Start With the Part and the Numbers

Every mold design starts with three inputs: the part design, the order quantity and schedule, and the target part cost. Skip any of them and the design floats free of reality. A mold for 5,000 parts a year looks nothing like a mold for 5 million — different steel, different cavitation, different automation.

The DFM review happens here, before any mold layout. Wall thickness, draft, ribs, tolerances, gate locations — every issue found now costs a conversation. Found after steel is cut, it costs a weld, a re-cut, and a week.

Machine and Cavitation

Choose the molding machine first, then the number of cavities — not the reverse. The machine’s clamp tonnage, shot size, and tie-bar spacing set the hard limits. Cavitation is the economic optimization inside those limits: enough cavities to hit the cycle-time and cost targets, few enough that the mold stays buildable and maintainable.

Estimate cycle time early from wall thickness and resin. Cooling dominates the cycle — roughly, cooling time grows with the square of wall thickness. That estimate feeds back into the cavity count: if the cycle is 40 seconds, you need twice the cavities of a 20-second cycle to make the same annual volume.

Select the mold base size from the cavity layout plus room for cooling, actions, and ejection. Undersized bases are a classic false economy — everything downstream gets cramped.

Fill First, Cool Second

Lay out the runner and gate system before the cooling. The melt delivery path decides where the steel can and cannot go — cooling channels route around runners, not through them. Design cooling first and you will redraw it around every runner change.

Runner and gate design aims at one thing: every cavity fills at the same time, with the same pressure and temperature history. Balanced filling is what makes multi-cavity molds produce identical parts. Gate type follows the part — pinpoint gates for automatic degating, fan gates for flat parts, hot runner valve gates for cosmetic surfaces.

Then lay out cooling channels for cavities and cores. Cooling is not “drill some water lines” — it is a thermal design. Channels follow the part geometry at consistent distance from the cavity surface, sized for turbulent flow. Cores get bubblers or baffles where straight channels cannot reach. Uneven cooling is the root cause of most warpage, so this step deserves real engineering, not leftover space.

Run CAE fill and cooling analysis before detailing. Simulation catches unbalanced filling, air traps, weld line positions, and hot spots while changes are still cheap. It does not replace experience, but it stops the expensive surprises.

Actions and Ejection

Side actions — slides and lifters for undercuts — come next. Each undercut needs a mechanism, and each mechanism needs space in the mold base, clearance to move, and cooling that does not collide with it. Action design interacts with everything: parting line, ejection, cooling.

Lay out the ejection system after the actions are placed. Ejector pins need solid steel to push against and clear paths through the ejector housing. Return pins, sprue pullers, and stripper plates all coordinate with the machine’s ejector rod pattern.

Venting goes in alongside ejection. Every cavity needs air escape at the last-to-fill areas and along the parting line. Shallow vents — typically 0.02–0.03 mm deep for most resins — let air out without flashing plastic.

Steel Safety and Detailing

Check mold parts for strength before finalizing. Cavity walls under injection pressure, slender core pins, thin steel between cooling channels and the cavity — each gets a sanity check. A core pin that deflects under fill pressure produces parts with wandering wall thickness, and no process adjustment fixes a bending pin.

Then the final detailing: leader pins and bushings for alignment, interlocks, clamping slots, water fittings, date stamps, and the bill of materials. Boring, essential, and the step most often rushed. A mold that cannot be clamped, plumbed, or aligned correctly is a mold that never runs right.

Prove It Before Cutting

Before any steel is cut, the design gets a final review against the original three inputs: does this mold make this part, at this volume, at this cost? Then the build starts — roughing, finishing, EDM, polishing, fitting — and the first trial (T1) tests every decision at once.

T1 is a diagnostic, not a celebration. Short shots, flash, sink, warpage, pin marks — each one traces back to a design decision. Set starting process conditions from the simulation, optimize from there, and document what the mold wants. A mold whose process window is understood at T1 runs for years. One that is “tweaked until it works” never quite settles.

The sequence exists because each step constrains the next. Respect the order and the design converges. Shuffle it and you iterate — on the most expensive medium there is, which is cut steel.