I'm working with a commercial device which uses a STM32F103RBT6 microcontroller. I need to replace the firmware to customize some functionalities, but before I would like to dump the current one to be able to restore it if it's needed.
I connected properly the board to my ST-Link v2 programmer, the device is powered by the programmer and in fact when I plug it on the USB the device starts correctly. The problem is that I don't manage to connect to the microcontroller using the ST-Link utility, I always get a "can't connect" error (error message), trying with different settings.
I need to reset the chip before? How?
Thanks!!
Please check you have connected the JTAG/SWD signals from ST-Link debugger to correct pins on the target.
Also ensure the board is power.
Then check settings in ST-Link utility, check JTAG/SWD connection, speed and connect under reset which assert the NRST pin on the target in order to allow connection on JTAG/SWD pins.
If this doesn't work then it can be also the case that the MCU is protected against reading by setting an option byte (look for RDP2 in reference manual). If this bit is set then you won't be able to connect.
Related
The dev boards for the ESP32 family of MCU use CP210x (or similar) "FTDI" chips to communicate with the MCU when flashing. CP210x presents a COM port to the host computer which runs esptool.py, a script which implements the Esressif communication protocol. Rather then use a CP210x, I would like to utilize a SAMD21 in its place.
I have managed to receive logs from my ESP32, to a serial monitor on my computer, via the SAMD21, over USB. The SAMD21 connects to the ESP32 via UART, with the standard ESP configuration of 115200 8N1. I can manually enter bootloader mode by holding the correct pins low at boot, and I get the log back confirming the correct bootloader mode.
When I run esptool.py, however, the connection fails, and I get a timeout. Likewise the esptool "monitor" fails to acknowledge the same logs which my terminal emulator easily detects.
What is the communication protocol between the host computer running esptool.py, the CP210x FTDI chip, and the ESP32, and how can I emulate the CP210x with a SAMD21? The definition of "FTDI", "TTL" and "RS232" are all a bit fuzzy, as far as I can determine with research online, so if anyone has experience in this arena, I would be very curious to hear your advice.
The code I have on my SAMD21 is just:
#include <Arduino.h>
void setup()
{
Serial.begin(115200);
Serial1.begin(115200);
}
void loop()
{
if(Serial.available())
{
Serial1.write(Serial.read());
}
if(Serial1.available())
{
Serial.write(Serial1.read());
}
}
Where Serial1 is the UART and Serial is the USB Serial connection.
esptool.py expects a serial port (known as COM port on Windows) to communicate with the ESP32. It doesn't care if the computer has an old-style serial port, is using a USB-to-serial bridge with a proprietary protocol (FTDI, CP210x and the like) or is using the standardized UBS protocol for serial communication (USB CDC ACM). This is left to the operating system and the installed drivers.
However, esptool.py modifies the baud rate and uses the RTS and DTR signals to reset the ESP32. It is also rather susceptible to timing issues with regards to the reset. If you set the ESP32 in boot mode manually, you should be able to get away without these.
However, the most likely cause is that the Arduino CDC implementation does not implement flow control. If esptool.py sends more data than fit into the internal buffer, it is likely discarded, instead of sending a NAK back so the host computer can retry later.
I don't fully understand the SAMD21 Arduino core to verify it. If so, I don't see how you can make it work with an Arduino program. You would need to resort to some other framework for programming the SAMD21.
Update
After more tests, it turns out the USB CDC implementation of the SAMD21 Arduino core correctly implements flow control. So no data will be lost.
In fact, I was able to successfully upload code to an ESP32 module. The setup was:
Arduino code built with PlatformIO. Instead of Serial, I've used SerialUSB as I'm unsure how to control the project settings available in the regular Arduino IDE.
For the ESP32, I've used a ESP32-WROOM-32 module on a minimal board (reset and boot button, 2 pull-up resistors).
I've connected the board via GND, 3.3V, TX, RX directly to the SAMD21 dev board.
I've verified that I can see the ESP32 log output in normal run mode and the "waiting for download" prompt in bootloader mode (after pressing BOOT and RESET).
