Mass Production of Small Button Parts (High-Cavitation Tooling)

The Challenge

Small button parts live or die on inter-cavity consistency. A 32-cavity tool only pays off economically if every cavity fills at the same time and cools at the same rate, because a single lagging cavity produces a short shot or a dimensionally off part that drags down the whole shot yield. At an annual volume of 5,000,000 units the margin for cavity-to-cavity variation is essentially zero, and the program tolerance of ±0.03 mm applied to every one of the 32 cavities on every shot.

The material pair added complexity. The buttons used POM for the moving member because of its low friction and excellent dimensional stability, and ABS for the housing element because of its surface finish and toughness. POM is a semi-crystalline resin that shrinks significantly, typically 1.8 to 2.5 percent, and is highly sensitive to process variation and to moisture, while ABS shrinks far less and flows differently. Balancing the fill of two materials with different shrink and flow inside one high-cavity layout is a non-trivial thermal and rheological problem.

Processing windows were tight. POM runs at a melt of roughly 190 to 210 °C with mold temperatures of 80 to 100 °C, and it is prone to degradation if dwell time or temperature is not controlled. Because the part is small, the absolute dimensional budget is tiny: a ±0.03 mm tolerance on a button only a few millimeters across means the process must be repeatable to within a fraction of a percent of feature size, shot after shot, over millions of cycles.

Tool life and maintenance were also on the line. A production tool built for a 5M-unit yearly run must resist wear at gates and ejection surfaces for the full program, and any cavity drift from wear would violate the ±0.03 mm spec late in life. The 12-week production-tooling schedule had to deliver a hardened, balanced tool ready for sustained volume, not a prototype.

Reject economics at this volume are brutal. At 5,000,000 units per year, a one percent inter-cavity yield loss is 50,000 scrapped buttons, so the tool and process had to be designed for yield rather than merely for function. The small part size also meant automated vision or gauge inspection had to be fast enough to keep pace with the press, and any manual sorting step would bottleneck the line.

Material qualification added a second axis. Running POM and ABS in related but separate elements meant two resin specifications, two drying regimes, and two sets of dimensional behaviors to qualify and hold. The program could not tolerate a supplier lot variation that shifted POM shrinkage outside the window, because that would move every cavity at once and break the ±0.03 mm spec across the whole shot.

Process capability had to be proven, not assumed. A 32-cavity tool can look perfect on the first shots and then drift as the machine, the material, and the tool reach steady state, so the program required a capability study that covered warm-up, steady state, and the first article across multiple cavities. Only by measuring cavity 1 and cavity 32 at the same time could we be sure the balanced runner was doing its job across the whole array.

The small feature sizes left no room for variation in the press. With a ±0.03 mm tolerance on a button only millimeters across, a melt-temperature drift of a few degrees or a pack-pressure change of a bar moves the dimension past spec, so the machine had to hold its settings far more tightly than for a large part. The control plan therefore specified narrow bands and the cell was set up to stay inside them without constant operator adjustment.

The program also had to respect the assembly it fed. The buttons located into a larger appliance assembly where their actuation feel and their positional accuracy mattered to the finished product, so the molded dimensions had to hold not just to the drawing but to the mating parts downstream. A button that was in tolerance alone but off relative to its neighbor would still read as a defect on the line, so the inter-cavity consistency served the whole assembly, not just the tool.

The Solution

We engineered a naturally balanced runner system so that all 32 cavities received melt at equal pressure and temperature with no artificial restriction tuning. A symmetrical hot or cold runner layout placed each cavity at an identical flow distance from the machine nozzle, which is the only reliable way to guarantee simultaneous fill across a high-cavity array. This removed the chronic imbalance that causes some cavities to short-fill while others flash.

Cavity-pressure monitoring was built into the process from the first run. Transducers in representative cavities tracked the pack and hold curve in real time, giving an early warning of any drift in viscosity, fill, or cooling before it produced an out-of-tolerance part. The monitoring data also served as the statistical process control record that proved the ±0.03 mm capability across the full tool.

