References

Status key: ✅ = citation verified against the source (DOI/URL fetched or datasheet in hand) · 📁 = internal project file · ⚠️ = partial citation · verify full details before print publication.

External sources

[1] Moravec, H. · Mind Children: The Future of Robot and Human Intelligence, Harvard University Press, 1988. Quote: "It is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility."
✅ Quote wording verified 2026-07-31 via Narayanan, "Fact-checking Moravec's paradox" (normaltech.ai), which sources it to the 1988 book. ⚠️ Page number not yet located · find before print (commonly cited as p. 15).

[2] International Federation of Robotics · press release "Global Robot Density in Factories Doubled in Seven Years," published 2024-11-20. Figures used: 162 robots per 10,000 employees globally in 2023; 74 units in 2016; (context: Korea 1,012 / Singapore 770 / China 470 / Germany 429 / Japan 419 / US 295).
✅ Fetched 2026-07-31: https://ifr.org/ifr-press-releases/news/global-robot-density-in-factories-doubled-in-seven-years
Derived in text: "98.4% … no robot beside it" = 1 − 162/10,000.

[3] Liang, W., Liu, H., Wang, K., Qian, Z., Ren, L., Ren, L. · "Comparative study of robotic artificial actuators and biological muscle," Advances in Mechanical Engineering, 12(6), 2020. DOI: 10.1177/1687814020933409.
✅ Fetched 2026-07-31 (open access, SAGE). Values used (Table 1): muscle 40% / 0.35 MPa / 40 kJ/m³ / 40%; PAM 25% / 1.16 MPa / 200 kJ/m³ / 49%; SMA 10% / 200 MPa / 10,000 kJ/m³ / ~10%; DEA 380% / 7.7 MPa / 3,400 kJ/m³ / 60–80%; CP 10% / 34 MPa / 100 kJ/m³; CNT 3% / 26 MPa / 1,000 kJ/m³; IPMC 40% / 3 MPa / 5.5 kJ/m³. Quotes used: "the strain and stress of this PAM are slightly higher than those of biological muscle"; "ultra-high stress to 200 MPa"; "giant strains of 380%"; hydraulics "1.6–2 kW/kg… exceed those of electric motors" (~100 W/kg for electric).

[4] Shi, M., Yeatman, E. M. · "A comparative review of artificial muscles for microsystem applications," Microsystems & Nanoengineering, 7:95, 2021.
✅ Fetched 2026-07-31 (open access, Nature). Quote used: "the main challenges for piezoelectric artificial muscles include high stiffness, small strain, and relatively high operating voltage."

[5] Zhou, X., Wu, S., Wang, X., Wang, Z., Zhu, Q., Sun, J., Huang, P., Wang, X., Huang, W., Lu, Q. · "Review on piezoelectric actuators: materials, classifications, applications, and recent trends," Frontiers of Mechanical Engineering, 19(1):6, 2024. DOI: 10.1007/s11465-023-0772-0.
✅ Full analysis: 📁 paper_analyses/001_Zhou_2024_Review_Piezoelectric_Actuators.md. Used for: direct/indirect taxonomy; Fig. 24 performance envelopes; 891.3 mm/s best standing-wave linear speed (their ref [129]); application examples (biomimetic fish, endoscopic devices, micro air vehicles); stack equation L₁ = n·d₃₃·U; absence of AM discussion in 211 refs; stick-slip = simplest indirect structure (replication assessment).

[6] Lin, Y., An, D., Lin, Z., Chen, X., Huang, W. · "Progress in high-performance stick-slip piezoelectric actuators: a review," International Journal of Smart and Nano Materials, 15(3):652–696, 2024. DOI: 10.1080/19475411.2024.2395293.
✅ Full analysis: 📁 paper_analyses/002_Lin_2024_SSPA_Review.md. Used for: 2–40 nm resolution range; 12–46 mm/s linear speeds; >10 kg load; 80–400 V typical drive; gap statements (no 3D-printed SSPAs; <30 V design unexplored; no energy-efficiency data in 184 refs); EDM-vs-SLA tolerance assessment (~sub-10 µm vs ~50 µm, "viable" for proof-of-concept).

[7] Li, J., Deng, J., Zhang, S., Chen, W., Zhao, J., Liu, Y. · "Developments and Challenges of Miniature Piezoelectric Robots: A Review," Advanced Science, 10:2305128, 2023. DOI: 10.1002/advs.202305128.
✅ Full analysis: 📁 paper_analyses/003_Li_2023_Miniature_Piezoelectric_Robots_Review.md. Used for: 342-reference scope; zero mobile stick-slip robots in Table 2; 60–200 V typical; power data absent from Table 2; AM "currently used ONLY for the structural body"; Figure 4 data points (mass, speed, working principle per Table 2 rows as extracted in the analysis file).

