Guided lead-screw feeder

FM08-R3 selected academic design

The selected architecture uses separate restrained cells, a guided lead-screw pusher and one indexed receiver. Dedicated 1U spacers stay in the cassette; a capped channel and closed far-side stops contain the spacecraft; receiver capture is proved before the pusher returns. FDR10 compares the top three concepts using common inputs and records the seven-architecture ranking.

FM08-R3 storage, feed, capture and release-preparation states — Native CAD geometry rendered in Blender. The sequence is prescribed kinematics; it is not a dynamic-contact or measured-release result.

Architecture	Weighted judgment /10	Decision
Guided lead screw	7.30	Selected academic architecture
Rack and pawl	7.10	Retained runner-up
Spring and two gates	6.90	Alternative; spring force and controlled reset require redesign

Weights: safety 35%, packaging/mass 25%, positioning 20%, cycle/power 10%, manufacture 10%. The ranking is sensitive to priorities; it does not estimate reliability. Selected-design evidence includes native CAD/STEP, nominal motion, declared analytical sizing/load/tolerance screens and executable four/twelve-cell interlock cases. Physical qualification and precision measurements remain future work.

Complete selected-design package · Alternative concept archive

Historical geometry and numerical studies

Evidence through 22 October 2026: FM06-R1 component/contact studies, FM07-R6 nominal rotary clearance and the separate common-boundary feeder comparison. Seven Gen6 concepts share a 200 mm lateral transfer task for passive 1U/3U envelope surrogates. The receiver, gate, captive transfer and arrest requirements are evaluated separately from the launcher motor.

No feeder is selected or released for manufacture. Native CAD and a clear pose do not demonstrate handoff, safe power-off restraint, joint strength, complete installed hardware or reliability. Gen5 remains the evaluated configuration with its original failures.

Complete-angle clearance and contact-model diagnosis

RVC01 evaluates the separate FM07-R6 candidate. A 0.25 mm gate-panel relief retains the hinge, locks and upper edge. Four complete rotary intervals pass the nominal native-distance screen; the smallest guarded conservative bound is 0.0098846 mm. Eight assembly and two gate STEP files have checked exchange records. This does not establish manufactured or loaded clearance.

Native gate to deck distance in millimetres and guarded complete-angle bounds in micrometres — Source: RVC01 native FreeCAD data. The panel relief and global guarded bound are different quantities. Near-closed R5 zero distance is surface contact with zero intersection volume.

Native CAD, STEP, interval method and report.

FCE13 adds six completed native Chrono diagnostic runs on FM06-R1. Instrumentation reproduces the original baseline byte for byte. Original PSOR, native BB and an exactly equivalent box-core representation do not satisfy the unchanged pose-refinement criteria. No correction is qualified; these contact results do not validate FM07-R6.

Common-time centre of mass and orientation refinement differences, with solver iteration-budget counts — Source: FCE13 native Chrono data. Shading marks feed deceleration; dashed lines retain the 0.05 mm and 0.05 degree screens. The differences are numerical sensitivity checks, not certified error bounds or contact-force ratings.

Source, raw runs, reproduction commands and checks.

Seven concepts on a common four-payload boundary

FDR09 models two 1U and two 3U payloads with one common receiver. Seven R1 native/STEP assemblies have separate exchange readback. A restricted lead-screw R2 trim clears the checked nominal payload corridor and identified housing overlap. Full carriage, loaded support, tolerance, fault and insert-custody gates remain open; no feeder was selected in that study.

Represented CAD material and separate unintegrated catalog option mass subtotals for seven four-payload feeder concepts — Source: FDR09-R1. The columns are partial accounting, not full installed-system mass or an eligible architecture ranking.

Ten native/STEP pairs, BOM, component sources, exact geometry findings and checks.

Release and empty-carrier reset geometry

FM07-R5 adds captive receiver-lock pins, a hinged forward exit gate and prescribed empty-carrier reset. The 1U and 3U nominal cycles have 46 checked states, 34 additional door-angle samples and 40 continuous linear sweeps. Twenty native/STEP assemblies and 74 part exchanges pass independent readback. Rotary motion remains discretely sampled; these checks do not establish loaded clearance or hardware reliability.

Native FM07 exit-open geometry — Actual FreeCAD geometry rendered in Blender Workbench. The 80 mm carrier travel is a handling coupon; no release speed follows.

Self-contained geometry, STEP parts, native application capture and report. All seven concepts remain subject to the controlled selection gates.

