Your PCB technology choices seem a little odd:
- 4-layer (sensible for a first attempt with RP2040, but not the cheapest)
- Via-in-pad. This is normally a very expensive option; fine if that's what you've actually paid for, maybe a problem if you've just put vias in the pads and had it made as a normal PTH board. Such boards won't solder well.
- Huge clearances on other layers around the via holes. This is sometimes necessary if the board is being built on the cheapest possible process, but seems inconsistent with your other 'upmarket' choices.
The big clearances might possibly have caused things to fail to connect on your plane layers, but I can't actually see a place where it matters on your plots: something odd seems to have happened to the groundplane under the RP2040, but it still appears to connect.
Very little actually needs to work for the device to enumerate on USB: the power supplies need to work, the crystal needs to oscillate at the right frequency, and the USB pins need to be connected. The flash doesn't matter for getting to this first stage - if it's blank or broken the RP2040 will go into USB boot mode.
As you say, the CC pins on the USB connector really ought to be connected, but it won't matter in your current configuration.
One possible issue with an external PSU is rise-time. You are relying on the on-chip power-on-reset (and the fact that you haven't connected the reset pin to anything, even a testpoint, is unhelpful on a prototype), and a common problem with slow rising power supplies is that the RP2040 starts booting before the flash is ready - but that's less likely to be your problem here (it might be the next problem having got code into the flash).
So for debugging I'd recommend checking the power supplies (particularly the 1.1V out from the on-chip regulator), and that the crystal is oscillating. You should also see brief activity on the flash pins - not important in itself but an indication that the chip is running.
As an indication that things are well, expect around 10mA being drawn from your 3.3V PSU. Much more or less than this is suspicious.
I assume you never powered your board with the incorrect 5V regulators fitted - obvious risk of damage if you did.
- 4-layer (sensible for a first attempt with RP2040, but not the cheapest)
- Via-in-pad. This is normally a very expensive option; fine if that's what you've actually paid for, maybe a problem if you've just put vias in the pads and had it made as a normal PTH board. Such boards won't solder well.
- Huge clearances on other layers around the via holes. This is sometimes necessary if the board is being built on the cheapest possible process, but seems inconsistent with your other 'upmarket' choices.
The big clearances might possibly have caused things to fail to connect on your plane layers, but I can't actually see a place where it matters on your plots: something odd seems to have happened to the groundplane under the RP2040, but it still appears to connect.
Very little actually needs to work for the device to enumerate on USB: the power supplies need to work, the crystal needs to oscillate at the right frequency, and the USB pins need to be connected. The flash doesn't matter for getting to this first stage - if it's blank or broken the RP2040 will go into USB boot mode.
As you say, the CC pins on the USB connector really ought to be connected, but it won't matter in your current configuration.
One possible issue with an external PSU is rise-time. You are relying on the on-chip power-on-reset (and the fact that you haven't connected the reset pin to anything, even a testpoint, is unhelpful on a prototype), and a common problem with slow rising power supplies is that the RP2040 starts booting before the flash is ready - but that's less likely to be your problem here (it might be the next problem having got code into the flash).
So for debugging I'd recommend checking the power supplies (particularly the 1.1V out from the on-chip regulator), and that the crystal is oscillating. You should also see brief activity on the flash pins - not important in itself but an indication that the chip is running.
As an indication that things are well, expect around 10mA being drawn from your 3.3V PSU. Much more or less than this is suspicious.
I assume you never powered your board with the incorrect 5V regulators fitted - obvious risk of damage if you did.
Statistics: Posted by arg001 — Tue Jul 30, 2024 7:21 pm