I've been recently having to do that a lot. Troubles me... I prefer
fully clocked events.
But system "architects" (what a laugh ;-) leave me with undefined
states I have to generate.
I usually...
Set D-Flop with the only "edge" they give me
Delay(s) operate other needed event markers
Final Delay sets R-S Flop
Which sends the RESET
AND a bunch of concluding conditions before RESET clears
Somewhat like the PFD's I've posted here and on my website.
...Jim Thompson
An ecl flop, especially a modern high-gain one like EL or EP, resets
itself nicely. An r-l-c is a good feedback delay, because it gives the
reset a relatively sharp edge and recovers quickly compared to an rc.
A string of gates is a pretty good reset delay inside an ic, which
would be inside an FPGA to us. One trick is to use a string of AND
gates,
in------+--------
| and----------
+-------- and-----------
in---- and---------- out
in-------
which gives a bunch of delay but recovers very quickly.
We sometimes design around FPGA internal cell delays, and we've been
lucky so far.
Sometimes async logic is unavoidable.
John