2026-10-30 8:10 AM: TH pin-in-paste design session
Decisions and notes captured in project/docs/th_pin_in_paste.md: station order N1 ball, N2 flux, N3 IPA, N4/N5 PnP, N6 paste auger on the needle; cartridges for ball and paste; paste for TH and large pads, balls for small; bottom side traces only, preheated; the trace designer optimizes for reflow parameters after testing. Head v5 waits for go.
2026-10-30 10:35 AM EDT: Head v5, N6 r5, syringe recommendation
- Approved FINAL: cap r5, end fitting r4, wrench r2. The bond note is in the README and docs.
- linear_cam_short_v5 is done. Six stations at a 20 mm pitch, the same stack on every station, cam 55 deg.
- Neighbor drop 6.43 / 6.62 (v4: 11.5).
- Springs: N1 0.100, N2–N5 0.0635, N6 0.146 N/mm. SF 3.50 / 3.52 / 3.85.
- Moving mass: 43.9 / 27.6 / 57.9 g. Head total 2284 g.
- Port reroute: insert port at 42.8 deg, slider channel at X +3.4, to clear the D3 guide rods.
- Checks: 0 clashes and 0 hits over 7 + 76 poses.
- TOP run17: needle down only at T >= 462. Z_BOARD 351.6, Z_IL 358, Z_IL_BACK 369.
- NEW: HOME needles below Z_IL touch the cassettes at T 332–422. Proposed rule T_HOME_LOW_MIN 432, approval pending.
- Exports: cad/2026_10_30_head_linear_cam_v5.step/.glb/.usdz, anim/2026_10_30_linear_cam_short_v5_anim.usdz/.mp4.
- N6 r5 drawing + model: drawings/2026_10_30_n6_paste_station_drawing_r5., cad/2026_10_30_n6_paste_station_r5..
- docs/th_pin_in_paste.md: recommends a disposable Luer-lock syringe feeding the auger, with the all-metal needle kept as standard.
- Site not regenerated. No snapshot.
2026-10-30 1:15 PM EDT: T_HOME_LOW_MIN enforced
- Rule approved: with the head at HOME and Z below Z_IL (358), T must be <= 322 or >= 432. It was proposed, now enforced.
- head_v5.py: new
motion_allowed(z, t, head, e_tool) holds the enforced rules (Z_IL, T_HOME_LOW_MIN, needle down T >= 462, E tip T >= 462).
- New top/run17_rule_check.py re-checks the run 17 sweep + transit results: 809 poses, 0 violations (27 HOME hits at T 332–422 without the rule). Output run17/rule_check.json.
- LINEAR_CAM_SHORT_V5.md updated. Site not regenerated.
2026-10-30 1:35 PM EDT: machine motion animation r3 (run 17 rules, head v5)
- New anim_src/machine_motion_r3 (copied from r2): timeline imports head_v5 (Z_IL 358, Z_IL_BACK 369, board Z 351.6, E nozzle Z 324.6 at T >= 462, needle down T >= 462, E tip T >= 462 below 358, T_HOME_LOW_MIN: HOME below 358 only at T <= 322 or T >= 432). Mesh from run17_v5_sweep (head v5 + E hot-air).
- Check: all 43 keyframes pass head_v5.motion_allowed() (head rule + E tip rule, asserted); 50 Hz sampled timeline 0 interlock/limit violations.
- Camera front-left 3/4 per docs/refs/machine_motion_anim_camera_ref.png; Okabe-Ito palette with legend; rule summary on screen.
- Output anim/2026_10_30_machine_motion_r3_anim.usdz + .mp4 (H.264 1600x900, 57.4 s); stills drawings/2026_10_30_machine_motion_r3_anim_still_47s.png / _still_21s.png. Old r2 files kept. Site not regenerated.
2026-10-30 2:30 PM EDT: twin tray + X rework plan (option C, plan only)
- the designer chose option C (two cassette pairs side by side + X-axis rework). Plan written: cad_src/tray/TWIN_TRAY_X_REWORK_PLAN.md; top-view layout drawings/2026_10_30_twin_tray_x_rework_layout.png (Okabe-Ito; generator drawings_src/twin_tray_layout/layout.py). No CAD changed.
- Finding: head v5 pick reach (N4 +10 / N5 +30) with carriage X 160-650 (head right end ~782 inside the right panel) is X 170-680 = 510, so 2x 300 does not fit. Plan uses 2 pairs of 240-wide cassettes (X 162-402 and 446-686, centre bay 402-446 for rotation plate/up camera/bin): 57 pockets/row, 912 total (was 576). Each cassette prints in one piece on the P1S 256.
- X drive: GT2 9 mm belt, vertical NEMA17 on the beam right end (chosen over a ~720 T8x8 whose motor would sit outside the frame). Fume riser moves to rear-right X 640-770 Y 625-685 (130x60 duct, inlet Z 355). Tray drive moves to rear fixed bearings + motor folded under the rail. Carrier 3 alu + 2x 20x20x2 angle on 4 MGN12H: sag 0.07 mm (3 alu alone 1.33). Opening 640 x 165 (X 105-745), door ~670.
- Margins: tray ~1.8 kg SF ~4.5; X belt SF ~5.6. T rules provisional for X > 425 until a run 18 sweep. Staged build: all machine work + left pair now, right pair cassettes later. 7 open questions (Q1-Q7) in the plan. Site not regenerated.
2026-10-30 4:15 PM EDT: run 18 twin tray + GT2 belt X built in CAD
- the designer approved Q1-Q7. New: subsystems/tray/tray_r18.py, subsystems/top/run18_model.py, run18_static_check.py, run18_sweep.py, run18_summary.py, top_run18.py, make_bom_run18.py; head_v5 X_MAX_SOFT 650. Old run files kept.
- Built: 4 x 240 mm cassettes (912 pockets), right pair mirrored; 6 mm two-layer carrier (angles clashed with the TH keep-out, fallback per Q4); GT2 belt X 160-650 + X_max endstop; rear tray drive; 640 door with 2 gas struts; control strip at Z 80 (E-stop moved: decision); harness reroutes (X motor, X_max, hot-air hose behind the chain at Y 690).
- Static check (27 poses incl. door 30/60/95): fixed the rear U-bracket vs rail, X motor/X_max routes, idler block window and panel holes; the rest are mates or existing door-seal overlaps.
- Sweep run 18: 2109 poses, X 160-650. Needle down T >= 462 and HOME T >= 432 clean at all X; transit Z 358 clean (gap 2.4). One new rule hit: HOME at board Z, T 322, x 480 (N2 tip 0.2 mm3 in cassette BR) -> proposal: HOME below Z_IL only with T >= 432.
- Gas strut check: 50 N is ~10x too strong; 10 N + magnetic catch proposed.
2026-10-30 4:46 PM EDT: Snap v21
- Version v21_2026-10-30, same layout as v20: project/versions/v21_2026-10-30/{cad,drawings,anim,project,TOOLS.md} (project without private/ and versions/), plus project/versions/OSSO_v21_2026-10-30.tar.gz.
- Today's work included: head v5 (linear_cam_short_v5) and N6 paste station r5; T_HOME_LOW_MIN / tray limits (run17 rule check, run18_b HOME T >= 432); twin tray run 18 and run18_b (GT2 belt X, 4 cassettes, door struts + magnetic catch, E-stop); machine motion r3 and r4 animations; heater plate TH grid drawing; interposer r5c lattice.
- Site not regenerated. No Mac sync. No Drive upload.