S-100 Memory Card - Input now or you give up the right to complain later
Grant Stockly
quite a few QFP and fine surface mount parts. Hardware modifications
at a later date may be hard. : )
I am designing the board around MITS S-100 specifications. I do not
have any other S-100 machines or experience with them.
If your IMSAI or whatever has additional I/O, then please tell me if
its I/O, I, O, the pin, and what it does. For example, I really have
no idea how PHANTOM works. I'm not really interested in digging up
the details on how these things work since the have no meaning to an
Altair. The I/O question is a biggie. Timing is not an issue since
that can be worked out at a later date in the CPLD. If there is an
Altair signal that is only an input and you want/need it to be I/O, I
would also need to know that now.
The board will have a landing for 128kbyte ($2.80) or 512kbyte ($5.10)
SRAM. It will also have a landing for 512k of FLASH ($4.62). The
price difference between 64kx8 and 512kx8 is $2, so it hardly makes
sense to choose 64kx8... Bank swapping can/will be worked out later
in the CPLD code.
This board MAY have a landing for a bunch of other optional functions,
but for those of you who don't care about those options, you don't
need to install them. For example, two 6850 type serial ports, 4x SD
cards to emulate Altair disk drive systems, mp3 decoder and
ethernet ; ).
Another point to discuss is if its worth having an 8MByte 8 pin soic
(very small, like pinkey finger nail sized) FLASH chip tied directly
to the CPLD. This would allow the 8080/Z80 to access the serial flash
without AVR interference. The 4 SD cards for Altair disk emulation
are not available directly to the 8080. Is it also worth having a 5th
SD card also directly connected to the CPLD? How many of the 8MByte
data flash chips should there be? The poor things are $2.70 a piece
and we could have 4-8 and hardly notice.
All of the chips will be tied together with a Xylinx MAX II CPLD, and
buffered to the S-100 bus with 8T97 or equivalent ICs.
As you can see I am very flexible. My goal is to create a
"SuperAltair" card with a lot of universal appeal. Sort of "add the
chips you want and nothing more" card. Should I throw in the
functionality of the GIDE while I'm at it? Let me know what you want
and I'll see if I can fit it in.
The entire card will be open source, except for the ROM monitor which
is IP to a fellow enthusiast. I will sell the PCBs for around $35.
The CPLD is $10, SRAM/FLASH as priced above, add a few $ for heat
sinks and voltage regulators and you could be started for under $60.
I will be soldering the crazy surface mount parts for those who are
not brave enough. ; )
Grant
Rolf Harrmann
Grant Stockly schrieb:
>The entire card will be open source, except for the ROM monitor which
>is IP to a fellow enthusiast. I will sell the PCBs for around $35.
>The CPLD is $10, SRAM/FLASH as priced above, add a few $ for heat
>sinks and voltage regulators and you could be started for under $60.
>I will be soldering the crazy surface mount parts for those who are
>not brave enough. ; )
thanks for your big Info's.
Rolf
Barry Watzman
Cromemco bank select would be nice, as would 24-bit addressing. Take a
look at the address decode & control circuits from one of the later
Godbout cards or the later SCP (seattle computer products) 16K static
cards. These are available online.
Grant Stockly
> Phantom is essential ... the card won't work in MOST systems without.
> Cromemco bank select would be nice, as would 24-bit addressing. Take a
> look at the address decode & control circuits from one of the later
> Godbout cards or the later SCP (seattle computer products) 16K static
> cards. These are available online.
I looked into trying to make it IEEE standard, but it would require
quite a few buffers. I think it needs to be an 8bit only card. Were
there any 8 bit processors that used 24bit addressing? Like the 8085?
Grant Stockly
> Cromemco bank select would be nice, as would 24-bit addressing. Take a
> look at the address decode & control circuits from one of the later
> Godbout cards or the later SCP (seattle computer products) 16K static
> cards. These are available online.
I have an excel/openoffice spreadsheet with Altair BUS and IEEE S-100
compared side by side. If anyone wants this file send me an e-mail.
Then you can just show me what you want to support. It has signal
name, description, direction, and number for each.
PHANTOM is something I want to support I guess. I know its something
a lot of other machines have.
no....@no.uce.bellatlantic.net
<gr...@stockly.com> wrote:
>THIS is your chance to get in any requirements. The board will have
>quite a few QFP and fine surface mount parts. Hardware modifications
>at a later date may be hard. : )
>
>I am designing the board around MITS S-100 specifications. I do not
>have any other S-100 machines or experience with them.
