{
 "generator": "demos/pinpoint_demo/make_page.js",
 "project": "quadcopter",
 "window": "quad-detailed",
 "sources": {
  "goals": "demos/quad_detailed/quadcopter.goals.json",
  "pins": "demos/pinpoint_demo/quadcopter.pins.json (captured 2026-11-08)",
  "versionCut": "2026-11-08T15:44:29Z",
  "findings": "demos/quad_detailed/manufacturing.json"
 },
 "goals": {
  "kind": "prompt",
  "title": "A realistic, detailed quadcopter in NativeCAD beside the RBCFD CFD reference body",
  "detail": "Lewis, 2026-11-07, on the 17-body rbcfd-quad: it looks extremely basic; a real quadcopter has something like 50-200 parts. Keep the CFD reference body as it is and add a detailed one at the same scale, so it can share the CFD context later.",
  "state": "input",
  "children": [
   {
    "kind": "requirement",
    "title": "50-200 separate parts, each with a meaningful name and a group",
    "detail": "Groups: frame, propulsion, avionics, power, payload, fasteners, gear.",
    "state": "met",
    "evidence": "200 named bodies: frame 17, propulsion 48 (the four prop spacers added 2026-11-08), avionics 48, power 13, payload 6, fasteners 56, gear 12 (demos/quad_detailed/out/manifest.json)."
   },
   {
    "kind": "requirement",
    "title": "Same scale as rbcfd-quad: rotor radius 120 mm, motor axes 190 mm out, quad-X, props 1/3 CW and 2/4 CCW",
    "state": "met",
    "evidence": "tests/test_quad_detailed.js D: every motor axis at r 190.000 on the diagonal, prop spans match a 240 mm rotor; same NACA 44xx prop law as rbcfd."
   },
   {
    "kind": "requirement",
    "title": "NativeCAD's own ops only, every part watertight and passing checkFull",
    "state": "met",
    "evidence": "200/200 watertight and ops/check.checkFull ok after the rotor-CFD relayout; built from extrude, revolve, sweep, loft and primitives, no imported mesh (build.js)."
   },
   {
    "kind": "requirement",
    "title": "What a real build has",
    "detail": "Plates with cutouts, arms with motor mounts, standoffs and screws, motors as bell + stator + shaft + base + screws, props with hub and nut, FC/ESC stack with standoffs and grommets, receiver, GPS mast, camera with mount, battery with strap and pads, landing gear with feet, antennas, wiring.",
    "state": "met",
    "evidence": "All present as their own bodies; see demos/quad_detailed/README.md."
   },
   {
    "kind": "requirement",
    "title": "rbcfd-quad stays the unchanged CFD reference",
    "state": "met",
    "evidence": "demos/rbcfd_quad untouched; test F checks its 17 bodies and propGap 38."
   },
   {
    "kind": "requirement",
    "title": "Exports: STEP AP242 plus a manifest of names, groups, volumes and bboxes",
    "state": "met",
    "evidence": "out/quad_detailed.step (AP242, every part named), manifest.json; the Pinpoint pack is imported from that STEP."
   },
   {
    "kind": "goal",
    "title": "No unintended interference between parts",
    "detail": "packages/assembly AssemblyClash over all 19,110 pairs, sampled overlap volumes.",
    "state": "met",
    "evidence": "Run 2 after fixes: what remains is by design (screws in tapped parts, shafts in bores, legs in feet, leads into battery and motors); minimum clearance 0.10 mm. 2026-11-08 relayout: every pair touching a moved part re-checked; only by-design contacts remain (leads into the battery and XT60, screws in the GPS base, the antenna loop on its tube); the strap slots now follow the battery width."
   },
   {
    "kind": "goal",
    "title": "Nothing but a motor's own parts inside its prop disc",
    "state": "met",
    "evidence": "tests/test_quad_detailed.js D, all four props (34/34 after the props were raised to 46 mm spacers); adjacent prop tips 28.7 mm apart."
