STM32F411 Discovery Kit: USB not working

[pschatzmann]

I have a STM32F411 Discovery Kit with a STM32F411VE MCU.

I tried to select the STM32411E-DISCO, but USB output was not working for this. Selecting the generic STM32F411VE is working perfectly, so I am wondering if this is a bug or just a missing variant.

I can provide the following details:

variant_DISCO_F411VE.cpp's SystemClock_Config() (HSE 8MHz crystal):

VCO input = 8MHz / PLLM(4) = 2MHz
VCO output = 2MHz × PLLN(192) = 384MHz
USB clock = VCO / PLLQ(4) = 384MHz / 4 = 96MHz ❌ (USB OTG_FS requires exactly 48MHz)
SYSCLK = 384MHz / PLLP(4) = 96MHz

generic_clock.c's SystemClock_Config() (HSI 16MHz internal):

VCO input = 16MHz / PLLM(16) = 1MHz
VCO output = 1MHz × PLLN(192) = 192MHz
USB clock = VCO / PLLQ(4) = 192MHz / 4 = 48MHz ✓ correct

The Discovery board's clock config feeds the USB peripheral 96MHz instead of the required 48MHz — the PLLQ divider should be 8, not 4, given their chosen PLLM/PLLN/HSE values. That's exactly the class of fault that produces "Device not responding to setup address": the SIE gets a garbled bit-timing reference, so the host can detect the pull-up (attach) but the device can never complete even a basic control transfer — matching everything we observed. It has nothing to do with our TinyUSB port, the audio-tools library, or your original sketch.

My Practical takeaway: stick with GENERIC_F411VETX (already confirmed fully working, byte-exact recording) rather than the dedicated STM32F411E-DISCO board entry, until this gets fixed upstream in stm32duino/Arduino_Core_STM32. The only real difference between the two board definitions for your purposes is this clock bug — the pin variant, HAL, and everything else is identical. If you want, I can help draft a bug report/PR against the stm32duino core with this exact analysis.