Felix,

Your decode is fine i do have some comments;

1) You will still be requiring power for the 8255's when an external "code fetch" occurs so i would suggest you input PSEN to the '138 #6 to avoid this.

2) The only protection from device contention is "well behaved" programatic execution. We all play fast and loose sometimes but I would suggest a full decode which would guarantee device exclusivity. Especially if the operator was in anyway unsure or inexperienced.

3) Quite often (depending upon your requirement) it is easier to have you devices and memory in memory concurrently. I have found 32K windows into larger RAMS is a good granularity for RAM access and allows I/O-Memory transfers to occur using P2/R0 and DPTR tandem transfers (if you have multiple DP's all the better).

Now would i have chosen to implement with two '138's? No, again the two 138's were a "given".

regards,
p



regards,
p