STM32F4107VG6 silicon rev. 1 - LSE don't start at low temp

[olg]

Posted on December 01, 2015 at 17:27

Hi !

I'm just wondering if anybody have seen similar behaviour with LSE on revision ''1'' silicon of  STM32F4107VG6.LSE starts perfectly at room temp but fails to start completely at temps below -15 celsius.Load capacitors and oscillator selection was  done strictly according to application note.We started experience this behaviour on silicon revision ''1'' chips. On previous silicon revision ''Z'' everything was ok.BR,Oleg

[gkennerknecht]

Posted on December 01, 2015 at 17:32 Hi, have a look at the BootPins. We have a similar Problemn with STM32F105VCT6. In case of low Temperatures (below -15°C) the Micro doesn´t ''start''. But it´s  a wrong Interpretation. We missed to tie the Boot Pins to gnd. After pinning the Pins to gnd, Startup from Flash works at temperatures down to -40°C. We think that the Boot Pins are floating at low temperatures and in some cases the Startup jumps to SRAM. Bye

[Tesla DeLorean]

Posted on December 01, 2015 at 19:12

The STM32 parts are very sensitive to the characteristics/gain of the LSE component/circuit implementation. It can result in very slow starting, or non-starting of the crystal, but this should be demonstrably different from it not executing any code. ie you should be able to output diagnostics via a USART or GPIO/LED, to identify it is running normally, and then attempting to start the LSE. Neither crystal is started automatically, which makes signs-of-life testing of free standing STM32 quite difficult.