How to Design an Unscrewing Mold: Motor, Rack and Stripper Action

A 32-cavity connector housing tool with M6 molded threads ran on collapsible cores. When volume passed 500,000 shots a year, the slotted sleeves took set at 0.08 mm bore growth. Each collapse cycle also left a spiral witness line no connector buyer accepts. An unscrewing mold rotates the core out of the thread instead of crushing around it, and this guide covers the motor, rack and stripper mechanics behind that release.

The scope is the tooling mechanism. Resin-side choices for molded threads sit in the molded thread guide, and every figure below needs checking against the thread spec and machine data before a drawing releases.

The Snapshot

• Unscrew above 1.5 diameters of thread depth. Past that engagement a collapsible core either marks the thread or fatigues in under 500,000 cycles, while a rotating core strip-counts to millions.
• Release torque runs 5 to 30 N-m. It depends on thread form, pitch and shrinkage grip, and the mold must carry that reaction through guide posts and stripper bushings, never through the ejector plate.
• Rotation speed is 1 to 3 turns per second. On a 45-second cycle a 2-turn thread releases in 0.7 to 2.0 seconds, which fits inside mold-open time only when the drive is sequenced correctly.
• Expect 30 to 80 percent more move work. That is the cost delta of an unscrewing tool over the equivalent cored or collapsible build, paid for in ground drive fits at 0.05 to 0.1 mm clearance.

When Unscrewing Beats a Collapsible Core

Four conditions push a thread out of collapsible territory and into a rotating core. Thread depth above 1.5 times the nominal diameter is the first. A slotted core collapses inward a fraction of a millimeter per cycle to release that much helix, and the slot roots see alternating stress every shot.

Slot fatigue shows as bore growth: 0.03 to 0.08 mm of permanent set that widens the thread root and loosens every part downstream. A continuous helix is the second condition. Fine-pitch threads below 1.5 mm pitch cannot be formed around a split because the segments smear the flank. The cosmetic requirement is the third: a collapse witness is a spiral scar that connectors, lens barrels and fittings cannot ship with.

Volume is the fourth. Above roughly 500,000 shots the extra move work amortizes away: the rotating core runs to millions of cycles, while a slotted sleeve, its replacement set and the bore-growth scrap do not.

Below those thresholds, collapsing or core breaking stays honest. A 0.5-turn lift thread on a solvent tank cap released at 200,000 shots a year does not deserve a gear train, and DFM should say so on the report rather than upsell the mechanism.

Motor-Driven Rotating Cores

The motor drive puts a compact electric or hydraulic motor under or behind the mold base and gears it to the unscrewing cores. It owns the long-thread territory: engagements of 2 turns and up, 10 mm and deeper, and any program that needs the release torque set precisely instead of inherited from ejector stroke geometry.

Electric core-drive units deliver 2 to 20 N-m continuous on a typical 16 to 32 mm bore tool, with 1 to 3 turns per second on command. The controller holds the core in the cavity during injection, then reverses it at mold open and stops at a programmed angle so the thread unwinds in 0.7 to 2.0 seconds.

Hydraulic rotary units answer when one core needs 25 to 30 N-m, for example an M24 glass-filled PA66 thread where shrinkage grip peaks there. They need oil, seals and plumbing space, and the leak paths around a shaft in 90 to 120 °C oil are why shops reach for electric units first.

Mounting goes through a boss in the A or B plate, never clamped to the ejector housing. At 80 MPa cavity pressure the axial thrust on an M8 thread core reaches several kilonewtons, so a hardened thrust washer and a 620-class bearing belong between the drive and the moving steel.

Service design matters too. Pull the unit without demounting the core, keep connector plugs on the operator side, and dowel-index the motor so a swap returns phase within 1 degree. Standard drive components ship in 3 to 7 days from stock, which keeps a failed unit a maintenance event.

Where the thread sits on a side axis, the motor drives a bevel or rack linkage instead of a straight shaft. Rotation then waits for the mold-open position signal, the same sequencing rule as the hydraulic core pull guide.

Rack and Pinion Strip Plates

The rack drive converts the ejector stroke into core rotation. Pinions sit under the rotating cores, a strip plate carrying the rack teeth travels with the ejector system, and 40 to 60 mm of ejector stroke turns a 40 to 60 mm pitch diameter pinion through the 360 to 720 degrees that a 1 to 2 turn thread needs.

