Add SLS process/powder-science/defects/standards corpus from deep-research pass

9 open-access papers (SSR foundations, energy-density review, PA12
crystallization kinetics, aging/reuse, orientation effects, powder
characterization), 4 SDS (nylon-12, nylon-12-gf, nylon-12-white via
mirror, tpu-90a), 3 archived Formlabs support articles (defect taxonomy,
curling, bed-temp tuning), SLS4All docs index, and 6 DRAFT distilled
agent-servable docs under drafts/ pending human review.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q4zfbNDFb7rd8ovKrUSys2

drafts/defect-to-parameter-map.md

@@ -0,0 +1,26 @@
+---
+title: Defect symptom → parameter correction map (Inova overrides vocabulary)
+summary: Observable SLS defects mapped to causes and to the Inova MK1 override knobs (bed/chamber temps, laser energy, recoater passes/speed) an agent can actually turn.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# Defect → Parameter Map
+
+| Symptom (observable) | Likely cause | Correction in OUR override vocabulary |
+|---|---|---|
+| Edges lift after recoat / during lasing (curling) | Lased-vs-bed ΔT too large; bed temp low; IR misreading | Raise bed/surface temp target (+1 °C steps); verify thermal readings before trusting more heat |
+| Curled edge caught by recoater | Late-stage curling | Pause-level decision: reduce energy % locally is not possible — consider object exclusion before recoater jam |
+| Grainy, fragile, crumbly parts | Under-sintering (energy below SSR mid) | Raise laserFillEnergyDensity / totalEnergyDensityPercent toward mid-SSR |
+| Caked, hard-to-clean part cake; bonded surroundings | Over-energy or bed too hot | Lower energy density first; then bed temp −1 °C steps |
+| Orange peel / poor top surface | Energy slightly low or powder aged (poor coalescence) | Small energy increase; check powder refresh history |
+| Ripples/banding on verticals | Bed/chamber temperature fluctuation | Inspect temperature traces; stabilize before parameter moves |
+| Streaks/gouges along recoat direction | Debris on recoater or dragged curl | Extra full recoat pass (dry) to clear; inspect frame images |
+| Short feed / incomplete powder coverage | Dosing insufficient | Increase powder dose factor / staged passes |
+| Warp discovered post-build (in-build invisible) | Cooldown too fast; crystallization gradients | Slower cooldown; keep bed heater on longer after finish |
+
+Sources: Formlabs Diagnosing-a-print-failure + Curling articles (archived in
+source/defects/), SSR literature for energy direction.

drafts/energy-density-cheatsheet.md

@@ -0,0 +1,29 @@
+---
+title: Energy density formulas and PA12 operating ranges
+summary: AED/VED definitions with units, PA12 ranges, the stable sintering region concept, EMR degradation ceiling, and why more energy stops helping.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# Energy Density Cheat Sheet
+
+- Areal energy density (Andrew number): AED = P/(v·h) [J/mm²]
+  (P laser power, v scan speed, h hatch spacing)
+- Volumetric energy density: VED = P/(v·h·t) [J/mm³] (t layer thickness)
+- PA12 studies operate around AED ≈ 0.02–0.1 J/mm². [Materials 14:1169]
+- Stable sintering region (SSR): energy interval between just-melts (Em) and
+  degradation onset (Edeg). Example measured SSR (VED): PA12/CF
+  0.046–0.410 J/mm³. Optimal quality sits MID-window, not at the minimum —
+  strength plateaus (PA12/CF ~50 MPa at 0.23–0.24 J/mm³) while porosity falls.
+  [SSR-optimization study]
+- Upper bound: Energy Melt Ratio (applied energy ÷ theoretical melt energy)
+  ≈ 6.2 before degradation for PA12. Over-energy drives surface >300 °C into
+  TGA mass-loss territory — properties DECLINE past the optimum (chain
+  degradation), so "more laser" is not a recovery move. [Vasquez 2011/2013]
+- Plausibility envelope (polymer L-PBF literature): laser 2–50 W (avg ~16 W),
+  scan ≥45 mm/s, layer 100–200 µm. Diode-laser benchtop reference (Sinterit
+  Lisa, 5 W, air): chamber 140 °C, print surface 178 °C, 0.1 mm layers.
+  [Materials 14:1169; SciRep 2025]

drafts/material-windows-comparison.md

@@ -0,0 +1,20 @@
+---
+title: Material sintering-window comparison (PA12 / PA11 / TPU)
+summary: Relative processing-window widths and handling differences across our powder family.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# Material Windows Comparison
+
+| Material | Sintering window | Notes for agents |
+|---|---|---|
+| PA12 / PA12 GF | ~20–25 °C (PA12/CF measured 21 °C) | The forgiving baseline; GF adds stiffness, narrows practical window slightly |
+| PA11 | ~11 °C — HALF of PA12's | Far less bed-temp margin; nitrogen strongly recommended (air degrades faster per Formlabs); expect tighter tolerance to drift |
+| TPU 90A | Wide, elastomer behavior | Different failure modes (elastic recovery vs curl); Formlabs runs it at ~50 % refresh |
+| PP | ~16 °C measured | Reference only (not in our powder set) |
+
+Sources: Materials 14:1169 survey; SSR-optimization study; Formlabs TDS/support.

