DOI : 10.17577/Most first-time molding projects are quoted before the part is designed, and the quote is treated as a fixed property of the design rather than as the output of a set of tooling choices. That framing is expensive. Cavitation, tool steel, feed system and tolerance capability each move unit cost, and they interact: a tighter tolerance bought at the cavity is often cheaper to hold with a gate and cooling decision made earlier.
This article sets out a decision framework with arithmetic, because the choices are ultimately economic and the economics can be calculated before a tool is cut. All figures below are illustrative but internally consistent; substitute your own machine rate, cycle and volumes and the structure holds.
The Cost Structure Of A Molded Part
A molded part’s price decomposes into four terms:
C_part = C_material + C_machine + C_tool/N + C_overhead
where C_material is shot weight times resin price, C_machine is machine hourly rate divided by hourly output, C_tool/N is tool cost amortised over the total committed volume N, and overhead covers labour, quality and packaging. The tool term is the only one that falls with volume, and it falls hyperbolically — doubling N halves it, which is why the same part quoted at 100k and 500k pieces looks like two different designs.
Cavitation: A Capacity Decision Disguised As A Cost Decision
Take a 45 g polypropylene housing at $2.20/kg, a 28 s cycle on a 180-tonne machine at $35/hour, overhead of $0.03 per part, and two tooling options: a single-cavity tool at $12,000 or a four-cavity tool at $42,000.
Material cost is fixed at 45 g × $2.20/kg = $0.099. Hourly output at 28 s is 128.6 parts per hour single-cavity and 514.3 parts per hour at four cavities. Machine cost per part is therefore $0.272 and $0.068 respectively — a fourfold difference, because the machine’s hourly rate is spread over four times the output while the cycle time is unchanged.
At 200,000 pieces the four-cavity part totals about $0.407 against $0.461 single-cavity, a saving of $0.054 per part, or $10,800 across the order. That does not recover the $30,000 of extra tooling. Break-even is around 540,000 pieces. Read naively, the cheap tool wins.
The naive reading is wrong, because the deciding constraint at that volume is not unit cost but machine hours. Two hundred thousand parts occupy about 1,556 hours on the single-cavity tool and 389 hours on the four-cavity one. At two shifts that is roughly five months versus five weeks. A supplier quoting single-cavity tooling for a 200k order is quietly proposing that you own the capacity constraint. Cavitation should therefore be justified on required throughput first and amortisation second, and the honest comparison includes expediting cost, safety stock and the risk of a single-cavity tool being a single point of failure.
Tool Steel And Shot Life
Shot life is a validated number, not a marketing claim, and it is set mostly by steel grade, hardness and the abrasive or corrosive character of the resin.
- P20 / 1.2311, pre-hardened around 30 HRC. The default for consumer parts in unfilled resin. Expected life measured in the high hundreds of thousands of shots, repairable, and the cheapest route to a tool that works.
- H13 / 1.2344, hardened to 48–52 HRC. For glass- or mineral-filled grades, high shot weights and long programmes. Higher cost, longer lead time, and heat treatment that must be planned rather than squeezed.
- Stainless grades such as S136 / 1.2083. For PVC, PC and other corrosive or acidic melts. The premium buys corrosion resistance, not wear resistance; specifying it for an unfilled PP part is money spent on the wrong failure mode.
- Aluminium, 7075-T6 or similar. Bridge and prototype tooling with a life of thousands to tens of thousands of shots and a fraction of the lead time. Its real value is schedule, and its real risk is that it gets used for production because it was never retired.
The specification error to avoid is buying shot life you will never use. A 500,000-shot steel on a 20,000-piece programme is a deposit on capacity that does not exist; a soft tool on a 300,000-piece programme is a maintenance liability with a warranty attached.
Feed System, Gate Choice And The Tolerance You Can Actually Hold
Dimensional capability is decided before the mold is cut, because it is decided by how the melt enters and cools. Three consequences are worth stating plainly.
First, gate position sets the flow path, and the flow path sets orientation. Fibre-filled and semi-crystalline resins shrink anisotropically along and across the flow; a part gated at one end holds length better than width. Tolerances written without reference to flow direction are unenforceable.
Second, a hot runner removes runner regrind but holds a larger thermal mass in the tool and complicates startup; a cold runner is simpler and cheaper to build but adds a scrap stream and a cooling gradient if the runner is unbalanced. For a small part in a high-cavitation tool, an unbalanced cold runner system is one of the most common causes of cavity-to-cavity variation, and it shows up as a capability failure that gets blamed on the machine.
Third, packing determines warpage. A part that is dimensionally correct at the gate and overcooled at the far end will bow, and no amount of post-mold gauging fixes a warp that is moulded in. Where flatness is a functional requirement, the injection molding tolerances a supplier quotes should be traceable to a documented process window rather than to a generic table.
Tolerance Capability And Specification Discipline
A drawing that specifies ±0.05 mm on every dimension is not a precise drawing; it is an unpriced one. Process capability is a property of a specific feature, a specific tool and a specific resin, and it should be expressed statistically: for a dimension in production, a process capability index above roughly 1.33 on 30 consecutive shots is a reasonable acceptance criterion, with a documented reaction plan if it falls below. Where a customer insists on a blanket tolerance, the correct response is a capability study on the first tool and a revised drawing on the second, not a blanket agreement that neither party can verify.
Three practical rules keep tolerances affordable:
- Tolerance the features that function, and general-tolerance the rest. A sealing groove and a cosmetic edge do not deserve the same number.
- Do not tolerance across a parting line tighter than the tool can repeat. Parting-line mismatch is a toolmaking outcome; specifying against it retroactively is a dispute, not a design.
- Separate the datum from the measurement. A datum scheme that cannot be fixtured means the first-article report describes the inspector’s setup, not the part.
A Validation Protocol That Prevents The Second Tool
The difference between a competent tooling programme and an expensive one is usually visible at first trial:
- T1 with a dimensional report covering every toleranced feature, not a sample of convenient ones.
- A 30-shot capability run on the critical dimensions after process settling, with the process window recorded — fill time, pack pressure and hold time, melt and tool temperature.
- A material certificate and a colour plaque, because “natural” and “black” are not specifications.
- Gate and runner condition documented before the tool leaves the toolmaker, so that a later transfer of the tool has a baseline.
- A named owner for the tool, and a written statement of who holds title, who pays for maintenance, and what happens at end of programme.
That last item is the one most often skipped, and it is the one that costs most when a supplier changes. Injection mold tooling is a capital asset that frequently changes hands without documentation, and a tool without a baseline drawing is a tool that gets replaced rather than repaired.
A Decision Sequence
For a new part, the sequence that minimises total cost is:
- Fix the functional requirements and the tolerance stack that serves them.
- Estimate annual and lifetime volume, including spares and revisions.
- Choose cavitation from required throughput, then check amortisation over lifetime volume.
- Choose steel from resin abrasiveness and total shots, not from habit.
- Choose gate and runner layout from flow analysis, then set tolerances the process can hold.
- Validate at T1 against a written report, and record the process window.
- Transfer the tool with its baseline, or accept that you will pay for it twice.
Run in that order, the arithmetic is unremarkable and the tool behaves predictably. Run in the reverse order — quote first, design later, tolerance last — and every subsequent decision is a correction, priced at tool-modification rates.
