STM32 MCUs Software development tools

The new version of CubeMx (6.15.0) nags me about enabling SMPS mode, and I would like to, but the tool seemingly does not have the capability to do so.Under PWR->Low Power there's a checkbox for "Power Savings Mode" which other forum posts led me to believe would need to be enabled. I have found two things about this checkbox. 1.) It is already enabled and 2.) You cannot disable it.Then, under "System Core" -> RCC ->Parameter Settings, there's a "Supply Source" parameter. It has only a "PWR_LDO_SUPPLY" option, and cannot be changed to any other value.Am I missing something here, or is the tool broken? 

Hello, I get the same error recently on 1.19.0 for my STM32U385RGTxQ. The only option available in Power Supply source is PWR_LDO_SUPPLY

From Cube 1.19.0, i keep getting the "For the best power consumption figure, enable SMPS" and although the Nucleo-U385RG-Q board uses a Q variant, I'm not able to switch to SMPS from CubeMX.

Hello,I am trying to understand the memory allocation for the FDCAN, I am using a nucleo H723ZG.With the latest CubeIDE 1.18.1 I seem to be able to configure 32 TX Buffers and 32 TX FIFOs:However the message RAM mapping in AN5348 - Rev 5 shows that there are only a maximum on 32 for both Buffers and FIFOs:Is there a mistake in the MX interface that should not allow more that 32 of Buffers + FIFOs combined, or the memory mapping has 2 sections of up to 32 elements for both buffers and FIFOs? Thanks,Marco

As alread reported a while ago, strange things happen when entering an expression in CubeMX for Prescaler and/or Autoreload. If, for example, 72-1 is entered for Prescaler and 6667-1 for Autoreload, the following code will be generated:STM32CubeMX: TIM Prescaler, Period bug   TIM_InitStruct.Prescaler = 72-LL_TIM_IC_FILTER_FDIV1_N2; ... TIM_InitStruct.Autoreload = 6667-LL_TIM_IC_FILTER_FDIV1_N2; LL_TIM_IC_FILTER_FDIV1_N2 does not make any sense here.Depending on the exact values, the Timer will be initialized with some different settings than desired.With simple constants (without calculations) the generated code will be correct.In contrast to the previous post, in my case the MCU was a STM32F303.CubeMX is Version: 6.15.0-RC7, Build: 20250701-1659 (UTC)

Hi, I'm trying to come up to speed with the VSCode workflow, coming from experience so far only using CubeIDE.  In CubeIDE, it seems that I was encouraged to think of having a "workspace", in which several projects co-existed. I enjoyed the ability to be able to have multiple projects "open" at once, to be able to compare code and easily understand/navigate to out-of-project source code, like centralized libraries and such. The VScode workflow, in contrast, seems to want me to only open a single project at once, otherwise CMake seems to get crabby. This is slightly annoying in that it requires me to open multiple windows in order to see multiple projects simultaneously, but notably more annoying when I am trying to use out-of-project code - I need to open a second window at a higher directory level, but can't actually build anything in this second window. It's kind of cluttered and confusing. It also seems to kind of push me away from the path of having a git repo t

STM32CubeProgrammer v2.21.0 I implement Open Bootloader middleware to communicate USART protocol over USB CDC on Mbed OS 6 I found when I open STM32CubeProgrammer and click Erasing & Programming tab first time, program will freeze. But after first time does not freeze. After Click Erasing & Programming tab. Program send Get ID for 100 packets per sec Graph with 1 Sec interval, Get ID packets (I only filter 0x02) Graph with 100ms interval, Get ID packets Before Click, 2 Get ID per 1 sec After Click, 513 Get ID packets with no delay I also try with FTDI and result is mostly same Graph with 100ms interval, Get ID packets 6-7 packet per 100ms And also 513 packets I also found when before and after buck send. Program change baud rate to 9600 and switch back to 115200 (I use 115200)

QuestionI implemented a custom bootloader with USB DFU. It worked without issues on my Windows 10 computer. After switching to a Windows 11 machine, however, my DFU device is no longer recognized.What I’ve triedConnected the device again on Windows 10. It appears in Device Manager, plays the connection/disconnection sound, and read/write operations work with STM32CubeProgrammer. This confirms that neither the bootloader nor the USB cable is the root cause.Installed STM32CubeProgrammer on Windows 11, but it reports “No DFU detected.”On Windows 11, there is absolutely no activity related to the DFU device in Device Manager. Other DFU devices, such as the STLink, are listed and working correctly.RequestWhat steps can I take to get my DFU bootloader detected on Windows 11?