drafts/pa12-sintering-window.md

@@ -0,0 +1,35 @@
+---
+title: PA12 sintering window — the numbers that govern bed temperature
+summary: PA12 melt/crystallization temperatures, bed setpoint window, in-build crystallization kinetics, and cooling-rate dependence — the core thermal facts for bed-temperature decisions.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# PA12 Sintering Window
+
+- Melting point ≈ 184.8–185 °C; heat of fusion ≈ 96.7–97.2 J/g; specific heat
+  ≈ 3250 J·kg⁻¹·K⁻¹ (EOS PA2200-class powder). [Vasquez 2011/2013]
+- Peak crystallization on cooling ≈ 146.5 °C — supercooling gives the workable
+  window between melt and recrystallization. [Vasquez 2011]
+- Part-bed setpoint: hold inside the window, close under melt onset —
+  168–174 °C typical for PA12 (EOS P100 used 172 °C; ~12 °C below peak melt).
+  [Vasquez 2013; MDPI Polymers 10:168]
+- The window is NOT static during a build: PA12 held isothermally at 168 °C
+  fully crystallizes within ~50 min — lower layers crystallize during the
+  build, so long builds accumulate shrinkage/warp risk even at temperature.
+  [MDPI Polymers 10:168]
+- Crystallization onset is cooling-rate dependent: ~170 °C at 0.2 K/min vs
+  ~155 °C at 20 K/min. Slow cooldowns start crystallizing almost immediately
+  below setpoint; fast local cooling delays onset but sharpens gradients.
+  [MDPI Polymers 10:168]
+- Material comparison: sintering windows across SLS polymers range ~11 °C
+  (PA11 — much less forgiving) to ~37 °C, avg ~23 °C. GF-filled PA12 behaves
+  like PA12 with a somewhat narrowed practical window (~21 °C measured for
+  PA12/CF). [Materials 14:1169 review; SSR-optimization study]
+
+Decision heuristics: curling right after lasing → lased-vs-bed differential
+too large (bed low, or IR/thermal fault). Caking/hard part-cake → bed too hot
+or over-energy. Bed drift matters more late in long builds.

drafts/parameter-envelope.md

@@ -0,0 +1,24 @@
+---
+title: Parameter plausibility envelope (diode-laser SLS)
+summary: Sanity bounds for laser power, scan speed, layer thickness, and temperatures, with the diode-laser benchtop reference point closest to the Inova MK1.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# Parameter Plausibility Envelope
+
+Literature envelope (100 most-cited polymer L-PBF studies): laser 2–50 W
+(avg ~16 W), scan speed from 45 mm/s up, layer thickness 100–200 µm.
+[Materials 14:1169]
+
+Closest published analog to the Inova MK1 (diode laser, ambient air, benchtop):
+Sinterit Lisa, 5 W diode — chamber 140 °C, print-surface 178 °C, 0.1 mm layer.
+[SciRep 2025]
+
+Use: sanity-check any proposed override against these bounds; a suggestion
+outside the envelope needs explicit justification. Our own verified profile
+ladder (14–40 mJ/mm energy densities, build_registry) remains the primary
+reference — this envelope is the outer fence, not the target.

drafts/powder-reuse-aging.md

@@ -0,0 +1,29 @@
+---
+title: Powder reuse and aging — what changes and how much
+summary: Quantified effects of PA12 reuse (strength loss, DSC shifts, window drift) an agent needs when powder history is part of parameter decisions.
+---
+
+> **DRAFT — PENDING HUMAN REVIEW.** Not served to agents until promoted into
+> `source/` (top level). Numbers below are extracted from the cited papers by
+> the 2026-07-16 deep-research pass and domain-checked, but adversarial
+> verification is still in progress. Review before promotion.
+
+# Powder Reuse & Aging (PA12)
+
+- Reuse without refresh is a first-order strength variable: 5 cycles unrefreshed
+  on a diode-laser machine → UTS −35 % (65.3→42.2 MPa), modulus −31 %.
+  [SciRep 2025]
+- DSC signatures shift only slightly with reuse (melt peak +1.2 °C, cryst peak
+  +0.8 °C after 5 cycles) — the SINTERING WINDOW stays roughly put even as
+  mechanical quality degrades; you cannot detect tired powder from bed-temp
+  behavior alone. Crystallinity drops (49.1→44.9 %). [SciRep 2025]
+- Part-cake ages during the build itself via solid-state polycondensation:
+  molecular weight rises with build time (dominant) and chamber temperature —
+  long builds age powder faster than hot ones. [Wudy & Drummer, Addit. Manuf.]
+- Thermal aging first RAISES melt temp (~+4.2 °C over first ~100 h at 170 °C,
+  secondary crystallization), then thermo-oxidative degradation lowers Tm/Xc
+  and embrittles. Moisture accelerates degradation — pre-drying helps.
+  [Polymers 14:2682]
+- Practical: Formlabs refresh rates (30–50 % by material per TDS) are the
+  operating assumption; deviations should lower the agent's expectation of
+  achievable mechanical outcomes rather than trigger energy compensation.

