Safety Standards Explained for CE / FCC / RoHS Compliance

Where Standards Actually Diverge — and Why It Costs You If You Pick the Wrong One Late #

A Taiwanese ODM we worked with in 2023 had a 296Wh portable power station ready for production. CE marking was their primary target. The product had passed internal testing, the BMS was tuned, and the factory had quoted IEC 62368-1 as the applicable standard. Six weeks before their EU launch, a notified body flagged that their specific configuration — with a field-replaceable battery module and rated output above 75V DC — also triggered IEC 62619 requirements for secondary lithium cells and batteries. Two separate test campaigns, two different labs, an additional 9 weeks, and roughly $14,000 in unplanned testing costs. Nothing was wrong with the product. The compliance path was simply misidentified at the start.

This is not a rare scenario. The Chinese export ecosystem produces excellent hardware, but compliance scoping is routinely left to trading companies or certification agents who work from product category rules of thumb rather than actual standard applicability analysis. For a design engineer responsible for getting a battery product to market, understanding which standard governs what — and why the boundaries exist — is foundational, not optional.

The core problem is that international battery safety standards were not designed as a unified system. They evolved in parallel across different regulatory bodies, addressing different failure modes, for different deployment contexts. IEC standards come from the International Electrotechnical Commission and are consensus-based, often adopted by EU member states into EN equivalents. UL standards are developed by Underwriters Laboratories and carry weight in North America and increasingly in Asian markets via buyer specification. UN38.3 sits in transport regulation, not product safety. GB/T standards are Chinese national standards, mandatory for domestic sale and sometimes used as a proxy for export quality — but they are not recognized as equivalent by EU or US conformity assessors. These four frameworks coexist, reference each other selectively, and in several areas directly contradict each other on pass/fail thresholds.

The Parameters That Actually Differentiate These Standards #

The clearest way to understand the standards landscape is to work from the failure modes each one was designed to prevent, then map backward to the tests.

IEC 62619:2022 governs secondary lithium cells and batteries for stationary and industrial use. Its scope explicitly excludes portable consumer devices and road vehicles — which is where many engineers go wrong by assuming it covers their portable power station. The standard mandates overcharge testing at 1.1× the maximum charge voltage, sustained for a minimum of 1 hour, with pass criteria requiring no fire, no explosion, and no electrolyte leakage. Cell-level temperature rise during overcharge must not exceed 10°C above the manufacturer’s maximum rated operating temperature. That 10°C delta is a hard threshold that shows up in BMS design requirements — if your thermal management can’t hold that margin during a fault, you will not pass.

UL 9540A is the test method for evaluating thermal runaway propagation in energy storage systems. This is not a product certification standard — it’s a test method, and fire authorities having jurisdiction (AHJs) in North American markets increasingly require 9540A test data before approving any BESS installation above a certain capacity threshold (typically 20kWh, though this varies by jurisdiction). The key output is a propagation analysis: does a single-cell thermal runaway trigger cascade failure across adjacent cells? Chinese factories almost never have 9540A data unless a North American customer has specifically required it. If your product is heading into commercial or industrial channels in the US, request this data explicitly during supplier qualification — its absence tells you something meaningful about how seriously the factory treats safety engineering.

UN38.3 governs transport classification of lithium batteries. Every lithium cell or pack moving internationally by air must have a passing UN38.3 test report. The test sequence includes altitude simulation (11.6 kPa for 6 hours), thermal cycling (−40°C to +75°C, 10 cycles), vibration, shock, external short circuit, impact/crush, overcharge, and forced discharge. What most engineers don’t check is whether the UN38.3 report covers their exact cell-pack configuration. A factory that tested a 2S2P configuration does not have a valid report for your 4S2P design. We log configuration mismatch as a Category B discrepancy in our internal CP-11 compliance screening checklist — it’s one of the most common findings during pre-shipment document review.

For portable consumer devices, IEC 62368-1:2023 (the audio/video, IT, and communication equipment standard) is the dominant EU framework, replacing IEC 60950-1 and IEC 62301. Its hazard-based safety engineering (HBSE) model evaluates energy sources and safeguards rather than prescribing specific component solutions. This matters for BMS engineers because the standard’s approach to safeguard classification — Class 1, 2, or 3 — directly determines how many independent protection layers your design needs. A Class 2 safeguard must be capable of preventing a Class 2 hazard independently. If your BMS relies on a single overvoltage cutoff with no secondary protection, that configuration will not satisfy the required safeguard independence for a sustained power source in the >42V range.

The column that matters most in that table is the last one. GB/T 36276 is not recognized by EU conformity assessors as equivalent to IEC 62619, even though the two standards share significant technical overlap. A Shenzhen-based pack factory showing you a GB/T 36276 certificate is giving you evidence of domestic compliance, not EU market access. We see this confusion regularly when buyers source through Alibaba or at Canton Fair without a compliance-aware intermediary in the loop.

The most commonly overlooked parameter across all these frameworks is cycle life validation method. IEC 62619 specifies capacity retention testing at 0.5C/0.5C rate, 25°C ambient, reporting capacity at 80% retention endpoint. That’s a controlled lab condition. In practice, if you’re integrating LFP cells from Chinese pack houses into a product that will cycle at variable rates in non-climate-controlled environments, the standard’s test conditions will overstate real-world longevity — and your warranty model will be wrong.

