mawa wrote:
What could be an issues is the brown out behavior/protection of the ATMEL processor chip when the 3.3 power is "flickering".
Don't worry about the microcontroller. Atmel's AVR chips are made on a nowadays-ancient 0.35µm process with a ridiculously thick gate oxide. They're pretty hard to fry, all the I/Os have ESD clamps, and Atmel has excellent I/O pad designers. I used to work with their FPGAs (which could be reconfigured
while active at cell granularity by the on-chip AVR, such a cool feature) and those things took a ridiculous amount of abuse. I never managed to break one.
Regarding brownout specifically, you can't damage a CMOS chip that way, at least not a 0.35µm device. In a static-CMOS topology the PFET threshold is relative to the supply rail, so it droops as the supply droops -- as long as you don't have voltage gradients internal to the chip you won't be able to get both the PFET+NFET conducting meaningful currents simultaneously (i.e. shorted) by simply monkeying with the supply voltage.
