ziggy wrote:
> Personally i dont think 'cloning' anything is the right idea, it should
> be something unique, designed from scratch.

If you start from scratch, then it's a lot more work to ever reach even
a modest level of functionality. It would be likely to become a project
that never gets finished.

> I also agree that we will have to make some concessions for the loss of
> 'classic' components over the years, but i would think we should get by
> with a bare minimum of 'modern' parts in this thing.

I agree. Though it is tempting to use "new improved" parts, in may cases
they are also single-sourced and availability is poor. Probably 9 out of
10 new ICs won't survive; their manufacturers will soon give up and drop
them.

I tend to stick with generic multiple-sourced parts that have been
around for 10 years or so. Once they have a broad user base, they tend
to *stay* around much longer. So the odds are good that I will still be
able to get a replacment part 10 years in the future.
--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

>I don't have a website, but would be happy to post the schematic or
>board layout. Anyone have any suggestions on where to put them?

I can offer either/both my yahoo and comcast personal websites.

yahoo :
pros: that site has been around for awhile, and will likely remain
that way.
cons : heavy advertising.

comcast :
pros: Less advertising, the site is (for the moment) 'cleaner'
(as in less cluttered), easy file management (for me!)
cons: It's dependant on my Dad's comcast account. If he
gets a twist in his shorts and cancels his subscription, I
lose my sub-account.

My thoughts about the project in general:

While I would love to dive into an ASIC project, I also
love simplicity, commonality, and ease-of-replacement:
If I release the magic smoke on an ASIC, I'm gonna
be pretty mad. If I smoke a Z80 or '138, I can get over it
much easier and faster.

A lot of machines were built without an ASIC- Kaypros,
the H8, and many others. Lot's of homebrewers out there
manage quite nicely without them as well.

I'm all for using a pre-existing design for a starting point;
but I don't wan't to shy away from rewriting
(maybe even large) portions of the code, if it will serve
our purposes- such as the idea of upgrading the H8's
1701 (? WD) in favor of the 765. As an aside, I have a
765A, and was wondering when I'd get around to using it.

I propose setting some 'baselines' - issues we can agree
on now - to get the ball rolling.

proposals -
Z80 @ 4MHz : is there a reason to be slower or faster?
Disk Controller: 765A (cuz I have one ;o) )
True I/O : as opposed to memory-mapped. [1]
Rsvd Ports : for options like color/grafix/sound [2]
CP/M : (was there ever any doubt?) 2.2 for simplicity?
Drives: 3.5" cuz they're thick in these parts.

[1] This is a wonderful feature of the Z80, that requires
only two to four gates to implement. And I am a byte
hog; the thought of sacrificing RAM addy's for hardware
just makes me itch.

[2] In this way I am kind of signaling that I am leaning
towards a tty design. But I like the thought of leaving
'room' for people to extend the base design in their own
way- implement their own customized peripherals and
the code to use them.

After writing all of this, I just realized the H8 was an 8080
based design. Which actually doesn't affect much as far as
my proposals are concerned, I don't think. The 8080 only has
512 I/O ports, which is still probably more than I'll *ever* use,
but the Z80 did some cool things with diffferent out(~~) instr's.
I guess the both biphase clock and the bivoltage supply of the 8080
scare me.

Lemme know whatcha think.

-Tarkin

On 15 Apr 2006 20:37:52 -0700, "Tarkin" wrote:

>serve
>our purposes- such as the idea of upgrading the H8's
>1701 (? WD) in favor of the 765

Upgrading? Who (or what) are you people, anyway?

Bill

who:

1-bought one of first Xerox 820s ever sold (1981)
2-discovered its many many limitations
3-found designer, called him, learned about Bigboards
4-built two BB IIs, one for friend, one for myself (1982)
5-used BB II extensively for several years

There is no better Z-80 implementation anywhere
than the BB II. It has all the support chips: the CTC chip;
the DMA chip; the PIO chip; the full blown SIO chip,
and is the classic 'everything in the Intel databook
implementation' with the addition of a SASI interface
and the ability to program it's own BIOS ROMs.

