HAL_Delay blocks code

[QuirinS]

Hello, I have a custom board wich is based on the evspin32g4-dual Board. I generated the code with motorcontrol work bench and could upload it with a j-link. I added an uart output in the startup and noticed that it kept printing it, so the board was constantly resetting. When I stepped through the code the problem was in HAL_Delay, as the HAL_GetTick() function always returned the same value, thus not leaving the function. Could there be an error with the external Crystal? I am using a ECS-240-8-37-CKM-TR with 8pF caps.

Thanks!

[Sarra.S]

Hello @QuirinS, 

The HAL_GetTick() function relies on the SysTick timer, which in turn depends on a stable clock source, so yes, probably it's an issue with the crystal. 

You can refer to AN2867 for oscillator design guidelines.

[LCE]

I would try to verify this by building a version using the internal oscillator (HSI).

Check the PLL settings to get the same system clock rate as with the HSE.

[QuirinS]

Thanks for your reply, I just checked the project and HSI is already in use. 

 

[TDK]

If the code is running, it's not likely a crystal problem.

Probably you're calling HAL_Delay in an interrupt context, or with interrupts disabled.

Set SysTick priority to 0 and everything else above that.

[QuirinS]

I was able to step over the delay and ran into a breakpoint in SysTick_Handler, so the reset is caused by something else.

 

[TDK]

 

If it's resetting, look at RCC->CSR after startup to determine why.

[LCE]

Well, then "run a little further"! ;) 

[QuirinS]

I checked the reset cause and the RCC_FLAG_PINRST was set, i disabled the reset pin by setting the nrst_mode = 2 with cube programmer but its still the same problem

[Tesla DeLorean]

Do you start a watch dog timer? Have one on your board?

Otherwise constantly resetting is more likely a hardware/board level issue.

Perhaps scope NRST/NRESET pin check if it is periodic, and what that period is.

Check supply isn't browning out

Check VCAP pins if present, should have around 4u7 F total on them, and 1.25 V

The Systick is normally enabled/started in HAL_Init() code.

[QuirinS]

The Buck converter was the problem, thus causing a brownout