From: Jim Brain (brain_at_jbrain.com)
Date: 2007-12-11 09:10:23
ruud.baltissen@abp.nl wrote:
>> I thought I might see if I could extend the design to
>> do 16 bit transfers without too many more ICs.
>>
>
> http://www.baltissen.org/images/1541ide.png
>
> And this one works as well! I'll send you the SCH in a PM. But read
> http://www.baltissen.org/newhtm/1541ide8.htm for the details why I don't
> want to use it for the moment.
>
I reread it tonight. Main issues were lack of RAM and having to deal
with 512 byte sectors, but I thought I read somewhere else you felt the
16 bit IDE circuit was to complex.
And, I'm OK with that. But, as one is prone to tinker, I played with
your IDE8 design tonight and I think I may have turned it into a 8/16
bit design that requires only 3 more ICs (2 573s and one '00) The board
design fits on a 2.5' by 3.85" board. Your existing code should work
with it (it does require a small tweak, as you need to shift the
addresses to the IDE channel <<1, as odd addresses are to pick up the
high byte of the IDE, but otherwise, the circuit works just like
yours). Since I gate the 74LS42 line through 1 more level of logic, it
might require ALS or H variants for the '139, '244, and '00 parts.
My idea is to shift all the address lines up by 1, using A0 to determine
whether one wants the low or hi byte. Bytes are transferred to/from IDE
on even address read/write. High byte read/write is latched into an
appropriate '573, as in your 16 bit design, and off addresses pull/push
from the "memory". Of course, if the 573's and 00 are not there, then
no high bytes are read/written.
My thought was that, this way, folks could get started on tinkering with
your code (suitably tweaked) on a board that mimics yours. But, after
getting some of the low level plumbing working, they could pop in a
couple ICs and go for 512 byte transfers. If they got that far, it's
trivial to add your 27256 and 62256 back into the design for plenty of
RAM and EPROM space.
And no, I think your buffers are very useful, and I kept them.
As for boards, I considered running 1 panel of the original design, and
1 of the new design (if someone looks it over). But, I mainly was just
tinkering this evening.
Jim
Message was sent through the cbm-hackers mailing list
Archive generated by hypermail pre-2.1.8.