The Evidence Ladder — What Each Standard Actually Qualifies

Last verified: July 17, 2026

Every pack-level DVP row carries an implicit evidence claim. Not "we tested this." The claim is: we tested this against a requirement that has authority over this application, at this level of the system, under conditions that represent the operating profile. Most DVP rows cannot support that full claim. The distance between what is cited and what is established is where qualification risk accumulates. It compounds without producing symptoms until a customer review or a regulator's question forces the accounting. By then the cost is measured in program months, not engineering hours.

This piece maps the pack-level evidence picture for four applications: eVTOL/eCOTL, drones/UAS, UUV/subsea, and humanoid robotics. The organizing scaffold is a six-rung evidence ladder, ordered by the strength of the qualification claim each rung supports:

1. Transport eligibility — the pack can be legally shipped.
2. Generic battery safety — the pack passes safety tests written for a broad category of lithium batteries, not for this application.
3. Borrowed adjacent baseline — a standard written for a different application is applied because nothing closer exists.
4. Installed-system requirement — a standard or regulatory document addresses the battery as installed in the target platform.
5. Platform DVP — the program's own test plan, filling gaps the external standards leave open.
6. Customer/regulator acceptance — type certificate, qualification letter, or equivalent.

The highest-value output for each application is the "does not establish" finding. A standard that fills rung 2 does not fill rung 4. A standard that fills rung 3 for one application may fill rung 4 for another. The rung depends on who is asking.

Rung 1 — Settled, Universal, Narrow

UN 38.3 applies identically to all four applications (per published standard). It establishes that the cell or battery can be transported. It does not establish operational safety, installed-system suitability, cycle-life adequacy, or mission-profile compatibility. Every application starts here. No application can stop here.

Rung 2 — IEC 62619: What It Covers and Where It Stops

IEC 62619:2022 (Edition 2.0, TC 21/SC 21A) is the generic industrial lithium battery safety standard most commonly cited across the applications below (per published standard catalog page). Its scope names stationary applications (telecom, UPS, energy storage) and motive applications (forklifts, AGVs, railway, marine). Road vehicles are explicitly excluded. Where a special-application IEC standard exists, it takes precedence (per IEC 62619:2022 scope, published catalog page).

The standard does real pack-level work. It includes a battery-system propagation test (Clause 7.3.3, with trigger methods in Annexes B and C), BMS-level functional safety tests (Clause 8: overcharge voltage control, overcharge current control, overheating control), and EMC (Clause 9) (per public preview, iTeh Standards / GSO). Most Clause 7.2 misuse tests apply at cell or cell-block level: external short circuit, impact, thermal abuse, overcharge, forced discharge. The drop test applies at both cell/cell-block and battery-system level (per public preview structure).

What IEC 62619 does not establish for any application: duty-cycle suitability, mission C-rates, cycle-life targets, mechanical shock profiles beyond the drop test, environmental exposure beyond thermal abuse, pressure tolerance, IP rating, or altitude derating. For applications within its named scope, it fills rung 2. For applications outside that scope, it is rung 3 at best, and the borrowing must be explicit in the DVP row.

One category error surfaces repeatedly and is worth naming directly. IEC 62133-2 covers portable applications. Its scope and failure-mode population are materially different from IEC 62619. Applying 62133-2 to an industrial drone pack, or 62619 to a consumer-class sUAS, conflates two standards written for different risk profiles. The DVP must specify which applies and defend the selection.

eVTOL/eCOTL — Two Jurisdictions, Two Evidence Architectures

Standards in scope: IEC 62619:2022, DO-311A, EASA MOC SC-VTOL (MOC-3 through MOC-5), 14 CFR 21.17(b) special conditions (Joby JAS4-1 per 89 FR 17230; Archer AM1 per 89 FR 45944).

The eVTOL evidence picture is the most structurally complex of the four applications. Two regulatory jurisdictions have taken materially different approaches to filling rung 4. A program targeting dual certification must satisfy both, and the two paths do not converge at the test-parameter level.

