STM32F446RET Core held in reset

[KZhan.6]

Hi ST,

I have a custom board using STM32F446RE. I followed the Nucleo board schematics for my JTAG connection except that I connect the NRST pin to vcc through a pull-up resistor(10k) to prevent inadvertent reset.

The other thing is that I have a TI processor on the board, communicating with the stm32 via uart4.

What has been happening in the past two week is that I can successfully flash my board dozen times and suddenly I get error:

Error in final launch sequence

Failed to execute MI command:

load C:\\GD\\NimbusRobotics\\02_Engineering\\03_Firmware\\acp-exp\\Debug\\acp.elf 

Error message from debugger back end:

Error erasing flash with vFlashErase packet

Error erasing flash with vFlashErase packet

After this, I am not able to connect with the stm32 chip any more even with stlink utility.

I tried to connect under reset and do a full chip erase on STlink utility, remove the pull-up resistor and jumper wire the nrst pin to ground just before full chip erase. This has not allowed me to connect to the stm32.

Can you please advise any potential issue on why stm32 could be locked and any solutions?

Thank you,

[T J]

are you using a BGA package ?

if the processor has fractured a pin, you would need to rework the PCB.

Prototypes have this issue through 1000's of pushing and pulling connectors during development.

Production units get this issue if the temperature cycle is too high. +75 down to -10 will fracture after some time.

are you using lead Free solder ?, also not good for prototypes.. for the same reason.

This is why I use the LQFPs exclusively.

[T J]

or else, your Boot0 pin is floating...

did you try to access through the boot loader ?

[Tesla DeLorean]

Make sure VDDA is connected

Make sure BOOT0 is pulled low

Check the voltages observed on NRESET, and VCAP pins

Does design have/use PDR_ON pin?

[baruch]

The STM32 chip has been replaced, no other changes. It is working fine. We have performed a numerous number of reprogramming cycles. Still unknown the cause of the problem and when it is going to strike again. We're ready to ship the "dead" chip to ST for autopsy if ST is interested, we are!

[Tesla DeLorean]

Talk to the FAE responsible for your account, have them review your circuit.

[T J]

If it is a BGA, you can reflow it, 99% I bet you it will work.

during the process of manufacture, I believe the chips are exposed to 500C

during the 90's :( we use to say,

If you spit at a chip and it spits back, then its too hot.

but they all work happily when they cool down.

Heat change from hot to cold to hot will cause the fracture.

Lead free solder makes it easier to fracture.

how thick is the PCB, you can order a thicker PCB if it is flexing or vibrating..

Sdram, LCD, Jpeg decoding, video streaming, over clocking all cause heat, What were you doing ?

Did you check the processor temperature ? I get temperature of 42C from the Little F0's my H7 isn't quite ready to show that..

[baruch]

How can I find who he is? Pittsburgh,PA

It is NOT BGA

[baruch]

The same issue happened twice and both times during flushing the code using whatever ST eclipse tool is with ST-Link V2.

Myself, personally, did NOT use either and has never seen it ever before with Keil+ULink