Altair is a small subset of S100 and while the first...
It means you will have problems with it if put in later machines
especially IEEE696 S100. Tarbell exercise redux, and that was from
the same era.
An example is pins 20 and 70 but there are more.
>If your IMSAI or whatever has additional I/O, then please tell me if
>its I/O, I, O, the pin, and what it does. For example, I really have
>no idea how PHANTOM works. I'm not really interested in digging up
>the details on how these things work since the have no meaning to an
>Altair. The I/O question is a biggie. Timing is not an issue since
>that can be worked out at a later date in the CPLD. If there is an
>Altair signal that is only an input and you want/need it to be I/O, I
>would also need to know that now.
There are as many or more IMSAIs than Altairs. Then there are all
the other front pannel machines that tended to be more of the
IMSAI pattern or ALTAIR 8800B pattern.
Phantom is real easy if memory is addressed on the card you pull (or
raise) Phantom to say whoever else responds to this
address disable yourself.
Generally ram cards are responders to Phantom and cards that
do things like Eprom boards are generators. Other generators
could be a memory mapped IO.
Phantom is memory only.
>The board will have a landing for 128kbyte ($2.80) or 512kbyte ($5.10)
>SRAM. It will also have a landing for 512k of FLASH ($4.62). The
>price difference between 64kx8 and 512kx8 is $2, so it hardly makes
>sense to choose 64kx8... Bank swapping can/will be worked out later
>in the CPLD code.
>
>This board MAY have a landing for a bunch of other optional functions,
>but for those of you who don't care about those options, you don't
>need to install them. For example, two 6850 type serial ports, 4x SD
>cards to emulate Altair disk drive systems, mp3 decoder and
>ethernet ; ).
Why MP3?
Why 4s SD? One SD has more capacity than all of the mass storage
available for Altiair even CDC hawk! Altair disk drive system
was barely 1MB total for four drives! I don't think I can buy a
SD or other memory device that small (I did find some 32mb CF
parts once). Only Flashroms are in likely to be around 128K to 1MB
or so.
IO Altairs usually had the SIO-B (single uart) and some had 2SIO
(2 6850). I've seen more with multiple SIO-Bs and other non-altair IO
(PT VDM-1 was very common).
Eithernet... how would any 8080/z80 system use that? It requires
software and that has to leave enough of the 64k space to be useful.
Of course if it had it's own CPU...
>Another point to discuss is if its worth having an 8MByte 8 pin soic
>(very small, like pinkey finger nail sized) FLASH chip tied directly
>to the CPLD. This would allow the 8080/Z80 to access the serial flash
>without AVR interference. The 4 SD cards for Altair disk emulation
>are not available directly to the 8080. Is it also worth having a 5th
>SD card also directly connected to the CPLD? How many of the 8MByte
>data flash chips should there be? The poor things are $2.70 a piece
>and we could have 4-8 and hardly notice.
>
>All of the chips will be tied together with a Xylinx MAX II CPLD, and
>buffered to the S-100 bus with 8T97 or equivalent ICs.
>
>As you can see I am very flexible. My goal is to create a
>"SuperAltair" card with a lot of universal appeal. Sort of "add the
>chips you want and nothing more" card. Should I throw in the
>functionality of the GIDE while I'm at it? Let me know what you want
>and I'll see if I can fit it in.
Then research the S100 bus so if it's really useful it will work on
any machine. Most later machines use LS244, LS241.
With a CPLD and standard S100 card size there room for a sink
and dishwasher. By the end of the S100 era a full card had:
Z80 4-8mhz
128K of ram sme had 256K
Shadowed boot rom in the 4Kb range
FDC (for 5.25 and 8" DD)
2 serial IO with programable baud rate support
parallel IO for printer
Full s100 interface for expansion
In short SOB, system in a board.
Using modern parts I'd expect:
Z180 or eZ80 20mhz or faster
1MB ram
1mb flash rom or more
2 serial ports
PPI for parallel IO /printer
Floppy for tradition
CF or other non rotating disk memory media
S100 bus ( with all that) to hold it upright.
Battery backup/battery power friendly
USB inerface with frontpannel emulation GUI
>The entire card will be open source, except for the ROM monitor which
>is IP to a fellow enthusiast. I will sell the PCBs for around $35.
