Monthly Archives: November 2014

Welcome to the Cube

So continuing my adventure with the new STM32Nucleo development board I began playing around with the STM32CubeMX code generator today. STM32CubeMX is a graphical software configuration tool that allows you to generate initialization code using a graphical wizard.

STM32CubeMXWith today’s microcontrollers becoming more and more advanced and offering multiple functionality it can often take a substantial amount of time just getting up and running. It seems most manufacturers these days offer some form of graphical code generation tool. Texas Instruments have their Pinmux utility,  Renesas have their Appilet tool and ST are no exception with their offering for the STM32 range STM32CubeMX.

The tool is fairly intuitive but like any other tool there is always a learning curve. I found few areas where I came slightly unstuck which was not necessarily any fault of the tool but with configuring the resulting project.

Generating and building a demo project

As a quick guide I intend going through creating an example project with STM32CubeMX and running it on the STM32Nucleo development board. I’ll keep the project simple and in keeping with recent projects we will simply toggle the LED. Since FreeRTOS is a middleware option we will use that as well.

  1. Download and install STM32CubeMX from the link here.
  2. Launch STM32CubeMX and you will be greeted with something similar to that show above.
  3. Click “New Project” project. You can select your processor you wish to use. You can use the drop down boxes to narrow your search or simply check the peripherals you require under the “Peripheral Selection” pane and only MCUs supporting those peripherals will be shown.
  4. Since we will be using the ST32NucleoF401RE board then it’s far simpler just to select it from the “Board Selector” tab.STM32CubeMX_edit
  5. The main window will now show a diagram of your chosen MCU along with the supported peripherals. You will notice that some of the pins have already been assigned signals. Example PA5 the GPIO pin LED LD2 is connected. Also PC13 which is connected to the blue push button. Other pins such as PA2 and PA3 are coloured orange. These pins are are used by USART2 which by default has no mode configured. Some peripherals may have symbols next to them indicating there is currently a conflict or issue with them. Hovering over these peripherals should show more information.
  6. Expand “USART2” by clicking on the “+” next to it. Set the mode to “Asynchronous”. Leave flow control disabled.
  7. Expand “RCC”. Set both the High and Low Speed Clocks to “Crystal/Ceramic Resonator”.
  8. Expand “FREERTOS”  and enable it.
  9. Click the “Clock Configuration” tab. The default configuration should be fine. The clock source should be the 16 MHz HSI RC resonator connected through the PLL Source Mux. /M divide by 16, *N multiply by 336, /P divide by 4. The System Clock Mux should be set to PLLCLK resulting in a SSYCLK of 84 Mhz. AHB Prescaler should be divide by 1, APB1 should be set to divide by 2 and APB2 divide by 1. The end result being an FCLK of 84 MHz.
  10. Next click the “Configuration” tab. You should have FREERTOS, RCC, SYS, and USART2 all enabled.
  11. Save the project settings by either clicking the disk icon in the tool bar.
  12. Start the code generator by clicking the gear icon in the tool bar.
  13. Enter a project name and choose a suitable location to save the generated files.
  14. Select the “MDK-ARM” as the desired tool chain.
  15. Click on “OK” start code generation. When complete close STM32CubeMX.
  16. Open Keil uVision and then load your newly generated project.
  17. Select “Select Device for Target…” from the Project menu.
  18. Type “STM32F401RET” into the search box. The only MCU shown appear is the “STM32F401RET”. Select it and press “OK”.
  19. Select “Rebuild all target files” from the Project menu to rebuild the entire project. It should build successfully with no errors and no warnings.

Modifying the demo project to do something

STM32CubeMX generates quite a few files. The tool embeds a series specific software platform which includes an STM32 abstraction layer to provide an API for accessing the STM32 peripherals. This is all added to the project when created. Again this appears to be something most micro controller manufacturers are doing these days.

So getting back to the code. Referring to the main() function in main.c the functions starts with a number of initialisation functions, the creation of a new thread “StartThread”, followed by a call to start the scheduler (FreeRTOS) and finally an infinite loop. At this point the scheduler is now in control of the program flow.

We only have one thread in this example “StartThread” so let’s modify it to toggle the LED LD2 every 500ms. Make the following changes to “StartThread” function as shown below :-

 static void StartThread(void const * argument)  
 {  
    /* USER CODE BEGIN 5 */  
    /* Infinite loop */  
    for(;;)  
    {  
       HAL_GPIO_TogglePin(GPIOA, GPIO_PIN_5);  
       osDelay(500);  
    }  
    /* USER CODE END 5 */   
 }  

This function now contains all call to HAL_GPIO_TogglePin() which is API hardware call that as its name suggests simply toggles an GPIO pin. In this case pin 5 (GPIO_PIN_5) on general purpose I/O port A (GPIOA). After toggling the pin the function simply waits for 500 ms before repeating the process. Build the target and it should build successfully with no errors and no warnings.

