Battery pack coating insulation resistance verification is the practice of judging whether the insulating coating applied to cell cases, aluminium trays and housings, busbar brackets and cold plates satisfies the pass thresholds set by vehicle and battery safety standards, together with the sealing, grounding and structural design. The object of the verification is not how high the resistivity of the coating is on its own, but how much dielectric margin the coating retains after edges, pinholes, moisture uptake and mechanical impact have consumed part of it.
TL;DR: five conclusions
- The acceptance criterion is Ω/V, not kV/mm. Clause 5.1.4.1 of GB 18384-2020 requires an insulation resistance of at least 100 Ω/V for d.c. circuits and 500 Ω/V for a.c. circuits at maximum operating voltage; where d.c. and a.c. circuits are conductively linked, 500 Ω/V applies. On an 800 V d.c. platform that converts to 80 kΩ (calculated in this article).
- The coating bulk is normally many orders of magnitude above that threshold (published data: dielectric strength ≥30 kV/mm, surface resistance >1013 Ω), but a pack is measured as a parallel network: one path in the 80 kΩ range is enough to fail the whole pack. Verification therefore shifts from material data to structure, process, ageing and mechanics.
- Coatings are now mechanically assessed safety components. GB 38031-2025 adds a bottom impact test with a 30 mm, 10 kg hemispherical striker at 150 J, three strikes, and requires no leakage, no housing rupture, no fire or explosion and continued compliance with the insulation resistance requirement.
- Design (type) withstand voltage and inline test voltage must be separated: 4,000–6,000 V d.c. for qualification at 120–250 μm, versus a lower inline figure — 2,700–5,000 V d.c. is commonly quoted for production testing — chosen to find defects without creating them.
- Holiday detection voltage is a window, not a maximum. The rule of thumb for thin coatings is roughly 100 V per 25 μm of dry film, set from the measured minimum film thickness, with repairs re-inspected until no breakdown occurs.
1. Key data (with units)
- Insulation thresholds: 100 Ω/V d.c., 500 Ω/V a.c., 500 Ω/V for conductively combined circuits (GB 18384-2020, clause 5.1.4.1).
- Platform conversion (calculated): 400 V d.c. → 40 kΩ; 800 V d.c. → 80 kΩ; at the a.c. figure of 500 Ω/V, 800 V → 400 kΩ.
- Coating data (published): insulating powders ≥30 kV/mm (IEC 60243-1), surface resistance >1013 Ω (IEC 60093), typical dry film 120–250 μm, withstand 4,000–6,000 V; UV-cured systems 25 kV/mm breakdown, about 50 μm per pass.
- Ageing conditions (published): 85 °C / 85 % RH for 1,000 h, 130 °C for 720 h, thermal shock −40 °C to 85 °C for 1,500 cycles (20 min each), 1,000 h electrolyte immersion.
- Edge coverage: 40–60 % on corners and beads, so nominal thickness must be derated there.
- Mechanical qualification: GB 38031-2025 bottom impact Ø30 mm, 10 kg, 150 J, 3 strikes. The revision draft of GB 18384 adds a whole-vehicle floor impact with a Ø150 mm solid steel hemisphere at about 35 km/h; afterwards no electrolyte leakage within 30 min, no fire or explosion within 2 h, and insulation resistance still per 5.1.4.1.
- Test scope: 7 cell tests and 17 pack/system tests in GB 38031-2025, including salt spray, damp heat cycling and immersion, plus a new fast-charge safety test (cells charging 20–80 % SOC in ≤15 min must survive 300 fast-charge cycles and an external short circuit; the industry reads this as roughly 120,000 km).
- Electrical boundaries: potential equalisation contact impedance ≤0.1 Ω and ≤0.2 Ω between two simultaneously touchable parts within 2.5 m; supplementary shields must withstand ≥10 kPa without visible plastic deformation (GB 18384-2020, clauses 5.1.4.3, 5.1.4.4).
- Ingress protection: IPXXD inside the passenger/cargo compartment, IPXXB outside; parts mounted below 300 mm above ground require IPX7, at or above 300 mm unshielded IPX5, shielded IPX4 (GB/T 4208-2017).
