The DFM Checklist: 20 Decisions to Make Before Tooling
A mold problem caught after the first trial costs $5K–$50K and weeks of delay. The 20 design-for-manufacturability decisions that catch it for free instead.
By Axon Labs Engineering
If a mold problem only surfaces after the first injection trial, it typically costs $5,000 to $50,000 and pushes the schedule back weeks — a bill a 30-minute design review could have avoided entirely. The same math holds on the board: 2026 data puts the average PCB respin at $28,500 and 16 days, and a single wrong footprint is enough to trigger one. Design for manufacturability isn't a phase. It's twenty specific decisions, and this is the checklist we run before any tool gets cut.
Key takeaways
A DFM issue caught after the first mold trial costs $5,000–$50,000 and weeks of delay; the same check at design review costs a 30-minute meeting.
In 2026, PCB projects average 2.9 respins at $28,500 and 16 days each — and incorrect footprints alone are a leading single cause.
Parting-line placement, draft angle and wall-thickness ratios must be locked before tooling starts — changing any of them after steel is cut is one of the most expensive mistakes in hardware.
Why does DFM matter more than the design itself?
DFM review checks whether a part can be filled, packed, cooled, ejected, inspected, assembled and repeated across production batches without scrap or rework — usually covering wall thickness, draft angle, ribs, bosses, gate location, tolerance stack-up and final assembly fit. A part that looks perfect in CAD and can't be reliably molded isn't a design; it's a liability with a rendering.
In 2026, design for manufacturing is a competitive differentiator: teams that apply DFM rules early see better yields, lower costs and more predictable schedules than teams that treat it as a final check. This is the same "cheapest DFM review is the one held before tooling" principle from our hardware development guide — this article is the actual checklist.
What does a missed DFM issue actually cost?
Illustrative cost comparison: a design-review DFM check versus a defect discovered at first mold trial. Source: RJC Mold, 2026.
The gap isn't subtle. A problem flagged in a design review costs a meeting. The same problem found at first-shot mold trial costs $5,000–$50,000 in tool rework and pushes the whole production schedule back by weeks — squarely inside the tooling lead-time critical path we cover separately. On the electronics side the number is just as concrete:
Industry averages for PCB respin frequency, cost and schedule impact in 2026.
The enclosure and molding checklist (10 decisions)
Draft angle, everywhere. 1–2° per side is the working rule of thumb, with a 2° minimum on most surfaces; light texture needs 3°, heavy texture 5° or more (Protolabs). A wall with zero draft either won't eject or will drag and scar on the way out.
Wall thickness, held constant. Vary it and you invite sink marks and warping as the thick sections cool slower than the thin ones.
Rib thickness at 50–60% of nominal wall. Full-thickness ribs telegraph as visible sink marks on the opposite face.
Rib and boss draft, 0.5–1.0° per side — smaller than the main wall draft, but still non-zero.
Parting line agreed at design review, not after tooling starts. Moving it once a mold exists is one of the most expensive changes in the entire program.
Gate location chosen for flow, not convenience. A poorly placed gate creates weld lines exactly where cosmetic surfaces or structural loads can't tolerate them.
Undercuts minimized or engineered with a clear release mechanism. Every undercut is a slide, a lifter, or a redesign — decide which, on purpose.
Snap-fits sized for real material fatigue, not just first-assembly feel. A snap that works once in the demo and cracks on the tenth field disassembly is a returns problem, not a tooling one.
PCBA height and connector alignment checked against the enclosure, not assumed. Battery installation, LED visibility, button feel and cable bend radius all live at this boundary and are cheap to fix in CAD, expensive to fix in steel.
Screw posts and sealing surfaces sized for the actual fastener and gasket, not a placeholder. Heat paths belong on this list too — thermal problems discovered after tooling means redesigning around a mold that already exists.
The PCB and assembly checklist (10 decisions)
Every footprint verified against the actual component datasheet. Incorrect footprints are among the single most common causes of prototype failure — one wrong footprint can force a full respin.