[8] Xing, S., Gao, J., Lu, T., Zhou, L., Zhang, J., Li, X., Tian, H. · "A Three-Legged Stick-Slip Piezoelectric Actuator Inspired by Haystack Unloading Device, and Its Driving Method," Results in Engineering, 26:105146, 2025. DOI: 10.1016/j.rineng.2025.105146.
✅ Full analysis: 📁 paper_analyses/004_ThreeLegged_SSPA_Haystack_Unloader.md. Used for: 0% backward displacement with 120° triangular drive (Fig. 18a); 580 µm/s at 900 Hz, 100 V (Fig. 10); 2 kg vertical load (Fig. 13); 55/96/187 nm steps at 1/2/4 V (Fig. 20); 3D-printed casing + commercial stacks.

[9] CoreMorrow · PSt150/5×5/20H piezo stack datasheet (EN/CN product tables). Values used: 20 µm ±15% stroke @ 0–150 V; stiffness 60 N/µm ±10%; blocking force 1600 N; 5.1×5.1×18 mm; 1.8 µF; 50 kHz resonance.
✅ Manufacturer-verified 2026-07-10 (recorded in project CLAUDE.md §2, with Piezomechanik catalog cross-check).
Derived in text from these values: blocking stress ≈ 61 MPa (1600 N ÷ 26.0 mm²); strain ≈ 0.11% (20 µm ÷ 18 mm); work density ≈ 26 kJ/m³ (½ × 60 N/µm × (20 µm)² ÷ 468 mm³); strain rate ≈ 110 %/s at 1 kHz (0.11% × 1000 Hz).

[10] Huang, W., Sun, M. · "Design, Analysis, and Experiment on a Novel Stick-Slip Piezoelectric Actuator with a Lever Mechanism," Micromachines, 10(12):863, 2019. DOI: 10.3390/mi10120863. (PK4FQP1 stack, 20 µm class, 1.43× steel lever); demonstrated steps 0.875 / 1.75 / 3.33 µm at 10 / 20 / 30 V.
✅ Full citation resolved via Crossref 2026-08-22 (open access, MDPI, CC BY 4.0); matches the record in amplifier_leg_v2_report.md §1. The three step values are from that report's extraction, not from Crossref metadata · they trace to the paper's own results section.

[11] Liao, H.-S., Werner, C., Slipets, R., Emil Larsen, P., Hwang, I.-S., Chang, T.-J., Ulrich Danzebrink, H., Huang, K.-Y., Hwu, E.-T. · "Low-cost, open-source XYZ nanopositioner for high-precision analytical applications," HardwareX, 11:e00317, 2022. DOI: 10.1016/j.ohx.2022.e00317. Design files: OSF, DOI 10.17605/osf.io/7fk3u. (Arduino Mega → DAC0800 ×3 → TDA2050 ×3; ~$50–80 in parts; ~15 V p-p unipolar output as measured on our build.)
✅ Full citation resolved via Crossref 2026-08-22; article open access under CC BY. The OSF design files (Gerbers, schematic, BOM, Arduino code) carry CC BY-SA 4.0 · verified via the OSF API on the same date — which is why this project's copies of the corrected Gerbers and the KiCad re-draw are released under CC BY-SA 4.0 (§M6). Redistribution permission: Prof. Hwu, email 2026-08-09 — "The online repository is not changeable, you can make yours so I can direct other people to use your fix." Co-authorship declined by his group, email 2026-08-17. Note for print: Crossref renders two co-author names as "Emil Larsen, Peter" and "Ulrich Danzebrink, Hans"; check the PDF's author list before typesetting.

[12] Adibnazari, I., Nagel, W. S., Leang, K. K. · "A 3D-printed 3-DOF tripedal microrobotic platform for unconstrained and omnidirectional sample positioning," International Journal of Intelligent Robotics and Applications, 2(4):425–435, 2018. DOI: 10.1007/s41315-018-0071-9. The only 3D-printed SSPA in Lin 2024's survey (their ref [141]).
✅ Full citation resolved 2026-08-22 via the reference list Lin 2024 [6] deposited with Crossref, then confirmed by DOI resolution. Note: the paper's own title does not use the phrase "stick-slip" · that classification is Lin 2024's.

[13] TableFlip Foundry · "The Cones of Calibration V3" (official page + reading guide).
✅ Fetched 2026-08-02: https://www.tableflipfoundry.com/3d-printing/the-cones-of-calibration-v3/ · creator confirmed as TableFlip Foundry (Young had recalled "Forge Labs"; corrected). Quotes used: "The success cones should print fully, while the failure cones are designed to collapse"; "The sword should fit perfectly into the Monster Skull but not into the fail holes of the Attunement Block"; Ale/Mug over- and under-exposure sentences; Attunement Block 6 mm. ⚠ STL redistribution terms not stated on the page — link, don't re-host, until checked.