- Dates: GB 38031-2025 issued 28 March 2025, mandatory for new type approvals from 1 July 2026 and for already approved models from 1 July 2027, fully replacing GB 38031-2020.
2. Three levels of verification: material, coating, system
2.1 Material level — comparison and batch control only
Dielectric strength (kV/mm) and volume/surface resistivity (Ω·cm, Ω) measured on flat reference panels per IEC 60243-1 and IEC 60093 say nothing about corners, pores, residual stress or assembly damage. They rank resin systems and monitor batches; they are not a delivery criterion.
2.2 Coating and part level — what you actually ship
Withstand (hipot) and continuity testing on the real part: apply the specified d.c. voltage, hold it, keep leakage current below the limit, and find zero holidays. Published claims such as "4,000 V at 120–250 μm", "6,000 V", or "1,000 V with leakage below 0.1 mA after 10 min at 1,000 ± 200 °C" belong here — closer to engineering reality than panel data, but still not vehicle compliance.
2.3 System level — Ω/V plus four further demands
The measured value is the parallel result of every insulation interface in the pack (GB 18384-2020, clause 6.2.1), combined with potential equalisation ≤0.1 Ω, 10 kPa enclosure strength, IP protection and high-voltage marking. The coating must be designed against that drawing: mounting points, earth faces and thermal interfaces must conduct, everything else must insulate. Unclear keep-out boundaries are the most common origin of series disputes.
3. Film thickness and margin: an arithmetical check
Using the published 30 kV/mm (30 V/μm) as an ideal extrapolation, against an 800 V working voltage and the 100 V per 25 μm holiday test rule. Calculated values for order-of-magnitude discussion; acceptance still requires part-level testing.
| Dry film | Ideal breakdown at 30 kV/mm | Margin vs 800 V | Holiday test (100 V/25 μm) | At 60 % edge coverage |
|---|---|---|---|---|
| 100 μm | 3.0 kV | 3.8× | 0.4 kV | 60 μm → 1.8 kV |
| 150 μm | 4.5 kV | 5.6× | 0.6 kV | 90 μm → 2.7 kV |
| 250 μm | 7.5 kV | 9.4× | 1.0 kV | 150 μm → 4.5 kV |
| 350 μm | 10.5 kV | 13.1× | 1.4 kV | 210 μm → 6.3 kV |
At 800 V a nominal 100 μm coating keeps less than a four-fold ideal margin before corner derating, which is why insulating powders are usually specified at 120–250 μm. Thicker is not automatically better: internal stress, incomplete cure, longer thermal path and cost take over. Note that the holiday voltage is only about a tenth of the ideal breakdown value, so a correctly set test does not damage sound coating — the real risk is setting it from nominal rather than measured minimum thickness.
4. Four implementation routes and their failure boundaries
| Route | Film | Insulation and durability (published) | Processing | Parts | Failure boundary |
|---|---|---|---|---|---|
| Insulating powder (epoxy / polyester) | 120–250 μm | ≥30 kV/mm, >1013 Ω, UL 94 V-0, 1,000 h damp heat, 720 h at 130 °C | Baking required; 40–60 % edge coverage | Cell cases, busbars, brackets, housings, cold plates | Heat distortion of thin castings, thin edges, thickness scatter |
| UV-cured insulating coating | ≈50 μm per pass | 25 kV/mm, >6 kV, 1,000 h electrolyte resistance, 1.5 W/(m·K) filled grades | Seconds-long cure, 100 % solids, optional 80 °C / 30 s assist | CTP/CTB structures, busbars, CCS surroundings, local areas | Shadowed areas, deep-cure deficit in thick builds, shrinkage stress |
| Liquid epoxy / polyurethane | 60–150 μm | Wide formulation window, easy primer–topcoat pairing, VOC control needed | Multi-pass with levelling, slower takt | Housings, complex internal cavities, weld repair | Sagging, solvent retention, viscosity drift between batches |
| PET blue film (comparison) | Tens of μm plus adhesive | Stable initial insulation, about 1.2 CNY/m² reported | Window cut-outs, folding, manual or semi-automatic wrapping | Prismatic cell wrapping | Adhesive lift at folds and seams, long-term insulation at temperature |
The selection question is which route produces a continuous, inspectable and re-inspectable film on a given geometry — not which datasheet is higher. In the dry state all four are far above 80 kΩ; the differences appear after edges, seams, ageing and impact.
