Chat

How can I help you?

Explain project

Write a project description

Find schematic issues

Files

Schematic

Layout

3D View

Autorouter Placement Preflight Diagnostic

Scope and status

Table

Item	Status	Result
Inspect live board-check output	Completed	Live check sees 342 placed components and 244 nets; only the U16 exposed-pad via advisory is returned.
Capture full autorouter failure payload	Completed from repeated project-recorded live attempts	Full autorouter starts on this same layout repeatedly failed during auto-confirm with HTTP 400 and generic `PLACEMENT_GENERIC`. The backend did not return an object UID, selector, coordinate, pad, or rule.
Identify likely placement trigger	Completed, strong circumstantial identification	J3 has the board's only component-bounds excursion and contains a child rectangle whose coordinate frames are interpreted inconsistently.
Verify J4–J7 recent movement	Completed	All four have explicit numeric Top-layer positions and in-bound pads/bodies; they are not still `position:auto`.
Distinguish U16 EP advisory	Completed	The advisory is informational routing preparation, not a placement error.
Check duplicates, stale traces, invalid layers, protected intrusions	Completed	No duplicate element/footprint population, 0 dangling traces, and no focused DRC findings for these categories.
Start a fresh full autoroute job	Unattempted by design	A full start may create a job and mutate copper if accepted, conflicting with the read-only requirement. No project geometry, rules, copper, schematic, or placement were changed.

Exact observed payloads

Live standalone board-check

Text
Board check: 342 components, 244 nets
[INFO] ROUTING_PIN_NEEDS_VIA_AT_SMD
Pin 'EP_e52e451139440b8a9931c77065c5b528' modified to allow via-at-SMD.

Repeated full autorouter start failure recorded in the project

Text
Auto-route job failed: Auto-layout failed to start. Please try again.
Errors:
- Auto-confirm API call failed: HTTP 400
Preflight placement anomalies (1):
  - [ERROR] PLACEMENT_GENERIC: Board fully placed. Move components to be placed outside the boundary.
    Fix: Review component placement for conflicts and adjust positions to resolve placement issues.

Historical job IDs include:

auto-layout-3f5758ec-5d78-4b36-8ade-856abcef24f8
auto-layout-1662cc09-eb83-45f4-8820-ba0f6dca2151
auto-layout-50ac2c55-e3b7-4cb5-906a-da6e731f29f8
auto-layout-10a4ec34-c308-4307-a01f-58c4b7f1c8ed
auto-layout-1107a8af-a334-4ff2-955b-f6857d9de10a
auto-layout-662dcb64-fbda-4bb7-b227-55c6529c4da0
auto-layout-6daf1e9e-f16d-4280-9b90-b40221b52e86
auto-layout-7fa8d5ba-cbe3-4599-b408-0a8ffad0a3a8
auto-layout-adfca284-b791-4ad4-8c97-a9baff09328c

The failure-time layout commit #c0a79be8 resolves to the same current layout state, so the geometry analyzed below is the geometry associated with the failure.

Primary suspected object

Table

Field	Evidence
Component	J3
Element UID	`521e2d81-25e7-4bf1-880c-1807177d3862`
Position / layer	`(-116, 57)` mm, Top, 0°
Suspect child	`J3 5mm Body Outline` rectangle
Child UID	`942c9ffc-52ae-43e2-8eba-9e6761b503db`
Child layer	Top Silkscreen
Child local data	`position: 7.5mm 0mm`; `startPosition: -9.17mm -1.62mm`; `endPosition: 9.17mm 1.62mm`
Board extents	X `[-122.5, 122.5]` mm; Y `[-90, 90]` mm
Reported J3 component bounds	X `[-125.32, -106.68]` mm; Y `[55.23, 58.77]` mm
Excursion	2.82 mm beyond the left board edge
Rendered geometry	The renderer applies the child `position`, placing the outline approximately X `[-117.67, -99.33]` mm, inside the board; J3 pads/body also render in-board.

This is the strongest trigger because J3 is the only live component whose reported aggregate placement bounds exceed the 245 × 180 mm board. The likely failure mode is that autorouter auto-confirm uses bounds derived from the child rectangle's start/end coordinates while ignoring, or differently applying, its local position; the normal renderer applies both. That explains why the board looks correctly placed and the standalone routing board-check is clean while auto-confirm reports a generic outside-boundary placement anomaly.

