TL;DR: Compliance for cycle life and degradation is not a single-standard problem — the standard you need depends on application class, market, and whether your buyer treats performance retention as a safety criterion.
TL;DR: Under IEC 62619:2022 Clause 7.3, a cell must retain at least 80% of rated capacity after the number of cycles declared by the manufacturer — but the declared cycle count is manufacturer-defined, which is the clause most sourcing teams miss.
What the Standards Actually Test: Cycle Life as a Safety Criterion #
Most design engineers treat cycle life testing as a performance specification and safety certification as a separate track. That separation is a mistake that surfaces late, usually at pre-shipment audit or market entry.
The reality is that degradation behavior sits inside the safety envelope of multiple international standards. Capacity fade affects state-of-charge accuracy, which affects overcharge risk, which is a thermal runaway precursor. Standards bodies have understood this for over a decade, and the current revision cycle reflects it.
Here is how the major frameworks divide the territory:
| Standard | Scope | Cycle Life Clause | Pass Criterion | Application Class |
|---|---|---|---|---|
| IEC 62619:2022 | Secondary Li cells/packs, stationary + industrial | Cl. 7.3 — endurance test | ≥80% capacity at declared cycle count | Stationary, industrial, EV adjacent |
| UL 1973 | Stationary storage, light EV | Sec. 44–46 — cycle endurance | No degradation-induced safety event; 70% threshold for listing | Stationary, grid-tied |
| UN38.3 | Transport safety — cells and packs | T.1–T.8; no dedicated cycle test | Capacity ≥90% initial after T-series conditioning | Transport, all classes |
| GB/T 36276-2023 | Chinese national — stationary Li BESS | Cl. 6.4 — cycle life | ≥80% at 500 cycles (0.5C/0.5C, 25°C) | Stationary, China market |
| IEC 62133-2:2017 | Portable Li secondary cells/packs | Cl. 8.3.3 — endurance | No safety failure; capacity retention not numerically defined | Portable consumer |
The table makes one conflict visible immediately: IEC 62133-2 governs portable packs but sets no numeric retention floor for cycle life. It only requires that no safety event occurs. For a product sold into Europe as a portable power station, that means your CE mark via EN 62133-2 gives your buyer no verified cycle life claim whatsoever. If your product spec sheet says “2000 cycles to 80%,” that number exists entirely outside the certification boundary.
This is where I’d prioritize a supplementary test report to IEC 62619 methodology even for portable-class products, especially if your buyer is a system integrator who resells into commercial applications. The methodology gap between “no safety event at 50 cycles” and “documented 80% retention at 1,000 cycles” is what separates a defensible datasheet from a liability.
For LFP cell selection decisions upstream of pack design, this standard boundary matters from day one — you need the cell-level cycle data to already be aligned with the standard your pack will be tested against.
Where Degradation Testing Breaks Down in Practice #
Three failure patterns recur across our supplier qualification work. Each starts with a different root cause but converges on the same outcome: a certified product that degrades faster than declared, with consequences downstream that trace back to how the standard was applied.
The first is test condition gaming. IEC 62619 Clause 7.3 requires endurance testing at conditions “declared by the manufacturer.” A Shenzhen-based pack house can declare 0.2C charge, 0.2C discharge at 20°C and pass 2,000 cycles on paper. Ship that product into a residential BESS application that cycles at 0.5C/0.5C in a 35°C ambient, and you will see 800-cycle real-world life against a 2,000-cycle certificate. The certificate is not fraudulent. The test condition is just irrelevant to the application.
In our QC-F12 sourcing audit procedure, we require factories to provide cycle data at at minimum 0.5C/0.5C and 1C/1C — both, not one. The delta between those two datasets tells us more about cell quality than the 0.2C number ever could. LFP cells from one Dongguan pack integrator we qualified in 2024 showed 91% retention at 2,000 cycles under 0.5C/0.5C, dropping to 78% under 1C/1C. That 13-point gap is a meaningful indicator of internal resistance growth and electrolyte quality.
The second failure pattern involves multi-standard products. A battery pack designed for a portable power station that also supports a stationary backup mode now falls under the scope of both IEC 62133-2 (portable) and potentially IEC 62619 (stationary). Some factories handle this by certifying to the less demanding standard and shipping to both applications. A European grid operator sourced 48V 20Ah backup packs in 2023 on the strength of an IEC 62133-2 certificate. Post-delivery capacity mapping showed 17 of 40 units fell below 80% rated capacity within 14 months of light cycling. The IEC 62133-2 cert was technically valid. The application scope was wrong. Total replacement cost: approximately $43,000, plus contractual penalties.
