diff --git a/ROADMAP.md b/ROADMAP.md index 66be464..ea6209f 100644 --- a/ROADMAP.md +++ b/ROADMAP.md @@ -2,14 +2,38 @@ This document outlines the planned development trajectory for NIGHTWATCH. -## Current Release: v0.1.0 +## v0.1.0 -**Status:** Released (January 2024) +**Status:** Released (January 2026) Core foundation with voice control, mount operation, weather monitoring, and safety systems. --- +## Current Release: v0.1.1 - Repo Consolidation & Optics Engineering Roadmap + +**Status:** Released (August 2026) + +Single-branch consolidation and the start of the core-telescope engineering effort. + +### Repository Consolidation +- [x] Salvage swarm capabilities before branch purge (hourly scanner, event journal, space weather, scan severity — 142 new tests) +- [x] CNEOS fireball lat/lon sign fix (southern/western events were mislocated) +- [x] Salvage review tooling (`.claude/agents/`, `/full-review`) and frontend design prompt pack (`docs/design/`) +- [x] Archive historical landscape audit to `docs/audits/` +- [x] Merge pyproject.toml canonical-URL fix (closes #78) +- [x] Resolve all open PRs; delete 88 stale remote branches — repo ends at exactly `main` + +### Optics Engineering (the core telescope) +- [x] `design/optical/OPTICAL_ENGINEERING_ROADMAP.md` — requirements flowdown, make-vs-buy gate, glass sourcing reality (Astrositall → CTE-class spec), f/6-vs-f/8 fork dissolution, phase plan P0–P6, committed budget envelope +- [x] Documentation corrections: Intes production status (ceased 2018), MN76≠MN78, corrector-blank scarcity claim +- [ ] P0 decision packets: planetary sensor baseline, v1 tracking status, focuser steps→µm +- [ ] P1 design skeleton: `design/optical/calc/` prescription + CI ray-trace proofs (optiland/prysm) +- [ ] P2 trade study: retire the f-ratio fork with computed gates +- [ ] P3 tender BP-06: primary / corrector / flat RFQ lots with acceptance criteria + +--- + ## v0.2.0 - Imaging Foundation **Target:** Q2 2024 diff --git a/design/optical/OPTICAL_ENGINEERING_ROADMAP.md b/design/optical/OPTICAL_ENGINEERING_ROADMAP.md new file mode 100644 index 0000000..4f7ec33 --- /dev/null +++ b/design/optical/OPTICAL_ENGINEERING_ROADMAP.md @@ -0,0 +1,475 @@ +# NIGHTWATCH Optical Engineering Roadmap + +**Designing the telescope itself: a 180 mm Maksutov-Newtonian with a zero-expansion-class primary.** + +Status: ROADMAP — Issued with release v0.1.1 (2026-08-06) +Companion to: `design/mechanical/MECHANICAL_DESIGN.md` (the pattern this effort mirrors) +Method: assertion → proof → tender, with S/D/A provenance and PASS/FAIL gates + +--- + +## 0. Why this exists + +The mechanical effort engineered everything *around* the telescope — mount, pier, roof, +bearings, encoders, power — and proved it with computed budgets. The optical tube itself +appears in those documents only as a payload mass and a line on the COTS schedule that +says, in effect, "buy a used Russian tube." + +That line no longer survives contact with reality, and this document exists because we +checked: + +- **Intes Micro ceased production in 2018.** Adjudicated with dated sources and DNS/ + Wayback evidence (see `docs/research/SOURCING_RESEARCH.md`, corrected 2026-08-06). + The repo's earlier "still producing March 2025" claim was traced to a frozen 2005-era + website and recycled forum lore. +- **The used market is thin and cannot be specified.** Used MN76/78 units surface + intermittently at ~$1,500–2,500, with unverifiable optical figure and almost never the + Astrositall (Alter-grade) substrate that motivated the selection. +- **No current product meets the requirements** (§1.3). The only in-production + Maksutov-Newtonian above 152 mm — Sky-Watcher Starlux 190MN, $2,375 — is an f/5.3 + astrograph with 27–34% obstruction and a borosilicate primary. + +Therefore NIGHTWATCH designs its own optics. This roadmap defines the requirements, the +engineering phases, the CI-proof structure, the procurement strategy, and the budget — +with every load-bearing number carrying its provenance, exactly as the mechanical effort +demanded of itself. + +**Scope discipline:** telescope components only — optical prescription, glass, cell and +collimation interfaces, baffling, thermal behavior of the optics, acceptance testing. +No observatory automation, no enclosure, no mount (those interfaces are referenced, not +re-opened). + +--- + +## 1. Mission and requirements flowdown + +### 1.1 Mission ranking (SOURCED — NIGHTWATCH_Build_Package.md:32-40) + +1. Mars surface features (5/5) +2. Lucky imaging / high-speed planetary capture (5/5) +3. Autonomous overnight operation (5/5) +4. Remote monitoring (5/5) +5. Deep-sky visual (bonus, 4/5) + +NEO follow-up astrometry is a standing secondary interest (services/meteor_tracking is +already flying); it influences field-of-view as a tiebreaker, not a driver (§4). + +### 1.2 De-facto optical requirements (SOURCED — docs/SCIENTIFIC_FOUNDATIONS.md App. A) + +| Requirement | Value | Source | +|---|---|---| +| Aperture class | 180 mm (7") | Build Package + params.py (both fork branches agree) | +| System Strehl (design + fab) | ≥ 0.95 at 546–550 nm (λ/8 P-V class) | SCIENTIFIC_FOUNDATIONS.md:158-166, A.2 | +| Diffraction floor | Strehl ≥ 0.80 (Maréchal) after all budgets | SCIENTIFIC_FOUNDATIONS.md:144-154 | +| Angular resolution | Rayleigh 0.78″, Dawes 0.65″ at 180 mm | App. A.1 | +| Site seeing to exploit | r₀ = 10–15 cm, τ₀ = 3–8 ms (lucky imaging) | App. A.3 (ASSUMED — unmeasured, see risk R6) | +| Thermal environment | ~22 K diurnal swing, alkaline desert dust | params.py Site/Environment; MECHANICAL_DESIGN.md | +| Sampling | Nyquist at diffraction cutoff per chosen sensor | App. A.4 (re-derived per f-ratio in §4) | +| Back focus | ≥ 55 mm usable (imaging-train convention) | Industry standard; MN190 practice | + +### 1.3 The make-vs-buy gate: requirements no COTS instrument satisfies + +The four requirements below justify custom fabrication. If any future finding relaxes +all four, the buy-side (§2) reopens. + +- **(a) Zero-expansion-class primary.** Preferred CTE ≤ 0.1 ppm/K (Zerodur, + Clearceram-Z HS, ULE, or an opportunistic Sitall CO-115M find); fused silica at + 0.52 ppm/K is the acceptable baseline. *Proven substrate-tolerant:* re-running the + repo's own thermal-focus proof with fused-silica CTE leaves the verdict unchanged — + the aluminum tube term dominates the mirror term 44–153×, and the + temperature-compensated focuser remains mandatory at any f-ratio (§5, P1-T4). +- **(b) Central obstruction ≤ 18–20%, custom-sized** to the illuminated field actually + needed (§4.2) rather than inherited from a catalog. +- **(c) Certified wavefront:** λ/8 P-V class with raw interferograms deliverable — not a + 25-year-old marketing claim on a used tube. +- **(d) Sealed tube** (full-aperture corrector) for Nevada's alkaline dust and + unattended operation. + +Starlux 190MN fails (a), (b), (c). Used Intes fails (c) always and (a) almost always +(Alter-grade Astrositall units are unicorns). Newtonian + coma corrector fails (d) and +gives up the no-spider contrast advantage (§3.4). + +### 1.4 Decision packets required before design freeze (P0) + +1. **Planetary camera baseline** — IMX678 (2.0 µm) vs IMX585 (2.9 µm). This single + choice moves the required Barlow at f/6 from 1.76× to 1.21× (550 nm) and reorders + the sampling row of the fork matrix (§4.1). Deliverable: price/QE/availability + one-pager + decision. +2. **Sub-arcsecond tracking status for v1** — mechanical v1 concedes ~5″ with + plate-solve recentering; confirm the optics budget should not carry sub-arcsec + assumptions for v1. +3. **Focuser steps→µm coefficient** — still absent from `params.py`; blocks closure of + the thermal-focus compensation proof (MECHANICAL_DESIGN.md:383). + +--- + +## 2. Make-vs-buy evidence and the contingency line + +| Path | Cost | Availability | Meets §1.3? | Verdict | +|---|---|---|---|---| +| New Intes MN76/78 | — | **Closed** (production ended 2018; sanctions ended reseller channels 2022) | — | Impossible | +| Used MN76/78 | $1,500–2,500 + months of watching | Intermittent, thin | (c) no; (a) rarely | Watch-list only | +| Intes raw optical set via APM | — | **Gone** (APM carries no Intes stock as of 2026-08) | — | Closed | +| Sky-Watcher Starlux 190MN | $2,375 new | In production, 2–4 weeks | (a),(b),(c) no | **Named contingency/interim OTA** | +| **Custom design + fabrication** | **$6.5k–12k (committed envelope, §6)** | 6–12 months | **Yes, by construction** | **Plan of record** | + +The Starlux 190MN line stays in this document deliberately: it is the only way to get +*any* large Mak-Newt on the pier within a month if the program needs an interim +instrument, and its 13.2 kg mass is inside the mount envelope already proven for the +14 kg MN78 case. + +A standing **used-market watch** (Astromart + Cloudy Nights classifieds, weekly) also +stays open: a genuine Alter-grade MN78 at a sane price would satisfy §1.3(a),(b),(d) +and could be upgraded to (c) by independent