Debugging a really odd issue the past few days. It seems that firmware auto-compiled in the Linux build server causes a couple Bus Pirates not to boot. Firmware compiled on my local Windows machine seem to work fine.The PICO SDK Windows installer uses arm-gcc 10.3, while the build server uses the latest. I spooled up a new build server and installed arm-gcc 10.3, but that doesn't fix the bug either. For now I'll post a...
..manual build from Windows, and look into using a Windows laptop/surface pro as build server long term.
@buspirate How well characterized are the tuning caps on your xtal? We had boards that wouldn't boot due to what we thought was a firmware problem, but it turned out that the RP2040 SDK has an extremely short timeout for crystal startup. Changing that was a quick software fix: https://www.dropbox.com/s/bcwjz9dnmoshg4h/2023-08-22%20blinkypendant%20startup%20issue.pdf?dl=0
@cibomahto That appears to have fixed it, thank you so much. I'm waiting to get a board with the bug myself so I can poke at it a bit, but a user said it fixed their issue. Wonder if there is a different default delay in my windows SDK vs the Linux build server.
@buspirate There shouldn't be any difference- it's set by that variable, which is pulled from the board profile by default. Feel free to bring a board by if you want to measure the clock tuning.