   },
   {
    "kind": "goal",
    "title": "Flies like a 9-inch long-range quad: all-up mass under 1 kg, thrust-to-weight 4 or more, CG on the thrust axis",
    "state": "met",
    "evidence": "After the rotor-CFD relayout (6S1P Li-ion 18650 pack, props on 34 mm spacers): 928 g all-up, CG 0.10 mm off the thrust axis (z 35.7 mm), T/W about 7.5 against a 2806.5-class motor at 1.75 kg static each (vendor-table order, not measured), hover 2.28 N a motor. Per-group densities reconstructed from the original 849 g check, which they reproduce to 852 g. (Before: 849 g, CG 0.33 mm off axis, T/W about 8.2.) The 46 mm spacers ([ticket]) add an estimated 19.5 g: about 948 g, CG ~1.5 mm higher, T/W about 7.35, hover 2.33 N a motor (estimate from the spacer / shaft volumes, not the full per-group check)."
   },
   {
    "kind": "goal",
    "title": "Airframe drag across forward-flight speeds and attitudes (GPU wind tunnel)",
    "detail": "tools/cfd/quad_detailed_drag.js: the 192 non-rotating parts, cut cells, k-omega SST, resident on the GPU; props left out (sub-cell blades, rotor CFD territory).",
    "state": "met",
    "evidence": "tools/cfd/quad_detailed_drag.js --grids, GPU, 10 m/s, 320 steps, six grids 42..112 (benchmarks/out/quad-detailed-grid-study.json). 46 mm spacers, with sub-cell parts as thin walls / drag sources ([ticket], PR 348): mean CdA 0.0139 m2 level, 0.0177 at 10 deg nose down, 0.0203 at 20 deg; half-range over mean 9.8 / 7.3 / 5.2 % (was 40 / 30 / 17 % when thin parts were whole voxels or absent). NOT grid-converged: CdA still falls with cell size (0.0152 at grid 42 to 0.0125 at grid 112), mostly isolated-part drag closures ignoring sheltering; [ticket] stays open. Spacer cost 46 vs 34 mm: +3.5 % level, +7.7 % at 10 deg, unresolved at 20 deg. Found and fixed on the way: [ticket], [ticket]."
   },
   {
    "kind": "requirement",
    "title": "Fits the RBCFD overset rotor zones (1.1 R by 0.15 R plus a 2.5-cell margin of plain fluid around each prop)",
    "detail": "Lewis, 2026-11-08: raise the props in the main design and compact the fuselage top rather than keep a CFD-only variant.",
    "state": "met",
    "evidence": "Props on 46 mm spacers (34 at first; raised for the clear rotor zone, [ticket]); 6S1P Li-ion 18650 pack 66 x 38 x 38 centred; GPS mast at x +42 with a 24 mm puck; VTX antenna up through the top plate hole; leads low. tools/cfd/rbcfd_quad_run.js --model quad_detailed --checkRun at 30 cells/R: the executor accepts the zones and the run-time boundary check PASSES every zone."