Minimal corrective action

Normalize the rectangle into one coordinate frame without visually moving it:

Keep J3 fixed at (-116, 57) mm.
For child rectangle 942c9ffc-52ae-43e2-8eba-9e6761b503db, change local X data from:
- positionX = 7.5mm, startX = -9.17mm, endX = 9.17mm
To the equivalent rendered geometry:
- positionX = 0mm, startX = -1.67mm, endX = 16.67mm
Keep Y values unchanged.
Re-run standalone board-check, inspect J3 bounds, then start autoroute. Expected confirmation: J3 reported bounds become in-board and the auto-confirm HTTP 400 clears.

If normalized bounds still trigger PLACEMENT_GENERIC, this should be treated as a platform coordinate/bounds defect and escalated with the object UID above.

Secondary candidate

J7 contains an active but blank VALUE text child:

UID: 87139867-060d-4379-8303-54a1bdb40818__1c4b5e5e-7160-4ed6-9049-66eac7f05c10.shapes_455c35e17d316a234ac1997ba6c47584
Local position: (0, -7.41) mm
World Y: approximately -90.50 mm, 0.50 mm below the board
Content: blank

Normal component bounds exclude this blank text, so it is less likely than J3. If the J3 normalization does not clear auto-confirm, disable or move this blank VALUE text in-board; doing so should have no visible effect.

Ruled-out or lower-likelihood suspects

J4–J7 parent placement

Table

Ref	Position (mm)	Layer / rotation	Result
J4	`(-120.77, -44.87)`	Top / 0°	Explicit numeric position; body and pads in-board.
J5	`(-120.77, -57.61)`	Top / 0°	Explicit numeric position; body and pads in-board.
J6	`(-120.77, -70.35)`	Top / 0°	Explicit numeric position; body and pads in-board.
J7	`(-120.77, -83.09)`	Top / 0°	Parent/body/pads in-board; only the separate blank VALUE text extends below the board.

The four connectors have component X bounds approximately [-122.10, -119.44] mm. J7's lowest body/silk geometry is approximately Y -89.50 mm and its Pin 3 pad edge is approximately Y -89.02 mm. Commit #0cf621b9 baked prior position:auto values for J4–J7; later commit #d76946de moved them to their current explicit positions.

U16 exposed-pad advisory

Component: U16
Pad UID: 8f3443bd-b661-4526-a58c-2799807ccfdf__3af3ce3a-5535-49f3-8259-d1903dea2569
Alias reported by check: EP_e52e451139440b8a9931c77065c5b528
Net: QI_RETURN_RAW

This is an [INFO] ROUTING_PIN_NEEDS_VIA_AT_SMD message. It indicates the routing engine adjusted via-at-SMD handling for an exposed pad. It does not report placement invalidity and appears in the otherwise clean standalone board-check, so it is not the autorouter start blocker.

Other checks

Layout has exactly 342 Element nodes and 342 Footprint nodes with unique designators; no duplicate element/footprint population was found.
Focused DRC returned no open findings for missing footprints, invalid layers, dangling traces, under-connected vias, protected intrusions, body overlap, component-outside-board, or edge placement.
Routing state reports 85 remaining airwires and 0 dangling/stale traces.
Twelve existing isolation-barrier zones are present. A prior controlled diagnostic removing all twelve did not change the HTTP 400 outcome, so they are not the cause.
No evidence of malformed parent layer or rotation was found; component inventory is Top-layer except intended Bottom-layer J20.

Confidence and backend uncertainty

Confidence: moderate-high that the J3 rectangle/bounds mismatch is the auto-confirm trigger. It is the only object-level condition that directly matches the backend's generic outside-boundary message on the identical failure-time layout.

Object-level causal proof is still unavailable because the auto-confirm HTTP 400 payload suppresses the offending UID/selector/coordinate. Engineering feedback was submitted requesting object-level anomaly details and consistent coordinate handling between rendered geometry, component-bounds calculation, and autorouter placement confirmation.

Final execution update — 2026-09-04

Scoped item status

Table

Required item	Status	Evidence
Read J3 live hierarchy and geometry	Completed	Read Element, Footprint, Model, four pads, suspect rectangle, applicable position/rotation rules, board outline, and rendered 2D coordinates before mutation.
Normalize J3 with zero visible/global movement	Completed	Folded the rectangle's `+7.5 mm` local offset into its endpoints, then shifted only J3's aggregate parent transform inward and algebraically compensated Footprint, Model, and rectangle child transforms.
Verify J3 after mutation	Completed	J3 aggregate bounds are X `[-121.99, -103.35] mm`, Y `[55.23, 58.77] mm`, inside the board. Pad world coordinates remain exactly Pin 1 `(-116,57)`, Pin 2 `(-111,57)`, Pin 3 `(-106,57)`, Pin 4 `(-101,57)` mm. Layer and rotation remain Top / 0°. The rectangle's global geometry remains X `[-117.67,-99.33]`, Y `[55.38,58.62]` mm by compensated transform.
Standalone autorouter preflight	Completed	Clean/non-blocking: only `[INFO] ROUTING_PIN_NEEDS_VIA_AT_SMD` on U16 EP / `QI_RETURN_RAW`. No `PLACEMENT_GENERIC` was returned by `run_auto_route_check`.
Full-board autoroute	Failed before routing began	Started only after clean standalone preflight. Auto-confirm returned HTTP 400 / `PLACEMENT_GENERIC` on every start, before copper mutation. No cleanup retry was justified because no routing pass started.
Routing-state checks and full DRC	Completed	Final routing state and focused/full DRC were re-read live after the failed start.