You need either a keyboard and a monitor OR a
serial terminal (or a computer that looks like one)

And, of course, what ever kind of storage you can
wire up - drives, tape, paper tape, whatever.

It has an extensive built-in monitor that does
almost everything but write programs for you.
Which includes, INing and OUTing whatever
bits it takes to get all your I/O working right.

You don't even need an operating system to
save and get bits from your storage device(s).

Meaning: you can start from SCRATCH and
work up your boot interface, then paste it
onto ANY Z-80 operating system image, from
any device you can read from, and then
save it out to ANY storage device you have.

If you don't have ANY, then burn yourself
a ROM of the thing. It's even got extra
sockets to stash hardprogrammed code.

They didn't leave anything out. Period.

Because of it's hardware interface capabilities
there is NO WAY you can do it all with ANY
emulator. The BB II is/was uniquely versatile.

Call it the pinnacle of Z-80 development systems.

(and, it is impossible to emulate the 1791/3 type
disk controller with a 765. IT CAN NOT BE DONE.)
IBM and the others chose the 765 at least in part
because of it's 'ability' to protect THEM and their
precious softwares from unauthorized copying.
Which included, making safety backups yourself.

With the 1793 you can copy ANYTHING, with the
only exception being those little laser holes. For
some reason they forgot to give it a turn-on-the-
laser output port. Probably so you wouldn't
accidentally burn up your disks. Anyway, not too
many disk drives came with them (hole burners).

And, maybe 'weak' bits. Not too sure about them.

On Sun, 16 Apr 2006 09:57:35 -0500, wild bill
wrote:
>
>There is no better Z-80 implementation anywhere
>than the BB II. It has all the support chips: the CTC chip;
>the DMA chip; the PIO chip; the full blown SIO chip,
>and is the classic 'everything in the Intel databook
>implementation' with the addition of a SASI interface
>and the ability to program it's own BIOS ROMs.
>

Well, okay, so try Zilog Databook.

Ever since, it's been Intel On My Mind.

With an occasional AMD breeze.....

Bill

On Sun, 16 Apr 2006 09:57:35 -0500, wild bill
wrote:

>On 15 Apr 2006 20:37:52 -0700, "Tarkin" wrote:
>
>>serve
>>our purposes- such as the idea of upgrading the H8's
>>1701 (? WD) in favor of the 765
>
>Upgrading? Who (or what) are you people, anyway?
>
>Bill
>
>who:
>
>1-bought one of first Xerox 820s ever sold (1981)
>2-discovered its many many limitations
>3-found designer, called him, learned about Bigboards
>4-built two BB IIs, one for friend, one for myself (1982)
>5-used BB II extensively for several years
>
>There is no better Z-80 implementation anywhere
>than the BB II. It has all the support chips: the CTC chip;
>the DMA chip; the PIO chip; the full blown SIO chip,
>and is the classic 'everything in the Intel databook
>implementation' with the addition of a SASI interface
>and the ability to program it's own BIOS ROMs.
>
>You need either a keyboard and a monitor OR a
>serial terminal (or a computer that looks like one)

H8 has a software driven front pannel.

>And, of course, what ever kind of storage you can
>wire up - drives, tape, paper tape, whatever.

H8 had disk, paper tape and even cassette.

>It has an extensive built-in monitor that does
>almost everything but write programs for you.
>Which includes, INing and OUTing whatever
>bits it takes to get all your I/O working right.
>
>You don't even need an operating system to
>save and get bits from your storage device(s).
>
>Meaning: you can start from SCRATCH and
>work up your boot interface, then paste it
>onto ANY Z-80 operating system image, from
>any device you can read from, and then
>save it out to ANY storage device you have.
>
>If you don't have ANY, then burn yourself
>a ROM of the thing. It's even got extra
>sockets to stash hardprogrammed code.
>
>They didn't leave anything out. Period.

Sounds like SB180, AmproLB+, Davige
and a raft of others.

>Because of it's hardware interface capabilities
>there is NO WAY you can do it all with ANY
>emulator. The BB II is/was uniquely versatile.
>
>Call it the pinnacle of Z-80 development systems.

You forget many others out there.

