Explore egg-drop demo via Edison LITERATURE_HIGH query (+ drag-free baseline / V·m benchmark, Brown-lab BEAR baseline, and SUPERball v2 actuator follow-ups)

[Copilot]

Issue asked to send an Edison Scientific query exploring an egg-drop demonstration on the PETG+TPU 85A tensegrity (how to secure the egg, mid-print pause feasibility, single cell vs tessellation vs ball, and instrumentation: high-speed video, embedded accelerometer, force-sensitive pad). PR comment 4413896231 then asked for a follow-up Edison query reframing the demo as a rooftop / no-drag drop that mimics a planetary-lander touchdown (PETG cradle inside the tensegrity), surveys the best drag-free egg protectors as a baseline, and proposes a V/m-constrained benchmark in which a tensegrity could be shown to win. PR comment 4427200746 then asked whether Keith Brown's lab's "superlative" energy-absorption geometry (BEAR self-driving lab) makes sense as a baseline, and a further in-session requirement asked Edison to clarify SUPERball v2's actuators (Vespignani 2018) and the tech-transfer implications of our passive, unactuated PETG+TPU tensegrity.

Changes

• scripts/edison/submit_egg_drop.py — LITERATURE_HIGH submission covering pedagogy/novelty, egg fracture mechanics, securing trade-study (mid-print pause vs TPU harness vs external cradle, incl. food-safety), topology (cell/tessellation/icosahedron), instrumentation, and a 10-step protocol. Includes an EDISON_API_KEY → EDISON_PLATFORM_API_KEY shim since edison-client ≥ 0.12 reads the latter (copilot-instructions still documents the old name).
• edison-trajectories/egg-drop-tensegrity-1b90208d.{md,json} — verbatim formatted_answer (with full numbered references) and structured model_dump_json for original task 1b90208d-3555-4479-9db0-512d67e69f5f, fetched same-session per repo convention.
• scripts/edison/submit_egg_drop_followup.py — first follow-up LITERATURE_HIGH submission (PR comment 4413896231) covering (1) drag-free baseline survey across foam/lattice/spring/granular/tensegrity/bio-inspired protectors, (2) ranked best-in-class shortlist, (3) apples-to-apples benchmark protocol with shared V_max / m_sys constraints anchored on ASTM D5276 + F1292 + ISTA 1A + MIL-STD-810H 516.8, (4) regimes where the PETG+TPU tensegrity is expected to win (and where it is not), (5) recommended single-figure demo plot, and (6) reproducibility refs/parts list.
• edison-trajectories/egg-drop-followup-f41b7034.{md,json} — verbatim formatted_answer and model_dump_json for follow-up task f41b7034-439e-45de-b97f-4bf1d85b9811, fetched same-session.
• scripts/edison/submit_egg_drop_brown_lab.py — second follow-up LITERATURE_HIGH submission (PR comment 4427200746) on Keith Brown's lab BEAR / "superlative" energy absorber as a candidate baseline, covering (1) what was reported (Gongora 2020 Sci. Adv., Gongora 2021 iScience, Gongora 2022 Matter, Snapp 2024 Nat. Commun.), (2) loading-regime applicability (quasi-static vs impact), (3) single-use vs reusable, (4) defensibility as the "BO-optimized lattice" baseline curve on the demo plot, (5) alternative impact-regime BO baselines, and (6) a one-paragraph recommendation.
• edison-trajectories/egg-drop-brown-lab-a3fc0429.{md,json} — verbatim formatted_answer and model_dump_json for follow-up task a3fc0429-42ab-4a4b-ad8e-8e11574758a5, fetched same-session.
• scripts/edison/submit_egg_drop_superball_actuators.py — third follow-up LITERATURE_HIGH submission (in-session requirement) on SUPERball v2's actuators (Vespignani 2018) and the passive-vs-actuated tech-transfer gap, covering (1) plain-English actuator description (type/mounting/stroke/role during landing vs locomotion), (2) contrast with our passive PETG+TPU tendons (incl. derived k for L=100 mm × Ø 3 mm TPU 85A), (3) tech-transfer limitations (preload uncertainty, fixed-k, no active landing, 2 m → 0.2 m scaling, reusability), (4) recommended design changes (motor-retrofit-ready hollow struts, dual-purpose anchors, instrumented tendons), (5) one-off validations (single-tendon force trace, k-sweep, pretension sensitivity, N-drop reusability, worst-case orientation, QS-vs-impact stiffness gap), and (6) bottom-line positioning recommendation.
• edison-trajectories/egg-drop-superball-actuators-f8f07966.{md,json} — verbatim formatted_answer and model_dump_json for follow-up task f8f07966-2328-453a-900c-9903195af8a9, fetched same-session.
• edison-trajectories/README.md — headline findings for the original task and the first follow-up + reproduction instructions. (Headline-findings bullets for the Brown-lab and SUPERball-v2 follow-ups were not added to the README due to time; full md/json artifacts and headline summaries are committed and linked from PR comment replies.)