5. Re-test points: after ageing and after impact
5.1 Three mandatory re-checks
Damp heat (85 °C / 85 % RH, 1,000 h) to watch surface resistance and adhesion; thermal shock (−40 °C to 85 °C, 1,500 cycles) to expose coating-to-aluminium expansion mismatch; bottom impact (Ø30 mm, 10 kg, 150 J, 3 strikes) to prove insulation continuity after deformation. GB 38031-2025 makes damp heat, salt spray, immersion and bottom impact mandatory at pack level, so judgement is intrinsically a combined-stress result.
5.2 Report pairs, not single numbers
For each qualification, report initial and aged values together: withstand voltage (kV), cross-cut adhesion (GB/T 9286; 5B is quoted for UV systems), pull-off adhesion (GB/T 5210), dry film thickness (GB/T 13452.2) and holiday count. Retention is more diagnostic than absolute aged values. On the corrosion side, published specifications limit creep from a scribe to ≤2 mm.
6. Production testing: how much voltage is right
6.1 Separate type voltage from inline voltage
Type qualification demonstrates margin at 4,000–6,000 V. Inline testing exists to catch defects, not to reject sound parts, so it is set well below the coating's own breakdown — 2,700–5,000 V d.c. is commonly quoted — with defined dwell time, leakage limit and disposition rules in the drawing and the inspection plan.
6.2 Holiday detection and keep-out areas
Low-voltage wet sponge (ASTM D6979) suits thin films: safe, but insensitive to deeper defects; d.c. spark testing (ASTM D5162, NACE/AMPP SP0188 procedures) suits thicker films and finds minute pinholes but can puncture weak areas if over-volted. Set voltage from measured minimum thickness, overlap scan passes, mark, repair and re-scan until silent. Conductive filled (thermally or electrically conductive) coatings are not compatible with this principle. Pre-treatment decides the rest: the final rinse's chloride and conductivity must be controlled, because residual salts form hygroscopic paths under the film, and masking of earth faces and thermal interfaces must be fixed at tooling stage.
7. Validation matrix
| Item | Method or basis | Typical condition | Criterion |
|---|---|---|---|
| Material screening | IEC 60243-1 / IEC 60093 | Reference panel | ≥30 kV/mm, >1013 Ω; selection and batch control only |
| Part withstand | Customer DV spec | 4,000–6,000 V d.c. | No breakdown, leakage within limit |
| Coating continuity | Wet sponge / d.c. spark | ≈100 V per 25 μm of measured minimum film | Zero holidays; re-test after repair |
| Pack insulation resistance | GB 18384-2020, 5.1.4.1 / 6.2.1 | Measured at maximum operating voltage | ≥100 Ω/V d.c. (800 V → 80 kΩ) |
| Damp heat | 85 °C / 85 % RH | 1,000 h | Withstand and adhesion still compliant, reported as pairs |
| Thermal shock | −40 °C to 85 °C | 1,500 cycles, 20 min each | No cracking or detachment, insulation maintained |
| Bottom impact | GB 38031-2025 | Ø30 mm, 10 kg, 150 J, 3 strikes | No leak, no rupture, no fire or explosion, insulation resistance met |
| Environment set | GB 38031-2025 (17 tests) | Salt spray, damp heat cycling, immersion | Judged per standard with insulation re-measured |
| Flammability | UL 94 | V-0 vertical burn | Material level; not a substitute for thermal propagation testing |
| Grounding | GB 18384-2020, 5.1.4.3 | Test current ≥0.2 A | ≤0.1 Ω to platform, ≤0.2 Ω within 2.5 m |
8. Six frequent mistakes
- Treating panel dielectric strength as a part-level withstand result.