source/defects/formlabs-curling-sls.md

@@ -0,0 +1,34 @@
+---
+title: Formlabs — Curling (SLS)
+summary: Curling symptoms, root causes (thermal differential between lased area and bed), and correction sequence including bed-temperature offset tuning.
+source_url: https://formlabs.com/support/Curling-SLS
+---
+
+> Archived via content extraction from the URL above (not a verbatim page copy). Retrieved 2026-07-16.
+
+# Curling (SLS)
+
+## What it looks like
+Printed layers or surface armor curl and rise from the print bed: visible after
+recoating (edges lifting), during lasing, or as curled material caught and
+dragged by the recoater roller (print-ending). Early-layer curling: the first
+several layers curl at the edges creating a cumulative warped area, mostly at
+the bottom of the build volume.
+
+## Root cause mechanism
+If the temperature difference between the lased area and the rest of the powder
+bed is too large, the lased area shrinks more than the surrounding powder.
+
+Primary causes: (1) dirty IR sensor obstructing temperature readings,
+(2) clogged exhaust filter/lines, (3) improperly tuned print-bed temperature.
+
+## Correction sequence
+1. Clean IR sensor lens + housing
+2. Clean exhaust filter, fan, lines
+3. Run the bed-temperature tuning print for the material
+4. Adjust bed-temperature offset from results; test print
+5. Escalate if persisting
+
+For surface-armor curling (PP example): bed offset −2 °C, step down 1 °C to
+−5 °C; else reset and reduce armor laser power 20–40 %; if bulk curling
+appears, step bed offset up +1 °C toward +5 °C; increase part spacing ≥6 mm.

source/defects/formlabs-diagnosing-print-failure-sls.md

@@ -0,0 +1,32 @@
+---
+title: Formlabs — Diagnosing a print failure (SLS)
+summary: Formlabs' 13 SLS defect modes with symptoms and key issue classes (dimpling, layer shifting, banding, warping, fragile parts, delamination, recoater artifacts, hard caking).
+source_url: https://formlabs.com/support/Diagnosing-a-print-failure-SLS
+---
+
+> Archived via content extraction from the URL above (not a verbatim page copy). Retrieved 2026-07-16.
+
+# SLS Print Failure Diagnosis (Formlabs Fuse)
+
+## Best practices for prevention
+- Use recommended powder refresh rates for the material
+- Adjust bed temperature for optimal quality
+- Clean optical cassette, IR sensor, IR sensor housing, and print enclosure
+- Wait for internal temperature to drop to 30 °C before removing the build chamber
+
+## Defect modes
+
+| Defect | Symptoms | Key issue |
+|---|---|---|
+| Dimpling & pitting | Sunken/collapsed top surfaces | Temperature control |
+| Layer shifting | Horizontal distortion lines/bulges at specific heights | Mechanical/calibration |
+| Horizontal banding | Periodic ripples on vertical surfaces | Temperature fluctuation |
+| Poor vertical finish | Ridged/wavy surfaces, esp. at build volume edges | Optical or thermal inconsistency |
+| Warping | Deformed/twisted parts | Cooling/stress |
+| Fragile parts | Crumbling parts, grainy texture | Sintering energy too low |
+| Bumps along Z | Distortion extending vertically | Laser path obstruction |
+| Blurred/missing features | Fine details rounded off or absent | Laser focus or powder contamination |
+| Delamination | Layer separation | Insufficient layer adhesion |
+| Powder contamination | Blown-out areas, embedded debris | Environment |
+| Recoater artifacts | Gouges/ridges along recoater motion direction | Dirty recoater roller |
+| Hard caking | Difficult powder separation/cleanup | Surface-armor bonding |

