I thought so too... | adafruit | Page 1

solid locust Feb 19, 2025, 3:56 PM

#

are you using the Pi Debug Probe UART wired to the Feather RX and TX pins?

floral delta Feb 19, 2025, 3:57 PM

#

Yes

#

I also tried with a USB-serial cable

solid locust Feb 19, 2025, 3:57 PM

#

what version of zephyr did you build with, or more specifically, the HEAD commit of your zephyr repo?

floral delta Feb 19, 2025, 3:59 PM

#

4.1.0-rc1

solid locust Feb 19, 2025, 4:00 PM

#

I've been doing git pull; west update fairly regularly. Currently I'm at

(.venv) $ git branch -v
* main c0d8bd17d11 doc: Fix the doc issue in samples/net/wifi/apsta_mode/README.rst

#

also, which OS are you using for the host PC and which program are you using for the serial monitor?

floral delta Feb 19, 2025, 4:01 PM

#

Ubuntu Linux and U use both screen and tio -- same results.

#

Aslos, would you expect it to work if I just copied the .uf2 file to the board? Taht does not appear to work. at lest I ge no response via the terminal program

#

I'll try updating

solid locust Feb 19, 2025, 4:03 PM

#

brb... gonna hook my Feather RP2350 back up and see if I can reproduce your results

floral delta Feb 19, 2025, 4:04 PM

#

now at commit 90919eb6

#

Jsut to be clear for a simple test -- If I do make flash then make uart it works fine, but then if I reset the board, it does not work in the screen session ay more

solid locust Feb 19, 2025, 4:07 PM

#

This is what it looked like for me with tio when I first connected the debug probe, started tio, then connected the Feather RP2350, then reset the Feather RP2350

floral delta Feb 19, 2025, 4:08 PM

#

hmmm -- let me try

solid locust Feb 19, 2025, 4:09 PM

#

are you starting screen with -fn to disable flow control? If you don't do that, all sorts of exciting things might happen.

floral delta Feb 19, 2025, 4:10 PM

#

I just tried tio as above and it does not work after a reset

#

and yes to -fn for screen

solid locust Feb 19, 2025, 4:12 PM

#

hmm... I have no clue. actually, come to think of it, I'm probably not running the same board def as you. There's an open PR here: https://github.com/adafruit/adafruit-zephyr-support/pull/4 for an improved RP2350 board def. You're welcome to try it, but it's a work in progress

GitHub

add basic Feather RP2350 support by samblenny · Pull Request #4 · a...

This is bare bones with just UART serial and I2C. There's some SPI config, but it's untested and quite possibly broken. I2C and UART do work well though. I tested this using a Raspb...

#

this stuff is all super-alpha at the moment

floral delta Feb 19, 2025, 4:14 PM

#

OK -- thanks -- I'll keeep at it, but from your experience, the debug prbe should worke like a USB-serail cable onece the board is flashed -- that is just use tio or screen to connect.

solid locust Feb 19, 2025, 4:15 PM

#

yeah. I haven't noticed any odd behavior with the debug probe with either of screen or tio. There have been times when the Zephyr shell didn't give me a prompt until I first typed some stuff.

floral delta Feb 19, 2025, 4:17 PM

#

OK -- I'll keep working at it. Thanks for your help and the guides. The "getting started guide" woked great with the qtpy -- and except for this issue with the uart, the feather is working well...

solid locust Feb 19, 2025, 4:18 PM

#

That's good to hear. I'm glad the guides are useful. If you figure out anything with the odd reset behavior, I'd be curious to hear what the problem was.

#

Also, have you done the standard checks for bad solder joints, bridges, damaged or flaky USB cables, etc?

floral delta Feb 19, 2025, 4:20 PM

#

Yes -- as I noted, it all works fine immediately after flashing -- I just can't use it without flashing it first.

#

Aslo - should it also work to just copy the .uf2 file over?

solid locust Feb 19, 2025, 4:22 PM

#

I don't know. There's some potential for weirdness as the flash partition tables aren't matched up yet between my board def and CircuitPython's usage. I'm actually working on that today in the PR I mentioned.

#

One thing you could try is erasing the whole flash chip with openocd before flashing the zephyr firmware. point would be to see if any residual circuitpython stuff might be affecting the behavior of the Zephyr code.

#

The RP2350 bootloader does some fancy stuff for detecting between a few different boot options (image_def, partition_table, etc)

floral delta Feb 19, 2025, 4:23 PM

#

AH OK -- do you have the command for full erase handy.?

solid locust Feb 19, 2025, 4:24 PM

#

brb... maybe I can find it

floral delta Feb 19, 2025, 4:24 PM

#

thanks -- but don't spend a lot of time on it -- I can google....

solid locust Feb 19, 2025, 4:27 PM

#

didn't find it. sure I saw that amidst all the references I've been going through. I have vague memories of mass_erase or something like that, perhaps?

floral delta Feb 19, 2025, 4:27 PM

#

No problem -- Thanks for your help

solid locust Feb 19, 2025, 4:27 PM

#

you're welcome. good luck

solid locust Feb 19, 2025, 5:49 PM

#

@floral delta One other thing I just thought of... I'm using the base model Feather RP2350 (no PSRAM). If you happen to have the one with 8MB PSRAM, that could potentially affect things.

floral delta Feb 19, 2025, 10:00 PM

#

solid locust <@343570430042308609> One other thing I just thought of... I'm using the base m...

Thanks -- I am using the base model too...sigh

solid locust Feb 20, 2025, 6:02 PM

#

floral delta Thanks -- I am using the base model too...sigh

Another possibility might be some kind of clock startup thing. Things you could try:

Verify that board behaves normally if you switch back to CircuitPython firmware (in this case, maybe the default Zephyr RP2350 clock startup code does not allow for a wide enough range of manufacturing variations?)
If CircuitPython firmware is also weird, are there any unusual supply voltage or ambient temperature issues that might affect the oscillator?
Does anything change with the reset behavior if you power the board from a phone charger or a LiPo pack?

#

This might be relevant: https://github.com/raspberrypi/pico-sdk/releases/tag/2.1.1

The following board configurations have been modified:

adafruit_feather_rp2350 - Increased the XOSC startup delay

floral delta Feb 20, 2025, 7:07 PM

#

solid locust Another possibility might be some kind of clock startup thing. Things you could ...

I tested the UART via CircuitPython and it works normally with the RPi Debug probe, so that all checks out. I tried manually updating some of the zephyr board files from your PR, but I may have missed some settings -- I'll try switching to your new board files later today to see if that helps. I can also try a second boars as soon as I put some headers on it.

#

The poer supply does not seem to make any difference

solid locust Feb 20, 2025, 7:10 PM

#

I'm really suspicious about that XOSC startup delay thing mentioned in the Pico SDK release

floral delta Feb 20, 2025, 7:12 PM

#

Will that update flow in to Zephyr -- or what I need to update for it?

solid locust Feb 20, 2025, 7:14 PM

#

No clue when Zephyr will pick up those changes. I think the release just came out a couple days ago? As a diagnostic thing though, it suggests that somebody was having trouble with clock startup on the same board you're using.

#I thought so too...