About the lack of sense in writing twice because the write pulse isn't long enough on a single cycle I agree completely. However, if your target is students, you need to lay some groundwork, else it seems you're just spouting off some opinion which may or may not be founded in solid theory.

It's not clear to me what you mean by "8 MHz part" or "12 MHz part," and if it's not clear to me, it certainly won't be clear to a student. You see, the notion of an "8 MHz part" with respect to the i8255, means one thing, while "12 MHz part" with respect to an 805x means something quite different.

The i8255 was not designed for use with the 805x series, and the notion of an "8 MHz part" didn't even exist at the time it was developed. The 12 MHz i805x is a 1 MHz part, i.e. it manages to generate a single bus cycle in 1 microsecond. The i8255 doesn't have a clock input, so defining it in terms of a clock may not be particularly helpful. It would be easier to address the O/P's issues if we knew precisely which 8255 version he's using, but that is apparently not available.

Now, we, the slightly more experienced forum participants, could do the O/P's work for him, but he won't learn anything if we do that, nor will we.

Creating a lot of noise about how inappropriate we may find the use of this device or that, probably won't help him, nor will it solve the problem. What has to happen is this.

(1) the O/P needs to attach his oscilloscope to the board set and tell us what the duration of the various chip selects at the point of origin is.

(2) he then needs to measure the duration of those same signals at their destination(s) turns out to be. This may not tell US much, but it should tell HIM a great deal.

(3) he should ensure that the strobes he creates with his logic are of reasonable "shape and size" as one might conclude from the MCU datasheet, taking the intervening logic into consideration.

(4) he should then compare the physically observed signal timing with the datasheets for the SPECIFIC DEVICES he's using and verify that the device requirements are being met, both at the external device and at the MCU, and draw some conclusions.

(5) he should then report the results of his empirical observations to us and and only then can we help him assess the conclusions he has reached from his observations. Without those observations, we can't do much, nor can he.

Note that most of this is up to HIM, and not us. We can't force him to do those things, but without the observations, all we can do is generate lots of speculation and that results in noise, not information.

RE


Moreover, since his problem appears to be at the RAM, some timing data would be useful.