01 ldo_breakout¶
Reference: AMS1117-3.3 typical application. First built from memory of the datasheet; read on 2026-11-14 (see below).
Stresses: nothing; it exists to prove the loop from board.py to Gerbers.
What broke
- ERC: 16 library warnings: the project named libraries it did not carry.
- Schematic parity: /VIN on the schematic vs VIN on the board.
- kicad-tools' own router routed one net and overwrote .kicad_pro.
- Freerouting necked a track to 0.187 mm against a 0.2 mm minimum.
Tool changes
- Vendor the used symbols and footprints into <board>/libs, trimmed.
- Global labels on every pin instead of local ones.
- Route through Freerouting (DSN out, SES in); rewrite .kicad_pro after kct.
- Widen any segment under min_track_width after import.
Left open: none worth recording.
Second pass: the datasheet, read¶
The datasheet asks for 22 uF solid tantalum on the output ("will ensure stability for all operating conditions"): the regulator relies on the output capacitor's series resistance, and the board had a ceramic. C2 is now a 22 uF tantalum in EIA-3528; the board was regenerated and passes the gate. The "10 uF in" was a misremembering of the adjust-pin bypass in Figure 1; the datasheet shows no input capacitor, so C1 stays as a choice. SOT-223 dissipation is capped at 1.2 W, about 0.7 A from 5 V.