A 4 mm cushion can still be a bad process
A cushion target can be technically correct and operationally useless. Consider two runs on the same press, mold, screw, and resin. In the first, the end-of-hold cushion moves between 3.9 and 4.1 mm. In the second, it swings from 2.8 to 5.2 mm. Both runs can average about 4.0 mm, but they do not represent the same process.
The first run is likely delivering nearly the same pressure-transfer condition every cycle. The second is changing the amount and compressibility of melt available between the screw and the cavity. That difference can appear as part-weight scatter, dimensional drift, sink marks, cosmetic inconsistency, or mechanical-property variation even though the controller still displays an acceptable average.
A target is a setpoint; stability is evidence. The most useful question is not whether cushion reaches a particular number. It is whether the screw reaches a repeatable end-of-hold position under repeatable material and machine conditions, and whether the molded part responds consistently.
Cushion is a pressure-transfer condition, not just a distance
On a conventional reciprocating-screw machine, molten resin accumulates in front of the screw during recovery. During injection and hold, the screw moves forward and the check ring closes to prevent the melt from flowing backward. The remaining pad of melt in front of the screw at the end of packing is the cushion.
That pad acts as a pressure-transfer bridge. As the resin in the cavity cools and shrinks, the screw must continue applying pressure through the melt pad to force additional material through the gate. If the cushion reaches zero, the screw bottoms out. The controller may still command hold pressure, but the screw no longer has a liquid column through which that pressure can be transmitted effectively.
A cushion that is unnecessarily large is not automatically safer. A large melt pad can absorb more of the screw movement through compression, delay the pressure response seen by the cavity, and increase residence time in the front of the barrel. The useful condition is a cushion that is large enough to prevent bottoming and small enough to respond predictably.
Some thin-wall or plunger-style processes are intentionally designed to run with little or no cushion. In those cases, bottoming is part of the validated process strategy, and other variables must control packing. The stability principle still applies, but the controlled signal may be transfer position, cavity pressure, or injection force rather than a nonzero cushion.
The number on the screen is only a proxy
Machine controllers do not all define cushion the same way. One controller may record the screw's farthest-forward position during fill. Another may report the screw position at the end of hold. A third may display both values under different names.
The distinction matters because the screw can retreat after the machine switches from injection to hold. Cavity pressure may push against the screw, or melt may leak backward through a partially seated check ring. The farthest-forward position and the end-of-hold position can therefore differ by a meaningful amount.
For packing analysis, end-of-hold cushion is often the more useful value because it shows how much melt remains after the pressure-transfer stage. A maximum-forward reading can look stable while the screw retreats by a different amount on every shot. If only the maximum-forward position is monitored, that variation can remain hidden.
Record both values when the controller makes them available. The difference between them, sometimes called hold-stage retreat, can reveal that the instability occurs after transfer rather than during metering. If the maximum-forward position is stable but the end-of-hold value moves, investigate the hold phase, pressure response, and sealing behavior. If both values move together, look more closely at shot preparation, recovery, material condition, and screw positioning.
Distance also does not equal a universal melt volume. The approximate cylindrical volume represented by 1 mm of axial screw movement is about 0.49 cubic centimeters on a 25 mm screw and about 2.83 cubic centimeters on a 60 mm screw. Screw-tip geometry, melt compression, and the actual position reference change the precise value, but the comparison is still important: the same 1 mm variation can represent almost six times as much melt on the larger screw.
That is why a cushion alarm copied from one machine to another can be misleading. The same numerical limit may be insignificant on one screw and highly consequential on another.
Why the average hides the failure
A single cushion number is a snapshot. Cushion variation is a time-dependent process behavior. The minimum, maximum, range, standard deviation, and sequence of the readings all matter.
A useful diagnostic baseline consists of 30 to 50 consecutive cycles taken after the mold, barrel, hydraulic system, dryer, and material have reached steady conditions. Record the cushion for every shot rather than collecting occasional readings. Then review the data as a run chart.
A process with readings of 3.9, 4.0, 4.1, 4.0, and 3.9 mm is behaving differently from one that produces 3.0, 5.0, 3.1, 4.9, and 3.0 mm, even if both have a similar average. Standard deviation helps quantify scatter, but it does not show whether the process is drifting, alternating, or shifting after a specific event. The order of the readings is part of the evidence.