- Optimising insulation while ignoring the opposite demands: 0.1 Ω bonding and 0.2 Ω equalisation. A badly masked earth face is more dangerous than a thin coating.
- Setting holiday voltage from nominal instead of measured minimum thickness, or applying spark testing to conductive filled coatings.
- Reading UL 94 V-0 as protection against thermal runaway; thermal propagation and external fire are separate GB 38031-2025 assessments.
- Testing only new parts; retention after damp heat, thermal shock, electrolyte and impact is the life criterion.
- Mixing type-test voltage with inline voltage in one acceptance rule.
Frequently asked questions
Q1: If the coating is orders of magnitude above 80 kΩ, why control thickness?
Because the pack measurement is a parallel network and the coating on a real part is not a uniform body: corners run at 40–60 % coverage, pinholes expose substrate, and absorbed moisture creates surface paths. Thickness control buys calculable margin at those weak points, not bulk resistivity.
Q2: Which standard governs the pass decision?
Start with the vehicle and battery standards that define the threshold (GB 18384-2020 clause 5.1.4.1 for Ω/V; GB 38031-2025 for mechanical and environmental tests), then use coatings and test-method standards (IEC 60243-1, IEC 60093, GB/T 9286, GB/T 5210, GB/T 13452.2, GB/T 1771) as means, and close with the customer's DV/PV specification.
Q3: How much degradation after damp heat is acceptable?
The national standard fixes the vehicle-level Ω/V threshold, not a coating retention figure; retention limits sit in OEM and cell-maker specifications. Insulation system evaluations in the UL 1446 family commonly require aged properties to stay above a stated share of the initial value, so the practical floor is "still passes part-level withstand with zero holidays".
Q4: Is a higher holiday voltage safer?
No. At roughly 100 V per 25 μm the test sits near a tenth of the ideal breakdown value, in a non-destructive sensitivity band; excess voltage punctures weak areas and leaves defects inside the film. Measure minimum thickness first, then set voltage, scan with overlap, mark, repair and re-test.
Q5: Can thermal conductivity and insulation coexist?
With insulating filler systems. A published UV case combines graphene nanoplatelets with boron nitride to reach 1.5 W/(m·K) while keeping 25 kV/mm breakdown. Carbon-rich fillers lower volume resistivity sharply, so re-verify insulation and abandon spark testing for conductive formulations.
Q6: We already wrap cells in PET film — why evaluate coatings?
Film needs cut-outs and folded edges, its pressure-sensitive adhesive can lift at temperature, and yield falls on large cells. Coatings give a continuous film compatible with CTP/CTB designs and automated application, provided edge coverage, cure depth and inspectability are demonstrated. They are comparable engineering options, not slogans.
Sources
- GB 38031-2025 Electric vehicles traction battery safety requirements, issued 28 March 2025, effective 1 July 2026 for new type approvals (1 July 2027 for approved models), replacing GB 38031-2020; 7 cell tests and 17 pack or system tests.
- GB 18384-2020 Electric vehicles safety requirements, clauses 5.1.3–5.1.5, 6.2 and Annex A; revision draft adding whole-vehicle floor impact.
- GB/T 4208-2017 (IP protection degrees); GB/T 19596 (electric vehicles terminology).
- IEC 60243-1 (electric breakdown strength), IEC 60093 (volume and surface resistivity), UL 94, UL 1446 with IEC 60216.
- GB/T 9286, GB/T 5210, GB/T 13452.2, GB/T 1771, GB/T 9274 (adhesion, thickness, salt spray, chemical resistance).
- NACE/AMPP SP0188, ASTM D5162, ASTM D6979 (holiday detection practice); MIIT explainer on GB 38031-2025; CIBF2026 supplier releases and insulating powder datasheets (all vendor-quoted figures marked as published data).
Last updated: 2026-10-21 ｜ Kexin New Materials (Guangdong) Co., Ltd. technical team. Standards are cited as published; figures marked "published" come from supplier technical releases and those marked "calculated" are worked examples. Validate at part and pack level against your own geometry, film thickness, platform voltage and customer DV/PV specification before series introduction.