Transform result

The final effective transform uses a J3 parent X of -112.67 mm and compensates the Footprint, Model, and standalone rectangle by -3.33 mm in local X. This intentionally changes only the internal transform hierarchy: the pads, body/model, silkscreen outline, electrical endpoints, and connector mating geometry stay at their pre-mutation global coordinates. J4–J7 were not modified.

Exact remaining autorouter backend payload

Text
Auto-route job failed: Auto-layout failed to start. Please try again.
Errors:
- Auto-confirm API call failed: HTTP 400
Preflight routing anomalies (1):
  - [INFO] UNSUPPORTED_ANOMALY: Pin 'EP_e52e451139440b8a9931c77065c5b528' modified to allow via-at-SMD.
Preflight placement anomalies (1):
  - [ERROR] PLACEMENT_GENERIC: Board fully placed. Move components to be placed outside the boundary.
    Fix: Review component placement for conflicts and adjust positions to resolve placement issues.

The final failed job ID was auto-layout-3a04428f-21ab-41b1-a4dd-33c86ebd69a4. Two additional bounded, surgical zero-movement transform refinements were made after the first full-start failure: first a strict 0.01 mm aggregate margin, then a keepout-aware 0.51 mm margin. The same backend payload persisted, so no further transform edits were made.

Final routing and DRC counts

Table

Check	Final count
Airwires	85
Copper overlaps / shorts	0
Trace overlaps	0
Clearance / pad-access / keepout violations	0
Dangling traces	0
Floating copper	0
Invalid-layer violations	0
Protected-copper intrusions	0
Component-outside-board DRC findings	0

The board is not fully routed because 85 airwires remain. Those airwires affect 67 nets:

+12V_MOTION, +12V_RELAY, +12V_WIRELESS, +5V_B_PROTECTED, +5V_IN_B, +5V_LOGIC, +5V_SELECTED, 12V_RETURN_RAW, BASE_FOOT_DOWN, BASE_FOOT_UP, BDM_BKGD, CAN_TX_PROTECTED, CFG7, CMD_FOOT_UP, CMD_FOWLER_UP, CMD_HEAD_DOWN, CMD_HEAD_UP, CMD_KNEE_DOWN, CMD_KNEE_UP, DC_RETURN_A, DC_RETURN_B, ESTOP_CLOSED_N, ESTOP_OPTO_BASE, ESTOP_SENSE_LED, FOOT_COMMON, FOOT_DOWN_ALLOW_N, FOOT_DOWN_NO, FOOT_UP_NO, FOOT_UP_PNP_BASE, FOWLER_COMMON, FOWLER_DOWN_NO, FOWLER_UP_NO, HEAD_COMMON, HEAD_DOWN_NO, HEAD_UP_NO, HEAD_WIPER_RAW, I2C_PULLUP_FEED, J23_HL1_BUS_A, J23_HR_SIG_A, J23_HR_SIG_B, KNEE_COMMON, KNEE_DOWN_NO, KNEE_UP_NO, KNEE_UP_PNP_BASE, LC_ADC_A0, LC_ADC_DOUT, LC_ADC_PDWN, LC_ADC_SCLK, LED_FOOT_DOWN, LED_FOWLER_DOWN, LED_HEAD_UP, LED_KNEE_DOWN, LIMIT_HEAD_BOTTOM_NC, LOGIC_GND, MCU_RESET_N, MODE_B_LOW, PLL_XFC, QI_RETURN_RAW, RELAY_RETURN, REQ_FOWLER_DOWN, REQ_FOWLER_UP, REQ_KNEE_DOWN, REQ_KNEE_UP, ROMCTL_HIGH, SENSOR_GND, SPI_MISO_PROTECTED, UART1_RX.

