TL;DR: The standard printed on a charger’s label and the standard that actually governs its safety testing are often two different documents — understanding that gap is where compliance risk lives.
TL;DR: A USB-C PD 3.1 charger operating at 48V/5A (240W) must clear at least 6 distinct standard layers before it qualifies for sale in the EU, US, and Japan simultaneously.
The Specification That Drives Compliance Outcomes: Peak Operating Voltage Under Fault Conditions #
Most design engineers anchor their compliance strategy around rated output power. That’s the wrong anchor. The specification that determines which standards apply, which test regimes trigger, and whether your product can cross borders without re-certification is peak voltage under fault conditions — not nominal output.
IEC 62368-1:2023 Clause 5.4.4 defines the hazard-based safety engineering (HBSE) framework that replaced IEC 60950-1 and IEC 62087 for audio/video and IT equipment including chargers. Under this framework, the threshold that changes your test obligations isn’t your rated 20V or 28V — it’s whether your product can produce voltages above 60VDC (or 42.4Vpeak AC) under single-fault conditions. USB-C PD 3.1 Extended Power Range (EPR) operates at up to 48V, which means fault-state voltages can breach the ES2/ES3 boundary in IEC 62368-1 and trigger a completely different set of safeguard requirements.
This matters in practice because a design that passes ES2 touch-current tests at 20V may fail at 48V EPR. We’ve seen qualification packages from Shenzhen-area charger factories where engineers ran all ES2 testing at the 20V PPS (Programmable Power Supply) operating point and simply noted “EPR not tested” in the report. That’s not a documentation shortcut — it’s a compliance gap that will surface during EU Notified Body review.
The second critical parameter is the overcurrent fault response time. UN 38.3 Section 38.3.4 doesn’t directly test charger ICs, but it governs the cells powering any portable charging device. If your charger design integrates a battery (common in GaN travel chargers with passthrough), the cell-level fault response time must be validated separately from the charger IC’s protection logic. These two timings don’t always coordinate. In our QC-07 incoming inspection procedure, any integrated charger+battery module gets fault-timing coordination checked as a discrete pass/fail step — because miscoordination between cell BMS cutoff and charger IC overcurrent response is one of the more common root causes of thermal events in this product class.
Supplier Qualification: What to Request and What the Response Tells You #
When you’re qualifying a USB-C PD charger supplier from Dongguan or Shenzhen, the first document to request is not the CE Declaration of Conformity. Request the full IEC 62368-1 test report with the specific clause-by-clause pass/fail table, issued by a recognized third-party lab (SGS, Bureau Veritas, TÜV Rheinland, or UL-listed lab). Ask them to highlight which operating mode — SPR or EPR — was tested at which voltage setpoints.
A supplier that responds within 48 hours with a complete clause-level report is telling you something valuable: they’ve been through this process multiple times, understand what buyers need, and have invested in proper lab relationships. A supplier who sends a one-page certificate with a logo and an expiry date is telling you they’re accustomed to buyers who don’t read the underlying data.
For UL certification specifically, ask for the UL File Number and cross-reference it at UL Product iQ before the conversation goes any further. We’ve audited 9 USB-C PD charger suppliers over the past 18 months and found 3 whose UL certificates referenced a file number that either listed a different product configuration or had lapsed status. Not forged certificates necessarily — just certificates that had drifted from the actual product through design revisions that were never re-submitted.
For Japan sales, ask whether they hold PSE (Electrical Appliance and Material Safety Act) certification under the Specially Controlled Electrical Appliances category. Many Chinese suppliers hold PSE for lower-power chargers but have not gone through the process for 100W+ PD 3.1 products. PSE testing is conducted by METI-registered bodies and is not interchangeable with CE or UL — a supplier who tells you “CE covers Japan too” has just told you they don’t understand their target market.
For the China domestic market and for GB/T compliance required under certain procurement contracts, ask for the GB/T 35590-2017 and GB/T 36951-2018 test reports. These cover mobile power supply safety and performance respectively. GB/T 36951 is frequently cited in Chinese government procurement specs but is almost never included in export documentation unless the buyer specifically asks. Its absence from a supplier’s file doesn’t mean non-compliance — it usually means the supplier hasn’t been asked before.
Cost-Performance Trade-Offs in Compliance Architecture #
The cost of building a compliant USB-C PD 3.1 EPR product varies significantly depending on how you sequence your certification strategy.
A single-market approach (EU only, via IEC 62368-1 + EN 55032 EMC + RoHS) runs roughly $8,000–$14,000 in third-party lab fees for a new product, based on quotes from TÜV SÜD and SGS received in Q1 2025. Adding UL 62368-1 for North America brings that to approximately $18,000–$26,000 because US labs run additional dielectric withstand and leakage current sequences not required under EN. PSE for Japan adds $4,000–$7,000 more and requires a Japanese-language instruction manual as a hard requirement.
