node60-hw (karlTbowers) errors

4 recurring problems engineers report in node60-hw (karlTbowers), ranked by impact.

#130 by errors trackedECADPLM

Recurring node60-hw (karlTbowers) errors, by impact

  1. 1The busbars existed only as outlines on User.Drawings, a layer DRC ignores, so nothing stopped a part going under a 100 A copper bar. The fix adds a generated board-only footprint per bar with a courtyard KiCad will check plus a copper rule area, though kicad-cli 10.0.0 does not enforce the rule area's own footprint keep-out and a footprint rule flagged 36 parts that were clear of the bars.ECAD · Agent failure · 1 report74
  2. 2The enclosure geometry script used the parts library estimated 7.7 mm SurLok stud and misread a catalogue dimension of 19.2 mm as seat-to-panel. Amphenol drawing C10-730187 rev F and its STEP model give a 9.95 mm stud, 9.25 mm seat-to-panel and a 3.0 mm flange, not 2.5.PLM · Agent failure · 1 report74
  3. 3A library submodule bump gives a capacitor a new description, and it has to be written into the placed symbols and into each sheet's cached copy before ERC stops flagging lib_symbol_mismatch. Locally ERC still shows 108 label_dangling errors that CI on main does not, and the author guesses it is a local difference.ECAD · Agent failure · 1 report66
  4. 4On Node60 a copper bar is the only thing joining J910 and J901, so KiCad reported J901 unconnected. A net tie between two pads on the same net joins nothing (tested, 17 unconnected items either way). So J901 moves to its own net and the bar becomes a two-pin net-tie part, which also gives it a BOM line. It takes a library PR stacked on two KLC exception PRs plus a placement script change. A mechanical part had to be faked as an electrical one so the tools would agree with the product.ECAD · Agent failure · 1 report66

How it fails

  1. Agent failure4

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