Decision Framework — Matching Standards to Your Actual Product Configuration #

If your product is a portable power station under 2,500Wh destined for EU consumer retail, IEC 62368-1 is your primary framework, and you should scope UN38.3 for transport from day one. The IEC 62619 question depends on whether your product is marketed or used in industrial contexts — if the datasheet says “for home backup use,” you’re likely in 62368-1 territory. If it says “rack-mountable” or targets commercial customers, 62619 applies and the design requirements shift.

If your product is a stationary BESS above 5kWh heading into the North American commercial market, you’re looking at UL 9540 (system-level certification), UL 9540A (thermal runaway test method, often required separately by AHJs), and UL 1973 (battery standard for stationary applications). These three are not redundant — UL 9540 references 9540A as a test input and UL 1973 covers the battery subsystem while 9540 covers the integrated system. Scoping all three simultaneously saves substantial schedule time versus discovering UL 1973 requirements after UL 9540 testing is already planned.

If your product ships by air in any configuration, UN38.3 is non-negotiable and must be scheduled before any other transport test campaign. Many factories in Dongguan and Shenzhen have in-house UN38.3 capability or standing arrangements with nearby third-party labs — turnaround is typically 4–6 weeks for a standard configuration, longer if cell configuration is novel. What varies is quality: some labs issue reports without completing the full 8-test sequence, counting on buyers not to verify. Always confirm the report lists all eight tests with individual pass/fail results, not just a summary pass statement.

On the question of whether to pursue CB Scheme certification under IEC standards before or after CE marking: opinions differ. Some European notified bodies accept CB test reports as the technical basis for CE marking decisions, which avoids duplicate testing. Others want their own lab data or require supplementary tests beyond the CB scope. Our practice is to route through a CB-recognized lab for the primary test campaign whenever the product has multi-market ambitions — the incremental cost is modest (typically $1,800–$3,500 additional fee depending on product complexity) and the schedule savings are real. For single-market products, direct national route is faster. There is no universally correct answer here; it depends on your certification timeline, the notified body’s specific requirements, and whether you’re confident the design won’t change post-testing.

One non-obvious recommendation with a specific boundary: always request that your BMS engineering team documents protection threshold settings as part of the certification package, not just the hardware schematic. Standards like IEC 62619 and UL 1973 have post-market change provisions — if you modify firmware protection thresholds after certification (which happens constantly during production optimization), you may trigger a re-certification obligation. That clause applies to changes that affect compliance-critical functions. Changes to SOC display accuracy generally don’t trigger it; changes to overvoltage cutoff thresholds almost certainly do.

Sourcing Guidance for Buyers #

When evaluating Chinese suppliers in the CE/FCC/RoHS compliance space, the first document to request is the UN38.3 test report with the exact cell and pack configuration matching your target design. Not a generic report, not one for a similar capacity. An exact match. A factory that can produce this on the first request, with legible lab accreditation details and individual test results for all eight tests, has a functioning compliance process. A factory that sends you a PDF with a summary table and no individual test data is showing you a document they obtained, not a process they own.

The qualification red flag specific to this category: factories that quote IEC certifications but cannot name the notified body or testing lab that issued them. In 2024, across 11 factory audits we conducted in the Pearl River Delta region, 4 produced CE declaration of conformity documents that referenced test reports we could not verify through the issuing lab’s public certificate lookup. That is a serious problem, not an administrative gap.

For incoming inspection, request electrolyte leakage and mechanical integrity checks on a sample of at least 32 units per production lot (based on ANSI/ASQ Z1.4 Level II sampling at AQL 1.0 for critical defects). Visually inspect for case deformation, weld quality at terminals, and label adhesion — these are proxies for handling quality during pack assembly that correlate with whether the factory’s safety process is actually running.

FAQ

Which standard takes precedence when IEC 62619 and IEC 62368-1 both seem to apply to my product?
Scope definitions are the deciding factor, not the product name. IEC 62368-1 applies to equipment within its defined categories (audio/video, IT, communications). IEC 62619 applies to the battery system itself when it’s used in stationary or industrial contexts. For a portable power station sold to consumers, 62368-1 typically governs the product; 62619 may additionally govern the battery subsystem if it meets the industrial use definition. When both apply, you satisfy both — they’re not mutually exclusive, and the more stringent requirement on any given parameter is the one that controls.

Does a passing UN38.3 report mean my product is safe for retail sale in the EU?
No, and conflating these is a recurring issue. UN38.3 establishes transport safety classification — it tells a freight carrier that your cells won’t self-ignite in cargo. It has no bearing on product safety certification for retail. CE marking under the relevant EU directive (typically the Radio Equipment Directive, Low Voltage Directive, or Battery Regulation depending on product type) requires separate conformity assessment against the applicable harmonized standard. UN38.3 may be referenced as supporting documentation in a CE technical file, but it does not substitute for product-level testing.

How often do Chinese factories successfully fake or borrow third-party certifications?
More than buyers expect, less than alarmists claim. Our honest assessment, based on document verification across roughly 60 supplier engagements over three years: around 15–20% of CE declarations we’ve reviewed had some form of discrepancy, ranging from expired test reports to configuration mismatches. Outright fabrication is less common than “shared certificates” where a factory applies a test report from one product configuration to a different one. Our dataset covers mainly Shenzhen and Dongguan pack assemblers in the 100–5,000Wh range — we don’t have comparable data for larger stationary BESS suppliers, where the compliance process tends to be more rigorous because the buyer is usually a utility or commercial integrator with in-house technical review.

Published by compactbess.com Technical Team | Request a sourcing consultation

---