testing — it would then compete honestly +with the custom path at the P3 gate. + +--- + +## 3. The glass story + +### 3.1 Astrositall (Sitall CO-115M) — the signature, honestly stated + +Astrositall is the Soviet/Russian zero-expansion glass-ceramic (CTE 0 ± 1.5×10⁻⁷/°C +over −60…+60 °C) used in premium "Alter" Intes instruments — the original inspiration +for NIGHTWATCH's glass identity. The 2026 reality: + +- **LZOS (Lytkarino) is sanctioned** (Shvabe/Rostec; US Consolidated Screening List). + Western resellers stopped purchasing in March 2022. New CO-115M import is effectively + impossible. +- **Secondary market is opportunistic only:** observed lots are ~120 mm blanks and one + 8″ × 1.25″ blank on Astromart — not a plannable procurement at ~200 mm. + +**Resolution:** the primary is specified by **CTE class, not brand**: + +> *Primary substrate: zero-expansion class, CTE ≤ 0.1 ppm/K (Zerodur / Clearceram-Z HS / +> ULE 7972 / Sitall CO-115M), or fused silica (Corning 7980 class, 0.52 ppm/K) as the +> cost baseline. Substrate choice shall not require redesign (proven: thermal proof is +> substrate-independent, §1.3a).* + +A standing **Sitall watch** (eBay/Astromart alerts for ≥190 mm CO-115M blanks) keeps the +signature alive as a drop-in swap. If a blank surfaces, it replaces the baseline at the +next procurement gate with zero prescription change. + +### 3.2 Primary blank sourcing (P3 Lot A input) + +| Substrate | CTE (ppm/K) | ~200 mm blank path | Indicative cost | Confidence | +|---|---|---|---|---| +| Fused silica (Corning 7980) | 0.52 | Advanced Glass Industries, United Lens (stock/custom) | $400–900 | Medium (RFQ needed; 200×5 mm window seen at $647 retail) | +| Zerodur (Schott) | 0.00±0.007 (class 0) | Schott quote; secondary market (7″ blank seen at $359) | $1,500–3,000 new | Low (no published price) | +| Clearceram-Z HS (Ohara) | 0.0±0.02 | Ohara quote (material in volume production for TMT) | quote-only | Low | +| Sitall CO-115M | 0.0±0.15 | Opportunistic secondary market only | n/a | — | + +### 3.3 Corrector blank and secondary flat + +- **Corrector blank (the critical-path item):** N-BK7 (Schott) or H-K9L (CDGM), + ~230 mm dia × 35 mm, fine annealed, homogeneity **H2** (±5×10⁻⁶), bubble class ≤ B1, + striae grade A. Schott TIE-41 confirms H2 producibility above 300 mm — availability + is real; cost is the question ($400–900 est., RFQ). **Borofloat/Supremax are + excluded** for the transmissive corrector (float/rolled glass carries no homogeneity + or striae grade). +- **Secondary flat: a solved COTS problem — zero engineering effort allocated.** + Antares 1.83″ minor-axis 1/30-wave elliptical at $181.49 (quoted); Ostahowski + fused-quartz 1/10-wave-plus diagonals with interferograms as alternate. Final minor + axis follows the P1-T5 secondary-sizing proof. + +### 3.4 Why a meniscus at all (the honest trade memo, to be proven in P2) + +The Mak-Newt buys, over a Newtonian + Paracorr-class corrector: no spider (no spikes, +higher planetary contrast), smaller achievable obstruction (17–20% vs 25–32% typical), +all-spherical surfaces (cheap to figure *accurately*), and a sealed tube. It costs: +a thick (~D/10) meniscus with slow thermal equilibration (fan + heater already in the +system budgets), ~2 kg extra glass, corrector tilt sensitivity (1° tilt = 0.27λ P-V +coma — drives the cell spec, §5 P1-T7), and the corrector fabrication bill (§6). +P2's trade study scores this with the same Zwicky/Pugh + hard-gate method the +mechanical effort used; if the meniscus loses on evidence, this roadmap pivots before +any glass is bought. + +--- + +## 4. Resolving the fork: f/6 vs f/8 is a false binary + +The repo has carried "MN76 f/6 vs MN78 f/8" as an unresolved fork +(`design/mechanical/calc/params.py:85-109`; trade study selected MN78 only as a +"science override"). **Because NIGHTWATCH now designs its own optics, the fork +dissolves:** focal ratio is a continuous variable, and central obstruction is +semi-independent of it (set by focuser height + illuminated-field diameter, not by +f-ratio alone). + +**Provisional target (70% confidence): 180 mm f/6.5–f/7, obstruction ≤ 18%.