I think part of the NAND mirroring approach is that you eschew the last 5 tries and do the switch once you reach 5 in some automated fashion in order to expedit

I think part of the NAND mirroring approach is that you eschew the last 5 tries and do the switch once you reach 5 in some automated fashion in order to expedite the process. If you're at the point that you're messing with the hardware, the replacement procedure via some emulation scheme or what not is going to be faster than waiting on the lockouts.