Arduino has multiple boards where the esp32 is on-board as WiFi adapter. To flash the esp32, there is a tool sketch called SerialNINAPassthrough in examples of the WiFi library, which should be uploaded into the main MCU of the board (SAMD21 on two of the official boards).
The SerialNINAPassthrough sketch handles the DTR and RTS signals sent by the esptool to reset the board into the flashing mode.
I have a device connected to a USB and it periodically sends data to my PC. During the time it is not sending data, the USB cable needs to be removed, else, the device won't work (a fail-safe feature, if the USB cable is connected and attached to pc, the device is in communication mode, if the USB cable is disconnected from pc, it is in stand-alone operation mode).
I was thinking to use the Devcon so I font has to use another hardware, I'll just disable the USB port during an operation mode, then enable it during communication mode. However, the periodic transfer of data can be set to every 5,10,15,30 minutes depending on the settings.
My question is, will it be okay to disable/enable the USB port periodically? Say every 5 minutes? Won't it cause any problem in the long run? Or would it be more efficient for me to use external hardware, a switch to connect/disconnect the USB? Appreciate any advice or thoughts.
If devcon works for you, I don't see why it would cause any problems in the long run. There is no inherent reason why the hardware should get damaged when you run some commands in your software to disable a USB port.
Using external hardware to accomplish the same thing would not be more efficient because you'd have to pay for the hardware and maintain it.
I have an STM32-discovery board and I am trying to program it with not using any cables.In the place where I am doing my internship, they first wanted me to program STM32-discovery with UART. I was able to do this by making the necessary connections and using the Flash loader demo. Now my next task is to add an ESP-07 wifi module on the STM32-discovery board, connect this module to the same network as my computer, and wirelessly program it from my computer. No other device is wanted in between (like Raspberry). I did some research on this topic but couldn't come to a conclusion. What I found; I can remote program by connecting the card to a Raspberry or a device called Codegrip. Is it possible to do this with only an ESP-07 without these devices? I will be glad if you just tell me what should I look for.
Yes, it's possible to reprogram the STM32 flash wirelessly if the STM32 is running a program that supports this capability. When you programmed the STM32 via the UART there was a program running on the STM32 that:
opened the UART port,
received the new program data via the UART (using some protocol),
and then programmed that new data onto the flash.
To do likewise wirelessly, the STM32 will need to be running a program that:
opens the Wi-Fi port,
receives the new program data via Wi-Fi (using some protocol),
and then programs that new data onto the flash.
You may have used the STM32 internal ROM bootloader to reprogram via the UART. And if so then you used the protocol required by that ROM bootloader. But the ROM bootloader probably does not support Wi-Fi. So you'll probably be creating your own bootloader program that can communicate over Wi-Fi. And you might be defining your own protocol for transferring the program data over Wi-Fi. Or maybe you can apply some established protocol such as FTP. Search for examples of bootloaders that support OTA (over the air) firmware updates.
There are two possible solutions.
Write a custom bootloader for the STM32 - the flash is organised with smaller blocks at the start to support that, so you would move your application to higher memory and have the bootloader either jump to the application to load a new application. The bootloader can then access the Wi-Fi module (and other interfaces) to get updates.
Write custom firmware for the ESP0-07 so that it receives and stores the STM32 image, then transfers it to the STM32 using the existing ROM serial bootloader. In this case you need the details of the bootloader protocol, and it would be useful if the ESP-07 had a GPIO connection to the STM32 reset line so that it can invoke the bootloader without a manual reset.
Either way, you need to write software for one or other of the devices.
You can use any standard bootloader and connect the Wireless module like ESP32 (Bluetooth and Wifi), ESP8266 (Bluetooth and Wifi), BT-05 (Bluetooth), HM-10 (Bluetooth), etc.
Then create the android application or web application and update the Firmware or application.
If you don't want to use the Standard Bootloader, you can implement your own bootloader and add this OTA feature to that.
We have added the Tutorials step by step. Please refer to this if you get time. There we have developed the custom bootloader and updated the Firmware.
