Solved

Hello,I'm working with the STM32N6570-DK board and using X-CUBE-AI to run quantized neural networks converted from ONNX models.I'm trying to better understand which parts of the model are executed in software vs. hardware, specifically on this board, and what determines that split.My Observations:1. ResNet-8 Model (custom quantized)When I validated a quantized ResNet-8 model, the output during target validation showed:Total number of epochs: 16 of which 2 implemented in softwareepoch ID HW/SW/EC Operation (SW only)epoch 15 -SW- ( Softmax )epoch 16 -SW- ( DequantizeLinear )So, almost everything was offloaded to hardware, except for the final softmax and dequantization — which I expected.2. Simple Model Quantized via ST Developer Cloud However, when I tested a very simple model (shown above), quantized through the ST Developer Cloud, I saw this:Total number of epochs: 5 of which 3 implemented in softwareepoch ID HW/SW/EC Operation (SW only)epoch 2 -SW- ( DequantizeLinear )epoch 3 -S

hi all, I am using STM32F733xx and i want to read the Vout i will be monitoring is between 1-5v, to read that i have used analogue pins PG13, PH3 ,PH2,PG14, PE2,PE3, PE6, PE4,PE5,PF1,PF0,PF2I already have the PCB so i cant change the design can anyone suggest any solution to this?Looking for some firmware based solution to this problem. Thanks in advance,Regards,Sal

Split from EEPROM Emulation: meaning of the VirtAddVarTab as this is a new question.The eeprom emulation in my simple trial sofware is working fine. The code is quite small.I tried to use this software inside a more complex program, larger than 64K, and the system hangs.From what I have understood only the first 16K are available for program with the suggested setup.It is possile to have a setup for eeprom that allows to use more flash program memory?

I'm developing a UI on STM32H750 using TouchGFX with a 1024x768 display. I'm experiencing significant performance issues with touch responsiveness when displaying graphs.My UI Structure:Base screen (screen1) with a menu buttonWhen menu button is pressed, a menu container is displayed on topThe menu container has a graph button which opens a graph container that overlays on topThe graph display uses a CacheableContainer with multiple static graphsThe Problem:When the graph is displayed, touch response becomes extremely slow. I've used CacheableContainer which helped slightly, but issues persist.In the simulator, I discovered that during every tick, the entire screen (full rect 0,0,1024,768) is being redrawn unnecessarily. The CPU usage spikes when the graph is displayed.Here's the call stack when debugging:#0 0x005d3da7 in touchgfx::CacheableContainer::setupDrawChain(touchgfx::Rect const&, touchgfx::Drawable**) ()#1 0x005d3aa0 in touchgfx::Container::setupDrawChain(touchgfx::Rect co

I have a scroll wheel languageSelectorWheel with several items. Every item is a container languageSelectorItemOn the container I have a toggle button. When I click any toggle button, I want to send a trigger on the view of the scroll wheel to execute a function. On the toggle button I have added an interaction onClick via TouchGFX that emits the trigger languageChanged. The trigger is set on the languageSelectorItem containerOn the view where the languageSelectorWheel is added, I have set the callbackvoid languageSettingsView::languageChanged() Unfortunatly, it seems it is not called when i toggle the button (settings a breakpoint on the line that call emitLanguageChangedCallback(); in languageSelectorBase where void languageSelectorBase::buttonCallbackHandler(const touchgfx::AbstractButton& src) is defined enters) My thought is that because the trigger is set on the container, it won't be passed to the view

We are currently seeking layout recommendations for the ST25R3916 and ST25R3918 chips.On the Transmitter Side:ST25R3916B → Power Amplifier → Matching Circuit → AntennaOn the Reciever Side:Antenna → Matching Circuit →Power Harvester → ST25R3918The implemented power amplifier is a Class E amplifier configured in push-pull mode  (We intend to transmit power along with NFC communication)The matching circuit has been designed using the ST25R Antenna Matching Tool.We are using a differential configuration, and as such, the matching elements have been placed symmetrically on both arms.A provision for a balun has been included between the power transmitter and the matching network.We have the following queries:The ST25R Antenna Matching Tool provides both single-ended and differential matching options. For the single-ended configuration, we understand that the PCB trace impedance should be matched to 50 ohms, whereas for the differential configuration, the PCB traces should foll

Hi,we try to load a small program to RAM area starting from address 0x20008000 with STMCubeIDE and USB DFU and select start address 0x20008000 ( In debugger we can see that program is load to 0x20008000).The download is successful, even so setting start address is successful, but the start of program is not successful.If we download and start the same program with ozone debugger there is no issue. It runs as expected.Used controller: STM32L45x/L46x, MCU, Flash size 512.Is there any idea what I can check?Best Regards,Claus

Hi everyone,I am working on LoRa point-to-point communication using the STM32WLE5CCU6 and trying to run the PingPong example. However, I haven't been successful so far.I would appreciate any guidance on whether I need to modify any pin configurations or make changes to the PingPong example before uploading it to the MCU. Additionally, I am using this IC in the RAK3172 module. Has anyone here developed their own circuit for it?Any help or suggestions would be greatly appreciated!Thank you.

