G431 - Unable to program OB on MCU - first time flash

[mechatron]

Among 20-odd PCBs populated with STM32G431CBUx controller, I am facing issue in programming one of them. All the others works fine, with no issues what so ever.

The RDP is set to 0xFF, and the BOOT_LOCK bit was set. I searched through the community posts and tried the following methods.

1. Set the BOOT0 pin High and enter through “Under Reset” mode → try to reprogram RDP to 0xAA.
2. Try resetting through NRST and then reprogram the option bytes

15:42:24 : ST-LINK SN  : 0670FF564948897767133441
15:42:24 : ST-LINK FW  : V2J33M25
15:42:24 : Board       : NUCLEO-F303RE
15:42:24 : Voltage     : 3.25V
15:42:25 : SWD freq    : 4000 KHz
15:42:25 : Connect mode: Under Reset
15:42:25 : Reset mode  : Hardware reset
15:42:25 : Device ID   : 0x468
15:42:25 : Revision ID : Rev X
15:42:25 : Debug in Low Power mode is not supported for this device.
15:42:25 : UPLOADING OPTION BYTES DATA ...
15:42:25 :   Bank          : 0x00
15:42:25 :   Address       : 0x40022020 
15:42:25 :   Size          : 84 Bytes
15:42:25 : UPLOADING ...
15:42:25 :   Size          : 1024 Bytes
15:42:25 :   Address       : 0x8000000
15:42:25 : Read progress:
15:42:25 : Error: Data read failed
15:43:13 : Option byte command : -ob  BOR_LEV=1  BOOT_LOCK=0  nSWBOOT0=0 
15:43:13 : qCmd : -ob  BOR_LEV=1  BOOT_LOCK=0  nSWBOOT0=0 
15:43:13 : PROGRAMMING OPTION BYTES AREA ...
15:43:13 :   Bank          : 0x00
15:43:13 :   Address       : 0x40022020 
15:43:13 :   Size          : 84 Bytes
15:43:14 : Reconnecting...
15:43:15 : Reconnected !
15:43:15 : UPLOADING OPTION BYTES DATA ...
15:43:15 :   Bank          : 0x00
15:43:15 :   Address       : 0x40022020 
15:43:15 :   Size          : 84 Bytes
15:43:15 : OPTION BYTE PROGRAMMING VERIFICATION:
15:43:15 : Error: Expected value for Option Byte "BOOT_LOCK": 0x0, found: 0x1
15:43:15 : Error: Option Byte Programming failed Or modified by application after OB_LAUNCH
15:43:15 : Time elapsed during option Bytes configuration: 00:00:01.184

When I tried to read the flash, I expected 0xFF. However, in few areas, I was reading different values.

Only once was I able to successfully program the Option bytes and then flash the intended .hex file but the core was locked when I tried to reset the MCU (power cycle).

And sometimes I get the following error. 

 

Does this mean the MCU flash is corrupt? How such issue can happen?

 

The PCB is powered through a stable and clean power supply (5V), with an on-board LDO converting it to 3.3V. The same HW architecture is also followed in multiple other PCBs with no such errors encountered. 

 

Can improper handling (ESD event during such handling) can result in such a behaviour?

Any leads will be appreciated.

 

[TDK]

If only 1 board out of 20 work, it’s typically a hardware problem. Perhaps insufficient/intermittent power, or a missing connection, or missing caps, etc.

I don’t think ESD would cause these symptoms.

The RDP is set to 0xFF, and the BOOT_LOCK bit was set. I searched through the community posts and tried the following methods.

Typically indicates the option bytes were not read correctly, rather than them actually being 0xFF.

 

Can you program the flash?

 

If it were my board, I would reflow the MCU and relevant components.

[mechatron]

The hardware checks out. I did not find any issue with power supply or any missing connections etc. Everything apart from the MCU checks out.

 

I cannot program the flash. I am getting read failure, indicating RDP being programmed to 0xAA. I could not reprogram the RDP as well.

 

[TDK]

I cannot program the flash. 

So the problem is not specific to the OB, but a general problem with programming flash in general.

The hardware checks out.

I would recheck assumptions here. Look at power draw when idle and when holding NRST and compare with good hardware. Something is different. Everything here indicates an issue with hardware. If it is an issue in hardware, changing software settings isn’t going to fix it.

[MM..1]

In your log is Nucleo F303, by this i mean you use it to program external G431.
Primary check you disconnect right onboard MCU from STLink lines
Secondary check your wires GND SWD CLK NRST , no power with this your board….
And of course G431 pcb require right power on MCU , otherwise as you see all is unstable.

[mechatron]

The PCB has its own dedicated power supply unit. The SWD, CLK and GND are the only signals shared by nucleo and the MCU. The same setup worked for 19x such PCBs and numerous other F3, G4 MCUs I programmed so far. No such error encountered.

[MM..1]

Then simply unsolder G431 on failed board and solder back or replace … problem isnt STLink
except you dont connect NRST required for under reset mode, but this isnt your issue