From: rhc00@juts.ccc.amdahl.com (Rick Chalfant)
Subject: Re: A Couple of Z80 Questions (Schematic description) (0/1)
Date: Fri, 1 Mar 1996 16:16:08 GMT

In article <4gp2t8$8rp@krel.iea.com>, chubbard@oneworld.owt.com
(Charles W. Hubbard) says:
>
>I am putting together a Z80 system as a learning project.
[snip]
>Question number two.  I would eventually like to put about 8k worth of
>EPROM and 8k (or so) of SRAM on my system.  Can anyone recommend
>specific part numbers for these components or, alternatively, part
>numbers of components to stay away from.  I'll be hunting for these
>chips myself and my plan is just to find some which have READ, WRITE,
>and ENABLE inputs compatible with the Z80 control signals.  Are there
>other parameters I need to keep in mind?
>

Here's a schematic for memory select on a Z80.
It's in .PCX format and about 32K in size.  It's
ZIPed and uuencoded for a final download size of 12K.

To print the file, use Windows Paintbrush.
Set your printer to "landscape" and print
using a scaling factor of 80 or 90 percent.

Circuit description:

In order to use slow EPROMs on a 4-mHz Z80, I had 
to cheat a little.  Normally you set up everything
and then use CS to select the correct device.  That's
too slow to meet the M1 timing for Z80 instruction
fetch.

So we do things backwards:  select the chip with CS
as soon as we see M1 and MREQ, then when/if we get
a RD or WR, we hit OE or WE, respectively, to complete
the operation.

This works fine with NEC and Hitachi 2716 EPROMs, but did 
NOT work with Intel devices.  Don't know why.  I'm a 
software type, not an EE.  The RAMs were Hitachi 6116-3.

If someone wants to explain why this is a terrible
design I won't be offended.  But I won't re-wrap it
either.

I included a 16-pin wire-wrap header on my board because
the 8Kx8 parts were expensive back when I built it.  It
was for a robot and I wasn't sure if 16K would be enough,
so I planned ahead.  The entire program fit in 6K, leaving
10K for RAM, so I never upgraded to the 8Kx8s.

One of the design considerations was low package count.
The other half of the '08 AND was used for PIO INT-ENBL
propagation and the inverter was left over from the
oscillator.  The other half of the '139 was used for
I/O chip select logic.

Cheers,
Rick.C
mhm 15x3

    Source: geocities.com/siliconvalley/peaks/3938

               ( geocities.com/siliconvalley/peaks)                   ( geocities.com/siliconvalley)