Geometry sets the ratio. A pinion with a 50 mm pitch diameter rolling a rack through a 50 mm stroke turns its core almost exactly one revolution, 360 degrees per 3.14 diameters of travel. Fine-pitch threads that need 1.5 turns therefore need either a smaller pinion or a stroke multiplier from a secondary rack stage.

Speed follows the ejector profile: at 150 to 250 mm per s advance a rack drive spins a core at 1 to 3 turns per second with no controller at all. That mechanical honesty keeps rack drives the default on short-thread closures, where release finishes before the part is visible.

The cost is rigidity. Rack teeth carry reaction torque through the strip plate, so they need ground profiles, HRC 58 to 62 hardness, and back-pressure that keeps the rack seated on the pinion through both strokes. A rack that unloads at stroke end lets the core hammer its stop, and a chipped tooth breaks the phase of every core behind it.

Rack drives tie release distance to ejector stroke, which couples thread design to the whole ejection layout. Check that coupling against the ejector system guide before committing the stroke.

Broken Beam and Torsion Release Actions

The broken-beam mold lets the part drive its own release. Interlocking beams around each rotating core hold the cores in phase while the plastic grips them. As the mold opens and the ejector advances, the part grips the beams, the plastic thread shears its tiny bond to the core tip, and continued ejection travel spins the assembly through 0.5 to 1.5 turns until the thread walks free.

Torsion spring versions store energy differently. A spring-loaded strip plate rotates the core backward through 30 to 90 degrees while the mold opens, then releases at a defined ejector position and the spring unwinds into the thread. The core returns ready for the next shot on the ejector return stroke, with no external drive at all.

These actions live in the low-torque band: 2 to 8 N-m, short pitch, and free-flowing resins like PP or HDPE where shrinkage grip stays mild. Ask an ABS or POM part that shrinks 1.5 to 2 percent onto a deep thread and a broken beam will stall half a turn out and shear the crest.

Their virtue is compactness. No motor, no oil, no rack length: a broken-beam cluster fits cores on 20 mm pitch spacing where a pinion would collide with its neighbor, which is often why a 48-cavity closure layout survives on the chosen platen.

Unscrewing Timing Inside the Cycle

Release timing has one hard rule: the thread may only start turning while the part is still gripped by the core or supported against a stripper. Rotation against a loose part does not unscrew the part, it screws the part around the core and welds a witness line into the crest.

On motor drives the sequence runs: mold opens to the full-open signal, drive starts, cores reverse 1 to 2 turns in 0.7 to 2.0 seconds at 1 to 3 turns per second, then the ejector strips the released part. Long threads rotate before ejector motion; short threads overlap it.

Rack drives self-sequence because the rotation is the ejector stroke. At 200 mm per s advance a 1.5-turn release occupies 0.35 to 0.45 seconds of the roughly 2 to 3 seconds of open stroke. Keep the unscrew inside the first half of ejector travel so the last millimeters cushion.

Broken-beam timing belongs to the stripper plate. Ejector advance must start the spin within about 2 mm of travel; later and the beams lose grip before the core clears. The fix is usually ejection force balance, checked against the mold spring selection guide.

Machine-side, confirm the press exposes mold-open position signals to the mold controller with under 0.5 seconds of latency. A drive that starts on a timer instead of a position signal either rotates into a still-closed mold or releases so late the thread shears.

Core Drive Geometry and Stripper Bushings

The rotating core carries two jobs: it forms the thread, and it survives being driven. The drive interface is either a spline or a square. Square drives are simple and handle the highest torques, but they concentrate stress at the corners and gall if the fit runs tight. Involute splines on a 0.05 to 0.1 mm clearance fit rotate sweeter and stay serviceable through millions of strip cycles.

That 0.05 to 0.1 mm diametral clearance is the whole art. Tighter, and thermal growth of the core at 80 to 120 °C seizes the drive mid-cycle. Looser, and the core tilts under injection thrust, the thread runs eccentric, and pitch error shows as first-fit torque spread outside the gauge band.

The stripper bushing is the second wear pair. It grips the core below the thread and strips the part at the end of rotation on the same 0.05 to 0.1 mm clearance in a hardened bore. Put it in wear steel at HRC 58 to 62, grind the bore to +/-0.002 mm, and make it replaceable on a shoulder with two cap screws.