Nucleo Board STM32F401RE , STM32CubeIDE 1.19 16x2 LCD confirmed to be working using Arduino UnoHi, I am just starting with STM32 and trying to figure out my way around it after using Arduino Uno for a long time for simple projects with sensors. I installed IDE , created a new project, selected my board from the first menu and so far everything is OK. I generated code without changing anything related to clock or the pin out. The problem is that even though I have not added any code to the project, the IDE opens a different c file every time I upload code, sometimes with a HAL in the name other times it opens this file startup_stm32f401xe.s and highlights this line ldr sp, =_estack /* set stack pointer */I went back and enabled the builtin green LED and added 2 lines in the while to toggle it every 500ms, then uploaded the code but weirdly even after the upload is finished sometimes the LED would start toggling on and off but on the next upload it wouldn`t. the exact

Edited by a ST moderator. In next time please use English language.   在STM32CUBEIDE中配置STM32F103C4完以后直接编译出现《没有 “huart2() at 0x20000108” 对应的源文件》不知道该如何把huart导入到编译器的路径下 Translation: After configuring the STM32F103C4 in STM32CUBEIDE, I encountered the error "No source file corresponding to 'huart2() at 0x20000108'" during compilation. I don't know how to import the huart2() file into the compiler's path.

I created default project with FreeRTOS in STM32CubeMx for the board Nucleo-H563ZI. Nothing more added.Hard Fault is invoked at vRestoreContexOfFirstTask.I was looking for the solution in Copilot and Perplexity but without success. Only I can say it's probably no problem with stack overflow.Jan

Hi, my team works with STM32CubeIDE and shares projects via github. We've been running into somewhat frequent errors where projects settings don't get shared properly when pulling the files down from github. On the flipside, there are a bunch of automatically generated files that change frequently with builds, code generation, etc, that bloat the git log unnecessarily if they are shared.Is there a standard gitignore that we should be using for STM32CubeIDE projects. Are there any rules of thumb for how to set up stuff like project folder structure, project variables, etc, such that they get properly shared?

I am flashing a code to write into the last sector of the flash memory of STM32F401 Nucleo Board, for the first time the code was successfully dumped into the flash but while dumping it the second time this error pops up, whenever i mass erase the whole flash it again works properly but the second it doesn't. I even tried using different sectors but the issue was still there.What might be a possible solution of this error.I am attaching the snapshot of the error and the code snippet as well.uint32_t Flash_Write_Data (uint32_t StartSectorAddress, uint32_t *Data, uint16_t numberofwords) { static FLASH_EraseInitTypeDef EraseInitStruct; uint32_t SECTORError; int sofar=0; /* Unlock the Flash to enable the flash control register access *************/ HAL_FLASH_Unlock(); /* Erase the user Flash area */ /* Get the number of sector to erase from 1st sector */ uint32_t StartSector = GetSector(StartSectorAddress); uint32_t EndSectorAddress = StartSectorAddress + numberofwords*4; uint32_t

 Hello,I updated to Mac15 (MacOS Sonoma Public Beta), which was probably a mistake. Now I get an error from STMCubeIDE when I try to run indicating "ST-Link Server is required to launch the debug session" and to Please download it on www.st.com.  I figured something changed in the OS that no longer makes the ST-Link Server run any more.However when I try to download it, I get essentially a broken link, or at least it simply wont download. Even worse, the web page with Safari seems entirely broken and I cant even get a download to start after accepting the license, but with FireFox I get further, I get but still wont download and the link seems broke.I can still flash using the STM32CubeProgrammer fortunately (using 2.13.1).Can anyone tell me how to download this ST-Link Server? Or if there is an issue with the last MacOS and this and there is no fix? 

HiWhen I migrated by project to ST Cube IDE v2 this morning, my CM4 project would not link. I had to right-click the project and "convert to C++"  (even though I had been using C++ for months with previous versions).Using a Nucleo STM32H755 as a target (dual core), my project is configured as follows:CM4 is using C++, and using Free RTOSCM7 is using C, and only using the HAL Steps TakenUpgrade the IDE from v1.19 to 2.0Install latest Cube MXOpen my workspace, and chose to "migrate" when promptedDouble click the .ioc file, fix an issue (related to CAN which I am not using) and re-generate the codeReturn to the IDE, and try to clean and build the CM4 project. I get linker errors. ResolutionRight-click the CM4 project and select "Convert to C++". (In v.19 the only available option was "Convert to C")Clean and build. Maybe this is a bug in the project migration process? Could it be specific to dual-core, where only the CM was configured for C++? CheersNick Outram&nb