>(and, it is impossible to emulate the 1791/3 type
>disk controller with a 765. IT CAN NOT BE DONE.)

Why would you.

>IBM and the others chose the 765 at least in part
>because of it's 'ability' to protect THEM and their
>precious softwares from unauthorized copying.
>Which included, making safety backups yourself.

Bogus and meaningless.

>With the 1793 you can copy ANYTHING, with the
>only exception being those little laser holes. For
>some reason they forgot to give it a turn-on-the-
>laser output port. Probably so you wouldn't
>accidentally burn up your disks. Anyway, not too
>many disk drives came with them (hole burners).

Not quite but, their extinct anyway.

>And, maybe 'weak' bits. Not too sure about them.

??

Allison

In article <1145158672.790102.135200@u72g2000cwu.googlegroups. com>,
"Tarkin" wrote:

> >I don't have a website, but would be happy to post the schematic or
> >board layout. Anyone have any suggestions on where to put them?
>
> I can offer either/both my yahoo and comcast personal websites.
>
> yahoo :
> pros: that site has been around for awhile, and will likely remain
> that way.
> cons : heavy advertising.
>
> comcast :
> pros: Less advertising, the site is (for the moment) 'cleaner'
> (as in less cluttered), easy file management (for me!)
> cons: It's dependant on my Dad's comcast account. If he
> gets a twist in his shorts and cancels his subscription, I
> lose my sub-account.
>
> My thoughts about the project in general:
>
> While I would love to dive into an ASIC project, I also
> love simplicity, commonality, and ease-of-replacement:
> If I release the magic smoke on an ASIC, I'm gonna
> be pretty mad. If I smoke a Z80 or '138, I can get over it
> much easier and faster.
>

A fpga is really tempting me as well, for about 150 bucks and i have
plenty of hardware to do most any z80 project ( or 6502, etc ) i want,
and if i is crew it up, just erase it and start over.

But, that sort of takes away from the whole retro idea and wouldnt be as
much fun..

>But, that sort of takes away from the whole retro
>idea and wouldnt be as much fun.

I wholeheartedly agree. I was tempted, as
well, after looking at the Spartan3 dev
board from http://digilent.com .

I have a decent machine running CRUX linux
2.0 (eveolution), w/ gcc 3.4, and all sorts o'
libxxx's; I also have the J2SE sdks for win
and linux; but I get my kicks from writing
'Hello world!' using cp/m, MAC, and ED.
(My main box, the one I usually compose
my thoughts to the groups on, is a WinXP box)

I am not entirely against the idea of using the BBII
as starting point, either; I've heard good things
about it before.

I am also totally for using modern equivalents
of some of the componenets from yesteryear:
one of the prev. posters mentioned octal vs.
quad latches (registers? something), and his
point about the lower power consumption of
the HC devices is spot on; I don't believe
that these enhancements will violate the
spirit of the OP's project idea.

I guess the idea of using an fpga suggests
(to me) using a more robust fpga, and
implementing all of it on a single chip, or
use an fpga + 2/3 support chips, and do
it all surface mount on a credit-card sized
card.

My interpretation of the OP's intent for the
proj was something that would use cheap,
readily-available components, easily assembled,
and that the design would be something a
reasonably bright 10-12 year-old could
assemble, grabbing an elder for assistance
w/ the programming & firmware.

TTFN,
Tarkin

Tarkin wrote:
> >But, that sort of takes away from the whole retro
> >idea and wouldnt be as much fun.
>
> I wholeheartedly agree. I was tempted, as
> well, after looking at the Spartan3 dev
> board from http://digilent.com .
>
> < sniped >
>
> TTFN,
> Tarkin

This is just another bunch of random ramblings on my part so take what
you want and leave what you don't want. :-)

I was thinking of sticking to the standard Z80 and its support chips,
WD279x FDC and the GIDE interface. All of the "glue logic" would be
implemented in a couple of CPLD (Atmel or Xilinx). The parts are not
that expensive, $6 to $12 each, and don't require any special
programming hardware. The parts are in system programmable through the
JTAG port. The development tools are free.