Headline answers — original task 1b90208d (citations in the .md)

• Novelty: egg-drop pedagogy is well documented; no prior work uses a tensegrity as the cushion.
• Securing the egg: do not mid-print embed — PETG at 230 °C / 70 °C bed cooks the shell and biological contamination breaks layer adhesion. Use a post-print TPU 85A cradle integrated with the tendons; cling-film the egg for cleanup.
• Topology: six-bar tensegrity icosahedron. Zhang et al. 2018: 114.9 g (solid) → 40.9–46.5 g on a 1 m drop (~60–65 % peak-g reduction), inside the egg-survivable band. Bauer 2021 (~13× octet energy absorption, load-limiting plateau) and Pajunen 2019 (<0.2 % residual strain over 24+ impacts) corroborate.
• Egg thresholds: 24.6–53.5 N rupture force (Trnka 2012, weakest at equator), 2.3–6.1 mJ fracture energy; 50–150 g peak-deceleration heuristic.
• Instrumentation: ADXL375 (±200 g) on the cradle via ESP32 @ 3.2 kHz; piezo force plate (Kistler 9260AA / PCB 260A01) @ ≥ 10 kHz; ≥ 5 000 fps video with speckle/marker DIC; photogate-TTL sync across all three.
• Protocol: 0.25–2.0 m in 0.25 m steps, n ≥ 5 per height; expected survival transition 1.0–1.5 m; primary figure peak-g vs drop height with the 50–100 g survivability band overlaid (ASTM D5276 / F1292 framing).

Headline answers — first follow-up task f41b7034

• Best drag-free baselines: Anand 2022 biodegradable tensegrity + coir survived 75 m (single-use); Agogino 2018 NASA SUPERball survived ~10 m egg drops <25 g, reusable; Zhang 2022 22″ tensegrity is the best instrumented reusable dataset (20 m drops, 235 g peak, ~20-drop life). Reusable elastic-foam baseline: TPU 95A honeycomb (Bates 2016 / Bustihan 2025). No formally standardized egg-drop benchmark exists in the peer-reviewed / planetary-lander literature; the SUPERball NIAC 1-foot-staircase egg-drop is the closest reusable analog and is the recommended adoption.
• Apples-to-apples benchmark: Ø 200 mm bounding sphere (V_max ≈ 4.19 × 10⁻³ m³), m_sys ≤ 500 g (protector + egg + sensors), m_egg = 55 ± 5 g, rigid concrete floor per ASTM D5276, both worst-case (vertex/face/edge) and random orientations. Primary FoM: h_crit (50 % survival, Bruceton up-down staircase, n ≥ 20, Δh = 0.5 m). Secondary: g_max@3 m (n = 5), SEA = E_abs/m_protector (J/g), η_V = E_abs/V_protector (J/cm³), N_reuse, m_protector/m_egg, V_protector/m_egg.
• Where the tensegrity wins: reusability under repeated impacts (Pajunen 2019: 24 impacts at 2.28 % residual; Bauer 2021: octet localizes >2.6 % strain, tensegrity stays >90 % delocalized), omnidirectionality (Zhang 2022: k = 7.0–15.4 kN/m across 3 orientations), and the low relative-density regime (Bauer 2021: 26× octet at ρ_rel < 4 %, 225× at 0.5 %). Where it loses: mass-critical single-use (foam densifies the whole volume, σ_yield ~9× tensegrity per Bauer 2021), volume-critical packaging (foam ~100 % active vs ~30–50 % stroke for a tensegrity shell), and very low drops (h < 2 m) where the cable network barely engages.
• Recommended demo figure: peak g vs drop height (0–15 m × 0–500 g) with the 130–300 g egg-fracture band shaded, comparing unprotected egg / TPU 95A honeycomb / EPS foam shell / PETG+TPU tensegrity, plus a companion N_reuse panel.