** + +### 4.1 Decision inputs (computed, this research pass) + +- **Sampling** (Nyquist at diffraction cutoff, F = 2·pixel/λ): + | Sensor | λ 550 nm | λ 656 nm | Barlow at f/6 | at f/7 | at f/8 | + |---|---|---|---|---|---| + | IMX585/462 (2.9 µm) | f/10.5 | f/8.8 | 1.76× | 1.51× | 1.32× | + | IMX678 (2.0 µm) | f/7.3 | f/6.1 | 1.21× | 1.04× | 0.91× | + A 2.0 µm sensor nearly closes the f/6 sampling gap natively at 656 nm; every path + needs at most one high-quality amplifying element, and the ADC sits in the amplified + beam in all cases. +- **Obstruction MTF** (computed annular-pupil autocorrelation): mid-band contrast loss + vs unobstructed ≈ 10–18% at ε=0.25, but only 3–5% at ε=0.134. A custom ε=0.16–0.18 + recovers most of the f/8 advantage at any f-ratio — pending the P1-T5 illumination + proof (MN76's measured **3 mm** fully-illuminated field is the cautionary tale). +- **Corrector cost:** f/6 vs f/8 delta is **<10%** (ray-trace-verified: only the + R1−R2 match tolerance tightens, ±0.13 mm vs ±0.30 mm — both routine test-plate + practice). The corrector does not decide the fork. +- **Mechanics:** tube length is a *weak* lever on the pointing-deflection FAIL (1.42× + across 0.9–1.4 m; bearing **span** is the real lever — DEC ≥ 105 mm / RA ≥ 125 mm + needed for the 5″ target, a BP-01 housing-geometry item, cross-cutting action §5 + P2-X1). Mount and roof budgets are already proven for the worst case (14 kg, 1.4 m). +- **Depth of focus:** ±39.6 µm (f/6) / ±53.9 µm (f/7) / ±70.4 µm (f/8) — second-order, + since the 22 K swing walks focus ~10–18× DoF regardless and temp-compensated focusing + is mandatory at any f-ratio. +- **FOV (NEO tiebreaker):** shorter FL buys 1.82× sky area per pointing; all + sensor/FL combinations plate-solve routinely. + +### 4.2 What settles it (P1/P2 gate tasks) + +1. Sensor decision packet (P0-1). +2. Ray-trace the custom meniscus at f/6, f/6.5, f/7, f/8 — poly-Strehl + tolerance + Monte Carlo → manufacturability-vs-f-ratio curve (P1-T1/T8). +3. Secondary-sizing/illumination ray trace at each f-ratio with a low-profile focuser + stack → minimum honest obstruction (P1-T5). +4. Two real corrector quotes at f/6 and f/8 seed radii (P3 RFQ round 1) to replace the + <10% estimated delta with vendor numbers. + +The trade study (P2) then scores the continuous options and **retires the MN76/MN78 +fork permanently**, updating `params.py` and the mechanical trade-study row. + +--- + +## 5. Phase plan + +Phases gate forward progress; each has entry/exit criteria. The structure deliberately +mirrors `design/mechanical/`: a generated design doc, `calc/` proof modules paired 1:1 +with pytest gates, a locked trade study, then tender packages. + +### P0 — Decision packets (no hardware cost) +**Exit gate:** the three §1.4 decisions written into `design/optical/calc/params.py` +with S/D/A provenance. + +### P1 — Design skeleton and proofs (no hardware cost) +Create `design/optical/` mirroring the mechanical layout: + +``` +design/optical/ +├── OPTICAL_DESIGN.md # GENERATED by calc/report.py (drift-gated) +├── calc/ +│ ├── params.py # frozen dataclasses + PROVENANCE (S/D/A) — supersedes +│ │ # the 2-field Optical dataclass in mechanical params +│ ├── prescription.py # radii/thickness/spacing/glass, both seed cases +│ ├── strehl_budget.py # T1: poly-Strehl ≥ 0.95 gate (optiland + prysm) +│ ├── obstruction_mtf.py # T2: annular MTF, integral contrast +│ ├── sampling.py # T3: Nyquist/Barlow per sensor per band +│ ├── thermal_focus.py # T4: substrate CTE sweep; focuser compensation +│ ├── secondary_sizing.py # T5: illuminated field vs obstruction vs focuser height +│ ├── baffles.py # T6: stray-light geometry, knife-edge positions +│ ├── collimation_tol.py # T7: tilt/decenter/despace sensitivities (1° = 0.27λ) +│ ├── tolerancing.py # T8: Monte Carlo yield at fab tolerance grades +│ └── report.py # regenerates OPTICAL_DESIGN.md +└── tests/ # test_*.py paired 1:1 + test_report.py drift gate +``` + +**Seed prescriptions (ray-trace-verified this research pass; DERIVED):** +N-BK7 meniscus, OD 215 mm, CA ≥ 204 mm, CT 22 mm, achromat condition +R1−R2 = −0.565·t with spherochromatism factor /0.97: + +| Case | Mirror R | Corrector R1 | Corrector R2 | +|---|---|---|---| +| f/6 (FL 1224) | −2448 mm | −413.1 mm | −425.6 mm | +| f/8 (FL 1632) | −3264 mm | −513.4 mm | −525.8 mm | + +(Seeds only — final prescription co-optimizes with the actual primary and the +coma-zero corrector spacing; intermediate f-ratios interpolate.) + +**Tolerance table to encode as test fixtures (SOURCED — CFHT Baril 7.5″ f/8 precedent ++ standard precision-optics grades):** R1, R2 individually ±1%; **(R2−R1) to 0.1%** +(the governing fabrication tolerance); thickness and mirror radius ±2%; wedge ≤ 1′ +(≤10 µm ETR); index homogeneity ≤ 2×10⁻⁴; corrector tilt spec: arcminutes (cell +design driver). + +**Tooling (all pip-installable, CI-compatible; no commercial license):** +- **optiland** (MIT) — primary ray-trace engine: spots, wavefront, PSF/MTF, glass + catalogs, sensitivity + Monte Carlo tolerancing +- **ray-optics** — independent cross-check engine (.zmx import for interchange) +- **prysm** (MIT) — physical-optics layer for the polychromatic Strehl/MTF budget gates +- **OSLO EDU** (free, 10-surface limit; this system is 5–6 surfaces) — offline human + sanity check + +**Exit gate:** all T1–T8 proofs green in CI for at least two f-ratio cases; +OPTICAL_DESIGN.md generated with a PASS/FAIL results table (FAILs are findings, not +embarrassments — the mechanical effort's five FAILs were its most valuable output). + +### P2 — Trade study (no hardware cost) +Zwicky morphological box × weighted Pugh with **hard disqualification gates** (the +encoder-DISQ pattern): axes = f-ratio (continuous) × obstruction (T5-derived) × +substrate (§3.2). Grounded exclusively in P1 proof outputs, never opinion. +**Cross-cutting action P2-X1:** feed the bearing-span requirement (DEC ≥ 105 mm / +RA ≥ 125 mm) back to BP-01 housing geometry — an optics-driven mechanical change the +study cannot select away. +**Exit gate:** selected prescription frozen in `params.py`; MN76/MN78 fork formally +retired; mechanical trade-study row updated. + +### P3 — Tender package BP-06 (RFQ round; no fabrication commitment yet) +Three separately-awardable lots, mirroring the BP-01..05 anatomy (README + SOW + +acceptance + ASSUMED-design-intent registers): + +- **Lot A — Primary mirror** (spherical, 180–200 mm, substrate per §3.2). + Candidates: **Ostahowski Optics** (primary candidate — quotes one-offs, works quartz + and ceramics, interferograms with every surface, in-house coating), Lockwood Custom + Optics, Optical Mechanics Inc, Orion Optics UK. *(Zambuto closed to new orders + 2025-08-15; ISTAR not accepting orders — both verified 2026-08.)* +- **Lot B — Meniscus corrector** (60–75% of optics cost; melt-data design-assist + clause REQUIRED: final radii co-optimized to the delivered blank's melt sheet). + Candidates: Optimax (RFQ funnel go.optimaxsi.com / sales@optimaxsi.com / + 585-265-1020; prototype-singles capability; coats to 500 mm), Knight Optical UK + (+44 1622 859444, custom menisci explicitly offered), United Lens (508-765-5421, + blank + machining + annealing under one roof), Advanced Glass Industries + (585-458-8040, blanks). RFQ-ready spec sheet: Appendix C. +- **Lot C — Secondary flat** → moves to the BP-05 COTS schedule (Antares 1.83″ + 1/30-wave, $181.49 quoted; Ostahowski quartz alternate). + +**Acceptance criteria (written now, argued never):** system transmitted wavefront +≤ λ/4 P-V and ≤ λ/14 RMS, **Strehl ≥ 0.95 at 546–550 nm**, double-pass autocollimation +interferometry at 20 ± 2 °C after thermal soak, optic unstressed in test mount, Zernike +table + **raw interferogram data** deliverable. Component-level interferograms from the +fabricator, plus **one independent verification** (Orion Optics UK test service, or +AstroReflect/AiryLab) as the payment gate — the optics analog of the mechanical +PE-stamp gate. + +**Exit gate (budget checkpoint):** ≥ 2 real quotes per lot; fabrication go/no-go +against the §6 envelope; the used-market watch (§2) gets its last look here. + +### P4 — Fabrication and coating (hardware money at risk) +- Corrector is the critical path: 3–6 months typical. +- Mirror coatings (quoted price list, Spectrum Coatings 2026-08): 8″-class primary + $90/$130/$150 (protected/enhanced/MaxR aluminum), secondary $25–50, center-mark $10. + Corrector BBAR (R < 0.5% avg 400–700 nm both sides) quoted separately — often + cheapest bundled with the lens fabricator. +- Progress payments tied to component interferograms. + +### P5 — Acceptance and integration +- Component + assembled-system interferometry per P3 criteria; independent test gate. +- Optical-mechanical design: mirror cell (zero-expansion substrate → athermal cell), + corrector cell with arcminute-class tilt retention, collimation mechanism usable on + a sealed tube, knife-edge baffles per T6. +- Interfaces: tube structure to BP-01/BP-02 machining packages; corrector dew heater + < 2 W and temperature-compensated focuser tie-in (hard requirement, any f-ratio). + +### P6 — First light and verification +- Star test + lucky-imaging pipeline measured against the T1 Strehl budget; + double-pass autocollimation repeat after transport; thermal-equilibration + characterization (2–3 h