Configuring the debugger

By default ST-Link is not the default debugger for generated project so you will need to configure it before you can successfully execute the project on target.

  1. Connect the STM32Nucleo board to your PC via CN1.
  2. Select “Options For Target” from the Project menu.
  3. Click the “Debug” tab to configure the debugger.target_config1
  4. Select “ST-Link Debugger” as the required debugger.
  5. Click the “Settings” next to the debugger selection.
  6. Select “SW” under the Port setting. The device should now be shown in the SWDIO box.target_config2
  7. Change the “Max Clock” to 4MHz.
  8. Click the “Flash Download” tab.target_config3
  9. Click the “Add” button.
  10. Select “STM32F4xx 512kB Flash” and press “Add”.
  11. OK all open dialogs to return back to the IDE main page.
  12. Select “Start/Stop Debug Session” from the Debug menu.
  13. The project will be download to the STM32Nucleo board and the debugger started. The debugger will then break on the first instruction in the function main().
  14. Press F5 project will begin executing and with any luck LED LD2 should be flashing away.

So there you have it concept to working prototype in next to no time. Well not quite but you can see how STM32CubeMX takes a lot of the initial pain out of configuring your hardware when starting from scratch. The resulting code may be bloated using the HAL and there may not be an abundance of information detailing the use of the HAL but it sure beats trudging through endless pages of the datasheet.

Advertisements

STM32 Nucleo Hello World…Part 2

So after successfully managing to get a simple hello work program up and running on my new STM32 Nucleo development board using mbed I decided I would try some of the other tool chains available. Don’t get me wrong mbed is an awesome platform and offers some really great features but when it comes to development I want to get down and dirty with the micro controller and mbed doesn’t offer that flexibility.

There are a number of other tool chains that support the nucleo boards including IAR Embedded Workbench for ARM from IAR Systems,  Microcontroller Development Kit for ARM from Keil and TrueSTUDIO from Atollic. Not sure about TrueSTUDIO but both EWARM and MDK-ARM are available as free 32K code limited versions. This should be more than enough for modest developments. There are other limitations. I don’t think the limited version of EWARM offers any optimization for instance but I can live with that.

Because I use IAR Embedded Workbench a lot at work (not the ARM version I might add) I decided to try out the Keil MDK-ARM tool chain to see how it compares with the IARs. Below are the steps I took to install it and get the blinky example up and running.

Installing Keil MDK-ARM

  1. Firstly download and install the latest version of MDK-ARM from here. You will need to fill in a few details before downloading but you wont need to apply for a evaluation license.
  2. After the installation has completed launch the newly installed Keil uVision IDE.
  3. Next you need to make sure the middle ware and other packages are up to date. Go to the “Project” menu, select “Manage” and then “Pack Installer”.
  4. Select the “Boards” tab on the right hand side and highlight the “NUCLEO-F401RE” board.
  5. A list of available packs should now appear on the left hand side.
  6. To update the packs click on the install button in the “Action” column of each pack in turn. Ensure the following are up to date. “ARM::CMSIS”, “Keil::MDK-Middleware”, “Keil::STM32F4xx_DFP” and “Keil::STM32NUCLEO_BSP”.
    keil_mdk2
  7. When complete close the pack installer.

Install the STLink USB driver

If you haven’t already done so you will need to download and install the STLink/V2-1 USB drivers from here.

Upgrade STLink

Although not essential its probably a good idea to update the STLink firmware on the STM32 Nucleo board to the latest version by downloading and running the STLink upgrade tool here.

  1. Connect your STM32 Nucleo board to your PC via the USB connector CN1.
  2. Run ST-LinkUpgrade.exe.
  3. Press the “Device Connect” button to connect to your STM32 Nucleo board.
  4. The current firmware version should be displayed and if older than the latest version the message “Upgrade the firmware to VX.XXX.XX” should be shown.stlink_upgrade
  5. Click “Yes>>>>” to perform the upgrade.

