Injection-Molded EV Battery Pack Housings: Sealing, Flame Rating, Structural Strength and HV Insulation

A battery pack housing can pass every dimension on its own drawing and still fail at the vehicle: the lid seals against a tray it was never measured with, the gasket loses compression where a long rim warps, the flame rating on a resin data sheet says nothing about the assembled enclosure, and the creepage path that looked adequate on a flat sketch disappears where a rib, a boss, or an insert interrupts the surface. Each failure is discovered late, when the pack and the vehicle mounting are already fixed.

Define the molded housing RFQ from the complete pack interface chain — tray and lid, cell carriers, busbars, terminal covers, cable supports, vent valves, seals, fasteners, and vehicle mounting — and close sealing, flame rating, structural strength, and high-voltage insulation as four separate decisions, each with its own input, evidence, and owner, before any mold steel is cut. A housing that meets its own drawing is not approval of the battery pack, and no single molded part carries an ingress rating, a flame approval, or a crash rating on its own.

The rest of this article works through that chain in the order a buyer should close it: housing scope and pack interfaces, sealing, flame rating, structural strength, high-voltage insulation, and validation and compliance evidence.

What exactly is in the molded battery-pack housing scope, and what sits next to it?

A battery-pack housing RFQ fails most often before any molding discussion begins, because the words tray, lid, cover, enclosure, and housing mean different scopes to different suppliers. One supplier quotes a lid only, another a tray with inserts, and a third assumes the seals, the fasteners, and the assembly are someone else’s problem, and the buyer discovers the difference only when the first samples arrive without the parts that make the enclosure work.

Name every molded component, purchased insert, seal, fastener, vent interface, and mounting feature in the RFQ, assign responsibility for molding, procurement, assembly, testing, and final pack validation, and define the tray-to-lid interface before any mold steel is discussed.

Start with a controlled assembly model showing the tray, the lid, the internal components, the seal path, and the vehicle mounting points: the housing’s datums, seal lands, and fastener pattern are set by what it has to contain and attach to. The cell carriers, the vent valves, and the high-voltage cable routing inside the pack are neighbouring scopes with their own decision content; this article names them as interfaces only. The EV battery terminal covers and cable supports inside the pack are live-part protection governed by insulation coordination; their RFQ requirements belong to that article. The plastic envelope around insert-molded busbars is a separate decision, covered in the insert-molded busbar article, and the low-voltage fuse and relay distribution inside the vehicle is the scope of the automotive fuse and relay box housing article. The charging-port housing at the vehicle exterior is another separate assembly scope, covered in the EV charging port housings and flaps article, and none of those scopes should be absorbed into a pack-housing quotation by default. Define the assembly sequence in both directions, including lid seating, fastener torque order, and service removal, and identify where an operator could misorient the lid or pinch a seal. The pack itself is not the part: the molded scope ends at the tray and lid, and pack-level safety approval, cell chemistry, battery management, and thermal-runaway containment belong to the vehicle customer and the pack integrator.

How should sealing be specified when no single part carries an ingress rating?

Sealing is where a pack-housing program most often confuses a part property with a system result: a buyer asks for a housing with a stated ingress code, and both sides discover at vehicle test that the code was never theirs to claim.

Do not assign an ingress rating to the molded housing alone; define the enclosure boundary, the seal system, the mated condition, and the test configuration, and require the assembled enclosure to be tested as specified.

Ingress protection belongs to a defined enclosure and test configuration, not to a tray, a lid, or a gasket in isolation. ISO 20653:2023 specifies IP-code designations and tests for enclosures of road-vehicle electrical equipment; the buyer should state the enclosure boundary, the mated or unmated condition, the seals, vents, drain state, mounting, preconditioning, and test responsibility before any supplier is asked to confirm a rating. The assembled enclosure must be tested as specified by the program, with the test configuration written down. The seal system needs the same discipline: the gasket groove, the gland where a cable or connector passes through the wall, the compression along the seal’s full length, the continuity at corners and joints, and the fastener spacing and tolerance analysis that hold it. Warpage of a long rim or a large lid can reduce local compression even when point dimensions pass, so the seal should be checked as a compressed length, with the measuring condition named on the drawing. Cable-exit sealing is its own interface — see the automotive sensor housing cable-exit sealing article — and the same separation applies here: the housing defines the gland and seal land, and the cable, connector, and exit seal are reviewed as a joint. Leak testing has its own fixtures, conditions, and false-failure modes — see the leak testing of molded automotive housings article — and the buyer should name the method, condition, acceptance limit, and test owner in the RFQ rather than assuming a dimensional report proves a seal. Separate the water and dust requirement from the vent requirement: a sealed enclosure still needs a defined pressure path, and vent valve seating belongs to the vent valve scope.

