Interiors & Exteriors | Design to Proto Launch | Global EV OEM
Program | Variable Platform Architecture (VPA) — Interiors & Exteriors |
Components | Cockpit, door trim, pillar trim, defrost ducts, mid header display, console air vents, liftgate applique, door claddings, sill valance, rocker panel |
Team | 2 → 4 engineers (scope expanded mid-program) |
Tools | 3DEXPERIENCE (CATIA), JIRA, Innovia |
Context
A global EV OEM and Tier-1 supplier needed CAD execution across a wide interior and exterior component scope on a Variable Platform Architecture program from design through prototype launch. Platform-sharing architecture leaves little room for late-stage design conflict as every component carries simultaneous feasibility, tooling, cost, and timeline constraints.
Challenge
- Multiple components in the initial CAD data were not feasible for injection moulding or mass production — insufficient draft angles, parting line conflicts on Class-A surfaces, and wall thickness issues flagging warpage risk
- Clip tower geometry on pillar trim lacked undercut relief, which was functional in CAD, but not mouldable without a secondary side action that hadn't been tooled in
- If these conflicts carried forward into supplier PPAP and tooling commitment, cost of change would have jumped by an order of magnitude
- The proto launch schedule had already absorbed one delay, design freeze was a hard stop, with no cycle time left for revision loops
Approach
Goken treated manufacturability as a design input from the first CAD iteration, not a downstream review step.
- Manufacturability-first CAD: Draft angle compliance (minimum 1.5° on Class-A, 3° on structural walls), parting surface strategy, wall thickness against resin flow thresholds, and weld line positioning were all evaluated as design constraints at every geometry stage
- Tolerance-driven assembly design: Clip tower spacing, boss diameter, and locating strategy were designed to tolerance stack-up requirements rather than nominal-only geometry, with gap & flush targets tracked across the full door system assembly
- Root-cause proto resolution: When a clip engagement issue surfaced during the door trim proto build, Goken traced it to a tolerance stack across three mating parts that hadn't been assessed in the original locating strategy
- Structured cross-functional execution: Issues tracked through JIRA with owner and target closure date; multiple design alternatives (e.g., two parting surface strategies for the liftgate applique) presented with tooling cost and lead time trade-offs at each decision point
Results
Area | Before Goken | After Goken |
Part Feasibility | Multiple components with draft/tooling conflicts | All components released to full manufacturing feasibility |
Proto Quality | Risk of non-representative proto parts | Production-intent parts; improved assembly fit & finish |
Part Cost | Elevated due to part count & complexity | Reduced via geometric integration & assembly simplification |
CAD Release Quality | Repetitive rework cycles, feasibility gaps at review | Fewer supplier-stage revision loops |
Engineering Governance | Reactive issue resolution | Structured issue ownership and cross-functional closure |
Schedule | Delayed proto launch risk | Proto delivery met within hard timeline |
Why Goken
- Goken builds tooling feasibility into the CAD model itself — draft angle, parting strategy, and wall thickness are design constraints from iteration one, not a checklist applied later
- When problems do surface at proto, root cause comes first, the clip engagement issue was traced to a three-part tolerance stack, not patched at the symptom level
- A 2-engineer team scaled to 4 mid-programs without losing structure, issue tracking, release prioritization, and supplier coordination held up under expanded scope
- The driver of success was Goken's standard delivery framework across OEM, Tier-1, and EV platform programs: no feasibility problem survives into proto, regardless of team size
If your platform program is carrying manufacturability risk into proto, or your design freeze has no room left for revision loops, talk to Goken Engineering team