source/machines/fuse-1/formlabs-bed-temp-tuning-nylon12gf.md

@@ -0,0 +1,20 @@
+---
+title: Formlabs — Bed temperature tuning (Fuse 1+ 30W, Nylon 12 GF)
+summary: The per-printer bed-temperature calibration procedure for Nylon 12 GF; seven-plaque diagnostic across offsets -3 to +3 C at 50% refresh.
+source_url: https://formlabs.com/support/Tuning-the-bed-temperature-of-your-Fuse-1-30W-printer-for-improved-part-yield-with-Nylon-12-GF-Powder
+---
+
+> Archived via content extraction from the URL above (not a verbatim page copy). Retrieved 2026-07-16.
+
+# Bed Temperature Tuning (Fuse 1+ 30W, Nylon 12 GF)
+
+Every Fuse 1+ 30W should be fine-tuned when first printing Nylon 12 GF —
+neighboring printers may need different settings (machine-to-machine thermal
+variation is expected, not exceptional).
+
+Procedure: reset offset to 0 → run the diagnostic print (3 kg powder at 50 %
+refresh, hopper to 7 L) → full cool → extract the seven plaques labeled
+−3/−2/−1/0/+1/+2/+3 °C → clean → line up and pick visually optimal offset.
+
+Relevance to Inova MK1: same powder, same physics — a per-machine bed-offset
+sweep expressed as printed plaques. The concept ports; absolute values do not.

source/machines/inova-mk1/sls4all-tech-docs-index.md

@@ -0,0 +1,19 @@
+---
+title: SLS4All Inova MK1 — official documentation index
+summary: Index of official SLS4All documentation for the Inova MK1 (BOM, schematics, assembly/operating instructions, software) with URLs.
+source_url: https://sls4all.com/tech-docs/
+---
+
+> Archived via content extraction from the URL above (not a verbatim page copy). Retrieved 2026-07-16.
+
+# SLS4All Tech Docs (Inova MK1)
+
+- Bill of Materials: https://sls4all.com/bill-of-materials-inova-mk1/
+- Fusion360 model, PCB schematics, wiring: https://sls4all.com/hardware/ (free login)
+- SLS4All Compact software: https://sls4all.com/sls4all-compact/ · https://github.com/sls4all/SLS4All.Compact
+- Assembly manuals: https://sls4all.com/assembly-manual/
+- Inova MK1 instructions: https://sls4all.com/download/inova-mk1-instructions/
+- EPP set instructions: https://sls4all.com/download/epp-set-instructions/
+
+The operating-instructions PDF is login-gated; archive a copy here when
+downloaded (do not archive the unofficial scribd mirror).

source/materials/nylon-12-gf/2201646-SDS-ENUS-0.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47fe5172f7a7a3e5459e8610a6fd666362c2c3dd404e66b3a21d4a0af34db247
+size 66367

source/materials/nylon-12-white/SDS-nylon-12-white-powder.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ef2250d413b238b2310551e49398279675fef25df1dd420d94d55703f32133e
+size 69402

source/materials/nylon-12/2001461-SDS-ENUS-0.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8a21b3738dec1dfdc1bc2936454b44393bade00fb91bdd3c93f7639f57feba5
+size 66996

source/materials/tpu-90a/SDS-ENUS-TPU-90A-Powder.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d839e84cff114c80202587ac34f53bcce943de5cfcfc24dc868261a0dcc246ca
+size 60992

source/powder-science/2018-polymers-pa12-crystallization-kinetics.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43c56dc055142dacee39c4ebabfe888796b3e02f7f88eae5bb53d79c753fdf96
+size 1510651

source/powder-science/2019-sff-polymer-powder-characterization.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38b61b1f2ff626061e1060ee22f02c27254f8e0c5879a3969118a2e303ddb048
+size 2175548

source/powder-science/2022-polymers-pa12-aging-reuse.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b10ad3543e600c9462826c7f4d9b4cba4f6c51f99fd08e02a62111d3a6d9f6c5
+size 6408425

source/powder-science/2025-scirep-reused-pa12-diode-sls.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6fa53a5e8f466fc177f27ca65f79f4a4a4553ec5003fe08477ad8096223886e
+size 6539226

source/process/2021-materials-polymer-lpbf-review.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57136634c7fb9163190b5a19d6017bb2735e488962b328afb02a9f312b8b24c5
+size 1701744

source/process/vasquez-2011-optimum-sintering-region-nylon12.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cc74ff2829294513e0b787a91316814170f84fbb599843c2182d17c29324c6b
+size 823029

source/process/vasquez-2013-quantifying-stable-sintering-region-pa12.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f51b26f115ffc811bb27536b84d2b4ddcaa476a01a905165e8aed7c63c45f3c
+size 1097164

source/standards/2024-polymers-pa12-orientation-tensile.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64afa86f8fa8ba974909b1ae7a99c112c84d907cbd33919d3089762f8136d6fe
+size 6599314

source/standards/am-coordinate-systems-orientation-module.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a70d1faffa248c5e7f699d0f0bf3ccb4ca62f693b1cae94d188fc0179768182a
+size 283725