How to Design Injection-Compression Molds: Clearance and Force

A 620 mm PC/ABS appliance bezel came off T1 with 4 mm of bow across the diagonal and a visible sink line over every boss. The same pack pressure that filled the far corners, about 120 MPa, was the force that loaded the distortion into the part. The program lost two weeks to relief cuts and process chasing. The tool was then rebuilt for injection-compression: same cavity steel, a designed air gap, and peak cavity pressure held to roughly 25 MPa. That one change cleared both defects, and this guide covers the tool architecture it depends on: underfill, stop blocks, vent clearance, the compression face, machine control and the defect map.

The Snapshot

• Underfill to 80 to 95 percent. The shot stops before the cavity fills, then a compression stroke of 0.1 to 0.5 mm closes the remaining air gap around the melt instead of forcing it through the channel.
• 15 to 40 MPa peak cavity pressure. A conventional pack on the same part runs 80 to 150 MPa, so orientation stress and molded-in shrink differential fall by half or more.
• 0.02 to 0.10 mm of vent clearance. Stop blocks hold the open gap; the compression face lands on that gap at full close and defines the final part thickness.
• Flatness inside 0.5 mm per metre. The repeatable result on door trims, appliance bezels and battery covers, with clamp tonnage down 30 to 50 percent against the conventional layout.

Where Conventional Packing Falls Short

On a large flat part, the mold is not defeated by the fill. It is defeated by the pack. To push melt to the end of a 600 mm flow path, a conventional tool packs at 80 to 150 MPa. That pressure stretches polymer chains into orientation while the skin freezes against the steel.

Orientation is stored energy. The part releases, looks fine on the bench, then the molecules relax over 24 hours and the panel bows. The thin-wall version of that failure chain is covered in the thin-wall warpage guide. Simply lowering pack pressure is not available on a glass-filled panel, because the far corners go short.

The second limit is the press. Clamp force has to resist full injection pressure across the projected area. A bezel with 0.6 square metres of projection at 110 MPa cavity pressure pulls roughly 6,700 tonnes of clamp demand before any safety margin. The machine is bought for the pack peak, and the part is then designed around it.

The third limit is the filler. Glass-filled melt shrinks differently in the flow direction than across it by 0.3 to 0.8 percent on many grades. That differential is worst where pack pressure was highest. Ribs and bosses sink because the last melt to freeze sits far from the gate, and holding pressure has to travel to reach it.

The Mechanics: Underfill and the Compression Stroke

Injection-compression starts the shot as a controlled short shot. The cavity is filled to only 80 to 95 percent of its volume at switch-over. A layer of air sits between the melt and the closing face.

A compression stroke of 0.1 to 0.5 mm then closes that air gap around the melt. The moving half squeezes the material up, down and outward into the remaining volume. Velocity stays low, typically 1 to 10 mm/s, so nothing is forced through a narrowing frozen channel.

Because the melt is displaced rather than forced, peak cavity pressure lands at 15 to 40 MPa. The conventional pack peak for the same part would sit at 80 to 150 MPa. Published process comparisons on large flat parts show orientation stress and shrink differential falling by half or more under that pressure profile.

Pressure also equalizes. A conventional pack leaves a gradient from gate to end of fill; compression loads every melt surface the same way at the same time. The part is replicated from the steel rather than stretched to reach it.

Hold the stroke until the gate freezes, then release. There is no second pressure ramp and no long pack profile. The part expands against the steel as it cools, and the shrinkage it would otherwise pull as a sink is compensated mechanically before release.

What Injection-Compression Buys

The first purchase is flatness. Door trims, appliance bezels and battery covers molded in injection-compression hold flatness inside 0.5 mm per metre, as a repeatable process result rather than a hand-fitted best shot. The mechanism is low residual stress plus a part that cools against uniform pressure instead of a gradient.

The second is shrink control on glass-filled grades. When the compression face arrives before the skin freezes, ribs and bosses stay at nominal without a sink mark. The flow-to-cross-flow shrink differential then moves toward the isotropic behavior of the unfilled resin.

The third is the press. The fill event is deliberately stopped at 80 to 95 percent, so flow length stops mattering at switch-over. A thin-wall section of a large part is completed by the stroke instead of by injection force. That is how a panel needing 3500 tonnes of clamp runs on a 1600 to 2500 tonne machine. Clamp demand drops 30 to 50 percent against the conventional layout.

The fourth is dimensional repeatability. Cavity pressure at 15 to 40 MPa deflects the tooling far less. That deflection is the silent variable behind shot-to-shot size drift on panels over 400 mm. A narrower pressure band shows up directly in the capability study at sampling.

The Air Gap Is the Design: Stop Blocks and Vent Clearance

In a conventional tool the cavity is the mold. In injection-compression, the cavity plus the air above it is the mold. The shut gap is 0.02 to 0.10 mm, set by stop blocks, not by clamp touch. Zero gap makes a conventional pack; too much gap lets part thickness drift with shot size.