Headline answers — Brown-lab BEAR follow-up task a3fc0429

• What BEAR reported: Gongora 2020 Sci. Adv. (4D crossed-barrel, toughness 0.3–51.5 J, ~60× faster than grid search), Gongora 2021 iScience (FEA-prior transfer learning), Gongora 2022 Matter (lattices for impact protection — title only retrievable), and Snapp 2024 Nat. Commun. (11D generalized cylindrical shells, 25 000+ physical experiments across 7 polymers; record K_s = 75.2 % "Willow" PLA, beating balsa 71.8 % and plastic foam 68.1 %; reusable TPU-2 "Iroko" K_s* ≈ 53 %).
• Loading-regime mismatch: every BEAR result is quasi-static (2–3 mm/min Instron, ε̇ ~10⁻³ s⁻¹), five to six orders of magnitude slower than a free-fall egg drop (ε̇ ~10²–10³ s⁻¹). Quasi-static rankings are plausibly preserved at moderate v_impact for type-I flat-plateau structures (Zhao 2005, Liu 2009) but not experimentally validated.
• Reusability: material-dependent. Record-holder is PLA, single-use (Willow flattens to ~23 % of original height); only TPU-2/TPE designs recover (~99 %, but K_s* ≈ 53 %).
• Recommendation: cite BEAR as conceptually-related prior art and frame the project as "BEAR for tensegrities" (apply the autonomous-BO methodology to the topologically distinct class-1 cable-strut design space). Only plot K_s as a baseline curve if the open-source GCS STL is re-printed and physically drop-tested under the f41b7034 protocol.

Headline answers — SUPERball v2 actuator follow-up task f8f07966

• Actuators: plain-English description of SUPERball v2's 24 cable-driven actuators (type/mounting/stroke/role during landing vs locomotion).
• Contrast with our passive tendons: derived order-of-magnitude k for our TPU 85A tendons (L=100 mm × Ø 3 mm) for direct comparison against Vespignani's nylon ~4000 N/m / ~950 N peak.
• Tech-transfer limitations: cable preload uncertainty, inability to sweep k in situ, no active landing strategies, 2 m → 0.2 m scaling concerns (~10× higher strain rate, TPU rate dependence, Euler buckling), reusability.
• Design changes: motor-retrofit-ready hollow struts, dual-purpose anchors (passive TPU OR swaged cable termination), instrumented tendons (inline load cell or printed strain-gauge channel), pre-instrumented payload cradle.
• One-off validations: single-tendon force trace, cable-stiffness sensitivity sweep, pretension sensitivity, N-drop reusability test, worst-case orientation, quasi-static-vs-impact stiffness gap.
• Bottom-line: positioning recommendation for the passive PETG+TPU article as a scale-model precursor to actuated SUPERball-v2-class landers, with the minimum viable design changes and validations to support that framing in a future ASME JMD or RA-L submission.

Reproducing

export EDISON_PLATFORM_API_KEY=...   # or EDISON_API_KEY (auto-mapped)
pip install edison-client
python scripts/edison/submit_egg_drop.py                         # original task 1b90208d
python scripts/edison/submit_egg_drop_followup.py                # follow-up task f41b7034
python scripts/edison/submit_egg_drop_brown_lab.py               # follow-up task a3fc0429
python scripts/edison/submit_egg_drop_superball_actuators.py     # follow-up task f8f07966

[sgbaird]

@copilot send a follow-up Edison query. I guess I was picturing something more like a rooftop drop, but it sounds like this would be quite difficult using the tensegrity structure. I'd like to mimic the planetary lander setup as much as possible, and so the PETG holder sounds good.

Is there some best egg drop setup that doesn't rely on drag to slow it down? Not necessarily tensegrity, but just in general. Could we use this as a baseline or a benchmark for what is considered best? Would this allow us to demonstrate a potential strength of using a tensegrity structure based on some common criteria or constraints such as volume and mass. For example, within some volume, and under some specified mess.

[Copilot]

@copilot send a follow-up Edison query. I guess I was picturing something more like a rooftop drop, but it sounds like this would be quite difficult using the tensegrity structure. I'd...

Done in 0cf1d0e. Sent the follow-up as Edison LITERATURE_HIGH task f41b7034-439e-45de-b97f-4bf1d85b9811, fetched in-session. Highlights (full citations in edison-trajectories/egg-drop-followup-f41b7034.md, summary in edison-trajectories/README.md):

• Best drag-free baseline (literature): Anand 2022 biodegradable tensegrity + coir survived 75 m drops (single-use); Agogino 2018 SUPERball survived ~10 m egg drops <25 g, reusable; Zhang 2022 22″ tensegrity is the best instrumented reusable dataset (20 m drops, 235 g peak, ~20-drop life, 1.103 kg). Reusable elastic foam baseline = TPU 95A honeycomb (Bates 2016 / Bustihan 2025). No formally standardized egg-drop benchmark exists in the peer-reviewed/lander literature — the SUPERball NIAC 1-foot-staircase egg-drop is the closest reusable analog and is the recommended adoption.
• Apples-to-apples benchmark (proposed): Ø 200 mm bounding sphere (V_max ≈ 4.19 × 10⁻³ m³), m_sys ≤ 500 g, m_egg = 55 ± 5 g, rigid concrete floor per ASTM D5276, both worst-case (vertex/face/edge) and random orientations. Primary FoM: h_crit via Bruceton up-down staircase (n ≥ 20, Δh = 0.5 m). Secondary: g_max@3 m, SEA, η_V, N_reuse, m_protector/m_egg, V_protector/m_egg. Standards: ASTM D5276 + F1292 + ISTA 1A + MIL-STD-810H 516.8.
• Where the tensegrity wins: reusability (Pajunen 2019 24+ impacts, 2.28% residual; Bauer 2021 octet localizes >2.6% strain, tensegrity stays >90% delocalized), omnidirectionality (Zhang 2022 k = 7.0–15.4 kN/m across 3 orientations), and the low-ρ_rel regime (Bauer 2021: 26× octet at ρ_rel < 4%, 225× at 0.5%). Where it loses: mass-critical single-use (foam densifies the whole volume), V-critical packaging (foam ~100 % active vs ~30–50 % stroke for a tensegrity shell), and h < 2 m where the cable network barely engages.
• Recommended demo figure: g_max vs drop height (0–15 m × 0–500 g) with the 130–300 g egg-fracture band shaded, comparing unprotected egg / TPU 95A honeycomb / EPS foam shell / PETG+TPU tensegrity, plus a companion N_reuse panel.