   },
   {
    "kind": "milestone",
    "title": "Share the RBCFD CFD context: run the rotor CFD on the detailed body",
    "detail": "The body now fits the overset method (props on 46 mm spacers, fuselage top re-laid so each zone + margin is plain fluid; run check passes every zone). Hover at 30 cells/R runs: 56.9 M cells, 3 revolutions in 222 s on the RTX 5060, finite (tools/cfd/rbcfd_quad_run.js --model quad_detailed). [ticket] (a solver fault: one axial receiver layer turned the zone into a centrifugal pump) is fixed. Rotor loads are NOT lift at 30 cells/R: a reversed-spin pair shows the axial force is spin-independent (reference quad -3.9 N either way; lift = odd part ~0 N centre-sampled, ~+0.27 N with --shellZones thin, against ~+2.61 N expected). [ticket] added --shellZones thin (sub-dx blades sealed to the tip, hub unchanged): this body gets a ~+0.15 N lift signal under the same spin-independent force. [ticket] found the spin-independent force: a draft the overset coupling grows when the hub / blade sit in the background fringe band over the disc (props held still still read -3.6 N); a zone that clears the band removes it (reference quad: +0.14 N real lift with sealed blades), and runs now warn and report the clear half-height. The spacers were raised 34 -> 46 mm so this body's clear 32 mm zone and its donor margin pass the run check (now the default zone); a reversed-spin GPU pair with --shellZones thin then reads +0.46..+0.53 N lift per rotor with the spin-independent part down to -0.01..+0.09 N (was ~0.5 N). Still ~18 % of the expected lift and torque ~8x high: 30 cells/R is below the ~80 cells/R lift needs, so loads stay not validated. Forward flight also waits on [ticket].",
    "state": "met",
    "evidence": "Rotor CFD runs on this body with actuator-line rotors (wind_tunnel rotorMethod actuatorLine, [ticket], PRs 338 + 346): no overset zone, 20 cells/R, 5400 rpm, 10 revolutions, airframe wall forces on (188 static parts). Hover: 5.54 M cells, 99 s on the RTX 3090; rotor thrust 2.435 / 2.379 / 2.397 / 2.417 N (total 9.63 N), airframe download 0.43 N (arms 0.40 N), net 9.20 N against 9.30 N weight (about 948 g): 1.1 % short at 5400 rpm. Forward flight 6 m/s, 5 deg nose down: 12.52 M cells, 136 s; front 2.633 / 2.634 N, rear 2.402 / 2.423 N (total 10.09 N), airframe drag 0.57 N. Records benchmarks/out/quad-detailed-actuator-hover.json and -ff6.json. The actuator line is within about 4 % of blade-element momentum on the reference prop (2.517 against 2.614 N); loads come from a section polar, not from resolved blades. Blade-resolved loads stay not validated: thrust is not grid-converged at 120 cells/R ([ticket]). The 46 mm spacers are not needed by this method; their interference effect was not measured."
   }
  ]
 },
 "pins": [
  {
   "id": "e502df70647f",
   "at": "2026-11-08T03:23:39.216Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "top_plate",
   "point": [
    -22,
    21,
    34
   ],
   "version": 1,
   "text": "[Check] Assembly clash over all 196 parts (19,110 pairs, AssemblyClash, sampled volumes). Run 1 found 5 real problems, all fixed in dc24b8105: battery straps cut into the top plate (118 mm3, strap inner corners), arm bolts 0.6 mm through their nuts into the motor wires, rx antenna tubes starting inside the receiver, a motor wire 0.09 mm from an ESC lead, and props with no shaft bore (311 mm3 each). Run 2: what is left is by design (screws in tapped parts, shafts in bores, legs in feet, leads into battery and motors). Minimum clearance 0.10 mm, motor_screw_3_3 head to arm_3."
  },
  {
   "id": "b4c43dc9b464",
   "at": "2026-11-08T03:23:39.224Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "battery",
   "point": [
    0,
    0,
    51
   ],
   "version": 1,
   "text": "[Check] Mass and balance with per-part material densities: all-up 849 g (battery 262 g as 6S LiPo effective, frame 209 g, propulsion 227 g, power 290 g, fasteners 49 g, avionics 38 g, gear 25 g, payload 14 g). CG (-0.3, 0.1, 27.9) mm: 0.33 mm off the thrust axis. Inertia Ixx 0.0061, Iyy 0.0064, Izz 0.0117 kg m2. Boxed electronics are solid stand-ins with an effective density."