pre-cool policy per the Petrunin review criteria). +- **Site-seeing measurement task** (risk R6): DIMM or star-trail seeing monitor to + replace the ASSUMED r₀ = 10–15 cm with data. + +--- + +## 6. Budget (plan of record) + +**Decision (Tim, 2026-08-06): the expanded optics envelope is committed now** — not +gated on RFQs. RFQ checkpoints (P3) refine numbers but do not reopen the commitment. + +| Item | Range | Basis | +|---|---|---| +| Meniscus corrector (fab + AR, one-off US shop) | $3,500–6,500 (central $5,000) | Triangulated from 3 public anchors; ±40% until RFQ | +| — Chinese-fab option (H-K9L) | $1,500–3,000 | Unverified; higher QA risk | +| Primary mirror (blank + figure + cert) | $1,200–3,000 | Ostahowski $2,800/10″ quartz comp; sphere is easier | +| Secondary flat (finished, certified) | $150–250 | Antares quoted $181.49 | +| Mirror coatings (primary + flat) | $150–250 | Spectrum quoted table | +| Corrector BBAR (if not bundled) | $400–1,500 | Estimate | +| Independent acceptance test | $300–1,500 | Estimate (AiryLab/Orion Optics class) | +| Design-assist / melt re-optimization | $1,000–5,000 contingent | Estimate; may be absorbed by Lot B vendor | +| **Optics envelope** | **$6,500–12,000** | **Committed** | + +Project context, stated honestly: the prior all-in target was $8,500–9,500 and the +mechanical selected configuration already consumed its contingency (+$2,290–5,790 in +mandatory remediations). With this envelope the **project total is acknowledged at +~$15,000–20,000.** The $900 eyepiece line and other visual accessories remain the +obvious donors if trimming is ever required. + +Schedule: corrector RFQ turnarounds 3–10 business days; fabrication 3–6 months; +custom set end-to-end **6–12 months**. P0–P2 cost nothing but engineering time and +should complete before any RFQ goes out. + +--- + +## 7. Risk register + +| # | Risk | Exposure | Mitigation | +|---|---|---|---| +| R1 | Corrector single-vendor concentration (Lot B is 60–75% of cost) | Schedule + cost | ≥ 2 quotes incl. one non-US channel; melt-assist clause; Starlux contingency line | +| R2 | Sitall unobtainium erodes the "unique glass" identity | Identity/scope | CTE-class spec (§3.1) makes substrate a drop-in; standing Sitall watch | +| R3 | Corrector cost band is estimator-grade (no signed quote yet) | Budget | P3 RFQ round anchors it; envelope carries the high end | +| R4 | Custom obstruction ≤ 18% unproven until T5 illumination trace | Performance | T5 is a P1 exit-gate proof, before any procurement | +| R5 | BK7 poly-Strehl ≥ 0.95 at f/6 is derived, not yet traced | Performance | T1 gate at multiple f-ratios; f-ratio floats until P2 | +| R6 | Site r₀ = 10–15 cm is ASSUMED, never measured | Mission | P6 seeing-monitor task; lucky-imaging yield model re-run with data | +| R7 | Bearing-span remediation (P2-X1) touches BP-01 machining | Interface | Cross-cutting action tracked in both documents | +| R8 | Budget expansion (~$15–20k total) outruns appetite | Program | Committed knowingly (Tim, 2026-08-06); P3 checkpoint is the natural re-look | + +--- + +## Appendix A — Seed prescription table (DERIVED, ray-trace-verified 2026-08-06) + +Glass N-BK7 (n_e = 1.51872, n_F = 1.52238, n_C = 1.51432); OD 215.0 +0/−0.2 mm; +CA ≥ 204 mm; CT 22.0 ± 0.2 mm; achromat condition R1−R2 = −0.565·t (with /0.97 +spherochromatism factor); corrector-mirror gap seed 0.85·f; residual F–C focus shift +~10–11 µm at the 0.7 zone (inside depth of focus); element mass ≈ 2.0 kg. + +| Case | Mirror R (mm) | R1 (mm) | R2 (mm) | Sag₁ (mm) | Edge thickness (mm) | +|---|---|---|---|---|---| +| f/6, FL 1224 | −2448 | −413.1 | −425.6 | 14.2 | 21.6 | +| f/8, FL 1632 | −3264 | −513.4 | −525.8 | 11.4 | 21.7 | + +Radius-match tolerance for diffraction-limited focus: ±0.13 mm (f/6) / ±0.30 mm (f/8); +common (pair) radius error: ±3.7 mm (f/6) / ±11 mm (f/8). R/D ≈ 2.0–2.5 both cases — +gentle spheres, routine test-plate work. + +## Appendix B — Verified vendor contacts (2026-08-06) + +| Vendor | Role | Contact | +|---|---|---| +| Ostahowski Optics | Primary (Lot A lead) | ostahowskioptics.com — quotes by email/phone | +| Lockwood Custom Optics | Primary alternate | loptics.com — 50% deposit model | +| Orion Optics UK | Primary alternate + **independent Zygo test service** | orionoptics.co.uk | +| Optimax | Corrector (Lot B lead) | go.optimaxsi.com/rfq · sales@optimaxsi.com · (585) 265-1020 | +| Knight Optical UK | Corrector | +44 (0)1622 859444 — custom menisci page | +| United Lens | Corrector/blank | (508) 765-5421 · info.unitedlens.com/request-a-quote | +| Advanced Glass Industries | Blanks | sales@advancedglass.net · (585) 458-8040 | +| Antares Optics | Secondary flat (COTS) | antaresoptics.com — live prices | +| Spectrum Coatings | Mirror coating | paul@spectrum-coatings.com · (386) 848-3924 | +| AstroReflect / AiryLab | Independent testing | astroreflect.com (free offer) / airylab.com | + +Removed from candidate lists (verified 2026-08): Zambuto (closed to new orders +2025-08-15), ISTAR Optical (not accepting orders), the Optimax online Estimator +(offline; DNS dead + 404). + +## Appendix C — RFQ-ready corrector specification + +> Full-aperture Maksutov meniscus corrector, qty 1 (option qty 2). Material SCHOTT +> N-BK7 or CDGM H-K9L, fine annealed, homogeneity H2, bubble ≤ B1, striae grade A. +> OD 215.0 +0/−0.2 mm; CA ≥ 204 mm; CT 22.0 ± 0.2 mm. Baseline (f/8): R1 = −513.4 mm, +> R2 = −525.8 mm; alternate (f/6): R1 = −413.1 mm, R2 = −425.6 mm. Radii to test plate +> ±0.1%; R1−R2 difference held to ±0.25 mm (f/8) / ±0.10 mm (f/6), coupled to CT per +> R1−R2 = −0.565·t. Irregularity ≤ λ/8 P-V per surface at 633 nm over CA; surface +> quality 60-40; wedge ≤ 1 arcmin ETD. BBAR both sides, R_avg < 0.5%, 400–700 nm. +> Note: seed radii are optimization starting points — final prescription to be +> co-optimized with the specific primary and delivered melt data (design-assist clause). + +## Appendix D — Documentation corrections executed with this roadmap + +1. `docs/INTES_MICRO_HISTORY.md` — "environmental regulations" blank-scarcity claim + softened to the evidenced cost/lead statement (Schott TIE-41; in-production 190MN). +2. `docs/research/SOURCING_RESEARCH.md` — MN76≠MN78 nomenclature corrected (f/6 vs + f/8, different models); the false "still producing March 2025" claim replaced with + the adjudicated 2018 cessation and its evidence trail; APM raw-optics channel marked + historical; direct-Russia action item closed (sanctions). + +--- + +*Method note: every number above is tagged by origin — SOURCED (repo or cited page), +DERIVED (computed this pass; ray traces and MTF autocorrelations re-runnable in P1 CI), +or estimate pending RFQ. The research dossier with per-claim URLs was produced by the +2026-08-06 multi-agent research workflow and is summarized in the PR that introduced +this file.* diff --git a/docs/INTES_MICRO_HISTORY.md b/docs/INTES_MICRO_HISTORY.md index 8435d82..43f4999 100644 --- a/docs/INTES_MICRO_HISTORY.md +++ b/docs/INTES_MICRO_HISTORY.md @@ -274,7 +274,9 @@ Additional factors contributed to the end of Russian amateur telescope productio ### Future Outlook -According to industry sources, **there will be no more large Maksutovs (>180mm) from Russian sources**. The primary constraint is that corrector plate blanks have become prohibitively expensive due to environmental regulations on glass production. The specialized optical glass required for large meniscus correctors is no longer economically viable to produce for the amateur telescope market. +**There will be no more large Maksutovs (>180mm) from Russian sources.** The Russian pipeline ended through the 2018 Intes Micro closure and, after 2022, sanctions on LZOS (part of Shvabe/Rostec) that stopped Western resellers from purchasing Russian optics at all. + +A caution on a widely-repeated claim: community lore holds that large meniscus corrector blanks became impossible to obtain "due to environmental regulations on glass production." This is overstated. Eco-glass reformulation (lead/arsenic-free melts) and shrinking demand did raise the cost of large fine-annealed blanks, but Schott's TIE-41 documents N-BK7 blanks producible above 300 mm at H2 homogeneity, and Sky-Watcher's 190 mm Maksutov-Newtonian (Starlux 190MN) remains in current production — proof that ~200 mm-class meniscus correctors are still economically manufacturable. Large corrector blanks are a cost and lead-time problem, not a physical or regulatory impossibility. *(Corrected 2026-08-06 during the v0.1.1 optics research pass; see design/optical/OPTICAL_ENGINEERING_ROADMAP.md.)* ### End of an Era diff --git a/docs/research/SOURCING_RESEARCH.md b/docs/research/SOURCING_RESEARCH.md index bc7a7e4..74710e2 100644 --- a/docs/research/SOURCING_RESEARCH.md +++ b/docs/research/SOURCING_RESEARCH.md @@ -9,23 +9,22 @@ ### Status: ACTIVE SEARCH REQUIRED — RARE VINTAGE OPTIC -> **Nomenclature Note:** The 7-inch (178 mm) f/6 Maksutov-Newtonian is officially designated **MN76** by Intes Micro. However, some dealers and Cloudy Nights forum threads refer to it as "MN78." Both designations refer to the same optical tube assembly. Search for both terms when sourcing. (Sources: Cloudy Nights discussions, Stellar Optical listings) +> **Nomenclature Correction (2026-08-06):** An earlier version of this note claimed MN76 and MN78 are the same tube. **They are different models.