Stop blocks are hardened pads on the fixed side that the moving platen face or a dedicated compression plate lands against. They carry the compression load and they define closed height. Their parallelism is what the part sees, so grind them to ±0.005 mm and make them shimmed or replaceable. A block that peens 0.02 mm over 200,000 shots adds 0.02 mm to every part.

Vents live in that gap. The open cavity parting line is itself the vent, which is why the process tolerates glass-filled resins that burn in a shallow land vent. The constraint is timing: melt must not bridge the vents before the stroke lands. If the land seals while the far side of the panel is still flowing, gas is trapped and the compression welds it into the part as a void or a burn.

Size the air gap from the volume shortfall. On a 620 by 420 mm panel with a 2.5 mm wall, filling to 90 percent of cavity volume leaves a gap of roughly 0.25 mm at switch-over. The shut gap of 0.05 mm at the stop blocks is the difference between underfill and closed position, and the controller commands the stroke distance.

Compression Face and Self-Centering Geometry

Put the compression face on the core side. When the moving half carries the compression plate and the fixed half the finished cavity, every polish, repair and texture decision stays on accessible steel, and the part releases toward the face that just replicated it.

The closing geometry is a taper, not a wall. A self-centering draft of 5 to 10 degrees on the compression land lets the moving half find home at full close even when the press arrives with a small positional offset. A vertical land instead collides with frozen flash at the first drift: a 0.1 mm chip at the land holds the plates 0.1 mm apart and the air gap is gone.

Keep the compression land wide enough to survive. A 6 to 10 mm land on hardened steel at HRC 50 to 54 holds its grind for hundreds of thousands of cycles; a narrow land wears into a radius and the vent function changes with it. The land is simultaneously the seal at shut position, so it follows the same shut-off discipline as a flash-critical conventional tool, with the difference that it only contacts at the end of the stroke.

Reserve clearance for the melt film. During compression, a thin layer is squeezed outward across the taper, and the angle plus the stop blocks let it thin out and break rather than weld a fin. Cores that the plate passes over carry the same taper and hardness.

Plate Stiffness, Support Pillars and Guide Wear

Cap plate deflection at 0.01 mm under full compression load. The part reads the plate gap as its thickness, so a plate that springs 0.05 mm at the center of a 600 mm tool prints a thick center and thin edges into every shot. Verify with a deflection calculation on the actual pillar grid before ordering steel.

Support pillars on a 100 to 150 mm grid carry the compression load from face to bolster instead of through the clamp bolts alone. On a large tool this is where the money goes: more pillars, more drilling, and a template that keeps them off every cooling circuit.

The stroke also changes wear. A conventional tool guides its halves at light load; an injection-compression tool drives the moving half through a loaded, short, repeated stroke hundreds of thousands of times, and soft bushings smear. Use through-hardened guide posts, oversized relative to a conventional layout, and inspect them earlier than the cavities.

Parallelism of the press platens is now a mold tolerance. On tie-bar-less machines the platen can bow under load by hundredths of a millimetre, and the compression face copies that bow into the panel. Specify platen parallelism at 0.02 mm or better across the tool area and verify it with a dial indicator at every tool change.

Machine and Control Requirements

Injection-compression needs position, not just pressure: a clamp programmable to stop at a commanded distance and hold it. In practice that means an electric or hybrid toggle-less clamp with closed-loop control. A legacy hydraulic toggle machine either slams the stop blocks or creeps short, and both failures show in the parts.

Two control architectures make the process repeatable. The direct route is a closed loop on cavity pressure: a flush transducer reads the 15 to 40 MPa window and the controller starts and shapes compression against pressure, not against a timer. The wiring and placement discipline is covered in the cavity pressure sensor guide. The proven alternative is a force or position profile qualified over thousands of shots.

Compression velocity runs 1 to 10 mm/s. Slow enough that the melt flows as a sheet with the gas ahead of it; fast enough that the stroke lands before the skin chills. On amorphous resins with a narrow softening band, the top of that range needs a hot mold.

Sizing changes too: injection duty and clamp duty drop, but position control and platen area stay mandatory. The buyer walks the same trade-offs as the mold-to-machine matching guide, with compression distance added to the checklist.

Building the Process Window

Start from the fill. Set shot size to reach 80 to 95 percent at switch-over, verified by weight or a pressure trace. Then set the compression start at or just before transfer, while the melt is still mobile across the whole cavity. A stroke that waits for the runner to freeze arrives against a skin and does nothing for the far corners.

Hold the closed position until gate freeze, then release. Gate freeze is a cooling-time problem with a compression twist: because cavity pressure is low, the part starts freezing the moment the stroke completes, and the hold time is set by the thickest section at the gate. The cooling time calculation guide covers the wall-thickness math; add the stroke duration to the result and that is the compression cycle.

← swipe to scroll →

Table note: bands are typical industry values for large flat thermoplastic panels on position-controlled presses. Confirm each one against the resin datasheet and cavity-pressure traces recorded at the trial.

Then tune vent timing in 2 percent fill steps while watching the end-of-fill corner. A blackened corner means the land sealed before the stroke landed; voids under the compression face mean the melt bridged the gap early. The fix is fill speed or mold temperature, never a deeper gap.