The evaluation board, STM32H743I-EVAL, is used. Where can I find the sample code?I would like the source code.

..

Hello everyone,I'm experimenting PCROP activation on stm32h743 following this ST-tutorial: https://youtu.be/7Wr90XFA8_Q?si=tDpqUtMdin9Oyn-PThe activation works fine but i cannot disable PCROP using CubeProgrammer.I did a RDP1 to RDP0 regression with DMWP1 checked but the PCROP is still activeAfter RDP regression i can se that the unprotected flash are is erase:But not the protected one:I can also not change the option bytes anymore, even RDP0 to RDP1 does not workHow can i recover from this situation ?@Jocelyn RICARD, @Fred 

Hello,I would like to confirm the correct procedure for measuring antenna impedance using a NanoVNA. The chip under test is the ST25R3916.I came across the post titled "How to Measure Antenna Parameters According to AN4974 with NanoVNA," which was very helpful in understanding the general procedure. However, I would like to request a few additional details to clarify certain aspects of the measurement:Following is my differential antenna circuit.Since the NanoVNA supports only single-ended measurements, I am considering the following approach:I plan to remove the voltage divider blocks (highlighted in red) and disconnect the ground at the center point marked with “?”. This way, the VNA will only see the differential network, without any path to ground.Could you please confirm if this approach for impedance measurement is correct?If not, I would appreciate your guidance on how to properly measure the impedance of this differential circuit using a NanoVNA. RegardsSuraj

I'm working on a bare-metal project using an STM32F407VET6 (generic “black board” from China - https://stm32-base.org/boards/STM32F407VET6-STM32-F4VE-V2.0.html). I'm trying to trigger ADC1 conversions using TIM2_TRGO (timer trigger output), but the ADC only works when using SWSTART. I've done extensive testing and debugging to isolate the problem. Here's what I found:What works:ADC1 works with SWSTART: I can read from channel 1 (PA1) manually and see the expected values.LED debug: PD12 lights up if the ADC reads a value above 1000.TIM2 generates update events: Confirmed by enabling the TIM2 UIF interrupt and toggling PD12 on each update.Trigger configuration:EXTSEL = 4 → TIM2_TRGOEXTEN = 1 → rising edgeForced trigger via TIM2->EGR |= TIM_EGR_UG does not trigger the ADC.VREF+ is connected to 3.3V, and VREF- to GND, with a 100nF decoupling cap.TIM2_TRGO is confirmed to work via IRQ debug (PD12 blinks).DMA is configured (when enabled), but never triggers → PD12 never toggles. Debu

Hi,Thanks in advance for your attention.I'm wondering if any examples are available about transmit and receive raw ethernet frames.The examples in STM32CubeH5 use the TCP/IP stack NetXDuo and transmit packets built on the upper layers of the stack and pass received packets to the upper layers.How could I transmit/receive single raw ethernet frames to handle a custom layer 2 protocol?Comparing with what is done by NetXDuo:In transmission (as in _nx_driver_hardware_packet_send) I could build the ETH_BufferTypeDef with the next pointer to NULL and send a single buffer?In reception (as in _nx_driver_hardware_packet_received) in HAL_ETH_ReadData, the HAL_ETH_RxLinkCallback must be implemented to chain all received packets from the Rx Descriptors? Is not possible to receive and process a single ethernet frame? It's confusing to me... And what is the purpose of HAL_ETH_RxAllocateCallback, should it be implemented?An alternative could be handling a custom layer 2 protocol in Net

Hello, I made a custom board using the STM32C031K4T6. I have tried several times now to connect to the chip using STM32 Cube IDE and the STM32 Programmer, both didn't work. The ST-Link gets recognized. I have tried different combinations regarding the Boot0 pin and the reset pin but I can't get it to work. I get the error message that no target was found. I also attached the schematic for the board.    

I want to add the STM32 sequencer utility (https://wiki.st.com/stm32mcu/wiki/Utility:Sequencer) into my STM32L072 project.I have downloaded the STM32Cube_FW_L0_V1.12.3 package for STM32L0 but there is no "sequencer" folder for the utility. Besides, there is no option in STM32CubeMX to add the sequencer utility.Where can I find the STM32 sequencer utility for STM32L0?Thanks! 

Custom board will not run without debugger.  With debugger removed and cycling power I have manually touch NRST to ground to get it to run.The NRST is only connected to SWD pin, but when debugger is removed it is not connected.  Theis is no ST-LINK on the board.  The NRST does measure 3.3 volts after power up.I do use BOOT0 pin as an output.  It is configured low with pull down in .IOC  (could this be issue?)Debugger is set to hardware reset, but I have tried all resets. Debugger is set to break at main (not sure if it is trying to break at main without debugger attached?)  I tried to not break at main and will not run with or without debugger.I also put HAL_DELAY(3000); after Hal_init(),  no difference.Here is how my clocks are set: Any suggestions on why it will not run?