  },
  {
   "id": "532f732e7244",
   "at": "2026-11-08T03:23:39.232Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "prop_1_cw",
   "point": [
    134.35,
    134.35,
    37
   ],
   "version": 1,
   "text": "[Check] Hover by momentum theory at 849 g: 2.1 N a motor, induced velocity 4.3 m/s, disc loading 46 N/m2, about 14 W shaft power a motor at figure of merit 0.65. Thrust-to-weight about 8.2 against a 2806.5-class motor at about 1.75 kg static on a 9x4.5 prop at 6S: vendor-table order, not measured. Motor here weighs 39 g against about 55 g for a real 2806.5."
  },
  {
   "id": "46e93a9a56ae",
   "at": "2026-11-08T03:23:39.237Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "prop_2_ccw",
   "point": [
    -134.35,
    134.35,
    37
   ],
   "version": 1,
   "text": "[Check] Prop discs: nothing but each motor's own parts inside its 240 mm disc (tests/test_quad_detailed.js D). Adjacent prop tips 28.7 mm apart. The first build had the rear receiver antennas reaching into the rear discs; they were lowered."
  },
  {
   "id": "dbc38a758dae",
   "at": "2026-11-08T03:23:39.243Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "enhancement",
   "part": "arm_1",
   "point": [
    95,
    95,
    6
   ],
   "version": 1,
   "text": "Next: run the RBCFD rotor CFD on this body (same scale and prop law as rbcfd-quad). Goals milestone goal-b79158106d2d."
  },
  {
   "id": "6b1a5c680efe",
   "at": "2026-11-08T14:37:35.166Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "arm_1",
   "point": [
    95,
    95,
    6
   ],
   "version": 1,
   "text": "[Check] Airframe drag on the GPU wind tunnel (tools/cfd/quad_detailed_drag.js): the 192 non-rotating parts (props left out: sub-cell blades, rotor CFD territory), cut cells, k-omega SST, every step resident on the RTX 5060, ~12 s a case. CdA 0.0091 m2 level, 0.0094 at 10 deg nose down, 0.0103 at 20 deg; drag 0.56 N at 10 m/s, 2.24 N at 20 m/s level. Grid 56: grids 42/56 agree within 0.5 %, finer grids swing as the 1.5-2 mm plates and wires resolve, so this is the coarse-grid airframe drag. The run found and fixed [ticket] (thin walls failed the cut-cell solve) and [ticket] (bc.inlet.AoA alone was ignored)."
  },
  {
   "id": "b9b50d5102c9",
   "at": "2026-11-08T14:37:35.173Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "question",
   "part": "top_plate",
   "point": [
    70,
    30,
    34
   ],
   "version": 1,
   "text": "Rotor CFD (pin dbc38a758dae) does not fit this body: the minimum overset zone around each prop (r 1.1 R = 132 mm, +-18 mm) takes in this top-plate corner (122 mm from the motor axis, z 33-35) and the front standoffs and screws (~133 mm). Bell, shaft, washer and nut spin with the prop and the stator sits inside the bell, so those can ride in the zone; the frame cannot. Raise the props (rbcfd_quad propGap style) or pull the plate corners and front standoffs ~15 mm in? Forward flight also waits on [ticket]; hover is the case to run first.",
   "resolvedAt": "2026-11-08T15:45:37.251Z",
   "closedBy": "Claude (co-author: Lewis)",
   "closer": "answered",
   "evidence": "Lewis chose to raise the props in the main design and compact the fuselage top; landed bd75fc8eb, run check passes every zone."
