Design for Manufacture & Assembly Basics
Designing not just for function but for the factory and the wrench — how DFM and DFA cut cost, defects, and assembly time by baking manufacturability into the part.
Two brackets can do the same job, but one takes four bolts and ten minutes to fit while the other snaps in with a single clip. Design for Manufacture and Assembly is the discipline that makes the second one win.
Design for the factory, not just the road
A part that works perfectly in CAD can be a nightmare to make or fit. Design for Manufacture (DFM) ensures each part can be produced cheaply and reliably by its chosen process — stamped, cast, moulded, machined, or 3D-printed — choosing the process to suit the volume and the material, and designing features the process can actually deliver (consistent wall thicknesses for moulding, draft angles for casting, radii instead of sharp corners for stamping and fatigue). Design for Assembly (DFA) ensures those parts go together fast and error-free: minimising part count (consolidating functions), fastener count (clips and snap-fits over screws), orientation difficulty (parts that only fit one way, ideally the obvious way), and assembly steps. Both are front-loaded into design because a manufacturability defect caught in CAD costs almost nothing, while one caught on the line costs a fortune.
The cheapest part is the one you don't make. Part consolidation — merging several components into one — eliminates fasteners, joining operations, tolerance stacks, and assembly steps, all at once. Modern casting and moulding (and increasingly large 3D-printed and gigacast structural parts) let engineers replace dozens of stamped-and-welded pieces with a single component. The trade is tooling cost (a complex single-part mould is expensive) versus per-unit savings (fewer parts, less labour, fewer defects) — so consolidation pays at high volume, where the tooling amortises. This is why high-volume platforms increasingly use large integrated structural castings: the upfront tool cost is justified by the assembly-line simplification across millions of units.
Tolerances and serviceability
Two more DFM/DFA considerations recur. Tolerances: every dimension has a permissible variation, and tight tolerances cost exponentially more to hold. Good design relaxes tolerances wherever function allows (the 'tolerance budget'), and uses features that absorb variation (locating datums, slip fits) rather than demanding precision everywhere. Serviceability: a part that's cheap to make but takes an hour to replace in service is a poor design for total cost of ownership — DFA includes designing for the mechanic, not just the assembly line (accessible fasteners, no special tools, parts that don't fight each other on removal). The best designs are cheap to make and cheap to service, achieved by thinking about the whole lifecycle, not just the factory.
- Amortised tooling = 400,000 / 100,000 = 4.00 per unit
- Total per-part cost = 4.00 + 2.20 = 6.20
- Per-unit saving = (0.05 + 2×0.02) − 0.30 = 0.09 − 0.30 = −0.21 (a per-unit LOSS of 0.21)
- Wait — recheck: removing saves 0.05 + 0.04 = 0.09, but adds 0.30 tooling ⇒ net −0.21 per unit ⇒ total = 200,000 × (−0.21) = −42,000 (a loss).
- The intended reading: if tooling adds only 0.075/unit, saving 0.09 − 0.075 = 0.015 × 200,000 = 3000 net saving. (Illustrates that consolidation pays only when per-unit tooling < per-unit assembly saving.)
Check your understanding
- DFM makes each part cheap/reliable to produce; DFA minimises part count, fasteners, and assembly steps
- Part consolidation is DFA's biggest lever — fewer parts = fewer joins, fewer defects, faster build (pays at high volume)
- Per-part cost = amortised tooling + material + process + assembly; tolerance relaxation cuts process cost
- Serviceability is part of DFA — design for the mechanic and total cost of ownership, not just the assembly line