The third pattern is GB/T substitution for export products. GB/T 36276-2023 is a rigorous standard, and its 500-cycle minimum at 0.5C/0.5C with ≤20% capacity loss is a real test. But it is a Chinese national standard, not a harmonized international standard. Some factories present GB/T 36276 test reports as equivalent to IEC 62619 for export sales. They are not equivalent. GB/T 36276 does not include the same abuse testing matrix as IEC 62619, and its test conditions are fixed rather than manufacturer-declared — which actually makes it less gameable in one respect but not interchangeable for EU or US market entry. The Safety & Certification category covers market-specific mandatory standard requirements in more detail.
When you receive a test report citing GB/T 36276 for a product destined for Germany, ask the factory directly: “Do you have an IEC 62619 report, or are you representing these as equivalent?” The answer tells you everything about their export compliance maturity.
Is UN38.3 a Cycle Life Standard? #
No. UN38.3 is a transport safety standard, not a performance standard, and it contains no cycle life requirement in the conventional sense.
The T-series tests in UN38.3 (Rev. 21, 2023) cover altitude simulation, thermal cycling, vibration, shock, external short circuit, impact/crush, overcharge, and forced discharge. The 90% capacity retention figure referenced in T.1 conditioning is a measurement checkpoint, not a lifecycle test. A cell can pass UN38.3 completely with 200 cycles of conditioning data and still fail at 500 cycles under real operating loads. Transport certification confirms the cell won’t vent or ignite in a cargo hold — nothing more.
The confusion arises because many factory datasheets list UN38.3 as evidence of product quality or durability. It is neither. For cycle life compliance, you need IEC 62619 or a cell-level test report against the buyer’s contractually specified test conditions.
Sourcing Guidance for Buyers #
When evaluating Chinese suppliers in this category, the first document to request is a cycle life test report that explicitly states: test current (charge and discharge), temperature, cutoff voltage, and the specific cell lot or pack serial number tested. Any report that omits test current or temperature is unverifiable. A factory that cannot produce a complete test report with those parameters has either not done the testing or is sharing a report from a different configuration.
One qualification red flag specific to cycle life and degradation: watch for test reports where the declared end-of-life criterion is capacity below 70% rather than the conventional 80%. Some factories use 70% to inflate their cycle count numbers. A 70% EOL threshold at 3,000 cycles is not comparable to an 80% EOL threshold at 2,000 cycles — the latter represents significantly better performance at the standard evaluation point.
For incoming inspection, pull a minimum sample of 3 units per 100-unit lot and run a 50-cycle accelerated capacity check at 1C/1C, 25°C, with voltage windows matching the product spec. At cycle 50, a healthy Grade-A LFP cell should show no more than 1.8% capacity loss from initial. Any unit showing more than 3.2% loss at cycle 50 under those conditions is either a Grade-B cell or has a BMS calibration fault affecting cutoff accuracy — both require hold-for-investigation status before acceptance.
Frequently Asked Questions #
Which standard is mandatory for selling a portable power station in the EU?
For lithium portable power stations sold in the EU, CE marking currently routes through EN 62133-2 (the harmonized version of IEC 62133-2) under the Low Voltage Directive or Radio Equipment Directive depending on features. The EU Battery Regulation (2023/1542) introduces mandatory cycle life performance declarations starting with industrial and EV batteries, with portable battery requirements phasing in — design engineers should be tracking the delegated acts timeline, as numeric cycle life thresholds are expected before 2027.
Can a cell that passes IEC 62619 also satisfy UL 1973?
It depends on the specific test clauses involved. IEC 62619 and UL 1973 overlap in scope for stationary applications but are not mutually substitutable for US market entry. UL 1973 is the pathway to UL Listing for stationary storage in North America, and UL Listing is required by many AHJs (Authorities Having Jurisdiction) for building permits and insurance. Holding an IEC 62619 certificate helps with technical documentation but does not replace the UL 1973 Listing. Budget for both if your product targets EU and North American stationary markets.
What does “80% capacity retention” actually mean in a test report?
It means that at the declared cycle count, the measured discharge capacity under the test conditions is at least 80% of the initial discharge capacity measured at cycle 1 (or after formation cycles, depending on the protocol). The critical detail is “under the test conditions” — capacity retention is not a fixed number for a cell, it’s a function of rate, temperature, and depth of discharge. An 80% retention figure at 0.2C/25°C is a different statement than 80% retention at 1C/35°C.
How does GB/T 36276-2023 differ from the 2018 version for cycle life?
The 2023 revision tightened the cycle life test to 500 minimum cycles at 0.5C/0.5C with explicit ≤20% capacity loss criterion, added a temperature variation test during cycling that the 2018 version lacked, and aligned more closely with IEC 62619 abuse test methodology. Products certified to GB/T 36276-2018 are not automatically compliant with the 2023 revision — factories supplying into Chinese national grid procurement tenders should have already requalified, but many Shenzhen-area pack houses supplying export products have not updated their test reports.
Is a shared UN38.3 certificate acceptable for a new cell configuration?