EASA provides a public, clause-level path. EASA MOC VTOL.2440 defines two approaches for propulsion-battery thermal-runaway evidence (per EASA Easy Access Rules for small category VCA). The non-propagation path requires demonstrating that a single-cell thermal-runaway event does not propagate. Test conditions include 55°C or maximum operating high-temperature stabilization, representative aircraft installation variables, and minimum 8-hour post-event monitoring with no propagation, rupture, fragment release, uncontrolled flame/emission escape, or warning/safety-function compromise. The CSFL containment path requires demonstrating that the aircraft can contain the consequences when at least 20% of propulsion-battery cells enter thermal runaway under worst-case combinations defined with the Agency. That 20% is a minimum test-severity condition, not a propagation limit.

MOC-4 SC-VTOL Issue 2 was finalized July 11, 2025 (per EASA SC-VTOL consultation page). MOC-5 SC-VTOL Issue 1 reached consultation closure October 10, 2025 but remains in comment-response status as of July 2026 (per EASA CRT document list). MOC-5 should not be cited as a finalized compliance layer.

DO-311A sits at rung 3 for propulsion batteries, not rung 4. EASA's published position is explicit: DO-311A's thermal-runaway containment test was developed for lithium batteries powering other aircraft systems or equipment and did not necessarily consider electric/hybrid propulsion batteries (per EASA Easy Access Rules). The MOC layers exist precisely because DO-311A's scope does not reach the failure modes that matter most for a propulsion battery in a powered-lift aircraft. Citing DO-311A alone as the eVTOL propulsion-battery qualification standard misrepresents the current regulatory posture in both jurisdictions.

FAA has taken a structurally different path. The Joby JAS4-1 final airworthiness criteria (89 FR 17230, effective April 8, 2024) and Archer AM1 final criteria (89 FR 45944, effective June 24, 2024) both type-certificate powered-lift as special-class aircraft under 14 CFR 21.17(b) (per Federal Register, published final rules). Both include energy-system requirements (JS4.2430 / AM1.2430) covering independence between energy-storage systems, usable-energy indication, safe energy isolation, and hazard minimization after emergency landing. Both include fire/overheat isolation requirements (JS4.2440 / AM1.2440).

Neither notice publishes pack-level thermal-runaway test parameters. No trigger method. No percent-cells-in-runaway threshold. No stabilization temperature, post-event monitoring duration, venting configuration, or no-propagation acceptance criteria. When commenters on the Joby notice specifically requested battery-level criteria for fire protection, propagation, crashworthiness, high-voltage disconnection, and toxic-gas leakage, the FAA responded:

and stated that those risks would be addressed through existing Subparts E and F, JS4.1529, Appendix A ICA requirements, and tailored means of compliance (per 89 FR 17230, FAA response to comments). The Archer notice contains the same structural response to the same category of comment (per 89 FR 45944).

The FAA's rung 4 is aircraft-level and performance-based. The battery-pack specifics live in the applicant's certification plan and issue papers, not in a public standard. For a dual-certification program, the evidence paths diverge at rung 4: EASA provides a public clause-level target, FAA provides an applicant-specific negotiation framework. In practice this means building to the more prescriptive path and demonstrating equivalence to the other.

Does not establish (either jurisdiction), by parameter class:

Parameter class	Status	Notes
Energy density / capacity	DVP-defined	No minimum or maximum Wh/kg or Wh/L threshold in published criteria
Power capability	DVP-defined	Mission-phase C-rates (editorial: vertical takeoff at 2.5–4.5C, cruise at 0.75–1.5C per published engineering analyses, not a governing standard) not specified in any published airworthiness criterion
Cycle life / calendar life	DVP-defined	No published cycle-count or calendar-degradation threshold under mission-representative profiles
Operating envelope	Partially governed	Altitude and temperature addressed at aircraft level; battery-specific derating DVP-defined
Mechanical (vibration)	DVP-defined	eVTOL rotor-harmonic vibration profiles not addressed in DO-311A or published special conditions
Safety (TR propagation)	Jurisdiction-split	EASA MOC VTOL.2440 provides clause-level criteria; FAA defers to applicant issue papers
Environmental protection	DVP-defined	IP rating, salt fog not addressed in published criteria
EMC	Governed (DO-160)	At aircraft-equipment level, not battery-specific
Transport	Governed (UN 38.3)	—

Drones/UAS — The Split

Standards in scope: UN 38.3, IEC 62619:2022, IEC 62133-2, ASTM F3005-22, MIL-STD-810 (military platforms).