A slow downward drift can indicate progressive wear or a changing viscosity condition. A random alternating pattern can indicate intermittent check-ring seating, nozzle leakage, inconsistent decompression, or unstable material feeding. A step change often points to a resin lot change, dryer event, maintenance action, or parameter adjustment. Treating all three patterns as the same problem wastes time.
Capability calculations also require caution. A process capability index calculated from a drifting process can create false confidence because the data may appear statistically narrow over a short window while the center is moving. Establish control only after the cushion behavior is stable.
Part response determines whether variation matters
Cushion position is a valuable signal, but it is not a direct cavity-pressure sensor and it is not a universal measure of part weight. A change in screw position can have different consequences depending on whether the check ring seals, how much melt remains in reserve, and how the material responds to pressure.
First confirm the controller's position convention. The examples below assume the common interpretation in which a lower reported cushion means less residual melt remains at the end of hold:
- Cushion falls while part weight and cavity pressure fall: This combination raises suspicion of lost melt, incomplete recovery, nozzle leakage, check-ring leakage, or a material-viscosity change. The screw may be moving farther forward without delivering the expected mass to the cavity.
- Cushion falls while part weight and cavity pressure rise: With a sound seal, the screw may be traveling farther forward and packing more material. The process may be producing heavier or more tightly packed parts even though the cushion number is moving downward.
- Cushion moves while part weight, cavity pressure, and critical dimensions remain stable: The change may be occurring in unused melt reserve, within the machine's positioning behavior, or in the measurement reference. It may not yet be a product problem, but the trend still deserves investigation.
- Cushion remains stable while part weight changes: Cushion is probably not the primary cause. Check material feeding, cavity balance, scale technique, moisture, regrind ratio, and other parts of the process.
These are directional clues, not component diagnoses. Part weight should be measured consistently on the same cavity and with the same runner condition. If cavity-pressure sensors are available, they provide a much stronger comparison than cushion alone because they show what the mold actually experienced.
A lower cushion reading does not automatically mean that more plastic reached the part. That assumption is one of the most common sources of incorrect troubleshooting. A worn check ring can allow the screw to travel farther forward while melt escapes backward, producing a smaller cushion and a lighter part at the same time.
Variation patterns point to different investigations
The shape of the trend should determine where the investigation begins.
A gradual widening of the cushion band over days or weeks points toward progressive mechanical or thermal change. A check ring may be seating less consistently as its surfaces wear. Screw-flight and barrel-bore clearance may be increasing. A heater band or thermocouple may be drifting, changing melt viscosity enough to expose a marginal seal.
A sudden shift that begins with a new resin lot, a change in regrind percentage, or a dryer interruption points toward material condition. Moisture, melt-flow variation, filler loading, and degradation can all change how the melt compresses and leaks under pressure without changing the controller setpoints.
A repeating up-and-down pattern from one shot to the next suggests an intermittent event rather than simple wear. Possible leads include a check ring that seats on some shots but not others, inconsistent nozzle contact, unstable decompression, or irregular pellet feeding. Averaging the readings can hide this pattern instead of solving it.
A trend that changes with startup time, injection speed, or hydraulic oil temperature points toward machine positioning and thermal behavior. Hydraulic valve response, oil viscosity, servo braking, and position-sensor performance can affect where the screw stops even when the material and mold are unchanged.
A practical cushion variation diagnosis follows this same logic: treat the reading as evidence of pressure-transfer behavior, then verify the sealing, thermal, material, and control conditions that could create it.
Why parameter changes can hide the real problem
When an operator sees the cushion falling from 4.2 to 3.2 mm, the fastest response is often to increase the metering stroke, raise back pressure, change barrel temperatures, or extend hold time. One of these changes may return the display to 4.0 mm, but that does not prove the process is healthy.
Suppose a check ring is beginning to leak. Adding more material during recovery can restore the final screw position because the extra charge compensates for some of the leakage. The leak itself remains, however, and its rate changes with injection pressure, melt temperature, material viscosity, and cycle time. The process may look acceptable during one hour and produce weight scatter during the next.
