery is designed to be installed inside equipment that provides overcharge protection, the overcharge protection may be verified at the equipment or battery level rather than inside the module — which is precisely why a module datasheet citing "UN 38.3 compliant" should be read as a transport statement, not a protection-performance claim.
Translate Cell Requirements into Numeric PCM Limits
The specification work happens before any test. For each module, a requirement table should state, at minimum:
- Overcharge detection voltage and its tolerance, plus the overcharge release voltage or release condition.
- Over-discharge detection voltage, release voltage and release method (voltage rise, charger connection, load removal).
- Over-current detection threshold, detection delay, and release condition.
- Short-circuit detection threshold and delay.
- Continuous and peak current rating, on-resistance budget, and thermal limits of the MOSFETs.
- Quiescent/standby current draw and its effect on shelf life.
- Charge and discharge temperature cut-off windows, and whether an NTC is required.
- Cell count and chemistry the module is designed for, and the designated test unit when the pack is split into sub-units.
Published protection-IC datasheets show how wide the legitimate design space is, and why a generic number is useless. Configurable ranges in widely available parts include overcharge detection from roughly 3.5 V to 4.8 V in 5 mV or 50 mV steps, with detection accuracy in the ten-to-twenty millivolt band and release accuracy looser (for example ±50 mV on some series). Over-current thresholds can be set in millivolts of sense voltage across a very fine step, and detection delays for over-current or short circuit commonly fall in the millisecond to sub-millisecond range depending on the function. None of that is a requirement — it is evidence that the requirement must be written per part number and per cell.
Two rules keep the requirement table honest. Every numeric limit must cite its source: the cell maker's specification, the protection IC datasheet revision, or the customer's product requirement. And every limit must have a stated test condition — temperature, load, and measurement point — because a trip voltage measured at 25 °C with a resistive load will not match one measured hot under a real charger.
What the Certification Actually Says
Certificates are frequently the weakest link in a PCM file, because a certificate number is easy to quote and hard to read. When a supplier says the module is "certified", verify these four things before accepting the claim:
- Standard and edition. IEC 62133-2:2017 and the 2017+A1:2021 consolidated version are different documents; so are superseded editions of older battery standards. The certificate must name the edition that is current for the target market.
- Scope text, not the title. Read whether the certified object is a cell, a battery, or a battery pack with a specified PCM. A cell certificate does not extend to the module, and a UL standard such as UL 2054 is not the US edition of either IEC standard — passing IEC 62133-2 does not establish UL 2054 compliance.
- Model and construction identity. Compare the certified model designation, cell type, cell count and PCB revision against what is being shipped. A certificate covering a different protection IC or a different MOSFET package does not cover the delivered article.
- Certificate validity. Check the issuing body's online directory rather than the PDF alone, and confirm the certificate has not been withdrawn, suspended or superseded.
If the module is supplied as a component inside a pack, the practical question is not "is the PCM certified" but "which certificate covers the assembly this PCM sits in, and is the PCM within the construction that was tested". That question belongs in the purchase specification, not in a first-article discussion.
Test Evidence to Request, and How to Read It
A workable evidence package for a production PCM has four layers:
1. Design and component identity
Approved schematic, PCB layout revision, protection IC part number and datasheet revision, MOSFET part number, sense resistor value and tolerance, and a bill of materials with approved-manufacturer equivalence rules. Without this, no test result can be tied to the article being bought.
2. Characterisation test data
Trip and release measurements across temperature, ideally at the extremes of the intended operating range, with sample size and measurement uncertainty stated. Data should show both detection and release, because a module that trips but never releases creates a dead pack, and one that releases too early defeats the purpose.
3. Reliability and abuse evidence
Short-circuit behaviour, over-current endurance, thermal rise at rated current, and — where the end product requires it — the battery-level tests from the applicable standard. Where a UN 38.3 test summary is supplied, note that the UN test summary format (T1–T8) describes the cell and battery tests; it should be read alongside, not instead of, module-level characterisation.
4. Production control records
End-of-line test limits and the equipment calibration status behind them, sample inspection results, non-conformance and rework records for the shipped lot, and change notification commitments. A PCM that passes type tests but ships with an uncalibrated end-of-line tester is an uncontrolled product.
Acceptance Criteria and Compliance Gaps
Acceptance criteria for a PCM should be written so that a third party can reproduce the judgement. A workable structure is: each requirement in the table has a limit, a test method reference, a sample size, and an accept/reject rule. Attributes such as solder joint quality, conformal coating coverage, labelling and traceability code are inspected, not measured, and need their own criteria.
The gaps that recur in practice are predictable:
- Overcharge protection assumed from a transport test. Accepting UN 38.3 as proof of overcharge protection leaves the module's actual trip window unverified.
- Release condition undocumented. Data sheets sometimes list detection values but not how the module recovers; recovery method affects field behaviour and must be specified.
- Temperature cut-off untested. Charge and discharge windows are often claimed but only verified at room temperature.
- Silent change. A substitute protection IC with the same footprint but a different threshold, accepted without notification, invalidates the whole qualification file.
- Scope drift in certificates. A certificate for a cell type reused as evidence for a module.
Keeping the Evidence Current
The PCM file has a short shelf life. Protection IC datasheets are revised, standards are amended (IEC 62133-2 gained Amendment 1 in 2021), and certificate directories update without notice. A simple review cycle works: re-verify certificate status and standard editions annually, re-verify the bill of materials at every design or supplier change, and re-run characterisation whenever a component change alters a threshold. Record who verified what and when, so that the next audit does not restart from zero.
The test of a sound PCM requirement set is not its length. It is whether a colleague can pick up the file, look at one module and one cell specification, and answer three questions without calling the supplier: what is this module supposed to do, what evidence shows it does it, and what would tell us it has stopped doing it.