
LPDC Pressure Curve: Filling, Feeding and Release
- Serdar Perçin

- Jul 27
- 6 min read
Short answer: an LPDC pressure curve should not be copied from another die, wheel size or furnace. It should be built from four physical duties: lift the melt through the stalk, fill the cavity with a stable front, maintain feeding while the relevant path remains open, and release without pulling liquid metal back from a still-sensitive region. The first useful calculation is not “how many millibars should we add?” but the pressure needed to overcome the metal-height difference, plus verified system losses. The curve is then tuned with actual fill time, die-temperature balance, vent behavior, part weight and defect location. If one of those signals changes, a curve that worked yesterday may no longer represent the same metal flow today.
Why the pressure curve is a metal-flow program
In low-pressure die casting, gas pressure applied above the bath displaces molten metal upward through the stalk and runner into the die. At the simplest level, the minimum pressure must overcome the metallostatic head: pressure rises with melt density, gravity and the vertical distance between bath surface and the highest liquid-metal point. Real equipment also adds pressure losses through the stalk, filters, bends, constrictions, leakage and dynamic acceleration. That is why a pressure value without metal level and geometry is incomplete.
The shape of the ramp controls how quickly the free surface moves. Too slow can produce excessive heat loss, hesitation and premature freezing. Too fast can create wave collision, splashing, bifilm entrainment and gas trapping. Research using water models, numerical flow simulation and experimental validation shows that free-surface behavior—not simply average fill time—matters. Studies of low-pressure filling also identify falling jets and rotational vortices as important oxide-film entrainment mechanisms. The engineering target is therefore a continuous, calm and adequately warm front, not the highest possible filling speed.
The four stages and what each one must achieve
1. Lift and approach
The early pressure rise brings metal from the bath toward the gate. It must follow changes in bath level. If the furnace level falls but the same initial pressure and timing remain, the metal may reach the cavity later even though the displayed recipe is unchanged. Conversely, an aggressive early rise can make the metal strike transitions or enter the cavity before vents and thermal conditions are ready.
2. Controlled cavity filling
The filling ramp controls front velocity through the runner and part. A single linear slope may be inadequate for a geometry that changes rapidly from hub to spokes or from thick bosses to thin walls. Breakpoints should correspond to physical events—reaching the gate, entering a larger section, splitting into branches or approaching the last-to-fill zone—not arbitrary percentages of machine time. Measured fill time and simulation can locate these events, but casting evidence must confirm them.
3. Feeding and solidification hold
After filling, pressure supports contact and feeding while liquid paths remain available. More peak pressure cannot feed an isolated hot spot after its neck has frozen. Holding longer also cannot repair a disconnected liquid pocket. The relevant question is whether the gate, runner and local feeding path remain permeable during the hot spot’s contraction. Thermal mapping, section thickness, cooling timing and defect morphology must be read together.
4. Release and return
Pressure should be released only after the casting can retain its metal without harmful backflow or loss of feeding. Releasing too early may contribute to shrinkage near the gate or a pressure-sensitive region. Holding unnecessarily long increases cycle time and may complicate ejection or furnace control without improving quality. The release point should be justified by solidification evidence, not habit.
Symptom-to-action diagnostic table
Symptom | Likely mechanism | Data to check | Safe first action |
Cold shut or incomplete thin section | Front loses temperature or continuity before meeting | Metal and die temperature, true fill time, venting, last-fill location | Restore thermal stability first; then test one bounded ramp change |
Oxide folds or scattered leak paths | Unstable surface, falling jet or vortex entrains films | Transition geometry, ramp breakpoints, simulation/video evidence, filter condition | Reduce the abrupt acceleration at the implicated transition |
Gas porosity concentrated at the last-fill zone | Air cannot escape or is enclosed by converging fronts | Vent condition, front meeting point, cavity pressure, fill balance | Correct vent and flow balance before increasing hold pressure |
Shrinkage near a thick hub or boss | Local hot spot becomes isolated from the pressured feed path | Gate freeze time, local modulus, cooling sequence, hold duration | Verify feeding-path openness; do not treat peak pressure alone as the cure |
Part weight or fill time drifts across the shift | Metal level, leakage, temperature or machine response has changed | Bath level, measured pressure, cycle temperature trend, sealing and stalk condition | Restore the baseline hardware and thermal state before retuning the recipe |
A practical validation sequence
Freeze the measurement definition. Decide exactly how fill start, cavity full and pressure hold are detected. A controller timestamp and the physical event may differ.
