Standalone application doesn't run unless NRST is forced to 0V

[Hicham Chaabane]

Posted on April 30, 2018 at 13:32

Hi guys,

I'm using the  

STM32L475RCT

 in a custom board. When using the JLink everything is working fine but when I discounnect the Jlink and power cycle the board the application doesn't work at all. Each time I have to force NRST to 0 and release it in ordre to make the application works.

I have 100nF capacitor between the NRST pin and GND as recommended in ST  datasheet(

6.3.15

).

I checked the voltage on the NRST pin and it seems correct : it goes from 0.358V to 3.1V  in ~4ms at board power up which follow the required low level voltage ( VIL(NRST)  =0.3?VDDIOx = 0.984V) and high level reset voltage  ( VIH(NRST) = 0.7?VDDIOx = 2.29V).

Any thoughts on where the problem might be ?

#stpm32-reset #reset #stm32l4 #nrst #reset-problem

[Tesla DeLorean]

Posted on April 30, 2018 at 14:31

I would look at what you are doing with the supply pins, especially VDDA, and if you have multiple supplies. Check initial rise times on a scope with respect to NRST.

Consider using a POR chip like the MCP120 to threshold on the slowest supply.

[Hicham Chaabane]

Posted on April 30, 2018 at 15:20

Thank you for your response,

I have a single 3.3V powering all the VDD and VDDA pins with 100nF capacitor for each pin.

I forget to mention that I checked both the V3.3 and NRST rise time and they were fine. here is a screenshot

I have the same circuit working with no issues on the STM32F417 !

[waclawek.jan]

Posted on April 30, 2018 at 15:41

Double-check if all the ground (including analog) and supply pins are connected - this includes checking for bad solder joints.

JW

[Tesla DeLorean]

Posted on April 30, 2018 at 15:42

>>I have the same circuit working with no issues on the STM32F417 !

Well not work here, so you'll need to change something.

I would suggest a 10K pull-up on the NRST pin, and a 10K pull-down on BOOT0

[Hicham Chaabane]

Posted on April 30, 2018 at 17:20

Thx,  I did and circuit seems okey. I have the same issue with the 4 prototype boards !

[Hicham Chaabane]

Posted on April 30, 2018 at 17:27

I have tried 10K pull up on the NRST and connecting the BOOT0 to GND (I do that on other boards) and still getting the same behavior!

Wonder If a delay between the rising of the V3.3 and NRST is necessary ? I doubled checked the curves and got the below figure :

Blue : NRST

red : 3.3V

When I force the NRST to 0 and turn on the power supply and just after I release manually  the NRST pin, the application run right away!

[Tesla DeLorean]

Posted on April 30, 2018 at 18:39

Are you sure it is not running, and just stuck somewhere early in your code?

I would try perhaps adding a delay in Reset_Handler, driving a GPIO, or jumping to the System Loader ROM, and seeing if I could get the USART to respond or USB/DFU coming up. Sorry, got very little insight into your design.

[T J]

Posted on May 02, 2018 at 02:53

I use a reset chip.

https://www.digikey.com.au/product-detail/en/microchip-technology/MIC803-29D4VM3-TR/576-3806-1-ND/2567031

 

2.93V before 1.2Second timer then it releases Reset.

never have a problem with Reset again.

[Hicham Chaabane]

Posted on May 02, 2018 at 10:50

Clive One wrote:

Are you sure it is not running, and just stuck somewhere early in your code?

I would try perhaps adding a delay in Reset_Handler, driving a GPIO, or jumping to the System Loader ROM, and seeing if I could get the USART to respond or USB/DFU coming up. Sorry, got very little insight into your design.

I tried to turn on an LED first thing in main() but it didn't work.

I introduced a delay aswell (in main) but still not working.

Are you interested in checking the design ?