Adom Custom Electronic Load

@@ -10,11 +10,12 @@ The architecture uses an IXTP15N50L2 linear MOSFET, four-terminal 20 mΩ shunt,  - A 104-component hierarchical schematic passes native KiCad ERC; all 52 intended nets match the native exported netlist. - Six limited current-servo sensitivity cases pass using TI's amplifier model and an approximate MOSFET. This is not full circuit or hardware validation.-- The portable C safety-policy core passes host tests. Flashable MCU firmware and USB integration are not complete.-- The saved two-layer routed board has zero native DRC errors and zero unconnected nets, with zero warnings in the latest saved-board check. Twelve power-track segments were widened and a parallel shunt via added after isolated checks; see the [power-width review](docs/power-width-review.md) and [shunt-path review](docs/shunt-path-review.md). All nine planned measurements have probe access, including dedicated fan-tach and Kelvin-minus pads. Current/thermal, USB, cooling and fabrication qualification remain incomplete.+- The portable C safety-policy core and instrument/ADC adapters pass host tests. An output-locked STM32 USB bring-up image compiles; functional load enabling, calibration and physical USB/interlock tests remain incomplete. No firmware has been flashed.+- The live two-layer board remains AIFlow board 19, with zero native DRC errors and opens and 15 recorded library warnings. Isolated rework candidates include the project libraries and pass without those warnings, but are not yet landed. The corrected fine-grid field screen passes the latest ground/drain paths and still rejects the shunt path. Current/thermal acceptance remains open; see the [completion plan](docs/completion-plan.md).+- AIFlow 0.1.82 is source-published, registry-published and installed in this workspace. It independently checks raw current/barrel evidence and individual thermal budgets. The corrected Fields solver is used only for isolated engineering studies while its owner reviews PR 9 and its dependencies. - The fan has a [component page with GLB motion states](https://wiki.adom.inc/adom/mf40200v1-1000u-g99). Its preview is simulated; shared viewer integration remains proposed. - The [thermal budget](docs/thermal-budget.md) now matches the 55 °C firmware trip ceiling. The earlier 25 W assumption is retained as an unqualified future option; R0 targets 5 W and still requires cooling validation.-- Build footage is being collected. Failed/static captures are identified in the shot log rather than included as action footage.+- There are 70 unique raw takes, 17 reviewed excerpts and 13 storyboard clips. Failed/static captures remain identified in the shot log; the final narrated 4–5 minute film is unfinished.  Read the [engineering status](docs/status.md), [placement review](docs/placement-review.md), [library review](docs/library-review.md), [cost breakdown](docs/cost/README.md), [firmware/control-panel contract](docs/firmware-contract.md), and [raw clip log](videos/shotlog.json). The [AI Flow report](docs/runs/board-codex-20261007-1749/README.md) records the run. 

@@ -0,0 +1,77 @@+{+  "status": "isolated-not-landed",+  "candidate": "checkpoints/unclipped-hotspots-v1/candidate-v45.kicad_pcb",+  "boardSha256": "21dbc64fa63c94e93e6d56fe45b03ee32856d2030e5881f5dc3c855f99f94408",+  "nativeDRC": {+    "errors": 0,+    "warnings": 0,+    "opens": 0,+    "sha256": "852391a9a8cc1676042aa5cdd8a909c4c11b62ed5dda986aa915cb60c08b7c95"+  },+  "solver": {+    "sha256": "202bb43774babdeb18f3e677090d559d01b9b62938bac8aa659326502f82f168",+    "sourceCommit": "e34bcddab94b3f14442bf511a8b6b35bcdb02237",+    "review": "https://wiki.adom.inc/adom/adom-fields/prs/9",+    "sharedRuntimeReplaced": false+  },+  "reports": [+    {+      "path": "checkpoints/unclipped-hotspots-v1/fields-v45-annular-0p1/fields.json",+      "sha256": "3940e2105695eb31d6c04fbec2d1347ede5ad4ec7888afb08e05698b31270262",+      "inputIdentity": {+        "algorithm": "sha256",+        "board": "21dbc64fa63c94e93e6d56fe45b03ee32856d2030e5881f5dc3c855f99f94408",+        "solverVersion": "0.6.2",+        "spec": "9050e3a86363f4958a3792dc41f3fb6a1ffbf09dd74681b343ec3048180c32d0"+      },+      "nets": [+        {+          "net": "LOAD_POS",+          "peakJ": 56.096134185791016,+          "peakAt": [+            127.14999999999999,+            84.75+          ],+          "peakLayer": "F.Cu",+          "residual": 9.59246368745904e-08+        },+        {+          "net": "SHUNT_P",+          "peakJ": 66.05319213867188,+          "peakAt": [+            127.14999999999999,+            83.25+          ],+          "peakLayer": "B.Cu",+          "residual": 9.407628619566938e-08+        },+        {+          "net": "GND",+          "peakJ": 58.3114013671875,+          "peakAt": [+            139.65,+            89.85000000000001+          ],+          "peakLayer": "F.Cu",+          "residual": 9.98356343597552e-08+        }+      ],+      "chips": [],+      "issues": [+        {+          "at": [+            127.14999999999999,+            83.25+          ],+          "fix": "Widen the neck or add a parallel path on the other layer with vias; the density map shows where the current squeezes.",+          "kind": "current-neck",+          "layer": "B.Cu",+          "net": "SHUNT_P",+          "reference": null,+          "severity": "medium",+          "what": "SHUNT_P: 66 A/mm\u00b2 at its worst point on B.Cu (3 A total; equivalent full-net width 1.30 mm at 35 \u00b5m; the threshold is 60 A/mm\u00b2)"+        }+      ]+    }+  ]+}