For years, ST seems to spit in the face of developers every time they try to “improve” the ecosystem. Even now they’re trying to move toward *** like AI chips or other *** that sounds great to shareholders, while we developers are left with the crumbs of their work. These chips are widely recognized around the world and probably have some of the best engineers in the industry behind them, and yet what we actually get as developers are slow programs programs with a thousand issues on Linux, badly developed no matter how you look at it.Recently, as Linux programmers, we ran into the problem that CubeProgrammer was using an “hack” I can’t call it anything else where it used an sprintf() on stdin to try to catch all files starting with 0x*. And as soon as this recursive hack stopped working, they blamed Linux for the issue (obviously). All these applications are written in Java and are ridiculously slow, to say the least. An Eclipse-based IDE in 2025 feels like a slap in the face to anyone

Hi! I have a problem. I faced with a problem of USB middleware initialization as a virtual port in CubeMX. After code generation USB init initializes in main.c within of one of task FREERTOS. If I initialize USB only as connectivity USB init replaces right. How to resolve it? Here details is: /* USER CODE END Header_start_task_can_handler */ void start_task_can_handler(void *argument) { /* init code for USB_DEVICE */ MX_USB_DEVICE_Init(); /* USER CODE BEGIN 5 */ app_start_task_can_handler(argument); /* Infinite loop */ for(;;) { osDelay(1); } /* USER CODE END 5 */ }

I would like to use USBX to make a CDC ECM USB device, but in STM32CubeMX, there is no option for CDC ECM or RNDIS in the USBX device class menu: The STM32U545RE on the other hand has options for CDC ECM and RNDIS:  Does STM32U385 support CDC ECM and RNDIS with USBX? If yes, is there something I need to do in order to show CDC ECM as an option in STM32CubeMX? Thanks!

When "Activating clock source" is ticked in RTC module, my MCU doesn't start up if I reset it from RST button. But as soon as I switch the RCC clock source to LSE external 32.768khz crystal and change RTC Clock Mux to LSE, it works without a hitch. Is this by design or there may be a bug?

As mentioned, in order to rule out the interference from the existing project, I created a new blank project. However, it still didn't work. This issue was not present in the previous version 6.13.As shown in the above picture, I have disabled spi1 but still cannot enable the quad spi mode of spi2. Please provide a solution as soon as possible, or please inform me how to roll back my project version to 6.14 or earlier. The project is urgent. Thank you.@

Hello everybody,Problem: I start a new project from scratch in CubeMX version 6.15.0 with STM32H7A3VG. First thing I do is to try to chose OCTOSPI1, Mode Quad-SPI. But it does not work. It's greyed out and I get the message:"Quad SPI disabled: Quad-SPI Mode is already activated".Same thing if I try to chose OCTOSPI2, Mode Quad-SPI.Same thing in STMCubeIDE 1.19.CubeMX version 6.14 allowed me to chose Quad-SPI.Do I miss something ? Is it a bug ?RegardsVmanne

 With the last versione of STM32CUBEIDE Version: 1.19.0 i can't configure the OCTOSPI as QUADSPI    

Hello everyone,I’m working on a project based on the STM32H563RIT, and I’m encountering an issue related to SWV trace when using STM32CubeProgrammer.The firmware runs correctly in both debug mode and standalone. The problem appears only when I connect to the target with STM32CubeProgrammer and attempt to read the SWV trace. When I do so, the MCU freezes during the initialization phase, specifically inside MX_GPDMA1_Init(), where the DMA used for SPI communication is configured.As part of the investigation, I also provided the main.c file, where the full list of initialized peripherals and their initialization order can be seen.For testing, I tried swapping the initialization order between MX_GPDMA1_Init() and MX_SPI2_Init(). With this modification, the firmware starts even while connected to CubeProgrammer, but later freezes when SPI is used in DMA mode, since SPI ends up being initialized before the DMA.My questions are:Why does connecting with CubeProgrammer and reading SWV cause the

We encountered an issue with the STM32L562E-DK board when attempting to clear TZEN by changing the RDP level from 0 to 1 and then back to 0. Using the STM32CubeProgrammer, we can connect to the device via ST-LINK. The memory blank check indicates flash is empty, but the first byte verification fails during download. Additionally, the PC register value is 0xFFFF_FFFE, and the SP register value is 0xFFFF_FFFC.Could you please provide guidance on resolving this issue?

Created a number of files in your glorious new IDE.After opening and closing a few projects in at attempt to find a hidden menu option, I can now see my .project files in finder but attempting to open them with the IDE via double clicking shows NO project data, just a blank space where the explorer should be.Is this a design 'quirk' with an odd workaround where I can recover the file somehow?is it an issue with V2.0.0 which needs resolving?Or is it just an inevitable fact of dealing with this buggy rubbish software?I actually wouldn't have minded MX moving to a different platform if it was designed to move the whole stack away from this eclipse junk, but I don't understand why you'd expend effort making IDE and MX less usable whilst not bothering to do anything about the buggy or just plain weird behaviour of the IDE