The counterargument to front-loading all certifications: if your launch market is EU and your volume in year one is under 5,000 units, the ROI on PSE certification before product-market fit is questionable. We’d prioritize IEC 62368-1 plus EMC first, then layer in UL and PSE once the design is frozen and volume justifies it. This holds for consumer-adjacent products. For B2B procurement — grid-tied portable BESS, medical-adjacent applications, or government contracts — start with the full stack regardless of initial volume, because a mid-project certification gap will cost more than the upfront investment.
The cost delta between a GaN charger design using a certified reference design (e.g., Navitas or GaN Systems application notes with pre-tested topologies) versus a fully custom design is measurable: typically 3–5 months less certification time and roughly $6,000–$9,000 less in re-test fees for EMI/EMC. Custom designs are not wrong — but the compliance cost needs to enter the NRE calculation from day one.
For a broader view of how cell-level compliance costs interact with pack-level certification, the Safety & Certification category covers incoming inspection thresholds and how to structure a compliance budget across cell, BMS, and pack layers.
Technical Deep-Dive: The Relationship Between USB PD Specification Layers and IEC/UL Standard Obligations #
This is where the compliance architecture gets genuinely complex, and where most design briefs I’ve reviewed underestimate scope.
The USB Power Delivery specification (currently PD 3.1, published by USB-IF) is a communication and power contract protocol. It is not a safety standard. It defines how devices negotiate voltage and current over the CC lines, how EPR is entered and exited, and what happens during hard reset. What it does not define is the thermal, electrical, and mechanical safety of the hardware implementing those negotiations.
Safety obligations for that hardware come from a stack of documents that reference each other in non-obvious ways:
USB-C PD Charger Safety Standard Interaction Map
| Standard | Scope | Relationship to PD Spec | Key Test Trigger |
|---|---|---|---|
| IEC 62368-1:2023 | Electrical/energy hazard for A/V and IT equipment | Mandatory safety framework; PD spec is silent on this | Fault-state voltage >60VDC triggers ES3 safeguard requirements |
| UL 62368-1 (3rd Ed.) | US equivalent with additional national differences | Adopted IEC 62368-1 structure but adds Annex DVP requirements | Dielectric withstand at 1.5× rated voltage + 1000V for 1 minute |
| IEC 61558-2-16 | Transformers in switch-mode power supplies | Referenced by IEC 62368-1 for isolation requirements | Creepage/clearance at operating altitude |
| EN 55032 / CISPR 32 | EMC emissions for multimedia equipment | Not referenced by PD spec; separately mandated by EU Radio Equipment Directive | Conducted emissions at 150kHz–30MHz; radiated at 30MHz–1GHz |
| GB/T 35590-2017 | China mobile power supply safety | Parallel standard; not harmonized with IEC 62368-1 | Drop test (1m, 6 faces), short circuit at 45°C ambient |
| UN 38.3 | Transport of lithium batteries/cells | Applies to battery cells inside the charger, not the charger itself | T.1–T.8 sequence; nail penetration test at T.6 |
The interaction between IEC 62368-1 and EN 55032 is frequently mishandled. Some suppliers present a single CE mark and assume EMC is covered under IEC 62368-1. It is not — IEC 62368-1 addresses electrical safety, not electromagnetic emissions. EN 55032 compliance requires a separate test campaign with its own pass thresholds (Class B: 30 dBµV/m at 10m for radiated emissions above 230MHz). A charger with a valid IEC 62368-1 certificate is still not CE-marked without EMC compliance.
The BMS Engineering category covers how the BMS firmware layer interacts with charger protocol compliance, specifically around CC/CV handoff and the role of the BMS in maintaining safe operating area during PD contract negotiation.
A specific data point from our lab work: in testing 4 EPR charger samples from Shenzhen-based manufacturers in late 2024 (240W, 48V/5A rated), all 4 passed IEC 62368-1 ES2 tests at 20V operating point. When re-tested at 48V EPR operating point per the same clause sequence, 2 of the 4 samples failed the touch current limit of 0.25mA (measured 0.31mA and 0.28mA respectively) due to insufficient Y-capacitor filtering in the isolation barrier design. Both failures were correctable with a component swap, but neither would have been caught by a standard 20V-only test protocol.
The open question I’m still tracking: IEC 62368-1:2023 was published in April 2023, and the transition period for EU (replacing the 2014 edition) closed in December 2024. Some national transpositions in Eastern Europe are running behind. Whether products certified under the 2014 edition retain full market access through 2025 under those jurisdictions is not yet uniformly enforced — this is worth confirming directly with your EU importer of record.