Thread relief at the core tip decides whether the part finds the helix or climbs it. Leave 0.2 to 0.5 mm of unthreaded land at the tip so the stripper edge contacts steel, not plastic. A part stripped against a live thread crest gets a 0.1 mm gouge that fails the cosmetic A-face spec at the next sampling.

Surface finish on the flanks pays for itself. Polish the thread form to Ra 0.2 to 0.4 micrometres and the release torque drops 20 to 30 percent against an EDM surface at Ra 1.6, because the melted plastic no longer keys into the micro-roughness. That finish difference is often the gap between a 5 N-m drive and a seized 20 N-m one.

Torque and Counter-Force Paths

Release torque is not one number. Take an M10 x 1.5 thread in 30 percent glass-filled PA66 shrunk onto the core: the helix friction at a 0.14 equivalent coefficient plus shrinkage grip of 30 to 60 MPa pushes the breakaway torque toward 20 to 25 N-m. The same thread in PP sits at 6 to 8 N-m. Design the drive for the resin at the top of its shrinkage band, not the datasheet average.

Newton's third law then needs a landing zone. Every N-m applied to unscrew the part reacts somewhere in the mold, and the only acceptable path runs through guide posts, support pillars and the mold base itself. A motor unit bolts to a boss that carries reaction into the A-plate; a rack drive carries it through the strip plate guideways into the B-plate.

Never route reaction through the ejector plate. An ejector plate is 25 to 40 mm of steel on guide pins built for axial force; a 25 N-m torque applied there twists the plate 0.05 to 0.15 mm across a 300 mm span, pins bind in their holes, and every return pin starts to wear oval inside a month.

Counter-force also appears as axial separation. Unscrewing a tapered thread generates an axial pull of 3 to 8 kN per core that tries to withdraw the core from the cavity. Retain the core with a shoulder and a hardened wear plate, and verify the retention against the peak drive torque, not the running torque, because breakaway spikes to 1.5 to 2 times the steady number.

Seal and lubrication discipline closes the loop. Where a rotating shaft crosses an oil gallery or a water line, use a shaft seal rated 150 °C minimum and route lubrication to the drive through a felt-loaded groove rather than a grease cup that migrates into the cavity and repels plating on the next paint line.

Multi-Cavity Phase Control

One core can stop at any angle. Thirty-two cores must stop at the same angle, or the stripper plate meets thread crests instead of core shoulders and shaves plastic off half the cavities every cycle. Phase control is what separates a single-spindle mechanism from a production unscrewing mold.

Geared train layouts solve it by construction. A common drive shaft through all pinions fixes relative phase through gear mesh, with tooth spacing inside 1 to 2 degrees across the row. Support the shaft every 100 to 150 mm so torsional wind-up under 25 N-m stays under half a degree.

Geneva and cam layouts solve it by timing. A 4-slot Geneva indexed by the mold-open stroke advances each core through 90-degree steps. Step angle and thread pitch must multiply out to a whole number of turns at release, a constraint that belongs on the drawing, not discovered at T1.

Independently motorized cores need a software interlock: start all drives on the same open-position signal and abort on any torque spike, because unwinding past one stalled core shears its thread while thirty-one parts release into the reject bin.

Verify phase the same way at the bench and at the press. Rotate the mold by hand through a full open-eject-return cycle with no plastic and watch every core pass a dial indicator set within 0.05 mm of the stripper land. Anything worse is a gear mesh or shim problem waiting to become a thread shear.

Failure Patterns in Unscrewing Tools

Late release shears threads. The signature is a crest torn on one side, appearing at the same clock position on every cavity: the rotation started after the part left the stripper support. The fix is sequencing, not a faster motor: rotation must begin while the part is still constrained axially.

Galling kills spline drives. It shows as a shiny wipe band on the spline flanks plus a 30 to 50 percent torque creep over 50,000 cycles. The cause is almost always clearance under 0.05 mm combined with a same-steel pair. Run a DLC or nickel-plated core against a different hardened steel in the hub, and check the fit against +/-0.005 mm mold tolerance with the wear allowance included in the print.

Stripped rack teeth come from torque spikes at stroke end: the core reached its stop while the ejector was still driving. Cushion the end of travel, set a slip-torque limit at 1.3 times the calculated release maximum, and inspect the first two teeth of every rack at 100,000-cycle intervals. Those two teeth carry the phase error of the whole row.