>The CPLD is $10, SRAM/FLASH as priced above, add a few $ for heat
>sinks and voltage regulators and you could be started for under $60.
>I will be soldering the crazy surface mount parts for those who are
>not brave enough. ; )
The SMTs can be nasty but the large footprint SMTs are pretty painless
once you master the trick.
My $0.02....
S100 systems generally need the following to work usefully
or for fun:
Attention to a flavor of S100 that is fairly universal.
CPU, a working one. (not saying the card have one
but you could...z80)
Boot/console rom, Rather than switches. Console rom
is a debug monitor with ability to load the system. the ability
to have a working dialog sans front pannel is very useful.
Ability to load "rom" images of classic Altair code
like MITS 4k, 8K and Extended BASIC, Software package 1
(ed/asm/debug) or mabe something good like PT ALS8 ???
64K ram, when rom is paged out. More ram is nice but only CP/M3
might use it and even then.... Ramdisk/Romdisk configuration that
does not impact ram map (port IO). If you have mass storage like
IDE/CF/SD/MMC ramdisk/romdisk is not needed.
Serial IO that works with boot/monitor - so the user can leave
existing IO alone or if the existing IO has failed or is absent...
Helps when debugging a system thats marginally alive or
for example a NS* that is not bootable for what ever reason.
I picked NS* because if it doesn't boot disk there is no
was to get code in there to debug via serial line. there are
many other non front pannel machines like that.
Some kind of mass storage IDE, CF (same thing different socket)
MMC, SD or USB. Pick one thats cheap and is also known to have
a good lifetime(read write and purchase life).
Software.... Something lacking for the first year of Altair ownership
(some shipped but really buggy!!!!!!!!!). None closed, only open
source as you will end up modding it for people that need xzzy
IO.
The rest is fun but required, useful even valuable?
Most need a complete memory IO board over super.
Allison
>Grant
no....@no.uce.bellatlantic.net
First 8085 is an 8080 that didn't need two more chips and +12V and
-5V, no real magic. Save for the added hardware for interrutps and
SID/SOD that made it pretty handy. It was in effect intels answer to
the 5V only Z80.
No there were very few 20/24bit CPUs on s100. There were plenty of
Z80s with MMUs that could easily map 1MB or more. IEE 696 allowed
for later 16bit or larger cpus and things like DMA. It also allowed
for Altair style split 8bit data bus or 16bit bidirectional data bus.
Most (all?) 8bitter only used the 8bit split data bus. So
implementing that is optional. the line SXTEEN* was there to indicate
that a device sould handle 16bit wide transfers. Mostly seen on
memory cards and boards like Compupro 68000 cpu.
Even the 8088 didn't hit 24bit addressing it was only 20bit.
Allison
Grant Stockly
The parts are only an extra $10. One of the first things I did with
an AVR was make an ethernet-MP3 player. For the cost of an ethernet,
another $7, the Altair can be connected to a LAN.
Imagine setting the Altair in a living room. It could connect to a
server using UDP and easily get the 16-25k per second needed to stream
MP3 audio. Use the sense switches for skip, pause, playlist
selection, etc.
IMO it gives the Altair a utility purpose beyond a computer. I've
talked to quite a few people about this. Some think its the craziest
idea they have ever heard of and ridicule me endlessly. Others think
its the craziest idea they have ever heard of...and like it. : ) The
ethernet decoder will most likely get shared between the AVR and
8080. Either the AVR will be tristated or the CPLD will be
tristated. That way complex software like a simple TCP/IP stack on
the AVR providing telnet to the Altair could be written. I don't
intend to implement any of that fancy telnet stuff (at least right
away), but I am very interested in the MP3 and ethernet for my own
purposes. : )
The ethernet and mp3 circuits occupy about 4 square inches, so they
can be ignored easily. We'll see after a few boards are sold how
popular an MP3 playing Altair is. : )
> Why 4s SD? One SD has more capacity than all of the mass storage
> available for Altiair even CDC hawk! Altair disk drive system
> was barely 1MB total for four drives! I don't think I can buy a
> SD or other memory device that small (I did find some 32mb CF
> parts once). Only Flashroms are in likely to be around 128K to 1MB
> or so.
SD/MMC cards can be felt and thought of as logical disks. If only one
SD or CF card was used in the system then you wouldn't be able to have
two "disks" mounted at the same time (unless of course both disks you
wanted were on the same card). And if you did have multiple disks on
one card there just isn;t a perfect way to select which image goes in
which "fake drive". Having four sockets VS one does not change the
price except for the cost of the socket and PCB real estate.
