The Process Window Does Not Travel With the Tool

The parameter sheet from T1 is not a process. It is a record of one machine, one afternoon, and one lot of resin.

We are handed these sheets regularly, usually when a tool moves out of a sampling cell into production, or from one plant to another after a supplier change. The sheet is a single column of numbers — melt temperature, mold temperature, injection velocity, hold pressure, cooling time — and everyone treats it as the thing that must be reproduced exactly. On the new press it produces a part that looks right, then drifts over the first week, and the argument starts about who changed something.

Nobody changed anything. The numbers were never portable.

What is machine-specific, and always was

Injection pressure as displayed is usually hydraulic or nozzle pressure. It becomes plastic pressure at the gate only after the intensification ratio of the unit, which is set by the screw diameter: same tool, different screw, different pressure where it matters.

The velocity setting fails the same way. "65%" is a percentage of that machine's maximum rate, so it converts to a different fill time on every press. Shot size in screw position depends on barrel capacity. The hold transfer position depends on cushion, and cushion depends on the check ring — a worn ring leaks material back, so the cushion you set is not the cushion the press holds.

Mold temperature is not a setpoint either, it is an outcome: whatever the controller can hold given its capacity, its flow rate and the water that plant has. A tool qualified at 80 °C in a sampling cell on one dedicated controller can sit at 72 °C in a plant where that circuit shares a chiller — and that shift moves shrinkage, warpage and cycle time with it.

What does travel

Five values describe what the polymer experienced rather than what the machine was told, and all of them are measurable on any press:

• Melt temperature, read from a purge with a probe, not from the barrel setpoint. On a large barrel running a short cycle the two can differ by 10 °C or more.

• Fill time in seconds, from first-stage start to transfer.

• Cavity pressure peak, if the tool or press has a sensor near the end of fill.

• Mold surface temperature, taken with a contact probe at the same points every time.

• Part weight per cavity, weighed at the end of hold, after gate freeze.

Part weight is the most useful of the five: cheap to measure, and sensitive to almost everything that goes wrong.

Ask for the window, not the setpoint

A setpoint says "this worked once". A window says what the process tolerates, and it is the only version of a process that survives a change of machine.

The format that works is a table with a nominal, a limit in both directions, and the failure mode at each edge — melt temperature ±10 °C, mold temperature ±5 °C, fill time ±15%, hold pressure ±10%. Below the low melt limit: short shots and cold slugs at the gate. Above it: flash, splay, sections staying soft. Fill too fast: shear marks, overpacked gates. Too slow: hesitation, visible flow marks.

Those limits illustrate the format, they are not universal. An unfilled resin in a part of moderate wall usually tolerates a wider fill-time band than a glass-filled thin-wall part, where shear history and fiber orientation do more of the work.

Re-qualifying on the production press

The order matters: each step assumes the previous one is stable.

1. Dry the resin to the same specification — temperature, dew point, time — and confirm it with a reading, not with the dryer display.

2. Soak the mold at temperature. A tool that sat cold is not at temperature at shot one, and the core runs behind the cavity for the first cycles.

3. Run fill-only shots with no pack, shortening until the part is just short of complete, and map which cavities fill last. That is the balance check, at the fill time you intend to run.

4. Build a pack curve: hold pressure against part weight, in steps, until weight stops rising. The plateau is gate freeze; beyond it you add stress and no material.

5. Walk the corners of the window — low melt with long fill, high melt with short fill, the same for mold temperature — and measure at each corner.

6. Measure the first article per cavity, after the rest period, against the drawing.

Where transfers actually fail

The tool was qualified in conditions production does not have: one material lot, no regrind, a technician watching every shot, cold water in a temperate month. Production adds regrind, three shifts, a warm water tower in August and nobody watching.

The quoted cycle was built on a mold temperature the plant cannot hold at volume, so the tool runs colder and gets blamed for the temperature controller.

And the balance was never re-checked. Cavity balance depends on fill rate as much as on gate geometry, so a tool that ran even on the qualifying press can run uneven elsewhere — and the first sign is one cavity weighing light three weeks in, long after the transfer was signed off.

What to put in the purchase order

One sentence does most of the work: the tool is supplied with a process window table — nominal and both limits for melt temperature, mold temperature, fill time, hold pressure and cooling time — with the failure mode at each limit, measured on the press that produced the first article.

Then run one shift at the corners of that window on the press that will hold production, before the tool ships. A tool that stays inside tolerance at all four corners can be moved again without starting over; one that only works at the middle was observed once, not qualified. Make both a condition of final acceptance, not a favour asked after the invoice.