whitequark
Possible additional information about cascade function
Atmel's documentation about the cascade function indicates that cascading happens in groups of 8 macrocells ("up to 40 product terms"). Experimentation with WinCUPL seems to indicate that the cascade chain is not cyclical, so it goes only in one direction, and from low-numbered macrocells to high-numbered macrocells.
Thus, for example, in the ATF1502AS, the cascade chains are:
1 → 2 → 3 → 4 → 5 → 6 → 7 → 8
9 → 10 → 11 → 12 → 13 → 14 → 15
16 → 17 → 18 → 19 → 20 → 21 → 22 → 23
24 → 25 → 26 → 27 → 28 → 29 → 30 → 31
Dunno if you want to include this in the documentation, or leave it out as unproven.
That's how it works, no need to prove anything--the dataseheet actually says that explicitly somewhere, it's just not really clear.
Hmm. At least in the Atmel datasheet, the description is solely limited to these words:
Within a single macrocell, all the product terms can be routed to the OR gate, creating a 5-input AND/OR sum term. With the addition of the CASIN from neighboring macrocells, this can be expanded to as many as 40 product terms with little additional delay.
This should logically be a part of the "functional outline" section, which I've yet to write. Please be patient.
No worries. It just wasn't obvious to me which direction the cascade lines went (they go from lower-numbered MCs up), or whether they are cyclical (they aren't).
Ah, you're right. I'm now not sure whether I read it somewhere or it was just "obvious" to me.