Associate III

Solved

STM32 won't start without physical reset.

Forum|Forum|5 years ago
January 6, 2021
19 replies
7539 views

Hi,

My mcu is STM32L476RGTx and I experiencing very strange behaviors to make my cpu start.

This mcu is on pcb I designed myself.

My boot0 pin is shorted to GND

And the reset pin pulled up to VCC.

I can program my chip with my st link (updated to the last version) an after each programming my chip is starting up as expected.

But as soon as I tried to make it boot from cold start it won't start. I have to manually press reset button and after that, it run as expected.

It append what ever i use as supply. 3V3 from programmer, 3V3 from local switching supply (LM43602)...

For the moment i do not use the 8mhz quartz on board.

Have you any idea I did not check yet or they're is any parameter that could prevent my mcu to run my program from flash but make it run as soon as I create a true reset event ?

Maybe is it BOD ? But how could I check that ?

anyway, Thanks for your help.

This topic has been closed for replies.

Best answer by dodonny

I put a 100uF capacitor between NRST and GND. Take 5 sc to charge. And boot fine with that !

(take care to the base time on the image below. Screen width take 14sc).

Same connection as previous. Top is VSS and bottom is NRST

EDIT: Removed the 100uF capacitor (was only for testing) an removed the 100nF capacitor between NRST and GND and work fine now...

So i'll let it like that.

That was my 100nF capacitor (between NRST and GND) which is the faulty one. Was here to protect RESET button and debonce NRST signal. But seems too big.

TDK

It could be that it's just stuck during startup somewhere. Maybe HSE failed to initialize in time or similar. Attach a debugger when it's in the stuck state to see where it is.

"If you feel a post has answered your question, please click ""Accept as Solution""."

dodonnyAuthor

Associate III

Hi, first thanks for your reply.

So, installed openOCD, and running under command line. Worked fine, connected to target and telnet cmd prompt.

> stm32l4x.cpu curstate
running

Event if my debug LED on my board is still off.

(For this very simple debug, the first action in my code is to change my led state to on. Witch is perfectly working after hardware reset)

I have to halt my device to get core dump. There it is :

> halt
target halted due to debug-request, current mode: Thread
xPSR: 0x61000000 pc: 0x1fff4956 msp: 0x200030a8
> reg
===== arm v7m registers
(0) r0 (/32): 0x48000000
(1) r1 (/32): 0x00000000
(2) r2 (/32): 0x00010100
(3) r3 (/32): 0x50000b00
(4) r4 (/32): 0x200007d8
(5) r5 (/32): 0x48000400
(6) r6 (/32): 0x40021040
(7) r7 (/32): 0x48000000
(8) r8 (/32): 0x48000800
(9) r9 (/32): 0xe000e010
(10) r10 (/32): 0x40013000
(11) r11 (/32): 0x40003800
(12) r12 (/32): 0x00000000
(13) sp (/32): 0x200030a8
(14) lr (/32): 0x1fff4947
(15) pc (/32): 0x1fff4956
(16) xPSR (/32): 0x61000000
(17) msp (/32): 0x200030a8
(18) psp (/32): 0x00000000
(20) primask (/1): 0x00
(21) basepri (/8): 0x00
(22) faultmask (/1): 0x00
(23) control (/3): 0x04
(24) d0 (/64): 0x0000000000000000
(25) d1 (/64): 0x0000000000000000
(26) d2 (/64): 0x0000000000000000
(27) d3 (/64): 0x0000000000000000
(28) d4 (/64): 0x0000000000000000
(29) d5 (/64): 0x0000000000000000
(30) d6 (/64): 0x0000000000000000
(31) d7 (/64): 0x0000000000000000
(32) d8 (/64): 0x0000000000000000
(33) d9 (/64): 0x0000000000000000
(34) d10 (/64): 0x0000000000000000
(35) d11 (/64): 0x0000000000000000
(36) d12 (/64): 0x0000000000000000
(37) d13 (/64): 0x0000000000000000
(38) d14 (/64): 0x0000000000000000
(39) d15 (/64): 0x0000000000000000
(40) fpscr (/32): 0x02000000
===== Cortex-M DWT registers

after a reset run command the I can get my software work. So I halted after that and get core dump. Here you can see:

> reset run
Unable to match requested speed 500 kHz, using 480 kHz
Unable to match requested speed 500 kHz, using 480 kHz
> halt
target halted due to debug-request, current mode: Thread
xPSR: 0x01000000 pc: 0x08002cae msp: 0x20017fc8
> reg
===== arm v7m registers
(0) r0 (/32): 0x48000400
(1) r1 (/32): 0x00000400
(2) r2 (/32): 0x00040000
(3) r3 (/32): 0x48000400
(4) r4 (/32): 0x00000000
(5) r5 (/32): 0x00000000
(6) r6 (/32): 0x00000000
(7) r7 (/32): 0x20017fc8
(8) r8 (/32): 0x00000000
(9) r9 (/32): 0x00000000
(10) r10 (/32): 0x00000000
(11) r11 (/32): 0x00000000
(12) r12 (/32): 0x00000000
(13) sp (/32): 0x20017fc8
(14) lr (/32): 0x08000f3d
(15) pc (/32): 0x08002cae
(16) xPSR (/32): 0x01000000
(17) msp (/32): 0x20017fc8
(18) psp (/32): 0x00000000
(20) primask (/1): 0x01
(21) basepri (/8): 0x00
(22) faultmask (/1): 0x00
(23) control (/3): 0x00
(24) d0 (/64): 0x0000000000000000
(25) d1 (/64): 0x0000000000000000
(26) d2 (/64): 0x0000000000000000
(27) d3 (/64): 0x0000000000000000
(28) d4 (/64): 0x0000000000000000
(29) d5 (/64): 0x0000000000000000
(30) d6 (/64): 0x0000000000000000
(31) d7 (/64): 0x0000000000000000
(32) d8 (/64): 0x0000000000000000
(33) d9 (/64): 0x0000000000000000
(34) d10 (/64): 0x0000000000000000
(35) d11 (/64): 0x0000000000000000
(36) d12 (/64): 0x0000000000000000
(37) d13 (/64): 0x0000000000000000
(38) d14 (/64): 0x0000000000000000
(39) d15 (/64): 0x0000000000000000
(40) fpscr (/32): 0x00000000
===== Cortex-M DWT registers

waclawek.jan

Super User

> (15) pc (/32): 0x1fff4956

That's in the bootloader.

Consult AN2606 and chapter Boot configuration in RM. They are slightly contradictory.

BFB2 bit in the user option bytes is set how?

> My boot0 pin is shorted to GND

Double-check, measuring directly on the physical pin coming out of the package.

JW

dodonnyAuthor

Associate III

Thanks again !

BFB2 is LOW (see picture).

And double checked, BOOT0 is shorted to GND. Measured value : 0V from GND and -3.3V FROM VCC. (is not floating)

waclawek.jan

Super User

This IMO looks good.

I assume your code is absolute minimal (no system init or library calls, just the usual C memory clear in the startup, and only the LED blinking with loopdelay in main()).

You can try to disconnect your power source an use an external one, to see if it makes any difference. You can try other BOD levels.

Other than that, I am out of ideas, sorry.

JW

dodonnyAuthor

Associate III

First, thanks anyway for your idea. I'll find, it's just a matter of time.

For the moment i'm checking basics. But soon, i'll be run out of it...

dodonnyAuthor

Associate III

Probably nothing to do mith my issue but is it normal that I can't program to the factory reset this option byte ?

> stm32l4x.cpu mdw 0x1FFF7800
0x1fff7800: ffbff8aa
 
> stm32l4x.cpu mww 0x1FFF7800 0xFFEFF8AA
> stm32l4x.cpu mdw 0x1FFF7800
0x1fff7800: ffbff8aa

But this seems normal, cause in the ST-LINK utility windows, some option are grayed. (DUALBANK ?)

dodonnyAuthor

Associate III

And mutliples times I get the BANK A and BANK B locked (even, if i never wanted to, and never check that option). So I had to set PCROP_RDP bit to 1 and switch security level from 0 to 1 to zero again... as explained in error message.