I have a project based on a Nucleo H743ZI2 board that communicates with a PC through the USB OTG port on the Nucleo.
Everything is working fine as long as, weirdly enough, I have the STlink USB cabled plugged in (which I have been using to work on the firmware, flash, debug, etc).
However, today I was getting ready to do actual tests with this project and used an external power source to power the H7. I initially tried a 3.3V external source which is available from one of the shields, and then tested a lab PSU at both 3.3V. Thinking it could be PSU related I also tried external 5V with the same results.
Basically, the device is not recognized by the computer operating system, unless the STlink USB is plugged in.
I can't really put my finger on this one; I initially thought it was a PSU related issue (that nucleo board has some restrictions when it comes to using external power sources) but at this point I am fairly confident I am failing to initialize something in my code. Not sure what though.
I'd be happy to share with my code, but wouldn't know where the start. I haven't done much as far as the USB OTG is concerned other than modifying the receive callback to process the data.
Any idea what I might be missing here?
Cheers
I made a STM32F042k6, battery powered device, that I would to flash vi DFU protocol.
In order to do so I disconnect the battery, I pull BOOT0 to VCC, and then plug the device in USB3 port of PC (no USB2 available). The device is not recognized and showed
VID:0000 PID:0002 (case I)
However, when I first connect battery, and then connect the device to PC it is recognized without a fuss (case II).
The problem I face is that I would like to initiate jump to bootloader program by command in software - but then I reach exactly the same condition as in case I.
Could you help me to identify what conditions for proper boot I am violating?
EDIT:
The way I have discovered the problem was when I implemented software jump to bootloader. It seemed not to work at all (VCP device dissapeared, and erroreous <> device appeard in system PID:0002 VID:0000).
I wanted to enter bootloader by pulling up BOOT0 pin, but I could see that I sucseed only when battery-power cycled. USB power cyccle resulted in the same "failed descriptor" device.
I suspect that when powered on via USB, the booloader has som ecommunication problems and cannot establish proper PID and VID. When powered via battery - this problem is not existing.
Do you have a discovery board?
I read in the STM32F042x4/x6 Reference Manual (DocID025832 Rev 5):
The STM32F042x4/x6 embeds a full-speed USB device peripheral compliant
with the USB specification version 2.0.
(part 3.19, p.27).
In the AN2606 (Rev 33), a table shows the USB configuration which is used by the DFU Bottolader in USB FS (Full-speed):
I assume that the USB in the bootloader is the 2.0, but USB 3.0 must be USB 2.0 compliant. So it is possible to use an USB 3.0 port.
I suggest you use the DfUSe Demo from ST. Its interface consists of a tab where the DFU devices available are displayed when detected. I try it on another STM32x by following these steps:
Open DfuSe Demo software.
Plug BOOT0 to Vdd to put the micro on DFU mode.
Plug the USER USB to your computer port.
The name of your device should appear in “Available DFU Devices” field.
You could try to use directly PA11 and PA12 as well (according to Table 12) instead of the USB port of your board.
EDIT:
Another solution to upgrade your firmware without driving boot0 is to jump directly into the DFU bootloader of ST inside your firmware. Else you could design your own IAP bootloader (In-Application Programming).
I hope that helps.
I faced this same issue with the STM32 Mini F401 and dug into it for several days. Unlike the OP I was unable to find any electrical conditions which reliably got the bootloader to do the right thing. I was able to program my board maybe once every 20 tries, which was really slowing down my process.
In the end I discovered that I can go to the Windows device manager, show hidden devices, and there will be an entry in Universal Serial Bus Devices called STM32 BOOTLOADER. When the issue is happening the board is connected in dfu mode but that device is not marked active in the device manager.
To resolve the issue consistently, you can manually delete the hidden STM32 BOOTLOADER entry, uninstalling the device (no need to delete the driver, just uninstall the device). Then reboot your board in DFU mode and it should connect as expected every time. Unfortunately I need to do this process every time I want to program, but that's far better than the inconsistent/intermittent failures before.