By PolyTakt Team · 21/10/2026
A process window is the range where the part is still acceptable, and it is unfinished until you also record where production will run inside it. This article covers which settings the earlier studies already fixed, how to map second-stage pressure, cooling time, and the temperature the material uses, and how to choose a working point without defaulting to the middle of the range.
A process window is the range in which the part is still acceptable.
It is unfinished when you only know the edges. It is finished when the sheet also says where, inside that range, the job will run, and what margin you kept on purpose.
The case for having a window at all has already been made. A single setpoint is a place the job happened to run. This article is the work that turns that place into a record: which edges are still open after the earlier studies, how to write them down, and how to choose the working point.
Three decisions should already be on the sheet.
Leave them there. Stepping speed, transfer, or cold-runner hold time again until something fails repeats work that already has a result.
Transfer is the one to be careful with. A band of a few millimetres can move the job from a deliberate short, through a complete outline, and into packing on first stage. Those are different processes. A transfer "range" that crosses them is not a window. It is three setups written as one.
Reopen an earlier study only when a failure forces you back. A part that flashes in the last of fill is a fill or clamp question, not a hold-pressure question.
On most jobs the boundary still worth mapping is second-stage pressure.
Pack development often already knows the low edge: the lowest pressure that clears sinks, voids, and the dimensions that pressure controls. The gate seal study assumes that pressure existed. What is often missing is the high edge, and the choice of where to live between the two.
With speed, transfer, temperatures, and hold time fixed:
The low edge is the last pressure that still makes an acceptable part. The high edge is the last pressure that still does. Write the failure beside each number. A limit with no failure mode is a percentage someone typed.
If cavities do not fail together, the light cavity sets the low edge and the heavy cavity sets the high edge. Balance is part of the window.
Those pressure edges are the cosmetic window: sink, short, flash, stick. A measured dimension can close that range. If the nominal of a dimension sits on the edge of a wide cosmetic window, the cavity steel is often what needs to move. Tightening the process around a dimension the tool cannot centre turns a wide window into an imaginary one.
Second-stage pressure is the edge most jobs are missing. Cooling time, and one temperature, belong with it. Almost everything else on the screen does not.
Cooling time. Time in the mould changes where the part leaves the specific-volume curve, so it changes shrinkage. Find the shortest cool that still holds the dimension and ejects cleanly. Time past that point is comfort, paid as cycle time. It is a limit you record, not only a cost of sitting high.
One temperature, chosen by the material. An amorphous resin has a melt range wide enough to move shrinkage. Mould temperature on that job is set to the low end that still fills and looks right. A crystalline resin uses mould temperature to form the crystallites, and the melt range is usually too narrow to be the factor. Polyolefins often need both. Add the second temperature only when this tool has shown that it moves the part. Thermal steps are slow: one direction, then a real return to the baseline.
If melt temperature does change, move transfer so the fill-only weight stays the one you recorded. Transfer is still not a factor. The fill is being held constant while the temperature is tested.
Hold time on a hot tip or valve gate. On a cold runner, time past gate seal adds no mass. On a hot tip or valve gate the land is small, and hold time can still move the part. That time stays on the sheet.
Available injection pressure, screw speed, and back pressure stay off it. Pressure available above what the fill speed needs does not change the part. Pressure set below what the speed needs makes the fill pressure-limited. Screw speed stays at the lowest rate that still homogenises the melt and recovers inside the cooling time. Back pressure stays at the minimum that meters a consistent charge and keeps the cosmetics. A dimensional change from raising it usually means pack is not finished, or the shear is damaging a filled material.
Clamp force joins the sheet only when fill-only shots already show the parting line moving. It is not a factor you owe on every tool.
The result is not a symmetric band around the setpoint you started with. The figures below are the shape of the record for one tool. They are not a starting recipe.
| Edge | What failed | |
|---|---|---|
| Hold pressure, low | 420 bar | Sink remains in the rib below this |
| Hold pressure, high | 640 bar | Parting line flashes above this |
| Cooling time | 14 s | Length leaves spec, and ejection marks appear, below this |
The acceptable range sits inside the failures, with room for a material lot to move. A window that ends on the failing shot is a window in name only.
Temperature is left off this example so the pressure and cooling edges stay easy to read. On a crystalline resin that line would be mould temperature. On an amorphous resin it would be melt temperature.
A narrow window can be an honest result. Tight dimensions, a difficult flow length, or a machine near its limit will do that. A wide window can mean a robust job. It can also mean the measurement was too coarse to see the edge. Unknown width is the result that causes trouble later: the process looks stable until the day it is not.
The window answers which settings make an acceptable part.
The working point is the setpoint production will run. Those are different decisions, and they belong on the same sheet. A window with no working point is a study. A setpoint with no window is a habit.
The middle of the range maximises distance to both failures. Choose it when lot and machine variation are poorly known, when a complaint is expensive, or when you have no other candidates to compare. On a wide, flat region the centre and a point nearer one edge can both make good parts. The centre often costs more heat, clamp tonnage, hold work, or cooling seconds, and the part does not get better in proportion.
Before you adopt the middle, write down what the high side is buying.
If the high side is comfort rather than a failure you have seen, a lower working point is the better record, provided the margin you kept is written down.
On the example above, 480 bar is a defensible working point if the last lots needed about 60 bar of extra pressure to clear the rib. 600 bar made the same part and spent the difference on the clamp and the hold. That difference is not a more robust process. It is unused range.
Enough margin is the amount that absorbs variation you already know about. Material lots, colour, reclaim, oil temperature, tool wear, the difference between shifts.
It is not half the window by policy.
If that variation is tight and you can see it, you can sit closer to the low edge, with the reason written down and a trigger to move back when a lot or a measurement says so. If the variation is wide and nobody is watching it, the middle is the honest place to run.
A validated process has a particular form of this record. The high and low you can defend are the challenges in the OQ. A nominal setpoint with tolerances added afterwards is a different document.
Confirm the working point with a consecutive run long enough to show the variation you are claiming the margin will absorb. If that run cannot see a lot change, say so on the sheet, and keep more margin until it can.
Store it with the process development record for the tool. The next time a lot changes, or someone asks whether the process is still inside what you approved, the answer should already be on that page.
Create the window from the edges the earlier studies left open: second-stage pressure, cooling time, and the temperature the material uses. Leave speed, transfer, available injection pressure, and cold-runner hold time where their own work put them. On a hot tip or valve gate, hold time is still an edge.
Then choose the working point in the same record. The middle is the right place when you cannot see the variation coming. It is an expensive habit when the region is flat and the high side buys nothing you can measure.
The boundaries and the place you intend to live are one piece of work. Document both before production has to discover either of them on a bad day.