Copy the Blinky example

  1. Open the Keil uVision IDE and go to the “Project” menu, select “Manage” and then “Pack Installer”.
  2. Select the “Boards” tab on the right hand side and highlight the “NUCLEO-F401RE” board.
  3. Select the “Examples” tab on the left hand side. The “Blinky” example should appear.
  4. Under the action tab click “Copy” to copy the blinky example project. You will need to provide a path to the destination folder.Blinky
  5. Uncheck the option “Launch uVision” to avoid relaunching the IDE since it is already open.
  6. Press the “OK” button and then close the pack installer.
  7. Open the example project by going to the “Project” menu, select “Open Project…” and navigating to your destination folder. Navigate to the “Blinky” project.
  8. Select and open the Keil uVision project file “Blinky.uvprojx”.
  9. The project will be opened and the “Project” pane on the left hand side should show all of the individual files grouped into folders.Blinky_Project
  10. Go to the “Project” menu and select “Rebuild all targets… ” to build the project. Assuming everything is set up correctly this should build with no errors and no warnings.
  11. Connect your STM32 Nucleo board to your PC via the USB connector CN1.
  12. Go to the “Debug” menu and select “Start/Stop Debug Session”. The generated image will now be programmed to the board.
  13. The debugger will now break at the start of the program (the first instruction in main()  which in this case is initialise the variable num with -1).
  14. Press F5 to start execution. LED “LD2” on the STM32 Nucleo board should start flashing. Using a terminal program connected to the enumerated COM port at 115200 you should also see the phrase “Hello World” being continually output.

So my first outing with the Keil MDK-ARM tool chain has been a success. First impressions of the IDE are very good. Well laid out and reasonably intuitive. However only once you start using a tool in anger do you start to realise its strengths and weaknesses so I guess only time will tell.

STM32 Nucleo Hello World

So managed to grab a bit of time to play with my new STM32 Nucleo development board yesterday. Just as everyone does when they get a new development board I started with the obligatory “Hello World” application. Now because these boards have mbed support and because I had never used mbed before I decided to give it a try.

For those people who have no idea what mbed is. The mbed platform is a free on-line development environment  for the 32 bit ARM Cortex-M range of microcontrollers. Code is written and compiled in this web based IDE. The IDE provides a private workspace with the ability to quickly import and share code with other mbed users. The mbed SDK provides a C/C++ software platform for creating firmware to run on your device. The SDK also provides a library of components that may be used in your own developments.

For anyone who’s interested I’ll quickly go through the steps I took to get the board up and running. Before getting started its probably a good idea to install the ST-Link USB drivers because these will be needed when you first plug in the board.

OK so you have the drivers installed lets get started :-

  1. Firstly you will need to create yourself an mbed developer account at developer.mbed.org/.
  2. Once registered and logged in click on the “Platform” button at the top of the developer page. Select your board, in my case the ST Nucleo F401RE. After selecting the board a detailed description page will open. Simply click “Add to your Compiler” to add this platform to your dashboard.
  3. Next click the “Open mbed Compiler” button to open the online compiler.
  4. Click “Import” to import an example project into your dashboard.
  5. Enter “nucleo” into the search box and click the search button. This should return a number of results.
  6. Highlight the “Nucleo_blink_led” program and click “Import!”. The import dialogue window will open. Leave these settings and press “Import” to continue importing the example program into “My programs”.
  7. To view the code click “main.cpp” to show it in the editor. So as you can see this is pretty noddy stuff. All the program does is turn the LED on for 200 ms and then off for a further second before repeating the process indefinitely.mbed_edit
  8. Click the “Compile” button to build the example program.
  9. The resulting binary image can now be downloaded directly to your device. The program will begin executing immediately.nucleo_drive_edit

So even from this basic example its clear to see mbed offers an extremely fast development path. Design concepts can be realised with minimal effort in next to no time. This example may only flash an LED but the principles are the same.

Not another development board

I fear I am suffering an addiction. An addiction to buying development boards. With ARM appearing to take over the world every manufacturer worth its salt seems to be introducing new ARM based development boards. On a recent spending spree on Farnell a new line of STM32 based boards caught my eye. The STM32 Nucleo range offers a great way to try these powerful micro controllers and comes in numerous flavours. One feature of these boards is that they are mbed compliant allowing for fast development with the on-line development environment. MFG_NUCLEOAnother great feature is that they come fitted with arduino headers so you can use any existing  arduino shields you have laying around. An integrated ST-Link debugger/programmer comes as standard so there is no need for any external tools. Supported development environments for these boards include IAR EWARM, Keil MDK, mbed and GCC based IDEs such as Atollic TrueStudio.

I opted for the STM32F401 one of the high powered variants. The STM32F401 comes fitted with a Cortex-M4 micro controller running at 84 MHz, has 512 kB of flash and bags of I/O including 4 dedicated SPIs. Can’t wait to start playing around with this board stay tuned for more updates.