CNC and molding: how each process forms a part

• 1
  Key takeaways
• 2
  How material removal works
• 3
  How a mold sets the shape
• 4
  Where each process stops working
• 5
  The crossover point in real numbers
• 6
  What this means for your design
• 7
  CNC and molding side by side

A CNC machine reads G-code and drives a rotating cutter along a programmed path. The tool touches the workpiece, shears away chips, and leaves the remaining material in the shape you drew. Nothing about the part exists until the cutter passes over it. Every surface is generated by a tool path, so every surface has a direction, a stepover, and a scallop height.

Because the tool must physically reach a surface, the geometry has to be machinable. Undercuts need a fourth or fifth axis, or a re-fixture. Deep pockets need a cutter long enough to reach the floor without chattering. A 3 mm end mill in a 60 mm deep pocket will deflect, so the wall will taper and the floor will not be flat.

Stock removal is a mechanical event. Heat goes into the chip, the tool, and the part. Aluminum 6061 conducts heat away fast, so it machines clean at 3,000 to 8,000 rpm with carbide. Titanium TC4 does the opposite. It holds heat at the cutting edge, so speeds drop and coolant flow matters more than spindle speed.

The payoff is direct control. Change the offset, and the part changes on the next cycle. That is why machining dominates prototypes, bridge tooling, and any run where the design is still moving.

• 1
  Reach matters more than toleranceA feature the tool cannot reach is a feature you cannot cut, regardless of what the print says.
• 2
  Tool deflection sets the real floorLong, thin cutters trade accuracy for reach. Short and stubby always holds tighter.
• 3
  Fixtures repeat the datumTwo operations mean two datums. Each re-fixture adds stack-up error.

Injection molding melts polymer pellets and injects them into a steel cavity at high pressure. The plastic fills the void, packs, and cools. As it cools it shrinks, and that shrinkage is the single biggest source of dimensional error. A 100 mm polypropylene part can shrink 1.5 to 2.5 mm depending on gate location and wall thickness.

The cavity does not move during the cycle, so geometry that would be impossible to cut can be cast into the steel by EDM or a CNC electrode. Ribs, bosses, snap fits, and texture all come out in one shot. Cycle time is the cost driver, not feature count.

Wall thickness decides whether the part fills. Thick sections cool slower, shrink more, and sink on the visible face. Thin sections freeze before the cavity packs, so the far end of the flow path starves. A uniform 2 mm wall with generous radii is far easier to mold than a 1 mm rib next to a 5 mm boss.

Tooling is the entry cost. A single-cavity prototype mold in aluminum can be cut in days and run a few hundred parts. A hardened multi-cavity steel tool takes weeks and costs orders of magnitude more, but it spreads that cost across hundreds of thousands of cycles.

• 1
  Shrinkage is not uniformFlow direction and cross-flow shrink at different rates. Gate placement decides which one dominates.
• 2
  Draft is not optionalWalls need 0.5° to 2° of draft so the part releases without dragging.
• 3
  Gates leave marksEvery gate is a witness line. Put it on a hidden face or plan a secondary op.

Machining gets expensive when the part is simple and the quantity is high. If you are making 50,000 identical brackets with no tight features, you are paying spindle time for something a die could stamp or a mold could shoot in seconds. The cut does not get cheaper with repetition the way a mold does.

Machining also struggles with soft, gummy materials and very thin walls. A 0.5 mm aluminum wall will sing and deflect under cutting force. You can support it with fixturing or fill it with wax, but at some point the part is more trouble than it is worth.

Molding stops making sense when the geometry keeps changing. Every design revision that touches the cavity means welding, re-cutting, or a new insert. Change the wall thickness three times and you have paid for three tool revisions.

Molding also fails on low volume with expensive resin. PEEK at high temperature needs a hot runner and a tool built for 400 °C. Running 20 parts that way costs far more than machining 20 parts from PEEK rod.

There is a middle zone. Bridge tooling in aluminum, vacuum casting from a silicone mold, and urethane casting all sit between the two. They borrow the mold idea but skip the hardened steel.

• 1
  Pick machining whenQuantity is under a few thousand, tolerances are tight, or the design is not frozen.
• 2
  Pick molding whenQuantity is high, the geometry is stable, and the resin is common.
• 3
  Pick bridge tooling whenYou need 50 to 500 parts in the final material before committing to steel.

Total cost is tooling plus parts. Machining has almost no tooling and a steady per-part price. Molding has heavy tooling and a very low per-part price. The two curves cross somewhere, and that crossing is the only number that matters.

If a mold costs 20 units of money and each molded part costs 1, while a machined part costs 8, the molded total passes the machined total at roughly 2,800 parts. Change the machined price to 15 and the crossing moves under 1,500. The math is sensitive to both numbers, so get a real quote on both sides before you commit.

Resin price moves the line too. A commodity polypropylene part crosses early. A glass-filled PEEK part crosses late, because the material cost dominates and the machining waste is a smaller fraction of a very expensive blank.

Setup time is the hidden variable. A five-axis job with two fixtures and a custom soft jaw carries more front-end engineering than a three-axis plate. That pushes the crossover up. A simple turned bushing crosses much earlier because the program is short.

• 1
  Amortize honestlyDivide tooling by the quantity you will actually order, not the quantity in the forecast.
• 2
  Count the revisionsOne cavity change can wipe out the savings from a thousand molded parts.

The process you pick changes the design rules you follow. Machined parts want generous internal radii, chamfered edges, and holes that align with a standard drill. Molded parts want uniform walls, draft on every vertical face, and ribs sized to a percentage of the wall.

Tolerance stacks differently. On a machined part you can tighten one critical dimension without touching the rest. On a molded part, tightening a dimension usually means controlling shrink, which means gate and cooling changes that affect the whole part.

Material choice follows the same split. Machining covers 6061, 7075, 316L, 17-4PH, TC4, and PEEK from bar or plate. Molding covers the thermoplastics that melt and flow, plus die casting for zinc and aluminum alloys.

A common mistake is designing a machined part and then handing the same STEP file to a molder. The wall thickness, draft, and radii that were fine for a cutter are often the exact features that cause a short shot or a sink mark.

The reverse happens too. A molded housing with 1 mm walls and no flat datum is hard to hold in a vise. Machining it means building soft jaws and accepting chatter on the thin sections.

• 1
  Redesign at the handoffTreat the DFM review as a design step, not a formality.
• 2
  Keep a machined datumEven molded parts need one accurate face for inspection and assembly.