4 is just a number I thought of that would most always be enough, or
too much. You could have two 8MB images and two 330k-1MB images.
This is how Altair Z80 handles it.
> IO Altairs usually had the SIO-B (single uart) and some had 2SIO
> (2 6850). I've seen more with multiple SIO-Bs and other non-altair IO
> (PT VDM-1 was very common).
I am either going to emulate the 6850s or use real 6850s. The CPLD
and AVR can emulate any I/O. It sounds like something I need to add
is a centronics interlace. Two RS232 ports and custom firmware and
the card can assume any serial device (or pretend to be parallel but
present serial to the real world). This is something I thought of in
the first brain storm, but thought hardware 6850s would be better.
Dave suggested software so now I'm back to thinking that software
would be a good idea...
> Then research the S100 bus so if it's really useful it will work on
> any machine. Most later machines use LS244, LS241.
I have the spreadsheet, and will try to conform to the IEEE standard,
but if people find the idea of a card like mine useful I would
appreciate it if they could make sure I know for sure the I/O
necessary. : ) 3 PCBs is $99, so if its really a big deal they could
just get their own board made I guess. My goal is to support the
Altair.
> Using modern parts I'd expect:
>
> Z180 or eZ80 20mhz or faster
> 1MB ram
> 1mb flash rom or more
> 2 serial ports
> PPI for parallel IO /printer
> Floppy for tradition
> CF or other non rotating disk memory media
> S100 bus ( with all that) to hold it upright.
> Battery backup/battery power friendly
> USB inerface with frontpannel emulation GUI
Howard pretty much has everything on that list in the Zaltair. I
thought of something similar last year. I'm glad he did it first.
> The SMTs can be nasty but the large footprint SMTs are pretty painless
> once you master the trick.
http://www.youtube.com/watch?v=wQXhny3R7lk
I never looked back. : ) Much less painful than the methods I've
used in the past! You hardly have to try! : )
no....@no.uce.bellatlantic.net
<gr...@stockly.com> wrote:
>> Why MP3?
>
>The parts are only an extra $10. One of the first things I did with
>an AVR was make an ethernet-MP3 player. For the cost of an ethernet,
>another $7, the Altair can be connected to a LAN.
Sorta like putting a 426 hemi in a 67VW bug, very cool, attention
getting but...
>Imagine setting the Altair in a living room. It could connect to a
>server using UDP and easily get the 16-25k per second needed to stream
>MP3 audio. Use the sense switches for skip, pause, playlist
>selection, etc.
>
>IMO it gives the Altair a utility purpose beyond a computer. I've
>talked to quite a few people about this. Some think its the craziest
>idea they have ever heard of and ridicule me endlessly. Others think
>its the craziest idea they have ever heard of...and like it. : ) The
>ethernet decoder will most likely get shared between the AVR and
>8080. Either the AVR will be tristated or the CPLD will be
>tristated. That way complex software like a simple TCP/IP stack on
>the AVR providing telnet to the Altair could be written. I don't
>intend to implement any of that fancy telnet stuff (at least right
>away), but I am very interested in the MP3 and ethernet for my own
>purposes. : )
If the Altair is that boring... Walks awayshaking head...
>The ethernet and mp3 circuits occupy about 4 square inches, so they
>can be ignored easily. We'll see after a few boards are sold how
>popular an MP3 playing Altair is. : )
From one perspective it's cool. From another if I needed MP3
I already have PCs all over and none eat near the space and power
or have noisy fans. The PCs can take an Wireless NIC can that beast?
If anything telnet as if it were a common serial port (from the 8080
side) would be handy for those that need to load files, remote
terminal from a PC.
Allison
Grant Stockly
> >talked to quite a few people about this. Some think its the craziest
> >idea they have ever heard of and ridicule me endlessly. Others think
> >its the craziest idea they have ever heard of...and like it. : ) The
> >ethernet decoder will most likely get shared between the AVR and
> >8080. Either the AVR will be tristated or the CPLD will be
> >tristated. That way complex software like a simple TCP/IP stack on
> >the AVR providing telnet to the Altair could be written. I don't
> >intend to implement any of that fancy telnet stuff (at least right
> >away), but I am very interested in the MP3 and ethernet for my own
> >purposes. : )
>
> If the Altair is that boring... Walks awayshaking head...
