Have you tried any of this at all? You do have an oscilloscope and a logic analyzer, don't you? Writing to external hardware via the SPP is extremely slow. Yes, it's faster than you can do it by hand, but you can get nearly a 10x performance increas with EPP (NOT ECP!!!)

The EPP hardware is REALLY difficult. It requires exactly one gate 1/4 of a NAND package, and two passives. Now, don't let me discourage you from trying this with SPP, because you may well manage it just fine, but I'd be concerned about the fact that the HMOS parts weren't static, hence, required a minimum clock pulse width. If you weren't aware of that, then perhaps you should rethink your approach. OTOH, if you've figured out the worst-case timing, and found it adequate, then go ahead. You can, of course, force the MCU to stop for a moment (define that in terms of allowable minima and maxima) which I gather is what you intend to do. I'd suggest that you ensure that you understand all the inherent timing requirements before attempting this, and ensure that you have adequate instrumentation to allow you to monitor the process so you'll be confident it's doing what you intend.




How do you know that it works? Have you examined it with oscilloscope and logic analyzer?


While that may be true of the SRAM, are you sure it applies to the EEPROM? Doesn't it have some page-boundary considerations you have to take into account?


Were I in your position, I'd be much more concerned with functionality than with appearance.



Have you considered that there's a minimum speed at which you can operate this part? You did say you're using the Intel HMOS part, didn't you? If you use a later, CMOS, part you won't have this concern. You can probably do what you want with the HMOS part, but you have to be quick, and the EPP mode goes about 8x -20x as fast as the maximum you can expect from the SPP. That's what I meant when referring to speed.


I think you'll find that this approach is much slower than than what you need.



How much rate-experimentation have you done with the SPP? I've used it some, and have found it quite "challenging" to attain 50 kBytes per second. OTOH, 2 MBps is the "classic" maximum transfer rate on an ISA-bus-based EPP card. If you have a PCI-based port, EPP is easily capable of 8 MBps (Bytes, not bits). It's up to your code and your OS, though. The hardware creates the data transfer handshakes, though, and that eliminates about 60% of the overhead, and the overhead, maintaining the control signals, and keeping track of the processes involved, is about 75% of the code burden, on a time basis.

I'd suggest you spend some time with a parallel port, an assembler with which you're comfortable, and an oscilloscope in order to convince yourself that you can operate the SPP quickly enough to do what you want. Remember, too, that not all the SPP bits are in the correct "sense" at the output. Some are inverted from the way in which you may want/need to use them.

I'd start by writing a simple program to toggle a bit on the SPP. You will, after all, have to produce a clock of some sort with which to effect your data transfers. That will be the rate-determining step for any process you want to control over the SPP.

RE