21 October: geometry, stiffness and native contact review

The separate FM06-R1 research revision contains 30 native assembly/STEP poses, 70 independently checked reference part exchanges and 28 continuous prescribed rigid-translation checks. A backstop junction missed by solid-validity checks is corrected. These are nominal geometry results, not manufacturing or loaded-clearance acceptance.

FM06-R1 cassette-to-sled transfer geometry rendered from native FreeCAD solids — Blender Workbench render of actual CAD; the payload remains retained. Installed actuation, rail-end interfaces and forward-stop release remain unresolved.

Earlier and revised receiver displacement under the same clamped-bore component-fixture loads — Twelve actual-FM06 analyses and forty historical receiver cases use idealized clamped pin bores. The current 3U one-dog mean displacement is 0.05998 mm at 1,000 N, versus 1.883 mm for FM02. Historical FM03 is a different geometry. This does not bound assembled stiffness or approve material strength.

Native analytical planar-gap controls and fixed-pose replay — FCE12 uses exact native material partitions and checked unit scaling. Fourteen planar controls detect imposed 10 µm–1 mm overlaps; the largest checked gap error is 4.28 nm. The calibration applies to these controls. Final COM refinement changes remain 1.509 mm for 1U and 1.466 mm for 3U; no precision or design-load rating follows.

The twelve-case capture includes ten endpoint sequences and two injected-fault aborts. Eighty-four selected full-solid poses bound the recorded geometric residuals; they do not certify continuous loaded clearance. Current contact-method report, source, controls and native outputs.

Self-contained report, native inputs, CAD/STEP, checks and retained failures. All seven were unselected at this historical study revision. Evaluated Gen5 is unchanged.

Earlier candidate geometry · 19 October

Independent STEP-derived plan and side views of the unselected three-unit lead-screw transfer reference — Native FreeCAD/OCC STEP readback; millimetre axes and equal geometric scales. This is an orthographic projection, not a solver screenshot. The lead screw is an analysis reference rather than a baseline downselect.

Genuine FreeCAD Python-hosted native viewport showing the FDR-08 lead-screw candidate assembly — Actual FreeCAD Python-hosted Qt/Coin viewport of the FDR-08 native assembly, captured without composited application chrome. Visible solids remain a candidate design; source document and screenshot hashes are retained.

The earlier FDR-08 package contains fourteen editable native assemblies and twenty-eight STEP states. Exact intersections are checked at nine transfer stations, one stowed pose and five gate heights per case. The rectangular payload sweep covers only straight translation against fixed feed-state geometry. Nonlinear indexing, handoff, return, loaded deformation and contact are outside that proof.

What the mechanisms demand

Transfer force and idealized screw torque across duration and friction assumptions — Parameterized acceleration, friction and output demand. A 2 mm lead needs 6,000/2,400/1,200 rpm for the modeled 2/5/10 s triangular strokes. No actuator is rated by this plot.

Candidate	Useful feature	Decisive unresolved concern
Lead screw	Reversible positive displacement	Nut, bearing, brake and vacuum-duty ratings; no reliance on friction self-lock
Spring / two gates	Stored feed energy	15 N end force fails higher resistance cases; governor and dual-barrier proof
Rack / pawl	Discrete indexing	Tooth engagement, bounce, holdback and partial strokes
Indexed drum	One shared transfer window	Loaded rotor indexing, envelope, independent cell locks and common-path blockage
Pantograph	Compact folded linkage	Near-flat force, joint play and independently sequenced platforms
Telescoping carriage	Compact retracted length	Stage order, overlap, positive stops and anti-separation
Belt shuttle	Low moving drive mass	Tension, creep, breakage and independent arrest

Retain adverse cases

Remaining assumed side gap for different payload lengths and alignment bands — The adverse 0.5° alignment/thermal/manufacturing stack loses clearance. Positive nominal gap in the tighter band is not a manufactured or loaded fit.

The capture-rack cantilever screen also fails adverse mean-arrest cases: its 1 mm tooth root reaches approximately 1,973 MPa nominal bending under the largest computed mean force. Actual peak force and a selected material allowable are unknown; this is a redesign finding, not a working positive lock.

Native structural output, within its fixture

Native CalculiX displacement and peak stress trends across three solid meshes for each floor size — CalculiX 2.23, six C3D8I cases, assumed clamped short edges and uniform 6 kg × 25 g load. Displacement trends refine; peak support stress is not converged. Actual joints, contacts and host mounts are absent, so no assembly strength or qualification pass is claimed.