No. UN38.3 Section 38.3.5 requires that test reports identify the specific cell type and configuration tested. A certificate obtained for a 200Ah prismatic cell cannot cover a 100Ah cell of the same chemistry — the physical and electrical parameters differ, and the applicable test conditions differ. If a factory presents a shared or family certificate covering multiple capacity variants under a single test report, request the raw test data and verify that the cell dimensions, capacity, and electrode configuration in the report match your specific sourced unit. Mismatched UN38.3 certs are more common than the industry acknowledges, particularly for cylindrical cells where factories pool test reports across grades.
Do I need separate certifications for the cell and the pack?
Generally, yes. Cell-level certification (UN38.3, IEC 62133-1 for the cell standard) covers the electrochemical unit. Pack-level certification (IEC 62619, IEC 62133-2, UL 1973) covers the integrated system including BMS protection behavior. A pack built with certified cells is not automatically a certified pack. The BMS, protection thresholds, enclosure, and thermal management all fall under pack-level evaluation. For BMS engineering decisions that affect pass/fail on pack-level safety testing, the protection threshold requirements in each standard need to be validated against your specific BMS configuration.
Which standard should I specify in a purchase contract for a Chinese-manufactured stationary BESS?
Specify IEC 62619:2022 as the primary standard with explicit test conditions written into the contract: 0.5C charge, 0.5C discharge, 25°C ±2°C, 80–20% SOC window, with capacity retention ≥80% at the contracted cycle count. Adding GB/T 36276-2023 as an additional requirement is reasonable for dual-market products but does not replace IEC 62619 for export compliance. Build in a contractual right to request raw test data (not just the test report summary), and specify that the cell lot serial numbers on the test report must match production lot serial numbers on delivery. Without that clause, you have no way to verify the report was run on the cells you actually receive.
Published by compactbess.com Technical Team | Request a sourcing consultation
On the GB/T 36276-2023 test conditions — 0.5C/0.5C at 25°C is a pretty controlled lab environment, so how are teams handling correlation to field data when ambient temps in a real off-grid deployment swing between -10°C and 45°C seasonally? We’ve seen capacity fade curves diverge significantly from the certified baseline by year two on sites with that kind of thermal range.
The manufacturer-declared cycle count loophole in IEC 62619 Cl. 7.3 is where the real cost engineering happens — a supplier can declare 500 cycles, pass at 80%, and that’s a compliant product even if your application needs 2000. We’ve seen sourcing teams accept that compliance paperwork at face value and then eat warranty replacement costs around month 18 on 48V/100Ah portable station SKUs because nobody caught that the declared count was set to what the cell could reliably hit, not what the product actually demands.
The UL 1973 Sec. 44-46 threshold mismatch with IEC 62619 caused us a real headache on a dual-certified project last year — our BMS firmware was parameterized around the 80% SOH threshold for capacity derating triggers, and when the UL listing required us to demonstrate no degradation-induced safety event down to 70%, the protection thresholds in the firmware had to be re-validated against a completely different end-of-life definition. Two separate threshold tables in the same firmware build, which the BMS vendor’s config tool didn’t natively support as separate certification profiles.
Cell matching tolerance is something we’ve had to contractually define ourselves because none of these standards actually specify it — UN38.3 T-series conditioning will pass a pack that would fail our internal impedance spread limit of ±5mΩ at 1kHz within a 20-cell parallel group. The standard only cares that capacity is still ≥90% post-conditioning, not whether the cells sharing that load started life with mismatched internal resistance that’ll accelerate divergence over 18 months of field cycling.
IEC 62133-2:2017 Cl. 8.3.3 is the one that catches portable power designers off guard in mil/def programs — 50 cycles at 0.2C/0.2C is the conditioning regime, and that rate bears almost no resemblance to the 2C+ pulse discharge profiles we’re qualifying against for man-portable loads. We’ve had units sail through 8.3.3 and then show accelerated lithium plating under actual field duty cycles within the first 90 days of deployment.
We had a 480kWh grid-tied BESS at a water treatment facility in the southwest — commissioned late 2020, UL 1973 listed — and the degradation-induced safety event threshold in Sec. 44-46 didn’t flag anything for the first 28 months because capacity was holding above 70%. What the listing test didn’t capture was that the pack was cycling in a 38°C average ambient with daily depth-of-discharge pushing 90%, and by month 31 we had three modules showing lithium plating on post-mortem teardown with SOC estimation error bad enough that the BMS was regularly pushing cells past their actual full-charge voltage. The 70% pass criterion basically gave us a false sense of remaining margin that the real operating conditions had already eaten through.
UN38.3 T.3 (altitude simulation) came back as a surprise failure on a pack we’d already cycled through T.1 and T.2 clean — turned out the capacity loss from just those two conditioning steps pushed our cells close enough to the edge that the electrolyte seal integrity was marginal under low pressure. Nobody on our team had modeled the interaction between pre-conditioning capacity loss and the 11.6 kPa chamber test, and the standard gives you no guidance on sequencing dependencies. Cost us about 10 weeks mid-2022 getting a revised cell lot re-qualified.