Programming a virgin STM32G0B1

[jlthompson]

I'm trying to program a virgin SYM32G0B1 with the code I've developed and successfully run on the NUCLEO-G0B1RE/NUG0B1RE$AU2 development board. I'm using the STM32CibeProgrammer 2.20.0 software and an FTDI cable (TTL-232RG-VIP-WE with the Vcc line connected to the 3.3 Vdc rail) connected to LPUART1 on port PC0/PC1. I have the BOOT0 signal (PA14) tied to ground.

Having no success with my new hardware, I decided to try the development kit as a sanity check, so changed JP2 to use E5V and connected a 5-volt supply. The dev board runs just fine with that. I connected the FTDI cable to PC0/PC1 on CN10, and shorted PA14 on CN7 to ground. Unfortunately, I also cannot program that MCU. The programmer's log shows:

16:41:26 : RTS low
16:41:26 : DTR Low
16:41:26 : Serial Port COM6 is successfully opened.
16:41:26 : Port configuration: parity = even, baudrate = 115200, data-bit = 8, stop-bit = 1.0, flow-control = off
16:41:30 : Timeout error occured while waiting for acknowledgement.
16:41:30 : Error: Activating device: KO. Please, verify the boot mode configuration and check the serial port configuration. Reset your device then try again...

I've verified the FTDI cable works (short TX to RX and look for echoed characters) and swapped Tx with Rx, just in case that was the culprit.

I'm sure I must have overlooked something, but what?

[TDK]

PC0/PC1 are not pins that are supported by the bootloader. Check AN2606 to see supported pins. PA9/PA10, PA2/PA3, and PC10/PC11 are supported.

https://www.st.com/resource/en/application_note/cd00167594-stm32-microcontroller-system-memory-boot-mode-stmicroelectronics.pdf

In order to boot into the bootloader, BOOT0 should be HIGH on power-on, not low. AN2606 has details on that as well.

[jlthompson]

I just discovered I couldn't use an LPUART port myself, so I switched to PA2/PA3 and tied BOOT0 high, followed by a power cycle. Advanced Serial Port Monitor, in Spy mode, does show some activity, notably one or two 0x7f's.Same for PA9/PA10 and PC10/PC11.

[Tesla DeLorean]

Needs to use EVEN parity

You get ONE shot at sending the 0x7F training pattern, and it should respond with an 0x79

If you don't get that, the pattern has not be accepted correctly, and you must reset and try again.

You can't have any traffic or noise on any other the UART RX pins supported by the loader, per AN2606. So things like GPS/GNSS squawking at start-up will be a problem.

[jlthompson]

I've tried many times with different times between resetting my target and trying to connect with the programmer, but no success. I hope a capture of the serial port will help diagnose the problem, so I've attached both an ASCII view of a failed attempt and its hex view (from the same connection attempt).

Please take a look at the yellow-highlighted areas in both images. A 0x7F is sent, and the response is 0x7F 0x1B, followed by another 0x7F being sent. The programmer's log shows:

17:57:25 : RTS low
17:57:25 : DTR Low
17:57:25 : Serial Port COM6 is successfully opened.
17:57:25 : Port configuration: parity = even, baudrate = 115200, data-bit = 8, stop-bit = 1.0, flow-control = off
17:57:25 : Error: Activating device: KO. Please, verify the boot mode configuration and check the serial port configuration. Reset your device then try again...
17:57:36 : RTS low
17:57:36 : DTR Low
17:57:36 : Serial Port COM6 is successfully opened.
17:57:36 : Port configuration: parity = even, baudrate = 115200, data-bit = 8, stop-bit = 1.0, flow-control = off
17:57:36 : Error: Activating device: KO. Please, verify the boot mode configuration and check the serial port configuration. Reset your device then try again...

The Programmer window look like this:

[Tesla DeLorean]

Not sure what all the other extraneous junk is in the terminal window.

Just use something simple like Real Term in HEX mode to send a simple 0x7F pattern. Don't use STM32 Cube Programmer. You're just trying to test the challenge response sequences. 

You need to use EVEN PARITY.

Like I said the 0x7F should yield an 0x79 response if working properly. 

[jlthompson]

I start by doing a complete chip erase using ST-LINK. I looked to see if I could reset the flash option bits to factory settings, but that option is not available for my device (0x467). The three BOOT bits are all set, I then disconnected ST-LINK and connected to my serial port (COM6) using RealTerm, using EVEN parity, 115200 baud, and 1 stop bit. I have BOOT0 tied high, Rx from my serial port (COM6) goes to PA2, and Tx to PA3. Next, I reset the MCU by momentarily grounding !NRST, then send 0xFE from RealTerm.

No response.