The drone evidence picture splits by platform class. The split begins at rung 2.

ASTM F3005-22, developed by ASTM F38.01, is the most commonly cited sUAS battery standard. It is five pages long. Its public scope explicitly states that it "does not define requirements for the systems in which sUAS battery packs may be utilized" and is subordinate to ASTM F2910 (per public ASTM/ANSI summary; full text is paywalled). Confirmed coverage from public summaries includes cells, mechanical design/assembly and mechanical safety, electrical design, pack maintenance, and maintenance-data records. No cycle-life test, environmental test matrix, thermal-runaway propagation test, EMC test, or IP/ingress test is visible in public summaries. The defensible statement: these parameter classes are not confirmed in accessible summaries. That is distinct from asserting they are absent from the full paywalled text.

F3005-22 fills rung 3 for sUAS battery design and maintenance evidence. It does not establish installed-system qualification, mission endurance, reserve energy, aircraft-level fire containment, operational approval, or FAA/EASA acceptance (per public scope statement).

For larger industrial or military drones, IEC 62619 provides a stronger rung 2 with its propagation test and BMS control requirements (per public preview, Clause 7.3.3 and Clause 8). Military drone programs typically borrow environmental and mechanical testing from MIL-STD-810 at rung 3, adding vibration, shock, temperature-altitude, and humidity profiles that no battery-specific standard addresses for this platform.

The split is structural. A commercial inspection sUAS and a military ISR drone share rung 1 and diverge from rung 2 upward. The military platform's evidence ladder has more populated rungs, but they are populated with borrowed standards, not with standards written for the application. More rungs filled is not the same as more rungs governed.

Does not establish (either class), by parameter class:

Parameter class	Status	Notes
Energy density / capacity	DVP-defined	Neither ASTM F3005 nor IEC 62619 sets application-specific thresholds
Power capability	DVP-defined	Mission-phase C-rates (hover, transit, payload operation) not specified
Cycle life / calendar life	DVP-defined	No published cycle-count requirement under representative mission profiles
Operating envelope	Borrowed (military)	MIL-STD-810 provides temperature, altitude, humidity profiles; not battery-specific
Mechanical	Partially borrowed	MIL-STD-810 vibration/shock for military; no equivalent for commercial sUAS
Safety (TR propagation)	Rung 2 only (IEC 62619)	Not addressed in ASTM F3005 public summaries; IEC 62619 propagation test is generic, not drone-specific
Environmental protection	DVP-defined	IP rating, salt fog not addressed in any cited standard for this application
EMC	DVP-defined	Not visible in ASTM F3005 public summaries; IEC 62619 Clause 9 is generic
Transport	Governed (UN 38.3)	—

UUV/Subsea — The Structural Void

Standards in scope: UN 38.3, IEC 62619:2022 (if motive/marine scope applies). No application-specific standard identified.

No public application-specific lithium battery-pack qualification standard was found in ISO, DNV, NATO, or IEEE catalogs for unmanned underwater vehicles or subsea autonomous systems. This finding is based on public catalog searches. It does not rule out nonpublic customer specifications, subscription-only DNV rule clauses, military program requirements, or proprietary DVPs. But the absence in public catalogs is itself a finding.

DNV publishes rules for underwater systems including ROVs, AUVs, and unmanned underwater systems, and states that where no standards exist, DNV rules deliver guidance or recommend best practices (per DNV public service page). DNV Battery(Safety) is mandatory for vessel installations above 20 kWh (per DNV decarbonization material), addressing thermal-runaway, fire, explosion, and toxicity risks. That framework addresses marine-vessel fire and explosion risk in atmospheric or near-atmospheric conditions. It does not address UUV pressure cycling, enclosure leakage under depth, thermal rejection in a pressure-compensated housing, or post-fault containment at depth. It supports a marine-vessel borrowed baseline at rung 3. DNV does not present it as a UUV/AUV battery-pack qualification standard. DNV-RP-0577 (2023) standardizes cell-level degradation testing and explicitly excludes safety, maximum capacity, maximum current, and energy density (per DNV public scope description). It closes no UUV pack-level row.