Defect Map From the Tooling Side

A stroke that cannot finish the part means compression started too early: the plate closes against a skin too stiff to flow and the far corners stay empty at full shut height. Fix the start time first; on the tool side it is often a cold spot, so pull a cooling line back or zone the control there.

Flash at the compression land means the gap was overpassed: stop blocks peened to a smaller shut height, or plates deflecting open under stroke. Measure the fin and work back to the gap that made it, the same as on a conventional tool.

Sink over bosses returns when compression arrives after the skin freezes; the stroke then clamps a solid shell while the melt at the boss keeps shrinking. Push the start earlier or raise mold temperature at the feature, and check that switch-over part weight is inside the fill band.

Swirl marks on glossy faces come from fast compression across a chilled surface. The moving face drags the skin radially before it re-wets. Slow the stroke at the last 0.1 mm, raise the face temperature by 5 to 10 degrees, or texture the compression plate on non-cosmetic tools so nothing can mirror a defect.

Weld-line weakness disappears as a family. Fill fronts never crash into each other, so the classic V-notch over a hidden rib stops existing, and the merge line is a reknit under 40 MPa contact rather than a fold under 150 MPa.

ICM Versus Conventional and Gas Assist

The three large-part processes solve the same money problem in different ways. Conventional buys fill with pressure. Gas assist buys it with an internal void. Injection-compression buys it by deleting the pressure requirement, which is why it wins wherever flatness, sink and glass orientation dominate the drawing.

← swipe to scroll →

Table note: comparison bands are typical published and industry values for large flat parts on engineering resins. The tooling premium for injection-compression is counted against the same cavity steel, not a larger tool.

Gas assist wins on hollow structural sections where mass saving pays; injection-compression wins on solid sheets where nothing may be hollow. A formal moldability review settles the choice before either tool is quoted.

When ICM Pays for Itself

The tooling premium is 10 to 20 percent against a conventional layout of the same cavity steel: stop blocks, the compression plate, harder guides, more drilled pillars. That is the whole list at the tool. The press savings start at 30 to 50 percent on clamp force, and on a new program the machine purchase often dwarfs the premium by an order of magnitude.

The scrap arithmetic decides the marginal case. When bow and sink together cost 3 to 8 percent of output on a Class A surface, the process that removes both pays the premium inside weeks. Programs with visible faces, mating flanges or battery covers that must seal to a gasket across 500 mm have no cheaper path to those numbers.

It does not pay on small thick parts, on tools already packed with unscrewing actions, or on shops without a position-controlled press. If the machine is not there and the part is neither flat nor cosmetic, a well-packed conventional tool is the right answer.

DieStrike flags panel programs that qualify for injection-compression inside the DFM pass, while the steel is still a drawing. Send the part data and the flatness requirement with the RFQ and the tooling concept comes back with the air-gap scheme already on it.

Injection-Compression Mold Design FAQ

Q1. What air gap does an ICM tool need?

0.02 to 0.10 mm at the shut position, sized from the volume shortfall between 80 to 95 percent fill and the full cavity. The gap is the vent, the gas escape path and the last material movement of the cycle, so it is a ground dimension, not a tolerance leftover.

Q2. Can I convert an existing conventional tool to injection-compression?

Often yes: the cavity steel usually survives, and the work is stop blocks, a core-side compression face with a 5 to 10 degree taper, and a hardened guide set that accepts the loaded stroke. The fixed condition is machine-side: a position-programmable clamp.

Q3. Why does my part still bow on an ICM tool?

Check the start time first: compression after transfer leaves the orientation profile of a conventional pack. Then check plate deflection against the 0.01 mm target and stop-block parallelism; a 0.03 mm bowed gap prints a bowed panel.

Q4. Does injection-compression work on glass-filled nylon and PBT?

Yes, and semi-crystalline grades benefit most because the stroke compensates crystallization shrink mechanically instead of with 100 MPa of pack. Keep fill at the low end of the 80 to 95 percent band on fast-crystallizing resins so the stroke lands while the melt is still mobile.

Q5. How much press tonnage does the process save?

30 to 50 percent on large flat parts. The clamp resists 15 to 40 MPa of cavity pressure plus the closing force of the stroke, instead of an 80 to 150 MPa pack peak across the full projected area.

The Bottom Line

Injection-compression is not a molding trick; it is a tool design that moves the thickness decision from melt pressure to ground steel. The numbers that carry it: underfill to 80 to 95 percent, a 0.1 to 0.5 mm stroke landing on stop blocks that hold a 0.02 to 0.10 mm gap, peak cavity pressure at 15 to 40 MPa, a 5 to 10 degree self-centering taper on a core-side compression face, plate deflection capped at 0.01 mm on a 100 to 150 mm pillar grid, and a position-programmable clamp running the stroke at 1 to 10 mm/s to gate freeze.

Hold those and large flat parts come back inside 0.5 mm per metre, sink-free at the bosses, on 30 to 50 percent less clamp tonnage, for a 10 to 20 percent tooling premium. Miss them and the process is a slow short shot with extra stop blocks.