Skip to main content
Visitor II
Oct 23, 2026 5:32 PM UTC
Question

STM32H5 (all lines): wrong PLLxRGE / VCI range generated – incorrect rules in the CubeMX RCC database (still in 6.18.1)

Hello,

I would like to report what looks like a bug in the STM32CubeMX MCU database for the whole STM32H5 family: the PLLx input frequency range (PLLxRGE) that CubeMX generates does not match the actual PLL reference frequency.

A similar case was already reported for the STM32H5E4 and PLL3 in this thread (internal ticket #0064521). As far as I can see, the problem is wider: it affects every STM32H5 RCC database file, plus some PLL1 and PLL2 rules, and it is still present in STM32CubeMX 6.18.1.

Disclosure: this was found by a Claude Opus AI agent while reviewing the code generated by STM32CubeMX for my project. I have checked the generated code and the database lines it points to, but I have not verified the behaviour on hardware (PLL lock/jitter), so please treat it as a report to be confirmed by your team.

Setup

  • MCU: STM32H563VIT6 (RCC IP STM32H5_rcc_v1_0)
  • STM32CubeMX 6.17.0 (DB.6.0.170) and 6.18.1 (DB.6.0.181), with STM32Cube FW_H5 V1.6.0 and V1.7.0: same result
  • HSE 25 MHz, USB clock from PLL3Q

Generated code (HAL_PCD_MspInit)

  • PLL3 M=25, N=240, Q=5 (1 MHz reference, VCO 240 MHz) → PLL3RGE = RCC_PLL3_VCIRANGE_3 and PLL3VCOSEL = RCC_PLL3_VCORANGE_WIDE. Expected: RCC_PLL3_VCIRANGE_0 (1–2 MHz).
  • PLL3 M=5, N=48, Q=5 (5 MHz reference) → PLL3RGE = RCC_PLL3_VCIRANGE_0 and RCC_PLL3_VCORANGE_MEDIUM. Expected: RCC_PLL3_VCIRANGE_2 (4–8 MHz).

The “PLL3 input frequency range” parameter is read-only in the RCC Parameter Settings (and hidden when the default rule applies), so it cannot be corrected from the GUI.

Root cause (as far as I can tell)
In db/mcu/IP/RCC-STM32H5_rcc_v1_0_Modes.xml the PLL3_VCI_Range conditions do not match their values/comments:

  • Line 8351: RCC_PLL3_VCIRANGE_0 (“Between 1 and 2 MHz”) has the condition 4000000 <= VCOInput3Freq_Value < 8000000
  • Line 8355: RCC_PLL3_VCIRANGE_1 (“Between 2 and 4 MHz”) has the condition 8000000 <= VCOInput3Freq_Value < 16000000
  • Lines 8359/8363 (_2 and _3) have the correct conditions but seem to be shadowed by the two above
  • Any value below 4 MHz falls through to the hidden default RCC_PLL3_VCIRANGE_3 (line 8366, “Less than 4 or greater than 16”)

As a result, no PLL3 reference frequency from 1 to 16 MHz gets the correct range (only exactly 16 MHz does). This is consistent with the H5E4 case in the thread above (4 MHz → VCIRANGE_0).

Same rules in the other files. I found identical conditions in:

  • RCC-STM32H5_rcc_v1_0_Modes.xml (H562/H563/H573)
  • RCC-STM32H5_rcc_v1_1_Modes.xml (H5E4/H5E5/H5F4/H5F5, lines 9467–9483)
  • RCC-STM32H5_rcc_v1_2_Modes.xml (lines 6590–6606)
  • RCC-STM32H5_rcc_v1_512_0_Modes.xml (lines 5568–5584)

Related issues in the same block (line numbers from rcc_v1_0):

  1. PLL1, line 8322: the RCC_PLL1_VCIRANGE_3 (8–16 MHz) condition is (f >= 8 MHz & f < 8 MHz) | f = 8 MHz, so it only matches exactly 8 MHz. Other inputs between 8 and 16 MHz only get the right value through the hidden default.
  2. PLL1, line 8372: the PLL1_VCO_SEL MEDIUM condition compares against RCC_PLL1VCIRANGE_0 (missing underscore), so it never matches and PLL1 always gets RCC_PLL1_VCORANGE_WIDE, even with a 1–2 MHz reference. The same typo appears in all five H5 RCC files.
  3. PLL2, line 8331: the RCC_PLL2_VCIRANGE_0 condition is (f > 4 MHz | f = 1 MHz) & f < 2 MHz, so it only matches exactly 1 MHz.

Questions

  • Can you confirm the issue, and whether ticket #0064521 covers all STM32H5 RCC database files and the PLL1/PLL2 rules above?
  • Is there a planned CubeMX/database version for the fix?
  • Is RCC_PLLxVCORANGE_WIDE allowed with a 1–2 MHz reference (RM0481), or is MEDIUM mandatory in that case?

Workarounds I am considering: taking the USB clock from PLL1Q with a 5 MHz reference (PLL1 M=5, N=96, P=2, Q=10), or overriding PLL3RGE/PLL3VCOSEL in user code. I can attach the .ioc if needed.

Thank you!

1 reply

Ghofrane GSOURI
ST Technical Moderator
Oct 26, 2026 8:36 AM UTC

Hello ​@JSanEst 

Thank you very much for your detailed description.

I am currently checking this behavior. I will get back to you ASAP.

THX

Ghofrane

To give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.