[14] ELEGOO · Mars 5 Ultra product page (specs).
✅ Fetched 2026-08-02: https://us.elegoo.com/products/mars-5-ultra-9k-7inch-monochrome-lcd-resin-3d-printer · build volume "153.36 × 77.76 × 165 mm"; "7-inch 9K Mono LCD"; XY "18 μm (8520 × 4320)"; firmware feature "Section-Based Exposure" (the up-to-8-zones detail is Young's usage recollection, not on the page).

[15] Resins used in §M4a: Anycubic ABS-Like Resin Pro 2.0 (gray); Sunlu ABS-Like (white).
⚠ Product names per Young's notes 2026-08-02; verify exact SKUs/listings before print publication. Measured on our printer: 2.6 s/layer optimal for the Anycubic gray (COC V3 method); Sunlu white ≈2.8 s/layer, 30 s bottoms, extended post-wash drying needed (Young, 2026-08-06).

[16] Sitting KAWS figure STL (§M4b test model), downloaded from the internet.
⚠ Young to supply the exact listing URL + check its license before publication. Note: KAWS figures are a living artist's copyrighted artwork · the article links to the source only, never re-hosts the STL, and the 3D-viewer embed stays local-file-only for this model (same link-don't-rehost policy as [11] and [13]).

[17] M Armory · "EVA Unit-01 (Articulated)," MakerWorld model 2756182, https://makerworld.com/en/models/2756182-eva-unit-01-articulated ; creator profile https://makerworld.com/en/@user_3828721104
✅ Model page located 2026-08-22 (15 articulated joints, ~30 cm, single-plate, designed for FDM). Profile confirmed earlier via Young's screenshot 2026-08-03. License: MakerWorld "Standard Digital File License" · no redistribution and no derivatives, so this project publishes photographs and its own measured clearance numbers only, and links rather than re-hosting. IP note: EVA Unit-01 is a copyrighted character design (Evangelion franchise); fan-made model — credit the creator, link only.

[18] kvnper · "Tolerance Test," MakerWorld model 22224, https://makerworld.com/en/models/22224-tolerance-test#profileId-20163
✅ Fetched 2026-08-03: round/square pegs into holes, clearances stepped 0.05–0.25 mm.

[19] 3D Maker Noob · "Clearance tolerance test," MakerWorld model 423905, https://makerworld.com/en/models/423905-clearance-tolerance-test#profileId-406713
✅ Fetched 2026-08-03: compact part-clearance test (inspired by Maker's Muse's tolerance tests, per the page).
Measured with [18]+[19] on our setup (Mars 5 Ultra, Anycubic gray @ 2.6 s): 0.15 mm clearance for separately-printed mating parts; 0.30 mm for print-in-place. Young, §M4c.

[20] Coupang listing · XCHENIMO (체니모) LED gel nail lamp, portable, white (the §M4d curing-station donor).
✅ Existence/price documented via Young's screenshot 2026-08-04: 3,990 KRW (78% off a 18,500 KRW list; 609 reviews). Young paid 5,990 KRW at purchase time (his notes). ⚠ Add the listing URL.

[21] ARCTIC MX-4 thermal paste, 4 g (heatsink mounting, §M4d).
⚠ Product per Young's notes; add purchase link/price if wanted in the cost box.

[22] Anycubic · "Wash & Cure 3" product page, https://store.anycubic.com/products/wash-cure-3
✅ Fetched 2026-08-04: sale $89.00, regular $129.00. Used as the commercial-alternative price anchor in §M4d.

[23] Wise mid-market rate, https://wise.com/us/currency-converter/usd-to-krw-rate
✅ Fetched 2026-08-04: "1 USD = 1,453 KRW". Derived in §M4d: 5,990 KRW ≈ $4.12; 3,990 KRW ≈ $2.75.

[24] Bambu Lab A1 mini · specs (§M5). Build volume "180 x 180 x 180 mm"; AMS Lite is an optional multicolor add-on (Young's unit: without).
✅ Fetched 2026-08-04 via 3DPros printer database (official bambulab.com tech-specs page blocked fetching). ⚠ Re-confirm against the official page before print publication.
Correction logged 2026-08-04: project files (CLAUDE.md §2, README §4) and earlier write-up sections said "Bambu Lab A1" · Young corrected this to A1 mini, no AMS. M2 + outline updated; CLAUDE.md/README rows still need the fix (doc-debt).

[25] Coupang · FM 라디오 조립 키트 (FM radio assembly kit; soldering practice, DIY), https://www.coupang.com/vp/products/8215678023?itemId=23597303202&vendorItemId=91764318217
✅ Listing URL provided by Young 2026-08-04 (tracking params stripped). §M3b practice kit.

