Three reasons rev1 boards never work

Footprints. Decoupling. Hubris. Pick any three. A short list, written from the smoking ruins of a prototype.

There’s a thing that happens roughly two days after you receive rev1 of any PCB. You assemble it, you power it on, nothing works, and you spend the next six hours discovering exactly which kind of fool you were. Here is the pattern.

Footprints

You used the library footprint. The library footprint was for the SOT-23-3 with pin 1 on the left. Your part has pin 1 on the right. The part is now installed backwards and the magic smoke escaped within 200ms of power-on.

Decoupling

There is exactly one 100nF cap on the 3V3 rail and it is three centimetres from the chip it’s supposed to be decoupling. The chip is, predictably, doing interesting things on every clock edge.

Hubris

You didn’t simulate the analog front end. You didn’t breadboard the motor driver. You went straight to a four-layer board because you wanted to, and now you’re going to do rev2 anyway.