Yes, so, the DLL is useful for distributing, deskewing, and multiplying a fixed clock, but not so much so for extracting clock from a given data stream, or, frankly, any sort of typical PLL application.

The family datasheet discusses stopping the clock, or changing its source, (Spartan-II DS001-2 v2.2, pg 23) but provides no "spritual guidance" as to how, exactly, a missing pulse would impact the DLL operation. I have, on occasion, considered "trying it" but have my doubts, as making it work in one instance says nothgin useful about any other instance. Failure is probably more enlightening, but still not reliable.

Without spec's it's hard to determine, or even guess, whether a simulation is reliable enough to bother with.

I might be inclined to use that Spartan-3E or a later -AN part with level shifters to the standard 5-volt logic that my target hardware uses. However, until I know more about these DLL's I can't rely on them, and I'm not going to learn enough from "playing" with 'em to make it worth the time. If only the manaufacturer would provide some details.

RE