To me, I just can't understand why someone wouldn't want it. You fall
under the first category of people. : ) I listen to the radio or a
CD every day while I am getting ready to go to work. Why not have the
Altair do it for me? That would be much better than a CD player.
Remember, I soldered 19,008 LEDs to over 20 13x13" PCBs to make a
7x3.5 foot bitmaped display...
> >The ethernet and mp3 circuits occupy about 4 square inches, so they
> >can be ignored easily. We'll see after a few boards are sold how
> >popular an MP3 playing Altair is. : )
>
> From one perspective it's cool. From another if I needed MP3
> I already have PCs all over and none eat near the space and power
> or have noisy fans. The PCs can take an Wireless NIC can that beast?
No fan in mine! : ) It could have a wireless NIC, sure. I'm not
going to add one. The non wireless chip is $3. There are wireless to
wired bridges available for under $50 if its important.
> If anything telnet as if it were a common serial port (from the 8080
> side) would be handy for those that need to load files, remote
> terminal from a PC.
I found a few code examples. It shouldn't be too hard.
no....@no.uce.bellatlantic.net
<gr...@stockly.com> wrote:
>> >IMO it gives the Altair a utility purpose beyond a computer. I've
>> >talked to quite a few people about this. Some think its the craziest
>> >idea they have ever heard of and ridicule me endlessly. Others think
>> >its the craziest idea they have ever heard of...and like it. : ) The
>> >ethernet decoder will most likely get shared between the AVR and
>> >8080. Either the AVR will be tristated or the CPLD will be
>> >tristated. That way complex software like a simple TCP/IP stack on
>> >the AVR providing telnet to the Altair could be written. I don't
>> >intend to implement any of that fancy telnet stuff (at least right
>> >away), but I am very interested in the MP3 and ethernet for my own
>> >purposes. : )
>>
>> If the Altair is that boring... Walks awayshaking head...
>
>To me, I just can't understand why someone wouldn't want it. You fall
>under the first category of people. : ) I listen to the radio or a
>CD every day while I am getting ready to go to work. Why not have the
>Altair do it for me? That would be much better than a CD player.
I like my radio, it fits in my pocket, runs off little batteries and
portable. ;)
If I want something that does none of the above I call that a PC.
>Remember, I soldered 19,008 LEDs to over 20 13x13" PCBs to make a
>7x3.5 foot bitmaped display...
Pondering that one... First item to mind is whay too much free time.
;)
Was it color and how does football look on it?
>> >The ethernet and mp3 circuits occupy about 4 square inches, so they
>> >can be ignored easily. We'll see after a few boards are sold how
>> >popular an MP3 playing Altair is. : )
>>
>> From one perspective it's cool. From another if I needed MP3
>> I already have PCs all over and none eat near the space and power
>> or have noisy fans. The PCs can take an Wireless NIC can that beast?
>
>No fan in mine! : )
Any Altair I used had enough boards that no fan meant failure was
pending warm up. Generally some air is rquired even if it's a small
fan.
>It could have a wireless NIC, sure. I'm not
>going to add one. The non wireless chip is $3. There are wireless to
>wired bridges available for under $50 if its important.
Wires not wires all the same when it's not portable.
>> If anything telnet as if it were a common serial port (from the 8080
>> side) would be handy for those that need to load files, remote
>> terminal from a PC.
>
>I found a few code examples. It shouldn't be too hard.
I presume the code is for the AVR. Burdening the 8080 with that
would be a project.
Allison
Barry Watzman
for bank select in some 8-bit systems, also it would support use with an
8088 cpu card.
Max Scane
"Grant Stockly" <gr...@stockly.com> wrote in message
news:ef219c2f-e09c-4694...@s12g2000prg.googlegroups.com...
> THIS is your chance to get in any requirements. The board will have
> quite a few QFP and fine surface mount parts. Hardware modifications
> at a later date may be hard. : )
>
> I am designing the board around MITS S-100 specifications. I do not
> have any other S-100 machines or experience with them.
>
> As you can see I am very flexible. My goal is to create a
> "SuperAltair" card with a lot of universal appeal. Sort of "add the
> chips you want and nothing more" card. Should I throw in the
> functionality of the GIDE while I'm at it? Let me know what you want
> and I'll see if I can fit it in.
>
> The entire card will be open source, except for the ROM monitor which
> is IP to a fellow enthusiast. I will sell the PCBs for around $35.
