Because of the larger real estate in the larger chips, it takes the data longer to move around. That's nothing new.

It's just that Toshiba and ST took their 55 ns 1Mx16 parts off the menu. We get into trouble because decisions are made in meetings. The participants are acutely aware that the parts currently available are 45 ns - 90 ns speed parts, because they're using the 29F400's. What they do, then, is base their decisions on the idea that the pin-compatible parts a generation (4x) later, will also be that fast, which, historically has been the case. Unfortunately, that didn't turn out to be true this time.

Now I'm trying to "fix" things. Unfortunately, the same regulatory requirements apply now that applied before, and I can't scab-on an extra set of parts because they won't be approved, and I can't make new boards because there isn't time to get through all the regulatory steps.

RE