** MN76 = 180 mm **f/6**, FL 1068 mm, ~20% advertised obstruction, ~13.6 kg (historical dealer price $2,449, stellaroptical.com). MN78 = 180 mm **f/8**, FL 1440 mm, 13.4% advertised obstruction, ~14 kg, tube ~1.4 m (dark-star.it test; ENS Optical listing #14991). Search for both terms when sourcing, but do not conflate the specs. (Verified against dealer spec pages and the Dark Star instrument test.) #### Current Availability The Intes Micro MN76 (7" f/6 Maksutov-Newtonian, sometimes listed as MN78) remains a sought-after planetary telescope. Key findings: -**Primary Sources:** -- **ENS Optical UK** (https://ensoptical.co.uk) - Lists MN76/MN78 7" with Crayford focuser -- **APM Telescopes Germany** (https://www.apm-telescopes.net) - Official Intes Micro distributor - - APM offers premium versions with Matthias Wirth tubes and Starlight Feather Touch focusers - - Also sells raw Intes Micro optics in custom configurations -- **Astromart Classifieds** (https://astromart.com) - Active used market listings (search both MN76 and MN78) +**Primary Sources (used/NOS market only):** +- **ENS Optical UK** (https://ensoptical.co.uk) - Has listed a used MN78 7" f/8 with Crayford at GBP 1,000 (out of stock as of 2026-08) +- **Widescreen Centre UK** (https://www.widescreen-centre.co.uk) - Still lists Intes-Micro new-old-stock, but nothing above 127 mm class (ALTER MN56/MN58 at GBP 1,100) +- **APM Telescopes Germany** (https://www.apm-telescopes.net) - *Historical* Intes distributor; as of 2026-08 its Mak-Newt category carries only Explore Scientific 152 mm and Sky-Watcher 190MN, and no Intes tubes or raw optical sets +- **Astromart Classifieds** (https://astromart.com) - Intermittent used listings (search both MN76 and MN78; account required to see prices) -**Production Status:** -- As of March 2025, Intes Micro still produces MN56 (127mm, 760mm FL) and MN76 (178mm, ~1068mm FL) -- MN76/MN78 availability requires direct inquiry to distributors -- Factory located in Russia; export logistics may be complicated +**Production Status (corrected 2026-08-06):** +- **Intes Micro ceased production in 2018** (announced via Dr. Simon Bennett, Widescreen Centre; corroborated by stargazerslounge.com/topic/317143 and Wayback evidence: intes.su content frozen at a 2005-copyright page through 2021, an empty IIS default page by 2024, dead DNS by 2026) +- An earlier claim here that Intes Micro was "still producing as of March 2025" was traced to a stale frozen website and AI-recycled forum lore; it is **false** +- The buy-side for a *new* 180 mm-class Intes-quality Mak-Newt is definitively closed; only the used market and the in-production Sky-Watcher Starlux 190MN ($2,375, 27-34% obstruction, borosilicate) remain - **Exceptional optical quality:** Hand-figured Russian optics routinely tested at 1/8 wave or better. The Mak-Newt design eliminates coma while maintaining a fast f/6 focal ratio ideal for lucky imaging. - **Closed tube design:** Critical for harsh environments like Nevada's high desert. The sealed tube eliminates thermal currents from the primary mirror and protects optics from dust, insects, and moisture. @@ -53,10 +52,10 @@ The MN76/MN78 appears on the used market infrequently—expect to monitor listin #### Action Items -- [ ] Email APM Telescopes (info@apm-telescopes.net) for MN76 pricing and lead time (note: may be listed as MN78) -- [ ] Check Astromart and Cloudy Nights classifieds weekly (search both MN76 and MN78) -- [ ] Contact ENS Optical UK for availability -- [ ] Investigate direct Russia purchase if needed +- [ ] Email APM Telescopes (info@apm-telescopes.net) asking whether any off-catalog Intes new-old-stock optics remain (Markus Ludes historically held Intes inventory) — used/NOS inquiry only; new production ended 2018 +- [ ] Check Astromart and Cloudy Nights classifieds weekly (search both MN76 and MN78; Astromart prices are login-gated) +- [ ] Contact ENS Optical UK and Widescreen Centre for used/NOS availability +- [ ] ~~Investigate direct Russia purchase if needed~~ Closed: LZOS/Russian optics are sanctioned (post-2022); no direct-purchase path exists ---