Probability of bit flips

[BGhaz.2]

Hello,

Are there any statistics on the probability of bit flips for the RAM on stm32f0x family?

Are there any dependencies based on factors such age of the micro or temperature?

[S.Ma]

2000 prior, some customers implement SW Xtra tricks to prevent chip corruption from ESD gun onto the board. Alpha particle is just another possible corruption source. For example, make checksum signature in ram, checking hw peripheral registerd or overwriting them (never read them like it was a memory register), watchdog, fill empty memory area with known values or sw reset code, check reset rootcause at boot, etc....

So to answer your question would be to FW harden the application, as the whole board affect the sensitivity to corruption. Usually more advanced technology node are more sensitive, hence they have hw checksums in ram such as ECC

[eewiz]

Not just memory. Cosmic ray or ESD may cause any transistor anywhere in the device to flip state.

Sometimes even causing permanent destruction.

[BGhaz.2]

Thanks for the reply. But does ST publishes any data on the probability of bit flips, or permanent failures of RAM? If ST doesn't publish such data, is there any publicly available information for the type of memory in ST micros?

[AScha.3]

look for "space products" , STM seem to have many, but not stm32 :

https://www.st.com/en/space-products.html#overview

[BGhaz.2]

OK, maybe others will find this helpful. I found out that in the FMEDA for the processor, ST gives the probability (in FIT) of both transient and permanent RAM failures.

Thanks for the helpful replies.