The official story [that I heard from Iomega] was that the heads were set and retracted too fast and it caused the heads to break. Wouldn't that cause all kinds of sub-track noise that would scramble the data? Recovery doesn't sound very likely.
I bought the original parallel interface Zip 100 for use at college. Today, I can't even give it away, though it works perfectly.
