2022-04-11 12:38 PM
I have two identical pre-production PCBs containing fresh, unprogrammed 32F413VGT micros from our distributor. One board will program just fine but CubeProgrammer says "can't erase device!" when I try to program the other board.
After first scanning the PCB for visible solder shorts and open pads, the next thing I verified was that the micro wasn't soldered into the PCB upside down. I tripped last year on this often-discussed snafu caused by the ejector dimple mimicking the pin #1 identification dot.
Other early efforts to diagnose this PCB:
--the redundant corner VDD power pins have the same (correct) voltages on both boards.
--both boards have the same ~1V signal on their V_CAP pins.
--the reset line is the same state on both.
--both unprogrammed boards draw the same IDD current, a value near zero.
Any suggestion re additional things for me to check?
Does my primary 8MHz crystal oscillator have to work in order to program a 32F micro? I think no because an operating board could have an internal RC oscillator, correct?
Should I park a 'scope probe on the SWDIO and SWCLK pins and look for identical wiggles during CubeProgrammer's efforts to program?
Thanks in advance for any help you can offer on this.
Jim
Indianapolis US
2022-04-11 01:56 PM
F4 will run from internal HSI 16 MHz source until code selects otherwise
Boot loader might use PLL and external sources for USB / CAN type usage
VCAPs should be around 1.25V
NRST should be close to VDD
Check if it can read memory and/or option bytes
2022-04-12 03:23 PM
Schematic around MCU would help. At least show/describe the NRST and BOOT0 pin connections.