I understand where you're coming from, and I appreciate the constraints of embedded system design, but this is a pretty extreme scenario. Management should ser

I understand where you're coming from, and I appreciate the constraints of embedded system design, but this is a pretty extreme scenario. Management should seriously consider whether the 0.01c saved on that 8Mb chip is worth the design overhead from very tight constraints. There is most likely a pin-compatible 16Mb chip that would eliminate all the pain. Yes, I know that in high volumes every fraction of a penny counts. But if you frequently find yourself engineering your way out of trivial constraints, you might be doing it wrong.