Manufacturing Process Selection | Six Routes Compared

Manufacturing Process Selection, Decided by Eight Part Features

When the drawing is fixed but the process is not, manufacturing process selection follows a fixed order rather than gut feel. Sort out size, wall thickness, volume and tolerance first, then weight material, surface, internal structure and tooling budget. Yujiaxin Tech applies this same order to every custom part and records the outcome in a decision matrix.

Six processes are parallel candidates, not a ladder. MIM, press-sinter PM, precision casting, gear hobbing, CNC and CIM each have their own sweet spot. the six manufacturing process routes side by side opens the field, then the matrix narrows it.

Drawing Fixed, Process Open: Why This Step Costs the Most

A wrong manufacturing process selection costs at the tool, not at the drawing. MIM molds run 15,000 to 100,000 dollars, press tools 5,000 to 50,000. Rework lands after the tool, not after the drawing.

Once the drawing enters custom production, the next decision is the route. Because rework is billed after the tool, you waste the most money when the tool is already built. Changing the route after the tool exists stalls the entire downstream machining chain.

Manufacturing Process Selection: Eight Inputs, Drop the Impossible First

The eight inputs are maximum envelope, minimum wall, annual volume, overall tolerance, material, surface, internal complexity and tooling budget. The first four are hard vetoes, the rest only rank. The input list is identical for every material and every part shape.

Sort first, then rank. A weighted score puts "tolerance impossible" and "surface slightly off" in the same bucket, but the first is a zero, and a wrong manufacturing process selection still looks fine on paper.

If two of the ten rules point to the same route, the call is made. On conflict, tolerance wins. The table above compresses the same rules into a compact form for quick vetting. precision components for each process selection outcome groups by material and structure for a first pass.

Volume Threshold: Where the Per-Piece Cost Curves Cross

CNC has an almost flat per-piece curve; one part or a thousand cost nearly the same. By powder metallurgy cost estimating, press-sinter cost falls with volume and crosses CNC at 5,000 to 10,000 parts per year. That is exactly why PM and MIM only pay off on long runs, where the tool cost spreads across many thousands of parts.

MIM and CNC cross at very different points across sources: 1,000 to 3,000, 3,000 to 8,000, 10,000 to 20,000 parts all appear. The more complex the part, the lower the threshold; this article uses the 3,000 to 20,000 industry band, and how manufacturing process selection shifts the quotation shows the second line.

Three Veto Lines: Tolerance, Wall and Internal Structure

Start manufacturing process selection with tolerance. Overall tolerance inside ±0.02 mm leaves only CNC; MIM as sintered sits at ±0.3% to ±0.5%, so a 10 mm feature is about ±0.03 to ±0.05 mm, and key dimensions need a second operation to reach ±0.02 mm.

Wall under 0.5 mm rules out the injection routes. MIM is best at 1.0 to 3.0 mm; over 5 mm risks voids, under 0.5 mm fails at the injection step. A thin wall calls for a different casting method, not rework on a finished part.

Side holes, internal threads and free-form cavities veto press-sinter, since PM presses in one axis only. Internal gears and gear teeth against a shoulder leave no relief for the parting cut; where the gear hobbing process selection stops working gives the reason. Parting across the section is technically impossible here.

Six Routes in Detail: Best Range and Exit Point

Press-Sinter PM and MIM

PM covers 0.1 to 100 g, the press direction limits the size, and it pays above ten thousand parts a year. Material grade, density and tolerance should follow MPIF powder metallurgy standards so quotes from different suppliers sit on the same basis; MIM spans 0.1 to 200 g, best wall 1.0 to 3.0 mm, density 95 to 99%, and exits on cross holes, internal threads and wall over 5 mm. Manufacturing process selection for these two routes rests on the same density and tolerance basis.

Precision Casting and CIM

Precision casting runs 5 g to 50 kg, minimum wall 0.5 to 1.5 mm, economic at 500 to 10,000 parts; under 500 the wax-pattern tooling never amortizes. CIM suits 2 to 50 mm parts, wall not under 0.5 mm, zirconia better above 0.8 mm, and only beats CNC clearly past 5,000 parts a year.

CNC and Gear Hobbing

CNC holds ±0.01 mm, down to ±0.005 mm, with 20 to 60% material use, and needs no tool. Without tooling, CNC fits prototypes and small batches whose route changes weekly. Hobbing beats CNC milling for external and helical gears above a hundred parts, but internal gears and teeth against a shoulder have no relief.

Three Parts Through the Matrix

Running three parts of different shape through the matrix teaches more than memorizing eight inputs, and it is the fastest way to learn manufacturing process selection. Yujiaxin Tech runs these walkthroughs on real orders and keeps them as training examples for engineers.

Two of the three examples share the same exit: tighten the tolerance one step. The overlap is no coincidence; it is the most common conflict among real-part inputs.

That overlap is the most frequent reason a decision flips mid-project. A recorded exit condition cuts rework and travels with the first written statement on the matrix.

When Manufacturing Process Selection Fails: Four Cases

The matrix is not omniscient. In four cases it returns the wrong answer, and the fix is always the same: send it back to engineering.

1. Volume is unstable, swinging between 3,000 and 10,000 parts, so the threshold loses meaning.
2. The drawing still changes; wall and internal structure are not frozen.
3. Key criteria conflict: tolerance points to CNC, internal structure to casting.
4. Supplier capacity does not match; the route stands but no one delivers on time.

All four share unstable inputs. Stabilize those first, and the matrix works again.

After the Decision: Verify and Start

With the route set, the first-article inspection belongs in the purchase agreement: dimensions, density, material analysis and process capability need clear limits. Yujiaxin Tech issues first articles under ISO 9001 and GJB 9001C, aligns automotive work to IATF 16949, and the Yujiaxin Tech quality control policy covers inspection and records.

The second step protects the exit line: if the first article misses the drawing tolerance, go back to the matrix rather than loosen the tolerance. send drawings and annual volume to our engineers, and they return a written rationale against the eight inputs.