There are times when you need a "transparent" latch. For example, with an 805x, the spec's don't clearly define, for all possible 805x's, when, with respect to the clock or ALE, the addresses become valid.

It's "for sure" that it happens sometime after the rising edge of ALE, but you can't rely on the relationship to the oscillator, and, in case of the on-board crystal oscillator, you can't use the thing safely anyway. Even so, if you don't know in advance which variant you're going to use, you can't be sure on which edge of the input clock you can use to gate with ALE. In some cases, you'll get two edges, only one of which is valid, which, best case, causes lots of noise, and more often causes invalid addresses to be sampled. The obvious thing to use is the ALE, but, since you don't know when during ALE, the addresses will become valid, you can't use the rising edge. Would you use the inverted ALE in such a case?

On the "normal" garden-variety 12-clocker, the ALE goes false at least one clock cycle before the nRD or nWR signals become active. One could register the addresses then, except that they're not present on P0 any longer. I'm not suggesting it can't be done, but I don't think it's as simple as you suggest.

How would you instantiate a latch that works, in the general case, as an address latch for use with 805x's?

RE