netsmith (noahdgrant) errors

16 recurring problems engineers report in netsmith (noahdgrant), ranked by impact.

#22 by errors trackedMCADECAD

Recurring netsmith (noahdgrant) errors, by impact

  1. 1On a real 12 V automotive board, four loads on the 3.3 V net had no current value before any part pick, so the regulator's output current was unset and the bead and shunt were judged on nothing, with the tool taking the first catalog row nothing refused.MCAD · Agent failure · 1 report74
  2. 2A 2N7002 (7 ohm at 5 V Vgs) was picked for a 3.3 V drive because nothing compared the datasheet's test voltage with the design's gate drive. On a real 12 V automotive board two modem driver rows in the BOM changed to AO3400A after the check went in.ECAD · Agent failure · 1 report66
  3. 3CI failed in this PR because the committed parity overlay was made by the old checker before another PR changed it on main.ECAD · Agent failure · 1 report66
  4. 4It reads both boards with one reader and reports per item offsets for outline, holes and parts, with must-match tolerances. Tests fail a hole 0.1 mm off, J2 0.2 mm off, SW1 rotated 90 degrees, and two pads on each other's nets. The overlay SVG is hand-written because the project's viewer only draws schematics.ECAD · Agent failure · 1 report66
  5. 5Review of overnight merges finds design hazards. A fuse's rating is used as the diode fault current and the comment calls that conservative, which is the opposite of what happens. A consigned part's LCSC number lands in bom.csv with nothing saying it is consigned.ECAD · Agent failure · 1 report66
  6. 6The author wrote five footprints straight from manufacturer drawings, logging what was read, derived and assumed for each. Two differ from KiCad's stock footprints: the HRO TYPE-C-31-M-12 in KiCad cites an older drawing with signal pads 1.45 mm long where Rev A gives a different length, and the Omron B3U-1000P switch differs too. The library footprint and the current datasheet are two sources for one part, with no link telling which revision the library used.ECAD · Agent failure · 1 report66
  7. 7The fab --order command solved the boards of an order from several files with no pick lock, so an order could pick rows that fab on the same file, or the lock beside it, would not, and the --locked flag was parsed and ignored for orders before. Now each board is held to its source file's lock, the order's bom.csv is tested byte-equal to single-file fab, and with --locked every mismatched lock is reported before anything is written.ECAD · Agent failure · 1 report66
  8. 8The LT8641 part declares 10 uA quiescent current from the datasheet but nothing reads it, so the unloaded converter draws 0 A and any sleep-current budget passes.ECAD · Agent failure · 1 report66
  9. 9The maintainer measured the tool's autorouter against hand-routed versions of three boards. RP2350: 158 vias vs 40, 428 right-angle corners vs 7, 2,160 mm of track vs 1,086, one 0.1 mm width everywhere vs 0.15 to 0.8 mm by hand. All three boards 'routed completely.' The epic is titled 'a router a layout engineer would trust,' which says where trust stands today.ECAD · Agent failure · 1 report66
  10. 10The netsmith project's first manual release run built on Linux and macOS but failed Windows with symbol library Device not found under a hardcoded KiCad 9.0 library path, because chocolatey installed KiCad 10.0.8 in a 10.0 folder. The README still says KiCad 9. Tool version drift silently breaks the board build.ECAD · Agent failure · 1 report66
  11. 11The placement check measured keepouts that parts carry, but the placer ignored them, so a part could land in another part's no-parts or height-limit zone.ECAD · Agent failure · 1 report66
  12. 12The router had been drawing a 0.15 mm via in a 0.51 mm pad, which JLCPCB charges extra for and which is wider than its standard-price via. The fab's capabilities page had nine rules the design rules never stated (hole-to-track, pad-to-pad, mask-to-track, slot widths). The author re-read the page by hand and typed each one into the process file.ECAD · Agent failure · 1 report66
  13. 13The router kept board clearance from every pad but ignored a pad's own clearance, so KiCad 9.0.7 fiducials (0.6 mm) got tracks too close and fab stopped on DRC.ECAD · Agent failure · 1 report66
  14. 14Two parts of one kind that differ by a pin name got two library entries with the same name, two power-output pins on one net were reproduced with KiCad's own ERC, and a placed part's in_bom flag was always yes whatever its symbol said, which matters because the schematic instance is what a bill of materials reads.ECAD · Agent failure · 1 report66
  15. 15Two reviews found eight real defects in a board parity checker that had already merged. A reference pad on a net passed against a board pad with no net, holes were compared without plating, size or net, and some of the checker's own gates and controls could not fail. Four tests failed on the unchanged code before the fix.ECAD · Agent failure · 1 report66
  16. 16With no KiCad install, netsmith symbol and footprint lookups joined the library name to an empty root and read whatever sat in the current folder as installed library parts. fab --locked also blamed the wrong flag in its error.ECAD · Agent failure · 1 report66

How it fails

  1. Agent failure16

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