Injection Pressure Profiling — What Does the Machine’s Own Pressure and Screw-Position Trace Tell You Without a Cavity Sensor, and Where Does That Signal Stop?
Last updated: 2026-09-03 · Reading time: about 6 min
Every shot already leaves a trace on the press: screw position and injection pressure against time, recorded with no cavity sensor. Read as one signature, that machine-side curve catches a 3–5 percent peak-pressure shift, transfer drift and check-ring wear before parts fall out — but it sees the machine, not the melt. That ceiling decides when a cavity sensor earns its cost.
See also related article.
Key facts
<>
Shrinkage — The dimensional reduction from mold cavity to cooled part. Amorphous resins: 0.4–0.7% (uniform). Semi-crystalline: 1.0–2.5% (directional). Mold makers apply shrinkage factor when machining cavity. Actual shrinkage varies with wall thickness, gate location and packing pressure. See also related guide.
What does the machine’s own trace record — and how is it different from a cavity curve?
Two curves come off the same shot. A cavity sensor — when the mold has one — reads the melt’s pressure inside the cavity. The press’s own trace reads the screw: where it is and how hard it pushes. Most shops have only the second one, and it is worth more than they use it for.
| Signal (machine-side) | What it measures | What it diagnoses | What it cannot see |
|---|---|---|---|
| Screw position vs time | Where the screw is each millisecond | Actual fill speed vs setpoint, cushion, transfer position | Melt pressure at the feature |
| Injection pressure (hydraulic/servo) | Force the screw pushes with | Viscosity and lot shift, fill resistance, peak headroom | Pack pressure delivered to the melt |
| Cushion — position at end of hold | Residual melt left in front of the screw | Non-return-valve wear, shot repeatability | Gate-seal timing |
| Peak pressure + transfer position | The fill-to-pack signature | Lot viscosity, a new flow restriction, transfer repeatability | Cavity-to-cavity balance |
Read one axis and you get half the story. A fill that hits position targets under rising pressure is a viscosity change; one that hits pressure but misses position is a slip — the non-return valve is not holding. The two axes either agree, or the argument is the diagnosis.
How do you read the four phases of the machine-side pressure curve?
The trace divides into four phases. Each has a machine-side reading, and each reading has a limit on how far it goes without a cavity sensor.
- Fill ramp. Pressure climbs as the screw drives melt through the gate into the cavity. The slope is a viscosity read: steeper than the golden trace means harder-to-flow material, a colder melt or a restriction; shallower means easier flow. Trustworthy for a relative shift, not for the melt front’s speed at one feature.
- Transfer. At the changeover the screw switches from velocity to pressure control and the trace spikes. Watch the position at transfer as closely as the pressure — if it moves shot to shot, the fill volume is moving, and every downstream number floats with it.
- Pack plateau. Pressure holds while the screw creeps forward to feed shrinkage. Machine-side, the plateau confirms the press is holding set pressure — not that it reached the melt. A creep that stops early can mean the gate sealed or the valve leaked; the trace cannot say which.
- Release and cushion. Pressure decays as the gate freezes and the part cools, and the screw retracts. The position left at the end is the cushion — your best no-sensor repeatability signal. An erratic or shrinking cushion flags shot-to-shot variation before any dimension shows it.
Machine-side injection pressure trace. The pressure the press applies at the screw — hydraulic or servo — plotted against time and screw position, recorded every shot with no mold sensor. It is an indirect signal: it includes friction, non-return-valve and hydraulic losses that never transfer to the melt, so it reads the machine well and the cavity only by inference.
How far can you trust it without a cavity sensor — and when do you escalate?
Machine-side profiling is the majority-shop instrument: free, already recorded, honest within its limits. The discipline is separating fact-readings from guesses, then windowing the facts.
“Nobody argues with a cavity-pressure curve, but most molds do not have one — and the press has been recording a usable signature the whole time. We read the screw-position and pressure trace first: peak, transfer position, cushion. It will not tell you what happened at the far corner of the cavity, and we say so plainly. It will tell you the shot stopped repeating — usually the first thing you need to know.”
— Project Engineer, aCatMould
Frequently asked questions
If the machine-side trace catches drift, why add cavity sensors at all?
Sensitivity and location. The machine trace sees the screw, so it says the shot changed; the cavity sensor sees the melt, so it says where and how much — gate seal, delivered pack, cavity balance. On a tight-tolerance or multi-cavity part, that difference is the sorting bill.
For a robust part on a stable process, the trace plus a shot-weight check catches most of what matters for free. The sensor earns its cost when tolerance is tight, cavities must balance, or the trace keeps flagging a change it cannot locate.
My cushion is stable and weight is steady, but parts still vary — what is the trace missing?
In-cavity events. A stable cushion and steady weight rule out gross machine-side repeatability loss, but say nothing about gate-seal timing, cooling balance, or a lot viscosity shift inside the window that moves a dimension without moving the weight.
That is the ceiling, not a flaw in the reading. When weight and cushion hold but dimensions move, the cause is inside the mold — escalate to a cavity sensor or a dimensional check; the machine trace has told you all it honestly can.
Can the injection pressure trace really catch check-ring (non-return valve) wear?
Yes — one of the trace’s best no-sensor uses. A worn non-return valve leaks back during injection and hold: an erratic or growing cushion, position slippage against the speed setpoint during fill, and shot-weight scatter on the scale.
The classic confirmation needs no sensor: run a fixed hold, weigh a run of shots, and watch cushion and weight scatter widen together. That pattern is the valve, not the resin — and it shows long before parts fall out of tolerance.
Key takeaways
- The press already records a machine-side signature every shot — screw position and injection pressure vs time — with no cavity sensor; read the two axes together, because they either agree or the disagreement is the diagnosis.
- Read it in four phases — fill ramp (viscosity), transfer (position repeatability), pack plateau (held, not delivered), release and cushion — and window the facts: peak pressure, transfer position, cushion.
- Trust it for machine health and shot repeatability, never for in-cavity fact: gate seal, delivered pack and cavity balance need a cavity sensor, and that ceiling is exactly when the sensor earns its cost.
Injection pressure profiling does not require instrumenting the mold. The press’s own screw-position and pressure trace, read as one signature and windowed shot to shot, catches viscosity shifts, transfer drift and check-ring wear before parts fall out. Its limit is just as clear: it sees the machine, not the melt, so on tight-tolerance work the trace points at the problem and a cavity sensor locates it. At aCatMould we set every process on recorded traces and hand you the curve behind the numbers.
Sources
RJG Inc — scientific and decoupled molding, process monitoring
Kistler — cavity pressure sensing and monitoring basics
Plastics Technology — injection molding process data and troubleshooting
Autodesk Moldflow — simulated fill and pack pressure
Society of Plastics Engineers — molding process guidance
Sources: Plastics Technology · Plastics Technology Online · Society of Plastics Engineers (SPE) · ThomasNet