Record metal level and temperature. The same command pressure does not represent the same lifting condition when bath level or density changes.
Compare commanded and measured pressure. Lag, overshoot or oscillation means the actual process differs from the recipe.
Map the defect to the filling sequence. Identify whether it lies near the gate, a front split, a meeting line, a vent or a hot spot.
Check die thermal balance. Use stable-cycle temperature data, not only the first shot or a single handheld reading.
Change one curve segment at a time. Define the expected mechanism, the bounded change, the quality metric and the stop condition before the trial.
Validate with more than visual inspection. Use part weight, X-ray or CT, leak test, sectioning and mechanical testing as appropriate to the risk.
A useful trial matrix separates lift pressure, fill-ramp shape, hold level and release time instead of moving all of them together. It also records furnace level, metal treatment, die temperature and cooling state. Otherwise a “successful pressure change” may only be the result of a warmer die or a different bath level.
Limits and safety
Pressure limits, furnace integrity, stalk condition, die locking, venting and molten-metal procedures take priority over optimization. Do not bypass machine interlocks or exceed equipment and tooling limits. A curve that improves one defect may worsen oxide entrainment, die erosion, flash, cycle stability or another region of the casting. For safety-critical components, parameter release must follow the plant’s control plan, traceability rules and required validation.
How Castella helps
Castella AI for Die Casting is a mobile, AI-supported engineering decision tool for HPDC, LPDC and sand casting. It organizes alloy, geometry, process and thermal inputs, then proposes parameter and troubleshooting directions that an engineer can review. For an LPDC pressure-curve problem, Castella can help connect defect location, wall thickness, feeding path, metal temperature, pressure stages and cooling sequence so the team tests a mechanism rather than guessing. It requires no new sensors, but it does not replace engineering validation, safety procedures or detailed simulation.
For a broader symptom-to-parameter reference, use the Casting Defect Troubleshooting Guide. Related LPDC solidification guidance is available in Advanced Strategies to Reduce Shrinkage Cavity in LPDC.
Frequently asked questions
What is the correct LPDC pressure curve?
There is no universal curve. The correct curve must overcome metallostatic head and line losses, fill the cavity without unstable free-surface motion, maintain feeding pressure through the relevant solidification window, and release only after the gate or feed path is sufficiently solid.
Should I raise pressure when an LPDC casting has a cold shut?
Not automatically. First check metal and die temperature, fill continuity, venting, transfer distance, and the exact position of the cold shut. A faster ramp may help a thermally marginal front, but it can also increase turbulence and oxide-film entrainment.
Which signal is most useful for validating the curve?
Use synchronized evidence: commanded and measured pressure, actual fill time, metal level, die-temperature trend, vent behavior, part weight, X-ray or CT location, and defect timing. A pressure trace alone cannot prove how the cavity filled.
How long should holding pressure remain active?
Long enough to support feeding while the relevant path is liquid or semi-solid, but not as a fixed plant-wide number. Gate section, local modulus, alloy, die temperature and cooling determine when pressure can stop influencing the hot spot.
Can AI determine LPDC pressure settings without trials?
AI can rank likely causes and propose bounded changes from alloy, geometry, thermal and process data. Final settings still require controlled trials, equipment limits, safety procedures, and engineering validation.
Technical sources
Process optimization of an A356 wheel hub by LPDC using simulation and experimental coupling.
Numerical simulation and experimental validation of free-surface flow in low-pressure casting.
Mold filling and oxide-film entrainment in low-pressure casting.
Solidification optimization using numerical simulation and machine learning.




Comments