No parts, no breadboard, and no notebook page for this one. There was no math to do and nothing to draw. The LED is already wired to D13 on the board.
This is the "does anything work at all" test. Board, cable, driver, upload toolchain, all proven before a breadboard shows up.
Stock example uses delay(1000). I dropped it to 100 on and 75 off so it reads
as a flicker instead of a blink. Somewhere under 50ms it stops looking like
blinking and starts looking dim, which is PWM before I knew what PWM was.
What I learned
setup()runs once at power on.loop()runs forever after thatpinModesets direction, so drive voltage vs read it.digitalWritesets the voltage itself, 5V or 0VLED_BUILTINis a name, not a number. It resolves to whatever pin the onboard LED is on for the board you picked. D13 on the UNO, D6 on a MKR1000- Upload flow is Tools then Board, Tools then Port, then Upload
- If upload fails it is almost always the wrong Port, not the code
delay()blocks. The board does nothing else while it waits. No button, no Serial- It is not sleeping either. 16MHz chip, so a
delay(200)burns roughly 3.2 million instructions spinning a counter doing nothing millis()is the real fix. It shows up in the traffic light- 100 on and 900 off is a "heartbeat pattern". Debugging trick to show the code is alive but idle