Design rules matched to your fabricator's real capabilities, not generic defaults — trace width, spacing and via specs that your board house can actually produce at yield, referenced against standards like IPC-2221C and IPC-2152.
DFM checks run iteratively through the design stages, not as a single gate at the end — catching an error at 60% layout is cheaper than catching it at 100%.
Test points placed for factory test access, not just debug convenience — the fixture that validates every unit off the line depends on them.
Thermal relief and copper pour reviewed for both manufacturability and function — a pour that's hard to solder is a yield problem hiding as a design choice.
Component placement checked for pick-and-place and reflow reality — tall components near short ones can create shadowing and uneven reflow.
Panelization planned before layout is finalized, not bolted on afterward — break-away tabs and fiducials belong in the original design.
Silkscreen and polarity markings verified legible at production scale — a mark that's clear at 400% zoom in CAD can vanish at 1:1 on the actual board.
Connector and mounting-hole tolerances stack-checked against the enclosure — the PCB and the case are one tolerance chain, not two independent drawings.
A named DFM reviewer signs off before Gerbers or tooling release — accountability, not just a checklist, is what actually catches the error before it becomes a $28,500 respin.
When should DFM review actually happen?
Not once, and not at the end. Run it at three points: a lightweight pass during concept freeze, a full pass at design review before any tooling PO is released, and a final verification pass against as-built CAD immediately before Gerbers or mold data ship. Skipping the middle pass is the single most common shortcut we see — and the one that shows up as a $28,500 line item four to six weeks later.
Review point
What it catches
Cost if skipped
Concept freeze
Fundamental moldability, gross tolerance conflicts
A factory can't build your intentions. It builds exactly what the drawing says — including the mistakes.
The bottom line
A DFM issue caught in design review costs a meeting. The same issue caught after tooling starts costs $5K–$50K on the enclosure side, or a $28.5K, 16-day respin on the board.
The checklist is specific, not aspirational: draft angle, wall thickness, parting line, footprint verification and a named sign-off — twenty decisions, not a vibe.
Run DFM review at three points, not one — concept freeze, pre-tooling design review, and final pre-release verification. The middle pass is the one teams skip, and the one that matters most.
If you're heading into tooling and want a second set of eyes on the twenty decisions above, start with a discovery & feasibility phase — a DFM pass now is cheaper than any of the alternatives on this page.
Frequently asked questions
What is DFM in hardware product development?
Design for manufacturability (DFM) is the review process that checks whether a part or board can actually be produced at yield — filled, cooled, ejected, soldered and assembled reliably — before tooling or fabrication begins. It covers draft angle, wall thickness, parting lines, footprint accuracy and tolerance stack-up, among other decisions.
How much does a missed DFM issue cost?
On enclosures, a defect caught after the first mold trial costs $5,000–$50,000 and weeks of schedule delay versus a near-zero-cost design review catch. On PCBs, 2026 data puts the average respin at $28,500 and 16 days — often triggered by something as small as one incorrect footprint.
What is the correct draft angle for injection molding?
The general rule of thumb is 1 degree of draft per inch of cavity depth, with 1–2 degrees working for most plastic parts and a 2-degree minimum recommended. Textured surfaces need more: 3 degrees for light texture, 5 degrees or more for heavy texture. Ribs and bosses typically use a smaller 0.5–1.0 degree draft.
When should a DFM review happen in the development process?
At three points, not one: a lightweight pass at concept freeze, a full review at design review before any tooling purchase order is released, and a final verification pass against as-built CAD immediately before Gerbers or mold data ship. Skipping the pre-tooling pass is the most common and most expensive shortcut.
What causes most PCB respins?
Incorrect footprints are among the leading single causes — even one wrong footprint can force a complete respin. Other common causes include design rules that don't match the fabricator's real capabilities and DFM checks run only once at the end instead of iteratively through the design stages.
If the product has to ship, talk to the team that builds for that outcome.
Senior engineer on the first call. NDA before technical detail. References available under NDA after qualification. Or start with a fixed-fee feasibility study.