[26] Coupang · 주사기 솔더 플럭스 페이스트 (syringe solder flux paste), https://www.coupang.com/vp/products/8740212682?vendorItemId=92633572403
✅ Listing URL provided by Young 2026-08-04. §M3b consumable.

[27] Coupang · 국산 송진 실납 70g / 신성 실납 (Korean rosin-core solder wire, 70 g, Shinsung), https://www.coupang.com/vp/products/6187858707?vendorItemId=94657738492
✅ Listing URL provided by Young 2026-08-04. §M3b consumable. (Coupang listings drift/expire; re-check before print.)

Internal project files quoted or relied on

• 📁 CLAUDE.md §2 — ground-truth table (stack datasheet row; 15 V p-p drive correction of 2026-07-17)
• 📁 research_context.md — piezo limitation list (0.1% strain, 10–25% hysteresis, creep, brittleness, 60–200 V)
• 📁 first_motion_playbook.md — stick/slip inequalities; 5 Hz visibility analysis
• 📁 benchmark_table.md — targets and "metrics nobody reports" gap list
• 📁 paper_analyses/001–004 — quote extractions for [5]–[8]
• 📁 sla_tolerance_and_fits_guide.md — printed-fit tolerance experience
• 📁 Bench log 2026-07-30/31 — first travel: 2.51 mm / 10 min at N≈0 (user-measured against a masking-tape zero mark). Derived: 4.18 µm/s; ≈0.84 µm/cycle at 5 Hz.
• 📁 run_log_2026-08-10_no_motion_session.md — 8 null trials, N = 0→1.0 N, both polarities; drive measured at the stack wires: 14.7 V p-p, flyback ≈42 µs (commanded 60 µs). Derived: ≈0.63 A amplifier current into 1.8 µF.
• 📁 run_log_2026-08-11_motion_returned_then_died.md — motion ×2 on a fresh gray flexure, ~1 mm / 10 min at N≈0 (Derived: 1.67 µm/s ≈ 0.33 µm/cycle), then six nulls surviving a plate flip and a factory-clean plate.
• 📁 run_log_2026-08-12-13_new_stack_nulls.md — new stack, new plate, third flexure print; three trials, no reproducible motion. Component-swap scoreboard.
• 📁 run_log_2026-08-20_v4_direction_reversal.md — direction reversal demonstrated: FWD 3.09 mm / 11 min, signal-reversed BWD 0.54 mm / 13 min, 14.7 V p-p at 4.98 Hz, near-contact preload. Derived: 4.68 and 0.69 µm/s; 0.94 and 0.14 µm/cycle; 6.8× asymmetry. Root cause of the August nulls: the stack must be clamped rigidly in its printed pocket. ⚠ Exact trial date recorded as "all today" (19th or 20th).
• 📁 Correspondence with Prof. E.-T. Hwu, 2026-08-09 → 08-19 (Gmail thread "Driver") — Gerber redistribution permission; co-authorship declined; the magnet-coupling and blade-track suggestions; and the only record of the 2026-08-16 run (23-minute run filmed at 60×, 3.17 mm in the first 13 minutes ≈ 4 µm/s while moving, stall in the final third, sawtooth amplitude down by the end, polarity flip not yet reversing direction).

Figure-to-source map

§M2 additions · internal sources

• 📁 Builder description, 2026-07-31 (primary source): custom rail = bearing bolts riding printed V-grooves on bottom (carries gravity) AND side (reacts lateral actuator preload); stack drives a printed dome-tip resin foot, no flexure in current rig; interferometer laser currently the 520 nm 5 mW diode module (→ λ/2 = 260 nm per fringe).
• 📁 phase1_30V_build_guide.md — drive-chain architecture (Timer 4, PORTA/C/L; TDA2050 as I→V + power; 15 kΩ ∥ 68 pF)
• 📁 interferometer_sensing_setup.md + displacement_metrology_guide.md — sensing chain, BPW34 biasing, 50 kΩ (5×10 kΩ) load
• 📁 design_notes.md §16 (friction pair, galling rationale) and §18 (preload calibration 0.099 N/tick)
• 📁 Bench scope measurements 2026-07-13 (amplitude vs. frequency into 1.8 µF) and 2026-07-17 (15 V p-p unipolar output correction), recorded in project CLAUDE.md §2
• ⚠ Open citations for §M2, stated as unrecorded in the article itself rather than guessed: NICEPOWER PSU model number; bearing-bolt count and type. (Stack mount geometry was answered 2026-08-06: printed pocket + clamping screw, no adhesive — and §M11 explains why that screw turned out to matter.)