QUERY · ISSUE

[PYBD] Switching Frequency completely locks up the board

openby ThomasChropened 2019-04-01updated 2019-04-30

port-stm32

Examplecode:
machine.freq(216000000)
More details in a Forum Thread: https://forum.micropython.org/viewtopic.php?f=6&t=6190

CANDIDATE · ISSUE

ports/stm32: Can't change CPU frequency on NUCLEO_F401RE board

closedby dhylandsopened 2019-05-01updated 2019-05-02

port-stm32

Trying to change the frequency using the latest master causes a FATAL ERROR.

I tested using pyb and machine modules and tried various frequencies like 42 MHz and 64 MHz).

All attempts result in this:

MicroPython v1.10-290-g8402c26cf-dirty on 2019-05-01; NUCLEO-F401RE with STM32F401xE
Type "help()" for more information.
>>> machine.freq(42000000)

FATAL ERROR:
can't change freq

Originally brought up in the forum: https://forum.micropython.org/viewtopic.php?f=12&t=6365&p=36215

Assessment

S Sonnet · high

Both issues report machine.freq() failures on the stm32 port, but they differ in board, MCU family, and failure mode. The query is PYBD (STM32H7, 216 MHz target, silent lockup), while the candidate is NUCLEO_F401RE (STM32F401xE, 42/64 MHz target, explicit FATAL ERROR). STM32H7 and STM32F4 have entirely different clock tree architectures, so the root cause is almost certainly distinct. The different failure signatures (hang vs. explicit error message from the clock-change code path) further support separate underlying bugs.

Suggested action

Link the two issues as related for context. Investigate and fix separately: the PYBD lockup likely involves the STM32H7 PLL/clock switching sequence, while the NUCLEO_F401RE FATAL ERROR points to a bug in the F4 frequency-validation or clock-reconfiguration logic.

[PYBD] Switching Frequency completely locks up the board

ports/stm32: Can't change CPU frequency on NUCLEO_F401RE board

Keyboard