What does a flame rating actually prove about a molded housing?

A flame rating is the most misread line in a pack-housing RFQ: a resin data sheet carries a flammability classification, the buyer copies it into the specification, and both sides proceed as if the housing, the pack, and the vehicle now share that rating, when it described none of them.

Treat a flame rating as a material-level property published for a stated grade, color, and thickness, require the grade-specific listing at the design thickness, and do not transfer the rating to the housing, the assembled pack, or the vehicle.

UL 94 ratings are small-scale laboratory results for a stated grade, color, and thickness, and acceptability is determined in the end product — a rating cannot by itself approve the housing, the joint, or the vehicle. That matters more here than for most parts: the housing is large, its wall thickness varies across ribs, bosses, and seal lands, and the rating at one thickness may not apply at another. The buyer should therefore name the exact grade, color, reinforcement, and thickness at which the rating was established, and require the listing or test evidence at that thickness — not a polymer family name or a rating quoted without its conditions. The duty-first method in the PA6 versus PA66 under-hood article applies here even though under-hood duty is not pack-housing duty: define the duty as a time-dependent envelope of temperature, cycles, peaks, and abnormal events, each with load, fluid, and moisture state, before comparing grades, and choose only at exact-grade level. No resin family or grade is presented here as the pack-housing material, because none has been verified; the customer requirement and the grade-specific listing govern. Thermal-runaway containment is a pack-level system claim, not a housing claim: a housing meeting its drawing does not prove the pack contains a cell event, and the buyer should not ask the housing supplier to certify containment. Mold design should protect the seal land and structural features first: review gate location, weld lines at insert roots and seal lands, sink near bosses, and cooling around long rims. A weld line downstream of a hole or between two gates is governed first by geometry, as Protolabs explains in its guidance on knit lines, so gate placement should move weld lines away from highly stressed or sealing features.

How are structural strength and crash load paths handled at the RFQ stage?

Structural strength is the pillar where a buyer is most tempted to ask the molder for a number the molder cannot supply. The housing carries load paths between the vehicle mounting, the tray, the lid, and the modules inside it, and it sees vibration, impact, and crash events the vehicle customer defines — but no crash load, vibration level, or acceptance limit for a pack housing has been verified as a general value, and inventing one would put a false number on a safety-relevant drawing.

Treat structural strength as a concept-level load-path decision at RFQ stage: name the mounting points, the load paths, the ribs and inserts that carry them, and the customer-defined test plan, and do not invent test loads, cycle counts, or acceptance limits.

Start with the load path, not with a wall thickness. Identify where the vehicle mounts the pack, how load travels from the mounting into the tray, across its floor and walls, into the lid and its fasteners, and around the modules and the cooling interface, and name which features are structural and which are cosmetic: a rib that looks structural may carry nothing if the load path bypasses it, and a boss that looks minor may be the only path between lid and tray. Rib design, insert placement, and fastener spacing should follow that path, and the buyer should supply the customer-defined test plan — events, sequence, conditioning, sample quantity, and acceptance criteria — rather than asking the supplier to guess them. A large-format housing adds a molding dimension: a long tray or wide lid warps from differential wall thickness, asymmetric cooling, or fiber orientation, and warpage harmless on a small bracket can break a seal land or shift a mounting point on a part this size, so the datum strategy, measuring condition, and functional gauges should be defined up front, with the tray measured in its assembled or restrained state, not only free-state on a bench. Press and mold capacity should be confirmed as a scope question: which press, mold size, and cavity plan does the supplier propose? Record the customer test plan as an open input with a decision date if it is not yet supplied.

What must the buyer supply before creepage and clearance can be checked?

The housing separates live parts from each other and from the vehicle chassis, and the distances that make that separation safe are computed from inputs only the buyer and the vehicle customer can supply.

Supply the insulation inputs — working voltage, insulation type, pollution degree, material group, rated impulse voltage, and the governing standard and edition — before any creepage or clearance distance is checked, and check the two distances separately.