@@ -23,3 +23,9 @@ Next: choose a toleranced sleeve/fastener or insulated clamp that fits the exact ## Short-sleeve centering sensitivity  A rigid 2D cylinder model gives 0.176 mm remaining nominal radial clearance at the far side of a 1.40 mm tab with a 2.8448 mm screw, the 0.81 mm sleeve and nominal catalog diameters. Smaller illustrative screw diameters of 2.75 and 2.60 mm reduce this to 0.129 and 0.066 mm. Thus full-depth coverage is not inherently necessary, and this washer remains a candidate. These are sensitivity calculations, not tolerance limits: real thread roots, discontinuous crests, eccentricity, deformation and molding variation are absent. Do not use the result as insulation qualification. [Model and assumptions](mechanical/short-sleeve-sensitivity.json), reproduced with `python3 tools/check_short_sleeve.py`.++## Nonconductive screw alternative (2026-10-10; not selected)++NBK SPE-M3-8-C provides a complete PEEK M3 × 0.5 screw with 8 mm under-head length, 5.5 mm head diameter and 3 mm head height. Its listed 0.21 N·m torsional figure is reference-only; NBK recommends half that for tightening (0.105 N·m), not the 0.452 N·m previously discussed for a metal screw/washer combination. The listed pack is 20 screws at $21.37 before delivery/tax. NBK explicitly warns that the resin’s 260 C rating does not establish the assembled screw’s operating limit. [Manufacturer dimensions and precautions](https://www.nbk1560.com/en-US/products/specialscrew/nedzicom/plasticscrew/SPE-C/SPE-M3-C/SPE-M3-8-C/).++A nonconductive screw removes the metal-shank shorting mechanism, but it does not establish tab-to-sink insulation, clamp pressure, long-term preload or creep. The interface pad, support path, screw engagement and thermal-cycle validation remain necessary. Treat this as a candidate alternative to the unresolved sleeve geometry, not an approved assembly or BOM change.

@@ -7,3 +7,7 @@ DMP6023LE-13 is a possible SOT223 P-channel replacement circuit: 60 V, 28 mΩ ma This would require a new circuit, not a diode footprint substitution: verify body-diode direction, add a source-to-gate clamp and a gate pull-down for the 24 V input, assess startup and low-voltage behavior, then regenerate the schematic, component assets, BOM, board and tests. A single MOSFET does not guarantee reverse-current blocking while enabled. Temperature-dependent resistance, fault operation, stock and cost remain unverified. No part has been purchased or added to the board.  Keep the existing diode until a complete comparison justifies the change. More copper or an engineered heat path remain alternatives.++## Sourcing snapshot++The 2026-10-10 PartsSearch response reports Mouser 621-DMP6023LE-13 stock of 20,154, $1.39 at quantity one and $0.517 at 100; DigiKey lists active status but zero stock in that response. This is a dated listing, not a quote or purchase. [Saved API evidence](../design/parts-search-dmp6023le.json). Documentation and circuit checks above still prevent adoption.

@@ -1,3 +1,15 @@+# Current checkpoint — 2026-10-10, fine-grid rework++The dashboard is at current analysis, step 12 of 19, with 11 stages done. Live board 19 is unchanged. The corrected annular plated-wall solver rejects the remaining shunt-path hotspot: candidate v40/v45 yields GND 58.31, LOAD_POS 56.10 and SHUNT_P 66.05 A/mm² at 0.1 mm, against the unchanged 60 screen. Candidate v52 widens the shunt bus and relocates nearby signal/drain copper; native DRC is zero errors/warnings/opens, and its fine-grid solve is running. No candidate is accepted or landed.++AIFlow 0.1.82 is pushed directly, published and installed here with 132 passing workspace tests, 57 preserved CLI commands and passing workflow checks. Colby’s installed version is not verified. Fields PRs 5, 6, 8 and 9 remain owner-reviewed source; shared Fields is unchanged.++Diode copper studies still exceed the 35 C still-air budget under the documented nominal operating-heat case. Larger areas create ground islands that must be repaired before acceptance. The MOSFET package/heatsink interface and actual airflow remain unresolved. No thermal or finish gate has passed.++70 raw takes, 17 reviewed excerpts and 13 storyboard clips are available. The latest review clip explicitly shows failed current findings. The final narrated film is not assembled. Firmware is output-locked and unflashed. See [completion plan](completion-plan.md), [operating heat cases](operating-heat-corners.md), [build clips](build-clips.md) and [release receipt](../design/aiflow-0.1.82-release.json).++## Earlier checkpoints (historical)+ # Current checkpoint — native pours accepted  Current-step gate is explicitly failed: its width-only heuristic flags DUT_POS 0.25 mm and DUT_FUSED 1.0 mm, and a same-input formal Fields report is missing. The saved main-path review proves separate continuous front-layer paths of 1.5 mm and 2.0 mm respectively (IPC-2221 estimates 8.58 C and 5.34 C at 3 A), but excludes branch loads/pad necks and does not override the gate. See design/step12-main-path-review.json. Next work is a terminal-aware capacity check and supported same-board Fields integration, not blindly widening the voltage-sense branch.