NATO STANAGs surfaced in public catalogs address interoperability and underwater communications. STANAG 4817 covers command-and-control frameworks for unmanned systems; STANAG 4748/JANUS covers digital underwater signaling (per NATO CMRE 2025 annual report). Neither addresses battery qualification.

IEEE public standards searches surfaced lithium-battery standards for UAS, EV propulsion, stationary installations, and submersible electrical-equipment enclosure integrity (per IEEE SA catalog). None identifies a UUV/AUV lithium battery-pack standard for pressure cycling, leakage, thermal rejection under pressure, or post-fault containment.

The evidence ladder for UUV/subsea has a populated rung 1 (UN 38.3), a partially applicable rung 2 (IEC 62619, if the pack falls within its motive/marine scope), and nothing above that. Every parameter that matters to a subsea battery is program-defined. Operational UUV depth envelopes range from approximately 100 m (~11 atm) for shallow-water survey platforms to 6,000 m (~600 atm) for full-ocean-depth systems (editorial estimate from published platform specifications). No standard addresses pressure cycling for lithium battery enclosures at any point in that range.

Does not establish (all sources checked), by parameter class:

Parameter class	Status	Notes
Energy density / capacity	DVP-defined	No application-specific threshold
Power capability	DVP-defined	Propulsion and hotel-load C-rates at depth not addressed
Cycle life / calendar life	DVP-defined	Degradation under pressure-cycling thermal profiles not addressed
Operating envelope (pressure)	DVP-defined	Pressure cycling to operational depth (editorial: 11–600 atm range), seal integrity under thermal cycling at pressure — no standard
Operating envelope (temperature)	DVP-defined	Thermal rejection in sealed/pressure-compensated housing without atmospheric convection — no standard
Mechanical	DVP-defined	Shock from launch/recovery, handling at sea state — no subsea battery standard
Safety (TR containment)	DVP-defined	Thermal-runaway containment in a sealed housing with no atmospheric venting path; post-fault behavior in a flooded compartment — no standard
Environmental protection	DVP-defined	Seal integrity, leak detection, corrosion under saltwater immersion — no standard
EMC	DVP-defined	Electromagnetic environment in a metallic pressure hull — no standard
Transport	Governed (UN 38.3)	—

No public working group or committee designation addressing UUV lithium battery-pack qualification was identified in any catalog checked. Programs operating in this space should treat their DVP as the primary qualification document and build it with the rigor of a standard, because functionally it is one.

Humanoid Robotics — Borrowed at Every Rung

Standards in scope: UN 38.3, IEC 62619:2022 (if industrial motive scope applies), ISO 13482 (personal care robot safety, not battery-specific).

The humanoid evidence picture is borrowed from adjacent domains at every rung above transport. The borrowing is thinner than it appears.

IEC 62619 provides rung 2 if the battery system falls within its industrial motive scope. Its propagation test (Clause 7.3.3), BMS control tests (Clause 8), and EMC (Clause 9) are real pack-level evidence (per public preview). But the standard's mechanical tests expose the gap most clearly. IEC 62619's drop test (Clause 7.2, applying at both cell/cell-block and battery-system level per public preview) specifies a controlled drop from a defined height onto a hard surface. A humanoid robot falling from standing height subjects its battery to a fundamentally different event: 1.5 to 1.8 meters for current platforms, multi-axis deceleration with rotational components, impact geometry determined by fall dynamics rather than test fixture geometry, and the potential for repeated falls across the product's service life. The IEC 62619 drop test was designed for a battery being handled during installation or maintenance. It was not designed for a battery inside a platform that falls as a foreseeable operating condition. The test geometry is wrong. The energy is wrong. The repetition count is wrong.

IEC 62619 does not establish: humanoid duty-cycle suitability, fall or contact shock profiles, user-proximity installed-system containment requirements, locomotion shock profiles, power-pulse profiles for dynamic balance recovery (editorial: these transients can exceed steady-state locomotion power by multiples, but no published standard quantifies the requirement), cycle-life targets under rapid charge-discharge transients characteristic of bipedal gait, or thermal management constraints in an enclosed anthropomorphic form factor with limited convective surface area.