ISO 6469-3:2021 specifies electrical safety requirements for voltage class B electric circuits of electrically propelled road vehicles; voltage class B covers direct-current circuits above 60 V and up to 1,500 V, so a traction pack housing sits squarely inside its scope. The standard requires protective measures providing both basic and fault protection, including double insulation, reinforced insulation, and protective barriers or enclosures, with or without equipotential bonding. Which measure applies is a customer decision, and the buyer should name the measure, the governing standard and edition, and the test responsibility in the RFQ. The insulation-coordination method comes from IEC 60664-1, a paywalled standard, so this article states the method and its inputs rather than reproducing it: creepage is the shortest path along an insulating surface, driven by working voltage, pollution degree, and material group, while clearance is the shortest path through air, driven by rated impulse voltage. The two must be checked separately because they fail by different mechanisms — a generous air gap does not cover a short surface path, and creepage must never be smaller than the required clearance. The only verified creepage figures available here are a generic method illustration from a reproduction of IEC 60664-1 Table F.5: for basic insulation, pollution degree 2, material group III, the table gives 2.0 mm at 400 V, doubling to 4.0 mm at 800 V — so a housing designed for a 400 V pack cannot be carried over to an 800 V pack without re-checking — and reinforced insulation doubles the basic-insulation distance. Those figures are a method illustration, not pack requirements, and they must never be used as design values for a real housing. The buyer must supply the real inputs: the working voltage of the actual pack, the pollution degree inside the enclosure, the material group of the chosen grade from its comparative tracking index, the insulation type, and the rated impulse voltage. The EV battery terminal covers and cable supports article covers the same method for the covers and supports inside the pack; this article does not repeat its worked method. Do not assign a creepage or clearance distance on the basis of a voltage class alone, and do not accept a material data sheet as proof of insulation: the data sheet describes the resin, and the distances describe the geometry, the environment, and the voltage together.

What validation and compliance evidence should exist before production release?

Validation is where a pack-housing program either earns its release or discovers, too late, that the evidence it collected proves the wrong thing: a dimensional report is not a sealing test, a resin data sheet is not a flame test, a material flame rating is not a pack test, and a bench fit check is not a conditioned vehicle test.

Ask for layered evidence — resin data, molded-part conformance, assembled-enclosure testing, and vehicle validation — with sample quantity, fixture, test sequence, conditioning, acceptance limit, and approval owner named for each layer, and do not accept one layer as proof of another.

Molded-part evidence should cover resin identity, process traceability, dimensions in the defined condition, appearance against approved samples, insert retention, assembly fit, and agreed conditioning, with cavity identity recorded. Assembled-enclosure evidence should cover seal compression and the customer-defined ingress test on the defined boundary, lid-to-tray fit through tolerance extremes and conditioning, fastener torque sequence and retention, and drain and vent behavior in the installed orientation. Durability evidence should follow the customer-defined cycle, load, temperature, humidity, and impact sequence, written as a test plan with sample allocation. Material compliance evidence belongs in the same plan: the housing’s resin, colorant, inserts, and labels need a controlled IMDS material tree, and the IMDS material declarations article explains the submission object, the four-gate review, and the timing to plan for — a rejected submission can hold part approval. PPAP or equivalent submission evidence should be scoped at project launch — governing requirements, submission reason and level, applicable elements, production-intent run, sample and laboratory rules, responsibilities, timing, and approval status — and the PPAP scope and responsibilities article sets out that scope for molded parts; a submitted package is not an approved part. The tolerance and datum strategy deserves its own line in the plan, because a tray measured free-state may not represent the installed seal land; see the automotive sensor housing tolerances article. AutoMoldingPro is operated by C.K. MOLD & ENGINEERING LIMITED, an IATF 16949 certified and ISO 9001 Certified manufacturer, and that certification supports a controlled molding and quality process; it does not by itself approve a specific battery-pack housing design, material, or vehicle, and every program still needs its own controlled drawing, material basis, and validation plan. Assign sample quantities, fixtures, test sequence, and acceptance authority before testing begins, and record any unknown requirement as an open input with a decision date.