Split from https://community.st.com/t5/stm32-mcus-motor-control/use-of-b-g431b-esc1-to-implement-my-own-algorithm/td-p/794170 as that is answered, and this is a new question.I needed to confirm this before making use of it.In the same thread, I have seen that these boards, although an STM32G431 is programmed, also have an STM32F103 on the daughtboard:https://www.st.com/en/evaluation-tools/b-g431b-esc1.html#featured_resources-0What is its function on the board?

I have read about using these boards in conjunction with firmwares configured in Xcube-MC.However, is it possible to implement these field control strategies or similar in proprietary code?

Hi, I am trying to send an integer from the PC (using Putty serial program) to the Nucleo board via USART interrupt but the board doesn't acquire the number correctly. In the code written here this happens: if I send any kind of number to the board like: 100, 200, 500, 700, 900, a different number arrive to the board and I can't see any changing in the funtcion:__HAL_TIM_SET_COMPARE(&htim1, TIM_CHANNEL_1, num);Seems that the value that I am passing by using a char message and then converted in uint8_t format to the function is converted in a bad number. In fact on the channel TM1_CH1 where I connect one channel of my digital scope, any kind of three digit number I enter in the serial always has the same effect. Any help? Attached some picture:/* USER CODE BEGIN Header */ /** ****************************************************************************** * @file : main.c * @brief : Main program body ****************************************************************************

Dear STMicroelectronics Team & Community,I am writing to report an apparent error in the STM32H72x/73x/74x/75x System Architecture Overview Application Note (AN4891) , specifically on page 28 , regarding the Flexible Memory Controller (FMC) section.The document states:"For STM32H72x/73x, only SDRAM bank 1 can be remapped from 0xD000 0000 to 0x6000 0000 ."However, based on the STM32H7 reference manual and other technical documentation, the default address for SDRAM Bank 1 is 0xC000 0000 , not 0xD000 0000 . Therefore, the correct remapping should be:"For STM32H72x/73x, only SDRAM bank 1 can be remapped from 0xC000 0000 to 0x6000 0000 ."This discrepancy could lead to confusion for developers working with the FMC and SDRAM memory mapping. I kindly request that this error be reviewed and corrected in the application note to ensure accuracy and avoid potential issues during development.Thank you for your attention to this matter. Please let me know if further clarification is needed or i

I'm working on a Motor Control Project built with MCSDK(6.3.1).After initial creation of the CubeMX ioc File by the MC-Workbench, a lot of additional configuration were manually made using CubeMX (~30 additional pins and their config, additional timers, etc.).Then the Source Code was generated.But when i now need to change parameters/configuration in the MC-Workbench and re-generate the ioc File, all the previously manually added config is lost and must be redone again which is a lot of work and prone to errors which are hard to detect. Is this the intended workflow? Maybe i'm doing something wrong?The workflow shown here: https://www.st.com/content/st_com/en/ecosystems/stm32-motor-control-ecosystem.htmlshows a feedback cycle starting again at step 4 but i need to cycle to step 3 without loosing information in the ioc. In the meantime i use a clean ioc without modifications and look at the diff made by changes by the MC-Workbench.Then i manually apply the diff in the ioc to t

Hi Team,  Could you please share the detailed function of INT and SYNC in VL53L8CXV0GC/1?  We would probably use one sensor or two. If case of one sensor SYNC input doesn't add any value to the design. If two sensor used they'll probably share the proximity data(Just confirms if the object present or not) will it a have significance there.    Regards,  Mukesh

Hi,how do I start a new project in VS Code (using ST's STM32 extension or other ways) from an existing .IOC file?I am asking because for writing new code for some new PCBs, I have the .IOC file which I configured when I made the PCBs and I want to reuse because it has many settings that I don't want to copy manually. Alternatively, is there a way to import an existing .IOC file data/settings into a new .IOC file (obviously making sure they both use the same micro part number). Thank you

Good dayI am using a NUCLEO-STMG0B1RE board and i am trying to handle a 4x3 matrix keypad inputs using interrupts to switch menus on an OLED screen. The keypad inputs worked when I used continuous polling for keypad inputs. The OLED needs to continuously update (at least once per second) with data from ADC. This also worked when I used continuous polling, but the keypad inputs was a bit sluggish and I am scared it will slow my processor down for future use.With the current code I am trying to do the following:when key pressed, corresponding column goes highinterrupt triggeredRising callback sets KeypadtriggerDetected = 1 and stores the columnKeypadtriggerDetected activates if statement in while(1) which pulls all the rows low, then sets them high one at a time.When the key pressed is identified from keymap, that key gets processed with updates the state the OLED menu should be inthat state then gets passed to displayMenuOLED to display the corresponding menuWith the current code I