The tool was hardened for the full production run. Cavity and core steels were selected for wear resistance at the gate and ejector interfaces, and the 32-cavity plate was machined and verified so that cavity-to-cavity geometry matched to within the tolerance band. Venting was optimized for the small parts to prevent burn and short shots, and ejection was tuned to avoid witness marks on the visible button face.

Material handling closed the loop. POM was dried and processed under controlled moisture, and the machine setup locked melt temperature, mold temperature, and cycle so that the semi-crystalline shrinkage stayed predictable. A short validation run confirmed inter-cavity consistency before the tool was released to the 5M-unit yearly volume.

Inspection and release were engineered for speed and statistical strength. Cavity-pressure signals were correlated to final dimensions so the process could be judged shot by shot, and periodic gauge and vision checks confirmed the ±0.03 mm features without stopping the press. First-article capability studies proved the balanced runner delivered the same part from cavity 1 to cavity 32.

Tool care was planned for the full yearly load. The hardened cavities and the documented wear plan meant the 5M-unit volume would not erode the ±0.03 mm features before a scheduled refurb, and spare inserts kept the tool available. A locked machine setup let a second press or a cavity refurb reproduce the same result, supporting the appliance manufacturer production plan.

We built the inspection around the process signal. Cavity-pressure transducers gave a real-time proxy for fill and pack, and by correlating that signal to the measured ±0.03 mm features we could judge every shot instead of sampling a few. The approach turned the 32-cavity tool from a black box into a monitored process where a drift in any cavity showed up before it produced scrap.

Tool maintenance was scheduled, not reactive. The hardened cavities and the documented wear map meant the 5,000,000-unit yearly load was planned against known wear rates, with spare inserts staged so a refresh never became an emergency. The locked setup also meant a second press could run the same tool and the same result, giving the appliance manufacturer capacity flexibility without re-qualification.

We tied the tool output to the downstream assembly. The 32-cavity result was checked against the mating features of the appliance, and the process window was set so the buttons arrived at the line ready to seat without sorting. The cavity-pressure monitoring and the documented setup meant the appliance manufacturer could plan the line around a predictable, consistent button rather than a variable one.

The Result

The 32-cavity program sustained 5,000,000 units per year at ±0.03 mm with stable cavity-to-cavity variation, proving the balanced runner and monitoring approach at full scale. Inter-cavity dimensional spread stayed inside the specification across the production run, which is the defining requirement for a high-cavitation tool of this size.

The hardened tool delivered the expected life without a mid-program respin, and the statistical process record gave the appliance manufacturer confidence to plan the button line at the committed annual volume. The 12-week production-tooling schedule was met and the parts moved straight into the customer assembly line.

First-pass yield held at the program target across the production run, and the balanced tool removed the cavity-to-cavity variation that would otherwise have forced 100 percent sorting. By holding ±0.03 mm on all 32 cavities, the program avoided the scrap and rework cost that high-cavitation tools typically incur.

The validated, documented process gave the appliance manufacturer a repeatable baseline for future button programs, shortening quotation and tooling lead time on follow-on variants. The 12-week production-tooling schedule delivered a tool ready for sustained volume rather than a prototype needing rework.

The 32-cavity tool proved its economics at scale. By holding ±0.03 mm on every cavity, the program avoided the per-cavity sorting and the scrap that high-cavitation tools usually suffer, and the 5,000,000-unit yearly volume was met from a balanced, monitored process. The hardened tool delivered the expected life without a mid-program respin.

The documented capability study and the validated process gave the appliance manufacturer a defensible basis for planning the button line, and a repeatable baseline for future high-cavitation programs. The 12-week production-tooling schedule was met, and the parts moved straight into the customer assembly with the inter-cavity consistency the design required.

The buttons arrived as a predictable, consistent component for the appliance line. Holding ±0.03 mm across all 32 cavities meant the 5,000,000-unit yearly volume seated into the downstream assembly without a sorting step, and the 12-week production tooling delivered the life and the consistency the program needed. The validated baseline shortened future button quotations for the same customer.