Companion to the Matter Stack Compendium (Parts I–VII). Same rules as the Magnetic Matter Movers Phase 0 checklist: print it, fill it in ink, photograph the filled pages + videos as your gate evidence. Unexplained deviation fails a gate; explained deviation is data.
Shared-hardware note: the Stage 1 acoustic layer IS the TinyLev from Magnetic Matter Movers Build Sheet 002 — one levitator serves both projects. Buy once; the polarity-test discipline here is the same skill as Halbach polarity marking in Build Sheet 001. These projects train each other.
☐ Makerfabs Acoustic Levitator kit ($70) — or scratch parts per Compendium §6.1 ☐ 16× 12 V electromagnets 25 mm ($64) · 16× IRLZ44N MOSFET modules ($18) · 1N5819 diodes 100-pk ($7) ☐ Arduino Mega clone ($18) · 12 V 10 A PSU ($16) ☐ Ferrotec EFH-1 50 ml ($30) · glass cell + lid ($10) · distilled water + salt ☐ Polystyrene beads 1–3 mm · teabag paper/mesh scoop · perfboard/terminals/18 AWG ($14) ☐ Gloves + plastic sheeting ($8) · mini projector ($60) · gantry print ☐ Verify owned: soldering iron, multimeter (mandatory), phone tripod
☐ Frame printed, sockets filed to snug press-fit ☐ Polarity test EVERY transducer (A0 + Serial Plotter); + leg paint-marked BEFORE soldering → Record: transducers tested __ / inverted-marking count __ (builders report this is common — your number is real data) ☐ Arrays populated, epoxied (not hot glue), bussed; continuity: every +→+ bus, −→− bus, no bus-to-bus short ☐ Nano flashed, L298N wired; Nano powered from USB or L298N 5 V — never both ☐ Optional scope check: both arrays in phase
☐ Saturated saline mixed & settled; cell filled ⅔ saline + 10–20 ml ferrofluid injected below surface; lid sealed ☐ Hand-magnet test through glass (spiking observed — calibrates intuition) — done gloved, over sheeting ☐ 4×4 grid mounted at ~30 mm pitch, coils labeled 1–16, leads to terminal strip ☐ Flyback diode across every coil, stripe to + lead — count them: ______ /16 before first power-up ☐ One MOSFET per coil; coil rail separate from Arduino rail, grounds tied at one point; rail fused ☐ Firmware duty cap set at __ % (spec: 60–70%)
| Measurement | Predicted / spec | Actual | Dev./notes |
|---|---|---|---|
| Bead levitation duration | ≥ 5 min stable | __ | ___ |
| Bead vertical move via phase shift | smooth, no drop | __ | ___ |
| Levitation supply voltage | works at 6–9 V | __ V | ___ |
| Single coil 50% duty | one smooth bump | __ | ___ |
| Traveling bump (cross-fade demo) | full orbit, no tearing | __ | ___ |
| Tear-speed (max bump travel speed) | characterize — YOUR number becomes the spec | __ mm/s | — |
| Coil current, one coil full duty | 0.3–0.5 A | __ A | ___ |
| Coil temp after 10 min @ 50% | warm, touchable | __ | ___ |
| Two-layer milestone | bead above + bump tracking below, one host | __ | ___ |
| Full-stack milestone | 3 layers synced from one script | __ | ___ |
Triage: no levitation → re-verify polarity of suspect transducers, try smaller/irregular bead, drop voltage slightly · bead ejects → voltage down · no bump → coil polarity/wiring, cell standoff >8 mm · bump jumps instead of glides → PWM cross-fade logic, not hardware · anything dies at MOSFET switch-off → a missing flyback diode, stop and audit all 16.
GATE S1 — pass: ☐ 5-min levitation video ☐ traveling-bump orbit video ☐ full-stack sync video ☐ table filled, tear-speed recorded ☐ diode count photo ☐ anomaly paragraph per >50% deviation. Unlocks Stage 2 purchasing.
