Somewhere in every electrification project there is a morning when a freshly built battery pack needs to go somewhere else: to a test lab, a vehicle builder, a customer. And that morning a logistics question becomes a compliance question, because lithium batteries are dangerous goods, and dangerous goods law asks for one specific thing before any carrier will touch the crate: evidence that the battery passed UN 38.3.
Snelle antwoorden
- What is UN 38.3? The section of the UN Manual of Tests and Criteria that defines the transport safety tests for lithium cells and batteries, required under road (ADR), sea (IMDG) and air (IATA) rules
- What are the tests? Eight tests, T1 to T8: altitude simulation, thermal cycling, vibration, shock, external short circuit, impact or crush, overcharge and forced discharge
- Who needs it? Anyone shipping lithium cells, packs or equipment containing them: manufacturers, integrators and importers. The obligation follows the battery, not the brand
- What is the test summary? A standard document summarising the UN 38.3 evidence for a battery, which manufacturers and distributors must make available down the supply chain
Why transport law tests your battery
Transport regulators care about one scenario: a battery in a cargo hold or trailer entering thermal runaway with no one able to intervene. The UN 38.3 series therefore recreates the conditions of transport rather than the conditions of use: the low pressure of an aircraft hold, the temperature swings of intercontinental shipping, the vibration and shocks of the road, and the electrical abuse cases that a damaged or mishandled battery might see. Pass criteria are strict and simple: no disassembly, no fire, no leakage beyond limits, and voltage retained where required.
Readers who know [ECE R100] will notice overlapping themes, vibration and short circuit among them, and wonder whether one programme can cover both. The honest answer: the philosophies differ. R100 qualifies a battery for a life inside a vehicle; UN 38.3 qualifies it for the journey between factories. The test levels, sample states of charge and pass criteria differ in the details, and authorities do not accept one as a substitute for the other. Plan both, and where possible plan them together, because the sample logistics overlap beautifully.
The eight tests at a glance
Test | Naam | Simulates |
T1 | Altitude simulation | Unpressurised air transport at altitude |
T2 | Thermal test | Extreme temperature cycling between hot and cold |
T3 | Vibration | Road and air transport vibration spectrum |
T4 | Shock | Handling drops and impacts in transit |
T5 | External short circuit | Conductive contact across terminals |
T6 | Impact / crush | Mechanical damage to cells in handling |
T7 | Overcharge | Charger fault on rechargeable batteries |
T8 | Forced discharge | Cell driven into reversal in a multi-cell pack |
The tests run in sequence on defined sample sets, and the sample count is what surprises most teams: depending on configuration, a full programme consumes cells and packs in double digits. For a large heavy-duty pack that is a real cost, which is why the regulation allows evidence to be built at cell and module level with assembly rules on top. If your cell supplier holds solid UN 38.3 evidence, and reputable suppliers do, your programme shrinks to what changed at pack level.
Prototypes: the exemption everyone should know
A prototype pack by definition has no completed UN 38.3 file, and development would be impossible if that meant no transport. The dangerous goods rules solve this with a special provision for pre-production batteries: transport is allowed under specific packaging and approval conditions, road rather than air, with competent authority involvement where required. It works, but it is paperwork with lead time. Our standing advice to every [lithium conversion] and prototype programme: appoint one owner for battery logistics on day one, because the fastest way to lose three project weeks is to discover the shipping rules on the morning of the shipment.
The mistakes that stop shipments
The same handful of errors accounts for most UN 38.3 trouble we see, and every one is avoidable. Assuming the cell evidence covers the pack: it covers the cells; the pack needs its own assessment the moment it adds energy, circuitry or mechanical structure. Testing the wrong revision: the file describes design revision B while revision C ships, a mismatch that surfaces at the worst possible desk, the carrier’s. Ignoring state of charge rules: air transport of lithium ion batteries carries a state-of-charge cap, and a pack charged for commissioning convenience can be an illegal shipment. Improvised packaging: dangerous goods packaging is certified, not sturdy; a robust crate without the right UN marking fails inspection exactly like a flimsy one. And forgetting the people: staff who prepare lithium shipments need dangerous goods training with a certificate behind it. An afternoon of process design covers all five; a stopped shipment at an air freight hub covers none of them and costs more.
Timeline, cost drivers and the test summary
A pack-level UN 38.3 programme, assuming cells with existing evidence, typically runs six to ten weeks from kickoff to report: sample builds and conditioning first, then the environmental and mechanical sequence, then the electrical abuse tests, then reporting. The cost drivers are refreshingly concrete. Sample count dominates, which is why cell and module evidence reuse matters so much. Pack size sets the equipment class: a forklift module fits standard chambers, a 600 kWh truck pack needs facilities that can vibrate, heat and short-circuit at that scale safely. And schedule risk sits almost entirely in sample readiness, because a slipped pack build cascades through a booked test sequence.
The output that the supply chain actually handles is the test summary: the standardised document that names the battery, the tested configuration and the passing results. Since it became mandatory to make available, carriers, customers and authorities have learned to ask for it, and a missing or vague test summary now stops shipments as effectively as a missing label. Keep it current per design revision, and version it the way you version the pack itself.
Plan the battery paperwork with the engineering
On the campus, battery test planning, vehicle integration and homologation sit at one table. UN 38.3 evidence, R100 approval and thermal validation can share samples and schedule instead of competing for them.
1. What is UN 38.3 testing in simple terms?
The set of eight safety tests a lithium battery must pass before it may be transported by road, sea or air. It proves the battery survives the physical and electrical stresses of transport without fire or failure.
2. Is UN 38.3 required by law?
Yes, indirectly: ADR, IMDG and IATA dangerous goods rules all require lithium batteries to be of a design that passed UN 38.3, and require the test summary to be available in the supply chain.
3. What is the difference between UN 38.3 and ECE R100?
UN 38.3 covers transport safety; ECE R100 covers the battery’s safety inside a type-approved vehicle. The test programmes overlap thematically but differ in detail, and both are normally required.
4. Can I ship a prototype battery without UN 38.3?
Yes, under the special provision for pre-production cells and batteries, with strict packaging and approval conditions and normally not by air. It requires preparation, so plan it early.
5. Do I need to retest after changing a pack design?
Significant changes, such as different cells, cell count, or mechanical design, trigger reassessment and usually partial retesting. Minor changes may be covered by rationale; the test house should judge each case.
OVER DE AUTEUR
IntegratR Ingenieursteam
EMC & e-mobiliteit engineering, Innovatiecampus Nijmegen
IntegratR runt een zware EMC-testfaciliteit en engineeringsteam op de Innovation Campus in Nijmegen, Nederland. Het team brengt emissievrije voertuigen en machines van prototype naar gecertificeerd, en werkt samen met OEM's, Tier-1-leveranciers en scale-ups in heel Europa. Artikelen worden geschreven vanaf de testvloer, niet vanuit theorie.