Sourcing Guidance for Buyers #
When evaluating Chinese suppliers in the USB-C PD fast charging space, the first document to request is the full IEC 62368-1 clause-level test report, not the summary certificate. A supplier who can produce this within 2 business days, with legible test conditions and actual measured values (not just “pass”), has invested in real compliance infrastructure. One who cannot is either using shared certificates across product variants or hasn’t been through a rigorous third-party review.
The qualification red flag specific to this category: any supplier who quotes EPR (240W, 48V) capability but whose compliance documentation only references SPR (100W, 20V) test conditions. EPR compliance is not a superset of SPR — it requires a separate test sequence, particularly for fault-state voltage and touch current. Accepting SPR-only test data for an EPR-capable product is a material compliance gap.
For incoming inspection, test a sample of 8 units per lot at 48V/5A output under single-fault condition (open one Y-capacitor) and measure touch current per IEC 62368-1 Annex B.3. Pass threshold is 0.25mA DC or 0.7mA peak AC. If you’re sourcing at volumes below 500 units per shipment, reduce sample to 5 units but apply a tighter accept threshold of zero failures. At higher volumes, use standard AQL 1.0 sampling — but never skip the fault-condition test entirely, because nominal-condition touch current tells you almost nothing about safety margin.
FAQ
Does CE marking alone cover USB-C PD charger compliance for sale in the EU?
CE marking confirms self-declared or third-party verified conformity to applicable EU directives, but it is not a single test. A USB-C PD charger sold in the EU needs to comply with the Low Voltage Directive (via IEC 62368-1), the EMC Directive (via EN 55032/CISPR 32), and the RoHS Directive as a minimum. A CE mark without all three underlying technical files is legally incomplete.
Is UN 38.3 required for a charger that has no internal battery?
No. UN 38.3 applies to lithium cells and batteries being transported, not to charger hardware that contains no electrochemical cells. If your charger is purely a power conversion device with no integrated battery, UN 38.3 is not in scope. The standard becomes relevant the moment you integrate a cell — even a small backup or buffering cell — into the charger assembly.
What is the difference between IEC 62368-1 and the older IEC 60950-1 for charger compliance?
IEC 62368-1 uses a hazard-based framework (classify the energy source, then apply the appropriate safeguard) versus IEC 60950-1’s prescriptive approach (specific rules for specific components). For USB-C PD chargers, the practical difference is that IEC 62368-1 more explicitly addresses higher-voltage operating points and fault-state scenarios, making it better suited to EPR designs. IEC 60950-1 was withdrawn as a compliance route for new product certifications in most markets as of 2023.
Can a Chinese manufacturer’s GB/T 35590 test report substitute for IEC 62368-1 in an EU submission?
No. GB/T 35590-2017 is a Chinese national standard and is not harmonized with EU directives. It covers partially overlapping test areas (short circuit, overcharge, drop) but uses different pass criteria and ambient conditions. An EU Notified Body will not accept it as equivalent. It depends on your market: for China domestic procurement or certain B2B contracts specifying GB/T compliance, it’s the right document. For EU CE marking, it’s supplementary at best.
How often does a USB-C PD charger need to be re-certified after a design change?
It depends on the change scope. Component substitutions that affect isolation barriers, EMI filtering, or overcurrent protection circuits typically require partial re-test under IEC 62368-1, and full re-test if the change crosses a rated voltage boundary (e.g., adding EPR support to an existing SPR-certified design). Cosmetic changes or firmware updates that don’t alter hardware protection behavior generally don’t trigger re-certification, but should be documented in the technical file. Our practice is to flag any BOM change to the charger IC, transformer, or Y-capacitor array as a mandatory re-evaluation trigger.
Published by compactbess.com Technical Team | Request a sourcing consultation
The ES2/ES3 boundary issue actually compounds in battery-backed UPS applications because the BMS overvoltage protection threshold has to account for PD negotiation transients — we had a LiFePO4 pack where the BMS was tripping OVP at 58.4V during EPR handshake on a 48V bus because the CC/CV transition pulse wasn’t filtered in firmware before the protection comparator sampled it. Tightened the OVP debounce window from 100µs to 800µs and the nuisance trips stopped without meaningfully delaying real fault response.
The “EPR not tested” gap is exactly what we flagged on a 140W GaN unit last spring — our pre-compliance review caught that the factory had run dielectric withstand at the 28V EPR entry point, not 48V, which would’ve failed the 1.5× + 1000V UL requirement by a meaningful margin.