Flash memory corruption

[Richard Head]

Is it normal for the flash memory to occasionally get corrupted on power-up?
I have found this with an MSP430 device as well as a PIC. If I slow down the dV/dt on power-up it seems to help on the PIC.

Dick

[dannyf]

I have never found it to be an issue.

On the flip side, I have experienced eeprom wearing out, but only in an endurance test environment.

[Richard Head]

Ok, after reading that I'm worried.
I never realised that it was such a common problem.
In my application the micro doesn't get powered down often, luckily.
There are a couple of 2.2uF tantalums across the rails, as well as a 100nF ceramic or two.

Dick

[amyk]

Haven't had this happen on various 8051s, AVRs, and PICs when they're used in an application where the flash is read-only. It probably depends on the specific model, but flash getting corrupted even when it's not being written to (I can understand if a write operation was happening) is a little disturbing.

[dannyf]

Ok, after reading that I'm worried.

I wouldn't worry much about it.

Putting aside design issues and chip issues, flash corruption isn't a big issue, even in high reliability applications (medical for example). RAM corruption is usually where you are more concerned about.

There are cases where an application gets wiped out: most notably in pickit2/3 and some jlink clones. But those are design issues with the particular hardware.

[Richard Head]

Mojo-chan

Thanks for the feedback. I don't know whether brown-out detection has been enabled or not. I'll look into it.
I was also wondering if I could simply connect a 1K resistor between Vcc and GND to ensure that the supply dropped reasonably quickly following a power down condition. Luckily low power consumption is not a factor in this application.

Dick