Don

In article <1144973073.754320.312890@i39g2000cwa.googlegroups. com>,
"Don" wrote:

> Tarkin wrote:
> > Thanks for the tip. I realized the bit about the
> > registers initializtion being specific to operating
> > parameters and port mode, but it's gotchas
> > like keeping a copy of the xmt status that I
> > suspected would cause hours of r&d and
> > forensic debugging.
> >
> > I am off to hunt down the AMD spec/docs.
> >
> > To birng us back on track, what's the consensus
> > on a z80-based homebrew? Book? Databug?
> > Shameless eastern-bloc rip-off of the p112?
> > Xerox-clone[1] (snicker)?
> >
> > Thanx,
> > Tarkin
> > [1] I am old enough to remeber when mimeographs
> > were replaced, the new process was called
> > "Xeroxing".
>
> I remember "using" mimeograph machines. Xerox didn't exist at that
> time. :-)
>
> My reason for suggesting the 820-1, or bigboard for that matter, is
> that it is a fairly stable starting point. I have know idea of the
> skills of the readers of this list so thought a known starting point
> would be good. I had an 820-1 many years ago. It reminded me of two
> aircraft carriers parked side by side. :-) That puppy was big!
>
> Anyway, the design could be simplified and then expanded on. My 820 was
> hacked to the max. I had doubled the speed from 2.5 to 5 MHz, added the
> double density adaptor, hard drive interface, 68000 co-processor and
> changed the video timing to use a hercules monitor. It looked more like
> a high rise apartment complex than a computer board. :-)
>
> Don

How about a modular approach?

Sort of like a PC102 concept. We design seperate boards with some sort
of 'universal' bus, and people choose what features they want, beyond
the basic cpu board.

One board for video, another for IDE..

Of course everything will need to be done with CPM in mind, but i would
think a modular approach would make it fit most anyones needs.

( spending this weekend myself digging for z80 books in my garage, and
see what chips i have left .. its been a while but i think this might be
just the excuse to dust off all my old equipment, and braincells, and
actually do something with real hardware again. Going thru life as a
software guy is hard for an old EE like me )

Or am i way off base here in this day and age and should just be content
with my z80 card in my IIGS?

Darn your socks!! Is it fun (the IIgs + card)?

I am having fun with the C128. I RTFM for
once, and realized you CAN use cp/m
in the 40-col tv mode. So I knocked around
with Mac and Ed. Fun stuff!!
(*Thank You ETG!!!!*)

I have a couple of proposals for a middle
ground on this.

If the CPLD's are that low, I am guessing
that their gate count is as well.. So design
the glue logic in such a manner as to allow
implementation as a single CPLD or a handful
of 74-series logic devices. Combined with
the idea of either reserved ports or addy's,
this would leave a design that is consistent
(from a software point of view), but gives
a lot of room for the indiviual implementor.
Leftover gates could be connected to the
reserved ports/addy's to provide custom
peripherals- like ISP or video or audio
or cassette controls. The core chips-
uP, disk controller, ram, rom, sio,
ctc, pio, etc all need their specific pins
tended to. It will simply be left to the
individual to choose which 'black box'
he sticks between those chips.

Don - got any links to specific vendor's
pages on CPLD's in that range, and/or
the tools? I'm going to google up on
JTAG & connectors for same. I guess
I should come out of the 80's.

"Take, on, meeeeee......."
(take on me)
"Taaaake meee onnn...."
(take on me)
"I'lll beeee gooooneeeeee......'

-Tarkin

Feel free to get a space on my website. I can create a personal page
for you or just for the project itself, as you like.
Simply let me know if you're interested in this space.
This also is giving me the idea that ALL of the works of this thread
can be hosted on my server if the people here likes the idea.
It would be nice to create a dedicated forum on the site too....

Oops! i forgot to say that my last post about web space, was in
response to Lee Hart:
> I don't have a website, but would be happy to post the schematic or
> board layout. Anyone have any suggestions on where to put them?

