Mechanical feasibility - 26 May 2026

The current optics fit the box, but they are not fabrication-ready.

EmberScope applies the same mechanical screen to the spherical, conic, and starter-freeform mirror candidates: package reserve, detector fit, clipping, manufacturability, tolerance path, and field-service constraints.

Compare candidates Read payload allocation Read procurement path Open evidence package

Dated checkpoint (26 May 2026) — the candidates below are no longer current.

The spherical, conic, and starter-freeform shapes screened here are early three-mirror packaging variants that were never working imagers. The mechanical screening method (package reserve, detector fit, clipping, manufacturability, tolerance path) remains the template, and its next application is the optimized four-mirror prescription of 10 June 2026 on the optics hub. All results are model outputs; nothing has been fabricated.

Mechanical screen

A candidate has to be mountable, sealed, cooled, cleaned, and aligned before optical performance matters.

72.0 mm Current source-spec package span, leaving 78 mm before the 150 mm cube boundary.

0.664 kg Current source-spec mass proxy, leaving about 1.34 kg for the complete payload allocation.

0.38 Current edge-field aperture-valid fraction at 1.5 deg; this is the main clearance warning.

9.1 mm Largest starter-freeform departure, on M3; too large to treat as low-risk without supplier review.

Evidence boundary

This is a candidate evaluation, not a CAD release.

The screen uses the current review-package parameters, prescriptions, and ray-trace metrics. The prescription still lacks a final LWIR window, filter, stop, baffles, detector carrier, alignment datums, and tolerance model.

The useful result is narrower than a build approval: all three candidates are compact enough to carry forward, while none should be sent for fabrication until clipping, freeform limits, and tolerance acceptance are built into the next optical iteration.

Candidate disposition

The simplest mirrors are easier to build; the freeform branch is the optical leader but carries the highest mechanical risk.

Candidate	Package result	Main risk	Disposition
Spherical	Historical POC screen: 65.5 mm span, 0.324 kg proxy, detector sanity pass.	Edge throughput falls to 0.38 and edge RMS spot is 12803 um.	Mechanically plausible bench-surrogate only.
Conic	Historical POC screen: 65.5 mm span, 0.324 kg proxy, detector sanity pass.	Same 0.38 edge throughput, with 12491 um edge RMS spot.	Keep as the low-complexity manufacturing comparator.
Starter freeform	Current source-spec package: 72.0 mm span, 0.664 kg proxy, detector sanity pass.	M2 and M3 freeform departures reach 4.6 mm and 9.1 mm; edge throughput is still 0.38.	Current optical leader, but not quote-ready.

Cross-candidate findings

The package margin is real, but final usability is still dominated by clearance, baffles, and tolerances.

Package reserve

All three optical cores fit well inside the 150 mm cube and leave room for housing, window cell, baffles, electronics, and service access.

Detector bay

All three candidates pass the focal-plane sanity check, so the detector remains mechanically plausible at this review level.

Field clipping

All three candidates need another optical/mechanical iteration because the 1.5 deg edge field only retains 38 percent aperture-valid rays.

Manufacturing risk

The spherical and conic variants are easier to quote, while the starter-freeform surface departures need manufacturability limits and supplier review.

Missing hardware

The current prescriptions do not yet include the window cell, blackened baffles, detector focus mechanism, mirror-cell datums, or thermal straps.

Build decision

Keep all three variants in the trade, but delay fabrication until edge coverage, tolerances, and mounting datums are explicit.

Next iteration rules

The next optical pass should include the mechanical constraints directly in the merit function.

Preserve at least 35 mm side reserve around the optical core, hold the optical-core mass proxy below 0.6 kg unless performance justifies otherwise, and restore edge-field aperture-valid fraction to at least 0.80 before calling any branch prototype-ready.

Export mirror datums, keep-out zones, window/filter thickness, detector carrier depth, baffle envelope, stop placement, and tolerance limits with each prescription so the mechanical review is no longer a separate late-stage filter.