☐ 64× electromagnets ($200 — or reuse 16, save ~$46) · 64× MOSFET modules ($60) · 4× PCA9685 ($20) · extra diodes ($7) ☐ 12 V 30 A PSU ($32) + fuse · heatsinks + fans ($28) · USB mic ($25) · 16 AWG distribution ($20)
☐ 64 coils at ~15 mm pitch; cell standoff 3–8 mm; PCA9685 addresses 0x40–0x43 jumpered, I²C chained ☐ Per-coil MOSFET + flyback retained — diode count: ______ /64 ☐ Bus bars 16 AWG+, rail fused; fans wired to run whenever coils energized ☐ Firmware: duty cap, simultaneity cap (“max hot coils”: __), per-coil dwell limit set
| Measurement | Predicted | Actual | Dev./notes |
|---|---|---|---|
| Separable terrain features | 8–12 (fluid physics, not coil count, sets ceiling) | __ | ___ |
| Re-measured tear-speed (finer grid) | differs from Stage 1 value | __ mm/s | ___ |
| Total current, typical scene | well under 30 A fused | __ A | ___ |
| Coil-bank temp, 30-min run w/ fans | stable, not climbing | __ | ___ |
| Sound-reactive demo | bass→center swell, onset→bead dart | __ | ___ |
| Existing scenes on denser grid | run unchanged (normalized coords) | __ | ___ |
GATE S2 — pass: ☐ feature-count photo with count annotated ☐ 30-min thermal log ☐ sound-reactive video ☐ table filled. Unlocks Stage 3 or the art-piece fork (a gallery-ready machine legitimately stops here at ~$724).
Compendium warning stands: attempting this first is the most common project-killer. Entry condition: Gate S2 passed.
☐ Flat array 64–256 transducers (Ultraino/OpenMPD route); every transducer polarity-tested — same discipline, larger scale ☐ Phase controller (FPGA/MCU + driver boards) brought up ☐ Static trap first — reproduce a stationary hover before any steering ☐ Then slow commanded moves → then fast trajectories → RGB LED sync
| Measurement | Predicted | Actual |
|---|---|---|
| Static trap with flat array | stable hover | __ |
| Slow 3-D steering | smooth, no drops | __ |
| Max reliable bead speed | ≤ 8.75 m/s (literature ceiling) | __ m/s |
| POV shape (circle/line) | visible figure | __ |
| Mid-air haptic point | felt on palm | __ |
| Audible sound from array | recognizable tone | __ |
GATE S3 — pass: ☐ static-trap video ☐ POV-figure photo (long exposure) ☐ measured max speed ☐ bead-loss/re-trap behavior noted.
Safety stage, not polish. Until complete, machine runs attended only.
☐ Enclosure + security fasteners · levitator guard cage · warning signage (magnet/ultrasound) ☐ Access-panel interlock cuts coil power + levitator · hardware thermal cutoff independent of firmware ☐ E-STOP latching, one action kills coils + levitator · fused/switched IEC inlet, earth bonding, strain relief ☐ Software: attract/idle loop · bead-loss graceful recovery · 150 ms link watchdogs · daily restart
GATE S4 — the sign-off tests (each verified to reach safe state, each its own checkbox): ☐ Open access panel mid-show → safe state ☐ Trip thermal cutoff → safe state ☐ Hit E-STOP → safe state, latches until reset ☐ Pull Mega USB → coils drop ≤150 ms · ☐ Pull Nano USB → levitator idles ☐ Unattended soak test: __ hours (spec: several minimum), zero interventions Evidence: video of each safe-state test + soak log. Unlocks public deployment and Stage 5.
☐ One station wired (knob/slider→ADS1115, button debounced) → values streaming to host ☐ Calibration table per pot (min __ / max __ / dead-zone applied) ☐ One station → one feature confirmed (knob=height, slider=x, button=lock/reset) ☐ Stations duplicated: __ total, each own feature + projected hue ☐ Collision handling behavior chosen (merge / repulsion) and demonstrated ☐ Fair-idle drift + no-visitors generative idle confirmed ☐ Ensemble rehearsal: 2–3 people, tuned for responsiveness vs. calm (the actual product test)
GATE S5 — pass: ☐ multi-person play video ☐ calibration table ☐ collision + idle demos. The roadmap’s destination: exhibit live.