Sorry.
-Piergiorgio

In article <1145328224.554297.3460@j33g2000cwa.googlegroups.co m>,
"Tarkin" wrote:

> Darn your socks!! Is it fun (the IIgs + card)?
>
> I am having fun with the C128. I RTFM for
> once, and realized you CAN use cp/m
> in the 40-col tv mode. So I knocked around
> with Mac and Ed. Fun stuff!!
> (*Thank You ETG!!!!*)
>
> I have a couple of proposals for a middle
> ground on this.
>
> If the CPLD's are that low, I am guessing
> that their gate count is as well.. So design
> the glue logic in such a manner as to allow
> implementation as a single CPLD or a handful
> of 74-series logic devices. Combined with
> the idea of either reserved ports or addy's,
> this would leave a design that is consistent
> (from a software point of view), but gives
> a lot of room for the indiviual implementor.
> Leftover gates could be connected to the
> reserved ports/addy's to provide custom
> peripherals- like ISP or video or audio
> or cassette controls. The core chips-
> uP, disk controller, ram, rom, sio,
> ctc, pio, etc all need their specific pins
> tended to. It will simply be left to the
> individual to choose which 'black box'
> he sticks between those chips.
>
> Don - got any links to specific vendor's
> pages on CPLD's in that range, and/or
> the tools? I'm going to google up on
> JTAG & connectors for same. I guess
> I should come out of the 80's.
>
> "Take, on, meeeeee......."
> (take on me)
> "Taaaake meee onnn...."
> (take on me)
> "I'lll beeee gooooneeeeee......'
>
> -Tarkin

I have a C128 as well And an osborne 1 ( though the keyboard is
acting up and isnt reconized at all... )

Xilinx has inexpensive CPLDs and FPGAs. Plus their design tools are free
for home use.

Tarkin wrote:
> Darn your socks!! Is it fun (the IIgs + card)?
>
>
>
> If the CPLD's are that low, I am guessing
> that their gate count is as well.. So design
> the glue logic in such a manner as to allow
> implementation as a single CPLD or a handful
> of 74-series logic devices. Combined with
> the idea of either reserved ports or addy's,
> this would leave a design that is consistent
> (from a software point of view), but gives
> a lot of room for the indiviual implementor.
> Leftover gates could be connected to the
> reserved ports/addy's to provide custom
> peripherals- like ISP or video or audio
> or cassette controls. The core chips-
> uP, disk controller, ram, rom, sio,
> ctc, pio, etc all need their specific pins
> tended to. It will simply be left to the
> individual to choose which 'black box'
> he sticks between those chips.
>
> Don - got any links to specific vendor's
> pages on CPLD's in that range, and/or
> the tools? I'm going to google up on
> JTAG & connectors for same. I guess
> I should come out of the 80's.
>
> "Take, on, meeeeee......."
> (take on me)
> "Taaaake meee onnn...."
> (take on me)
> "I'lll beeee gooooneeeeee......'
>
> -Tarkin

A modular approach would be an excellent idea. Take the video circuit
like that on the bigboard or xerox as an example. All those counters
and glue logic to generate the sync and video signals could be drawn as
a single sub-circuit using your favorite TTL logic. That same circuit
could also be run through a logic compiler to produce the data file
needed to program a single CPLD. Now you have two ways to go to get to
the same result.

The GIDE is another perfect example. The 74xx646 is getting hard to
find and the design uses two of them plus a 16V8 and 20V8 GAL. The
whole thing would fit in a single CPLD.

All of the memory and I/O decoder logic is another place that a CPLD
could be used to good advantage.

I will start another discussion and share what I know about CPLD's.

Don

P.S. Please excuse my spelling and bad grammer. Not to bother anybody
with my problems but I have dyslexia which makes my spelling bad and my
ability to express myself in words next to impossible.

Yes Don! That's what I meant exactly.

ziggy: I have two Kaypros that I am hoping
to Frankenstein into one working beasty.
The keyboard issue on your Osb is probly
due to either corrosion or connector
mechanical failure, IIRC some of the
previous discussions from this very board.
A dis-assembly and inspection of the kb and
connectors may well be in order.
(>>for an old EE like me oops!
why am I giving you advice on hardware?
)


CPLD's in the 6-12 $ range....how sharp is the
learning curve on VHDL? (or what's that other
whiz-bang language?). You guys are tempting
me w/ the dark side....