Autoroute continuation attempt — 2026-09-04 05:40 UTC

The requested initial full-board autoroute was attempted after another live clean standalone preflight and confirmation of 85 airwires / 0 dangling traces. It again failed before routing or copper mutation with HTTP 400 and the same hidden PLACEMENT_GENERIC auto-confirm payload. Job ID: auto-layout-2842a81f-1ef6-4af9-806c-ab5f69ec628b.

No cleanup autoroute retry was run because the initial pass did not start or make progress. Post-attempt routing state and full DRC are unchanged: 85 airwires across the same 67 nets; 0 copper overlaps/shorts; 0 trace overlaps; 0 clearance/pad-access/keepout violations; 0 dangling traces; 0 floating copper. Component placement and existing copper were not changed.

Focused serializer cleanup and final retry — 2026-09-04 05:57 UTC

Scoped item status

Table

Required item	Status	Evidence
Inspect duplicate J3 rulesets and compute effective values	Completed	`2d62d1a8…` and `4d331f6b…` were exact duplicates: rectangle position `(-3.33,0) mm`, endpoints `(-1.67,-1.62)` to `(16.67,1.62) mm`, rotation 0°. `78ed65e7…` used position `(0,0)` with the same endpoints. Rendered J3 position/pads showed the parent rule `f0a80751…` at `(-112.67,57) mm` and child compensation `-3.33 mm` were the effective visual transforms; the duplicate rectangle rules affected aggregate-bound serialization despite identical rendered pads/body.
Consolidate J3 transform overrides	Completed	Deleted redundant parent `7b3eb362…`, footprint `a2ddeead…`, model `ba86ac02…`, and rectangle rules `4d331f6b…` / `78ed65e7…`. Retained one canonical parent (`f0a80751…`), footprint (`9d9ffca8…`), model (`073daf21…`), and rectangle (`2d62d1a8…`) ruleset. The rectangle rule was normalized to position `(0,0)` and endpoints `(-5.00,-1.62)` to `(13.34,1.62) mm`, algebraically preserving its previous global outline under the `-112.67 mm` parent.
Verify zero visual movement	Completed	J3 remains Top / 0° with rendered center `(-112.67,57) mm`; pad world coordinates remain exactly Pin 1 `(-116,57)`, Pin 2 `(-111,57)`, Pin 3 `(-106,57)`, Pin 4 `(-101,57) mm`. Existing copper endpoints remained coincident. Model/footprint compensation remains `-3.33 mm` local X. No other component or copper was changed.
Run standalone preflight twice	Completed	Both runs returned only INFO `ROUTING_PIN_NEEDS_VIA_AT_SMD` for U16 EP; neither returned `PLACEMENT_GENERIC`.
Make one final autoroute attempt	Failed before routing began	Job `auto-layout-65d1b01b-ce35-42a9-9851-72ef6011acc9` failed in hidden auto-confirm with HTTP 400 and `PLACEMENT_GENERIC`. No route was started and no copper was added or removed. Per scope, no further broad retry was made.
Run final routing-state and full DRC	Completed	Live routing state and selected full PCB DRC categories were read after the failed start.

Final retry payload

Text
Auto-route job failed: Auto-layout failed to start. Please try again.
Job ID: auto-layout-65d1b01b-ce35-42a9-9851-72ef6011acc9
Errors:
- Auto-confirm API call failed: HTTP 400
Preflight routing anomalies (1):
  - [INFO] UNSUPPORTED_ANOMALY: Pin 'EP_e52e451139440b8a9931c77065c5b528' modified to allow via-at-SMD.
Preflight placement anomalies (1):
  - [ERROR] PLACEMENT_GENERIC: Board fully placed. Move components to be placed outside the boundary.

Exact final state

Table

Check	Final count
Airwires	85
Affected nets	67
Copper overlaps / shorts	0
Trace overlaps	0
Clearance / pad-access / keepout violations	0
Dangling traces	0
Floating copper	0
Invalid-layer violations	0
Protected-copper intrusions	0
Component-outside-board DRC findings	0

The board remains not fully routed. The 67 affected nets are unchanged:

Scope and status
Exact observed payloads
Live standalone board-check
Repeated full autorouter start failure recorded in the project
Primary suspected object
Minimal corrective action
Secondary candidate
Ruled-out or lower-likelihood suspects
J4–J7 parent placement
U16 exposed-pad advisory
Other checks
Confidence and backend uncertainty
Final execution update — 2026-09-04
Scoped item status
Transform result
Exact remaining autorouter backend payload
Final routing and DRC counts
Autoroute continuation attempt — 2026-09-04 05:40 UTC
Focused serializer cleanup and final retry — 2026-09-04 05:57 UTC
Scoped item status
Final retry payload
Exact final state