Increasing back pressure can improve melt uniformity when the original setting was too low, but excessive back pressure can cause screw-flight slippage, additional shear heating, or unstable recovery. Raising barrel temperature can temporarily reduce resistance and alter the cushion, but it may also lower viscosity enough to worsen backflow through a worn valve. These changes can move the symptom without correcting the cause.
A better practice is to preserve the validated settings, document the trend, and change one controlled variable at a time. Separate two questions:
- Can the cushion be returned to its nominal position?
- Has pressure transfer and part response become repeatable again?
Only the second question establishes a real correction.
Alarm limits must come from the part, not a generic rule
Generic guidelines such as a 3–6 mm starting cushion or a ±0.05 in. alarm window can be useful during setup, but neither is a universal acceptance criterion. The correct limit depends on screw diameter, material, mold design, machine capability, and the product's sensitivity to packing changes.
A 0.5 mm movement close to zero can be far more serious than a 0.5 mm movement in the middle of a generous cushion. The first condition may bring the screw close to bottoming; the second may simply change an unused melt reserve. The relationship between cushion and part quality is often nonlinear near the lower boundary.
Set application-specific limits through a controlled study:
- Establish a stable nominal cushion using the machine and material supplier's recommendations as a starting point.
- Measure part weight, critical dimensions, appearance, and cavity pressure across a stable baseline.
- Make small, safe changes around the nominal condition and record how the product responds.
- Identify the cushion range in which the critical outputs remain within their own control limits.
- Set a warning limit before the product limit and an action limit at the point where investigation or containment is required.
For a commodity container, a wider cushion window may have no meaningful effect on performance. For a medical housing, optical component, or safety-critical automotive part, a much narrower window may be necessary. The machine display should not decide the limit by itself.
The alarm must also be wider than the real repeatability of the measurement system. A limit of ±0.1 mm is not useful if the position sensor, reference, and controller filtering cannot reliably distinguish a 0.1 mm change from normal measurement noise.
A disciplined press-floor investigation
A stable cushion is built by making the signal trustworthy and then linking it to the product. A practical sequence is:
- Define the measurement. Record whether cushion means maximum forward position, end-of-hold position, or another controller value. Confirm the sign convention and reference point.
- Collect an unchanged baseline. Run enough consecutive cycles after thermal and material stabilization to reveal random scatter, drift, and repeating patterns. Avoid adjusting the process while collecting the baseline.
- Capture companion data. At minimum, record cushion, transfer position, peak injection pressure, actual hold pressure, recovery time, cycle time, part or shot weight, and critical dimensions. Add cavity pressure, actual barrel temperatures, hydraulic oil temperature, resin lot, dryer condition, and regrind percentage when available.
- Compare the signal with the product. Determine whether cushion movement tracks weight, cavity pressure, dimensions, or cosmetic defects. This separates a meaningful pressure-transfer change from a position change that has not yet affected the part.
- Follow the trend pattern. For progressive drift, inspect the check ring, screw tip, barrel bore, nozzle seat, heater hardware, and position system. For a step change, review material, dryer, maintenance, and parameter history. For random alternation, investigate intermittent sealing, feeding, decompression, and measurement behavior.
- Verify the suspected cause. Do not replace a check ring solely because the cushion moved. Use an appropriate leak test, inspection, dimensional measurement, or controlled material and temperature check to confirm the hypothesis.
- Prove the correction over time. After a repair or parameter change, repeat the same data collection and confirm stable cushion and product response across enough cycles to cover normal production conditions. One good shot is not proof of a restored process.
If a conventional process reaches zero cushion, parts from the affected period should be contained until packing performance is verified. A zero reading may be acceptable only when the process was deliberately designed and validated to run that way.
The number to protect is repeatable pressure transfer
A stable 5 mm cushion can produce better parts than a variable 4 mm cushion. The correct target is the value that leaves enough melt to transmit hold pressure while keeping the screw position, cavity response, and part characteristics repeatable.
The strongest process-control question is therefore simple: does the screw reach the same end-of-hold condition, and does the part show the same weight, dimensions, and cavity pressure afterward? If the answer is no, an average cushion value is only masking the problem. Cushion size establishes a working point; cushion stability reveals whether the machine is still capable of holding it.