aethelgard (ClearForge-LLC) errors

5 failures reported against aethelgard (ClearForge-LLC), sorted by how they fail and ranked by impact.

#90 by errors trackedElectronics & PCB

Recurring aethelgard (ClearForge-LLC) errors, by impact

  1. 1A vacuous proof was enough to pass prove, and the reader had a critical bypass and must now tokenize exactly as KiCad does.Electronics & PCB · Wrong verdict · 1 report66
  2. 2Agent-run spike on off-board connectors: the work order assumed JST GH, the Pixhawk DS-009 spec says the debug ports are SH. A device-less SPI bus produced three one-pin nets, and the tool's validate, solve and emit all exited 0. KiCad would flag each one, but the KiCad gate run was blocked, so nothing caught it.Electronics & PCB · Wrong verdict · 1 report66
  3. 3ClearForge's aethelgard spike ran KiCad 10.0.7 at 4 and 6 layers and found the stackup is inert: KiCad needs no stackup, the gate discards the setup block, and DRC and the Gerbers ignore it. The team now requires a board section with layers, the stackup as bounds with sources, and pulls prepreg thicknesses from JLCPCB and OSH Park pages. The stackup is the one place board, fab and SI meet, and the ECAD file treats it as a comment.Electronics & PCB · Silent failure · 1 report66
  4. 4D1 board already fails its own 100 pF line limit on main, and no test runs the check on it While sizing I2C/SPI drive strength from extracted line capacitance, the team found 7 of 8 I2C lines on D1 at 87.8 to 119.6 pF against its declared 100 pF, and nothing in CI runs extract on that board. The IMU fixture has no KiCad symbol, so extraction gives it no capacitance at all. The MCU's pin capacitance is a typical value only. Datasheet numbers, library symbols and the board each hold a piece of the answer.Electronics & PCB · Silent failure · 1 report66
  5. 5Rule RUL-03 says the gate supplies every file that carries DRC rules from the spec and never uses the emitter's copies; the same PR notes Renode's STM32F7 I2C model has no I2C errors at all, so simulated firmware can't hit a NACK.Electronics & PCB · Missing capability · 1 report66

How it fails

  1. Wrong verdict2
  2. Silent failure2
  3. Missing capability1

Sourced from public community reports. Costs and delays on individual entries are modeled estimates, not measurements.