ISO 13482 (safety requirements for personal care robots) provides context for what the battery's installed-system containment must protect against, but it is a hazard framework, not a battery standard. It addresses hazard categories relevant to robots operating in close proximity to humans in uncontrolled environments, including contact hazards, entrapment, and the consequences of robot instability (editorial: ISO 13482 is cited here for its hazard framework; it does not specify battery test parameters, trigger methods, or acceptance criteria). A DVP for a humanoid robot battery should reference ISO 13482's hazard categories to define what the containment design must prevent, then set battery-specific test parameters to demonstrate that prevention. No published standard bridges that gap.

Does not establish (all sources checked), by parameter class:

Parameter class	Status	Notes
Energy density / capacity	DVP-defined	No humanoid-specific threshold
Power capability	DVP-defined	Locomotion C-rates, balance-recovery transient peaks, manipulation power profiles — no standard
Cycle life / calendar life	DVP-defined	Cycle count under humanoid duty profiles (frequent partial discharge, high-rate transients) — no standard
Operating envelope	Partially borrowed	IEC 62619 thermal abuse covers generic temperature range; humanoid-specific thermal constraints (enclosed torso, limited venting) DVP-defined
Mechanical (fall shock)	DVP-defined	IEC 62619 drop test does not replicate humanoid fall geometry, energy, or repetition; no standard addresses this
Mechanical (vibration)	DVP-defined	Locomotion-induced vibration spectrum not addressed in any cited standard
Safety (TR containment)	Rung 2 only (IEC 62619)	Propagation test exists but does not address user-proximity containment in an uncontrolled environment
Environmental protection	DVP-defined	IP rating for outdoor humanoid operation — no standard
EMC	Rung 2 only (IEC 62619 Clause 9)	Generic, not humanoid-specific
Transport	Governed (UN 38.3)	—

No standards body has published a working group designation for humanoid robot battery qualification. Every parameter class that distinguishes a humanoid battery from a generic industrial battery is DVP-defined.

The 6T Contrast

MIL-PRF-32565 is the active performance specification for rechargeable sealed 24V lithium-ion 6T batteries for military ground vehicles (document date August 4, 2022; next review due August 3, 2027; per DLA public metadata). It defines the form factor, the electrical performance requirements, the environmental test matrix, and the acceptance criteria for a named use case. QPL-32565 exists as the acceptance mechanism, though no public product rows are visible in accessible metadata.

This is what a settled rung 4 looks like. A standard written for a specific application, with test parameters derived from that application's operating profile, maintained by an authority with jurisdiction over the platform. The 6T lane has its own unresolved questions: the relationship between MIL-PRF-32565 compliance and Section 4865 FEOC supply-chain requirements remains unsettled, and QPL visibility is incomplete. But the specification itself fills the installed-system rung with application-specific evidence. None of the four applications mapped above has an equivalent.

DVP Implications

The pattern across all four applications is consistent. The bottom rungs are populated by standards answering questions other than "is this battery qualified for this application?" The middle rungs are void, borrowed, or jurisdiction-dependent. The top rungs are program-specific.

Each DVP row should carry an evidence-source designation:

• Governs — a standard written for this application and parameter class (per published standard scope).
• Complements — a standard that addresses this parameter but not for this application.
• Transport-only — UN 38.3.
• Borrowed — applied from an adjacent domain; borrowing made explicit, source domain named.
• DVP-defined — no external standard; the program sets the requirement and the acceptance criteria.
• Regulator-negotiated — the requirement exists but is defined through an applicant-specific process (e.g., FAA special conditions, issue papers).

The gaps documented above are not failures of the standards system. They are the current state of applications that are newer than the standards infrastructure built to govern them. Naming a gap precisely is the first step toward a defensible DVP row. The step that costs eighteen months is the other one: papering over the gap with a borrowed citation that implies more coverage than it provides, then discovering the shortfall when a qualification reviewer asks what the citation actually establishes.