Actual CalculiX nodal displacement displayed on its input mesh in the native Gmsh viewport — Gmsh4.14.1 native viewport export of actual CalculiX input topology and DAT displacement norm in metres; undeformed geometry and a sequential colour scale. The assumed clamped-floor boundary remains explicit.

Read the engineering record

Report, local source/CAD archive, material BOM/CG/inertia, issue dispositions and native results. All seven research candidates in Labs. Safety-led ordinal scores remain hypotheses with overlapping intervals; incomplete mandatory evidence prevents a final selection.

FM-02: revised captive handoff and actuation screens

20 October 2026. A 12 mm underfloor receiving frame and two endpoint crosspins replace the tall latch. The restraints now belong to a connected sled. Twenty-eight native/STEP poses and 26 continuous prescribed rigid-translation phases have no represented penetration. This is captive handling; launch, free separation and complete reset are not simulated.

Genuine FreeCAD Qt/Coin capture of the FM-02 3U cassette, fork, storage crossbolt and receiving sled — Actual native FreeCAD geometry. Source-sized drive envelopes and provisional interfaces are visible; no manufacturing or flight acceptance.

Six equal-scale projections of actual FM-02 STEP handling states — Prescribed poses from exchange-file readback. The final state translates the captive sled 80 mm; it does not launch the payload. View the prescribed-pose animation (no dynamic or time fidelity).

All 972 sampled ground-drive decisions agree with polynomial extrema/integrals. The study retains 115/324 nonpositive fit residuals, 12/54 winding-temperature failures, 96 brake no-stop cases and 72 finite runouts above the proposed allowance. The integrated drive and braking candidate are not installed in this CAD.

Handling-time margin for two assumed auxiliary-speed policies in a twelve-release coast schedule — Later coasts offer 92.182 s. A 103.0 s assumed handling sequence fails eleven releases; a 74.5 s hypothesis fits. Actuator performance, indexing, attitude and full campaign integration remain unverified.

Self-contained nine-page report, source, native CAD/STEP, reference sheets, logs, failed cases and provenance. Axial pin retention, actual actuation, loaded contacts, complete deployment/reset and installed four/twelve-cell integration remain open. All seven were unselected at this historical study revision.

CT-01: continuous prescribed-cycle follow-up

19 October 2026. A separate captive-transfer coupon checks receiver capture and fork withdrawal across a complete prescribed rectilinear sequence. The gate retracts down through a floor slot; this requires 120 mm gate travel and new under-deck accommodation. Receiver locks occupy separate motion bands.

Four STEP side projections on common millimetre axes, showing stored, transfer, receiver capture and fork return — Independent STEP-derived geometry. Twenty continuous 1U/3U phase cases have no positive penetration in represented nominal rectangular primitives; 60 native poses independently check solids. These results exclude real actuator joints, contact, loaded tolerances and external sled custody.

Actual FreeCAD native application viewport of the CT-01 receiver-capture pose — Genuine Python-hosted FreeCAD Qt/Coin viewport; source document and capture hashes are retained. No composited application interface or physical hardware is shown.

Stop-runout sensitivity with failed travel cases and a 24 millimetre proposed allowance — Assumed catch, compliance, damping and response.226 of972 cases exceed the proposed total travel allowance. A restricted drive-removed envelope is a validation hypothesis, not a rated safe stop. The compliance mechanism is not integrated into this geometry.

Native CalculiX tooth-fixture displacement and peak-stress mesh trends — Six isolated 16×6×3 mm tooth fixtures, assumed steel and fixed root. Peak stress remains unconverged; no actual pawl impact, material allowable or structural qualification follows.

Local source,23 native/STEP pairs including the separate contact coupon, reports and retained failures. False lock indication, a gate falling onto a straddling payload and missing actual sled custody remain physical counterexamples. No feeder is selected.

Limited orbital replenishment

The RPL-01 concept defines cooperative-cartridge datums, capture/retention, isolation, contamination, operations and faults. 108 capture sensitivities and 36 insertion cases retain 18 failed fits. Current integration does not support an operational reload claim; ground loading remains the campaign assumption.

Capture-energy force and insertion-gap sensitivity for a hypothetical cooperative cartridge — Concept sensitivity inputs, not a docking, robotics or supplier-rated performance result. Mean energy-equivalent force does not predict peak contact load.

Feasibility report, provisional interface and source calculations.