Sometimes without any reproduceable reason my project with STM32F767 does not power up. If this occurs Read Out Protection is set (which I definitely do not have done) and Option Bytes are scrambled: BOOT_ADD0 and BOOT_ADD1 are set to 0xffff. In this case STM32 seems to be 'dead' (does nothing). But I can access it via STLink and reanimate it with a 'complete erase' following reflashing firmware and new setting of the Option Bytes. This is not reliable behaviour for releasing the project so I wonder what can be the cause for this. Vcc is 3.3V stable, coming from a LDO from 5V supply. Capacitors near to MCU Vcc pins are present. Maybe it is important that I am using RTC circuit with connected CR2032 as RTC battery. Any help is welcome - I do not have any idea where to begin investigations.
Many thanks for your reply. VDDA is connected via inductor 47uH/3.7Ri to VDD and Vref via 47R to VDD. All caps are placed and connected properly and do have rommended specifications. So i will try replace inductor and resistor with 0R jumpers and check again...
VSSA is already connected to ground. I will observe behaviour of circuit the next days. Not so easy: sometimes it was functional for hundreds poweroff/on - and then like by chance issue occurs. But of course inner resistant of inductor leads to a difference in increasing/decreasing of voltage at VDDA in comparison to VDD. And 47R to Vref give a much bigger difference. We will see... Rest is working well so I hope it is such a small problem.
Sadly that does not help. All the lines you mentioned are now connected to VDD or VSS directly. Capacitors at VCAP are 2.2uF tantalum. I can force the issue by using bouncing connection to power supply. I have another project running nearly the same firmware that uses STM32F4 - that does not have the issue 