  },
  {
   "id": "7033c47f840a",
   "at": "2026-11-08T15:45:37.246Z",
   "by": "Claude (co-author: Lewis)",
   "kind": "note",
   "part": "prop_spacer_1",
   "point": [
    134.35,
    134.35,
    46
   ],
   "version": 2,
   "text": "[Check] Re-laid for the RBCFD rotor zones (Lewis's call, landed bd75fc8eb): props on 34 mm spacers, 6S1P Li-ion 18650 pack 66x38x38 centred, GPS mast at x +42 (24 mm puck), VTX antenna up through the top-plate hole, leads low. 200 parts watertight + checkFull, test_quad_detailed 34/34, clash on moved parts only by-design. 928 g, CG 0.10 mm off the thrust axis, T/W ~7.5. RBCFD run check passes every zone; hover at 30 cells/R ran 10 revolutions in 612 s on the RTX 5060 (56.9 M cells, finite). Per-rotor loads read ~0 thrust here AND on the reference quad with either spin direction: [ticket], not trusted. Airframe drag now CdA 0.0129 m2 level (was 0.0091): the spacers cost ~40 %."
  }
 ],
 "findings": [
  {
   "id": "bottom_plate:min_wall_thickness",
   "part": "bottom_plate",
   "point": [
    0,
    27.3,
    3
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 1.4 mm; the guide recommends at least 6 mm, twice the sheet thickness."
  },
  {
   "id": "bottom_plate:min_internal_corner_radius",
   "part": "bottom_plate",
   "point": [
    43.556,
    28,
    3
   ],
   "title": "Minimum internal corner radius",
   "text": "8 inside corners drawn sharp; the cut leaves a 0.81 mm radius there."
  },
  {
   "id": "top_plate:min_wall_thickness",
   "part": "top_plate",
   "point": [
    0,
    28.3,
    35
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 3.41 mm; the guide recommends at least 4 mm, twice the sheet thickness."
  },
  {
   "id": "top_plate:min_hole_diameter",
   "part": "top_plate",
   "point": [
    37,
    0,
    35
   ],
   "title": "Minimum hole diameter",
   "text": "2 holes of 2.2 mm; the guide recommends at least 3.18 mm."
  },
  {
   "id": "arm_1:min_wall_thickness",
   "part": "arm_1",
   "point": [
    134.35,
    140.55,
    9
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 3.4 mm; the guide recommends at least 12 mm, twice the sheet thickness."
  },
  {
   "id": "arm_1:stock_sheet",
   "part": "arm_1",
   "point": [
    51.619,
    51.619,
    9
   ],
   "title": "Listed sheet thickness",
   "text": "6 mm is not a sheet the service lists; the nearest are 5 and 8 mm."
  },
  {
   "id": "arm_2:min_wall_thickness",
   "part": "arm_2",
   "point": [
    -140.55,
    134.35,
    9
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 3.4 mm; the guide recommends at least 12 mm, twice the sheet thickness."
  },
  {
   "id": "arm_2:stock_sheet",
   "part": "arm_2",
   "point": [
    -51.619,
    51.619,
    9
   ],
   "title": "Listed sheet thickness",
   "text": "6 mm is not a sheet the service lists; the nearest are 5 and 8 mm."
  },
  {
   "id": "arm_3:min_wall_thickness",
   "part": "arm_3",
   "point": [
    -134.35,
    -140.55,
    9
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 3.4 mm; the guide recommends at least 12 mm, twice the sheet thickness."
  },
  {
   "id": "arm_3:stock_sheet",
   "part": "arm_3",
   "point": [
    -51.619,
    -51.619,
    9
   ],
   "title": "Listed sheet thickness",
   "text": "6 mm is not a sheet the service lists; the nearest are 5 and 8 mm."
  },
  {
   "id": "arm_4:min_wall_thickness",
   "part": "arm_4",
   "point": [
    140.55,
    -134.35,
    9
   ],
   "title": "Minimum web between cuts",
   "text": "Thinnest web 3.4 mm; the guide recommends at least 12 mm, twice the sheet thickness."
  },
  {
   "id": "arm_4:stock_sheet",
   "part": "arm_4",
   "point": [
    51.619,
    -51.619,
    9
   ],
   "title": "Listed sheet thickness",
   "text": "6 mm is not a sheet the service lists; the nearest are 5 and 8 mm."
  }
 ]
}