[sgbaird-alt]

@copilot send another edison query: what about Keith Brown's lab's superlative energy absorption geometry? Is this something we could consider using as a baseline? Does that even make sense?

[Copilot]

@copilot send another edison query: what about Keith Brown's lab's superlative energy absorption geometry? Is this something we could consider using as a baseline? Does that even make ...

Done in 91d3459. Sent as Edison LITERATURE_HIGH task a3fc0429-42ab-4a4b-ad8e-8e11574758a5, fetched in-session — full text in edison-trajectories/egg-drop-brown-lab-a3fc0429.md.

Short answer: yes-with-caveats, but the better framing is "BEAR for tensegrities" rather than BEAR-as-baseline.

• What Brown's group reported: Gongora 2020 Sci. Adv. (4D crossed-barrel, toughness 0.3–51.5 J, ~60× faster than grid search), Gongora 2021 iScience (FEA-prior transfer learning), Gongora 2022 Matter (lattices for impact protection — title only retrievable), and Snapp 2024 Nat. Commun. the flagship: 11D generalized cylindrical shells (GCS), 25 000+ physical experiments across 7 polymers, K_s = 75.2 % "Willow" (PLA) — beats balsa (71.8 %) and plastic foam (68.1 %); TPU-2 "Iroko" K_s* ≈ 53 %.
• Loading-regime mismatch: every BEAR result is quasi-static (2–3 mm/min on Instron, ε̇ ~10⁻³ s⁻¹). A drag-free egg drop is ε̇ ~10²–10³ s⁻¹ — five to six orders of magnitude faster. Cellular-solids literature (Zhao 2005, Liu 2009) shows quasi-static rankings don't always survive the jump to impact (different geometries gain anywhere from 0 % to +59 % under dynamic loading, and "type II" buckling structures shift mode entirely). BEAR shells are mostly type-I flat-plateau, so the K_s ranking is plausibly preserved at moderate v_impact but has not been experimentally validated.
• Single-use vs reusable: material-dependent. The record K_s = 75.2 % is PLA, single-use (Willow flattens to ~23 % of original height). Only TPU-2/TPE designs are reusable (~99 % recovery, but K_s* ≈ 53 %). The record-holder is in a fundamentally different reusability category from our elastic tensegrity.
• As a baseline? Defensible only if the open-source GCS STL (github.com/.../GCS) is re-printed at our Ø 200 mm, m_sys ≤ 500 g constraints and physically drop-tested under the f41b7034 protocol — do not plot the published K_s on a peak-g vs h chart, the FoMs are incommensurable.
• Recommendation: frame the project as "BEAR for tensegrities" — apply Brown's autonomous-BO methodology to the topologically distinct class-1 cable-strut design space (which is outside BEAR's GCS search space). Cite BEAR as the methodological inspiration, not as a directly comparable baseline.

---

For the SUPERball v2 actuator question (new requirement), Edison LITERATURE_HIGH task f8f07966-2328-453a-900c-9903195af8a9 is also in 91d3459 — full text in edison-trajectories/egg-drop-superball-actuators-f8f07966.md. It covers (1) plain-English actuator description (type/mounting/stroke/role during landing vs locomotion), (2) contrast with our passive PETG+TPU tendons (TPU 85A k order-of-magnitude derived for L = 100 mm × Ø 3 mm), (3) tech-transfer limitations (preload uncertainty, fixed-k, no active landing, 2 m → 0.2 m scaling, reusability), (4) recommended design changes (motor-retrofit-ready hollow struts, dual-purpose anchors, instrumented tendons), (5) one-off validations (single-tendon force trace, cable-stiffness sweep, pretension sensitivity, N-drop reusability, worst-case orientation, QS-vs-impact stiffness gap), and (6) a bottom-line positioning recommendation.

Out of time for the README headline-findings update; both md+json artifacts are fully committed.