While trying to read up on the BB, one site
quoted someone as saying the Kaypro was so
similar (to the BB) that it was considered by
some to be an unlicensed copy. If that's the
case, or if the video systems are even vaguely
similar, then I am for tty. The Kaypro's video
section makes my heard spin when I try to
comprehend it. Video is one place I'd
definitely go in for using a CPLD or
even a microcontroller (a la VCR Pong),
so as to keep it's headaches separate from
the rest of the system. Divide and conquer,
I say.

In any event, I feel that either a tty, a manual-
entry (switches/keypad)+lcd, or both should
be in place as a development bootstrap and
emergency system saver. As a not-quite-
expert linux user who likes to run as root,
I usually keep a small partition on my linux
machines w/ a 'stock' distro like slackware 8
or mandrake 7.2 or RH 6 in case I screw up a
script and lock myself out of my (main) system.
I also keep bootable cd toolkits (linuxcare BBT)
and flopppies (tomsrtbt). I may not use those
failsafes regularly, but they darn sure are handy
when the unexpected happens.

TTFN,
Tarkin

p.s before I get flamed by 'professional'
linux users, the machines that I run as root
on are *not* production servers or mission-
critical; I have no sensitive data on them.
Running as a separate user and having to su
everytime I wanna clip my toenails just
irritates me.

Don wrote:

> A modular approach would be an excellent idea. Take the video circuit
> like that on the bigboard or xerox as an example. All those counters
> and glue logic to generate the sync and video signals could be drawn as
> a single sub-circuit using your favorite TTL logic. That same circuit
> could also be run through a logic compiler to produce the data file
> needed to program a single CPLD. Now you have two ways to go to get to
> the same result.

> The GIDE is another perfect example. The 74xx646 is getting hard to
> find and the design uses two of them plus a 16V8 and 20V8 GAL. The
> whole thing would fit in a single CPLD.

What's the deal with the 74x646? Fairchild is currently making them.
You can get them from Mouser, but not Jameco.


--
David Griffith
dgriffi@cs.csbuak.edu <-- Switch the 'b' and 'u'

"Dave Griffith" wrote in message
news:3Yj1g.12385$4L1.12152@newssvr11.news.prodigy. com...
> What's the deal with the 74x646? Fairchild is currently making
> them.
> You can get them from Mouser, but not Jameco.

Jameco carries the 74ABT646 version - I've purchased 40-50 of them.
I've used 646 chips from the F, ALS, HC, and ABT families - all work
as intended in the GIDE.


Terry

"Don" writes:

>A modular approach would be an excellent idea. Take the video circuit


>The GIDE is another perfect example. The 74xx646 is getting hard to
>find and the design uses two of them plus a 16V8 and 20V8 GAL.


>All of the memory and I/O decoder logic is another place that a CPLD
>could be used to good advantage.

How about the Floppy-Controller?
It might be difficult to get a real 40-pin Chip these days; even
a version for the IBM-PC with some other IO may be hard to get...

Would it be possible to do a 765 like AND a WD-like controller
into one CPLD with the possibility to dynamically re-program that
chip 'on the fly'? I preferred th WD chip in those old days. And
I once had to read a Floppy with 128 Bytes in DD, which would
have been impossible on the 765.

Just asking, Holger

Terry Gulczynski wrote:
> "Dave Griffith" wrote in message
> news:3Yj1g.12385$4L1.12152@newssvr11.news.prodigy. com...
> > What's the deal with the 74x646? Fairchild is currently making
> > them.
> > You can get them from Mouser, but not Jameco.
>
> Jameco carries the 74ABT646 version - I've purchased 40-50 of them.
> I've used 646 chips from the F, ALS, HC, and ABT families - all work
> as intended in the GIDE.
>
>
> Terry

Terry and Dave. I stand corrected on the 74xx646. I typically look for
the LS, HC & HCT parts which I didn't find until I looked at the
B.G.Micro catalog. I did find the parts previously mentioned in the
DIGI-KEY and Jameco catalogs. While Mouser is a good source of parts I
have never used them so didn't look there.

The one thing I like about these groups is that while one person my not
have all the answers or may have incomplete information, collectively I
think we have all the bases covered.

Don