Companion to Build Sheets 001 (Inductrack) and 002 (Droplet Thruster). Print this; fill it in ink. The filled document + photos/video is your gate evidence — bring it back and we review it together, line by line.
Phase 0 closes when both builds pass their gates below. “Pass” means: the actual-vs-predicted tables are filled, deviations are explained (not ignored), and the evidence artifacts exist. Deviating from prediction does not fail the gate — unexplained deviation does. A 2× miss with a good story is data; a perfect match with no photos is nothing.
Gate evidence package (what you bring back): ☐ Filled tables from §3 and §5 (photo of this document is fine) ☐ Video: 001 hovering ≥60 s · 002 displacement + snap-back ☐ Photos: Halbach jig mid-assembly · enclosure interior with beam dump · goggles on head ☐ Anomaly notes: one paragraph per deviation >50% from prediction
Build 001 — Inductrack (~$80–120) ☐ 10× ½” N42 cube magnets (8 + 2 spares) ☐ Alloy-rim bike wheel w/ axle (Variant A) — used is fine, must be aluminum rim (magnet test: doesn’t stick) ☐ 775-class motor or brushless + ESC, ≥50 W, with speed control ☐ O-ring/belt for friction drive · epoxy (JB Weld) · 2× guide rods + PTFE tape ☐ Print queue: magnet jig, cart body (STL/design per Sheet 001 §5)
Build 002 — Droplet (~$150) ☐ TinyLev kit or parts (72× 40 kHz transducers, Arduino Nano, L298N, frame print) ☐ ≤5 mW red laser (starter) — this is the ONLY laser you need to pass Gate 002 ☐ 100–500 mW 450 nm module + OD4+ goggles rated 445–455 nm (may defer both to post-gate) ☐ 1 mL syringe + blunt 25 ga needle · india ink · foamcore/felt for enclosure + beam dump ☐ Optional: photodiode + resistor ($2) · laser power meter ($25)
Instruments (shared) ☐ Feeler gauges · phone tripod · mm-grid card (print one) · multimeter with current mode (or clamp meter) ☐ Phone apps: tachometer (acoustic or strobe), slow-mo camera
☐ Jig printed; ONE magnet test-fitted dry ☐ All 10 magnets polarity-marked (red dot = N face) BEFORE any assembly ☐ Array assembled in jig, one cube at a time, pusher block not fingers, clamped through full epoxy cure ☐ Strong-side check: washer snaps hard to one face, barely holds on the other — strong face marked ☐ Wheel mounted, spins smoothly to 1,500 RPM, plywood guard over lower half ☐ Guide rails mounted: cart free vertically ~25 mm, constrained laterally ☐ Array epoxied strong-face-down into cart; ballast to ~250 g total (record actual: __ g)
Table A — Levitation onset & gap (Variant A predictions; B in parentheses)
| Measurement | Predicted | Actual | Dev. |
|---|---|---|---|
| First cart motion (RPM) | ~40–60 (40) | __ | ___ |
| Clean liftoff (RPM) | ~60 (54) | __ | ___ |
| Gap @ 500 RPM (mm) | 1–3 (3–6) | __ | ___ |
| Gap @ 1,000 RPM (mm) | 2–6 (6–12) | __ | ___ |
| Gap @ 1,500 RPM (mm) | 3–8 (8–15) | __ | ___ |
Table B — The flat-drag signature (motor input watts = V × A, minus no-cart baseline at same RPM)
| RPM | Baseline W (no cart) | W with cart | Drag W (diff) |
|---|---|---|---|
| 500 | __ | __ | __ |
| 1,000 | __ | __ | __ |
| 1,500 | __ | __ | __ |
Predicted: drag column ≈ flat, 30–60 W (A) / ~28 W (B). The plateau in Table A and the flat line in Table B are the physics deliverables.
Troubleshooting triage (if outside 2× of prediction):
GATE 001 — pass conditions: ☐ Stable hover ≥60 s at ≤1,500 RPM (video) ☐ Table A filled, gap plateaus with RPM ☐ Table B filled, drag flat within ±30% over a 2–3× speed range ☐ Any >50% deviation has an anomaly paragraph
☐ TinyLev assembled; foam bead traps rock-solid (do not proceed to liquid until true) ☐ Enclosure built: grid card behind trap, camera port, horizontal beam path into felt dump ☐ ≤5 mW red laser rigidly mounted, aligned through node using trapped bead, locked down ☐ Dyed water mixed (~1 drop ink/mL); droplet loading practiced until 3 consecutive stable loads ☐ IF proceeding to 3B later: enclosure interlock, goggles for everyone present, key control — photo evidence required before any high-power run
Table C — Trap stiffness calibration (tilt method, Sheet 002 §7.2)
| Tilt angle | Force mg·sinθ (μN) | Displacement (mm) | k (μN/mm) |
|---|---|---|---|
| 2° | 1.4 | __ | __ |
| 4° | 2.9 | __ | __ |
| 6° | 4.3 | __ | __ |
Predicted k: 1–10 μN/mm. Three rows should agree within ~30% (that’s your instrument error bar).
Table D — Thrust runs (displacement × k = thrust; laser incident 5 mW, absorbed unknown → that’s what you’re measuring)
| Run | Displacement (mm) | Thrust (μN) | Direction horizontal? | Snap-back on off? |
|---|---|---|---|---|
| 1 | __ | __ | Y / N | Y / N |
| 2 | __ | __ | Y / N | Y / N |
| 3 | __ | __ | Y / N | Y / N |
Predicted thrust at 5 mW incident: 0.15–0.4 μN (absorption 30–80%). Implied absorption fraction = measured thrust ÷ 0.5 μN: __
Troubleshooting triage:
GATE 002 — pass conditions: ☐ Table C filled, k self-consistent within ~30% ☐ Table D: 3 runs, horizontal direction confirmed, snap-back confirmed (video of one run) ☐ Implied absorption fraction between 0.1 and 1.0 (sanity bound) ☐ Any >50% deviation has an anomaly paragraph ☐ Stretch (not required): offset map (Sheet 002 §7.6) — 5 offsets, force-vs-offset sketch
Bring the evidence package back to this chat. We review deviations together (that conversation IS the troubleshooting session), then Phase 1 opens with, in order:
Rule of the ladder: no step skips. The gate exists because Phase 1’s credibility rests on Phase 0’s error bars.