Wear of the stripper bushing bore is the slow failure. At 0.15 mm of diametral wear the core tilts, thread eccentricity climbs past the gauge band and first-fit torque variation widens. Measure the bore every 500,000 cycles; the bushing is a 3 to 7 day replacement, which is why the design makes it the only wear item.

Hot-runner interference is the surprise failure: a drop tube sitting 1 to 2 mm from a rotating core looks clean in CAD and chatters off in production. Keep 5 mm of clear radius around every rotating feature and check it against the selected valve gate envelope.

Decision Table: Unscrew, Collapse, Break or Assemble

The table below is the DFM shortlist we work from. It assumes continuous production on a standard two-platen or toggle machine and a thread that must meet its gauge band out of mold.

← swipe to scroll →

Table note: bands are shop-typical for threaded parts in semicrystalline engineering resins on 40 to 80 second cycles. Re-weight them against your resin shrinkage, gauge spec and annual volume before the tool commit.

Cost Delta and Program Planning

An unscrewing tool costs 30 to 80 percent more than the equivalent cored or collapsible build, and almost all of it lives in the moving half. The fixed cavity and core forming surfaces are identical; the delta buys drive shafts, strip plates, spline hubs, stripper bushings, bosses machined into the mold base, and the grinding time that holds 0.05 to 0.1 mm clearance fits inside +/-0.005 mm mold tolerance.

Schedule absorbs part of it too. Rotating-core fitting adds 3 to 5 working days, and a T1 realistically needs two iterations of release-torque trials before the timing map freezes. On our standard 2 to 5 week tool lead time, an unscrewing family tool sits in the upper band by design.

Then run the counter-case honestly. A collapsible tool needing a 4,000 USD core-sleeve set at 400,000 shots, plus two days of downtime and a scrap window while new sleeves fit, can hand the unscrewing tool payback inside 900,000 shots. Model it per program; do not adopt a blanket rule.

The purchase-order line that protects all of this is the maintenance clause: stripper bushings, spline hubs and rack tip teeth priced as spares with the tool, logged at 100,000-cycle intervals. A tool that ships with its wear parts costs 3 to 7 days of replacement forever; one that ships without them costs a new mold at the first wear failure.

Unscrewing Mold FAQ

Q1. How deep can a molded thread go before unscrewing is mandatory?

Above 1.5 times the nominal diameter of engagement, or about 2 full turns on most pitches, collapsible and core-break options stop holding gauge. Below that line they remain valid and cheaper.

Q2. How much torque does releasing a molded thread take?

Typically 5 to 30 N-m end to end. A half-turn PP closure may break away at 5 N-m while an M24 glass-filled nylon fitting peaks near 25 N-m; the drive and the reaction path get sized for the peak, not the average.

Q3. What clearance does a rotating core need?

0.05 to 0.1 mm diametral on the spline and stripper bushing fits. Tighter seizures at mold temperature, looser lets the core tilt and the thread run eccentric.

Q4. How fast do unscrewing cores rotate?

1 to 3 turns per second, which clears a 1 to 2 turn thread in 0.7 to 2.0 seconds. The mechanism only hurts the cycle if the timing overlaps cooling or clamp, which a good open-stroke map prevents.

Q5. Can an existing collapsible-core mold be retrofitted to unscrew?

Only when the mold base has plate thickness for bosses and the ejector layout survives a strip plate. On most tools a fresh rotating-core insert set costs 30 to 80 percent of the original move work, so a new tool is usually the cleaner answer.

The Bottom Line

An unscrewing mold is a torque problem wearing a thread. Pick the drive from the release number, size the reaction path through guide posts and mold-base bosses, hold the rotating fits at 0.05 to 0.1 mm with the stripper bushing as the only sacrificial part, and sequence rotation strictly while the part is still supported.

Do those four things and the mechanism disappears into the cycle: 1 to 3 turns per second, threads to gauge, and millions of shots between regrinds. Skip any one of them and the tool announces itself through sheared crests, galled splines or a twisted ejector plate.

DieStrike builds unscrewing, rack-drive and rotating-core tooling to IATF 16949 with +/-0.005 mm mold tolerance, +/-0.002 mm on EDM and jig-grinding, hardened steels to HRC 62 and 120+ machines behind the schedule, with tier-1 closure and connector programs from TE, Amphenol, Luxshare and Dongshan lines among our references. Send the thread print and the annual volume and the DFM report will price unscrew against collapse for your case.