Battery-pack housing buyer review

Review area RFQ input Evidence before release
Housing scope Tray and lid scope, inserts, seals, fasteners, assembly responsibility Controlled assembly model and scope statement
Pack interfaces Cell carriers, busbars, terminal covers, cable supports, vent valves, cooling interface Interface drawings and responsibility matrix
Sealing Enclosure boundary, gasket and gland concept, compression, test configuration Assembled enclosure test to the customer requirement
Flame rating Exact grade, color, reinforcement, thickness, customer flame requirement Grade-specific listing at the design thickness
Structural strength Mounting points, load paths, ribs, inserts, customer test plan Customer-defined test plan with acceptance criteria
HV insulation Working voltage, insulation type, pollution degree, material group, impulse voltage Creepage and clearance check against the governing standard
Material Duty envelope, exact grade, change control, traceability Technical data, listing, incoming controls
Tolerances and datums Datum strategy, measuring condition, functional gauges Dimensional stack and installed fit check
Large-format molding Press and mold size, cavity plan, gate and weld-line placement Capacity and cavity-specific control plan
IMDS Resin, colorant, inserts, labels, recipient and customer rules Accepted IMDS submission for the part and revision
PPAP Submission level, elements, production-intent run, approval authority Approved PPAP package for the molded scope
Validation Layered test plan, sample quantity, fixtures, conditioning Test reports per layer with named approval owner

Use the table to separate the evidence layers and name the owner of each decision: molded-part supplier, busbar or terminal-cover supplier, cell-carrier supplier, pack integrator, body engineering, validation laboratory, or vehicle customer. Link every RFQ input to a drawing characteristic, test method, sample condition, and acceptance authority, so that a dimensional report is never mistaken for proof of the complete enclosure’s sealing, flame performance, structural strength, or high-voltage insulation.

Illustrative scenario: a buyer supplies only a lid CAD model and asks for a housing quote against one ingress code and one flame rating. The molder cannot determine the tray, the enclosure boundary, the seals, the fastener pattern, the internal components, the vent valve seating, the cooling interface, the vehicle mounting, the load paths, or the test configuration from that package, and a useful response is not a guessed certification price. The buyer should provide the assembly model, the interface drawings, the sealing concept and test configuration, the material basis with its flame listing, the insulation inputs for the actual pack voltage, the load paths and customer test plan, and the responsibility matrix, and the supplier can then quote the molded tray and lid and the agreed evidence without claiming approval of the complete pack. This scenario is illustrative; it is not a customer case. The numbers in this article are not design values: the only figures given are the IEC 60664-1 Table F.5 method illustration, and every value in a real RFQ must be replaced with the governing customer requirement and the current edition of the applicable standard.

Questions to ask the supplier

  • What is included: tray, lid, inserts, seals, fasteners, assembly, or molded parts only?
  • Which pack interfaces govern the housing: cell carriers, busbars, terminal covers, cable supports, vent valves, cooling interface?
  • How is the enclosure boundary defined — which seal lands, compression, glands, vents, and drains define it?
  • Which exact grade, color, reinforcement, and thickness were proposed, and what listing supports the flame rating at that thickness?
  • What are the mounting points and load paths, and which customer test plan proves the structural scope?
  • What are the insulation inputs — working voltage, insulation type, pollution degree, material group, rated impulse voltage — and who checks creepage and clearance?
  • Where are the gates and weld lines, and how were they placed relative to seal lands, insert roots, and ribs?
  • How are dimensions checked on a large housing: free-state or assembled, conditioned or as-molded, and on which functional gauge?
  • What is the control plan for cavity identity, critical characteristics, seal-land dimensions, and change notification?
  • What IMDS submission and PPAP scope, sample quantity, and approval authority apply to this program?

Information to submit for a project-specific review

Send the STEP assembly model and 2D critical features, the tray-to-lid interface and seal concept, the pack interfaces and their drawings, the material basis with its flame listing, the enclosure boundary and ingress test configuration, the insulation inputs for the actual pack voltage, the load paths and customer test plan, the appearance method and reference samples, annual demand and volumes, and the validation responsibility matrix. Mark any unresolved point as an open input with a decision date. AutoMoldingPro can review molded battery-pack housing parts within its EV electrical components application scope and its automotive plastic injection molding production route; send the controlled package through the request-a-quote page or by email to jerry@ckmold.com.

Conclusion

Close the housing scope, enclosure boundary, flame evidence, load paths, insulation inputs, and validation plan before tooling, and the tray and lid become a defined molded scope instead of a late pack problem. Send the controlled package for a project-specific review when the assembly model and requirements are ready.

References

Facebook
X
LinkedIn

Wait, We Have Something Special for You!

Join our mailing list and receive a 10% discount on your next mold or CNC project.

Request a Quick Quote

Send your drawings and detailed requirements via:
Email: jerry@ckmold.com

Or fill out the contact form below:

We will contact you within one working day. Please pay attention to the email with the suffix “@ckmold.com”