> The CPLD is $10, SRAM/FLASH as priced above, add a few $ for heat
> sinks and voltage regulators and you could be started for under $60.
> I will be soldering the crazy surface mount parts for those who are
> not brave enough. ; )
>
> Grant
GIDE would be a worthwhile addition and would give you access to hard disks
or even a CD drive.
If you are going to add I/O devices may I suggest that you add an IRQ
controller
(possibly in the CPLD) or at a minimum support connection to the S-100 IRQ
lines
Same goes for DMA.
How about adding a USB slave chip. Something like the FTDI FT245BM (usb to
parallel)
would give you a high speed link for file transfer or even cp/net to a Linux
box.
If your on-board memory is I/O mapped don't forget to include an auto
incrementing address register
so that you could use the Z80's block transfer instructions (OTIR,INIR) or
even a DMA controller.
glen herrmannsfeldt
(snip)
> If your on-board memory is I/O mapped don't forget to
> include an auto incrementing address register
> so that you could use the Z80's block transfer
> instructions (OTIR,INIR) or even a DMA controller.
I was thinking about doing the ethernet that way, with its
own buffer as I/O mapped memory and autoincrement address
register.
-- glen
no....@no.uce.bellatlantic.net
<Watzma...@neo.rr.com> wrote:
>I'm not suggesting making it an 8/16 card, but 24-bit addressing is used
>for bank select in some 8-bit systems, also it would support use with an
>8088 cpu card.
good point as most of the later (CCS, Cromemco, Compupro, Inthica
Intersystems, ...) went to some form of 24 bit addressing extension.
Allison
no....@no.uce.bellatlantic.net
wrote:
>
>"Grant Stockly" <gr...@stockly.com> wrote in message
>news:ef219c2f-e09c-4694...@s12g2000prg.googlegroups.com...
>> THIS is your chance to get in any requirements. The board will have
>> quite a few QFP and fine surface mount parts. Hardware modifications
>> at a later date may be hard. : )
>>
>> I am designing the board around MITS S-100 specifications. I do not
>> have any other S-100 machines or experience with them.
>>
><snip>
>
>> As you can see I am very flexible. My goal is to create a
>> "SuperAltair" card with a lot of universal appeal. Sort of "add the
>> chips you want and nothing more" card. Should I throw in the
>> functionality of the GIDE while I'm at it? Let me know what you want
>> and I'll see if I can fit it in.
>>
>> The entire card will be open source, except for the ROM monitor which
>> is IP to a fellow enthusiast. I will sell the PCBs for around $35.
>> The CPLD is $10, SRAM/FLASH as priced above, add a few $ for heat
>> sinks and voltage regulators and you could be started for under $60.
>> I will be soldering the crazy surface mount parts for those who are
>> not brave enough. ; )
>>
>> Grant
>
>GIDE would be a worthwhile addition and would give you access to hard disks
>or even a CD drive.
GIDE is not required, however the essence being an IDE or CF interface
would be good.
GIDE was a specific idea to make a board that provided IDE (plus TOY
clock) on a small board that went under a Z80.
>If you are going to add I/O devices may I suggest that you add an IRQ
>controller
>(possibly in the CPLD) or at a minimum support connection to the S-100 IRQ
>lines
Do the full banana 8259 or similar.
>Same goes for DMA.
S100 DMA is very tricky and Altair flavor (and other early S100)
didn't always behave well. Nice to have especially if it can also
do fast blockmove (bliter).
>How about adding a USB slave chip. Something like the FTDI FT245BM (usb to
>parallel)
>would give you a high speed link for file transfer or even cp/net to a Linux
>box.
Ethternet and USB...
>If your on-board memory is I/O mapped don't forget to include an auto
>incrementing address register
>so that you could use the Z80's block transfer instructions (OTIR,INIR) or
>even a DMA controller.
Good idea. Compupro did this with MDRIVE ramdisk and it's easy to use
and faster.
Allison
Herb Johnson
I too have a list of Altair bus and IEEE-696 lines, among others, on
my Web site at:
http://www.retrotechnology.com/herbs_stuff/s100bus.html
..but that's not enough to describe their functions for design. There
are many S-100 manuals available which describe all the designs
discussed. I offer them from my Web site as good photocopies. Others
offer them for free download as PDF'ed scans.
It's always been my philosophy, that many of these manuals provide
MUCH MORE technical
description than can often be found in on-line discussion. Morrow and
Compupro / Godbout manuals have
good technical descriptions, but sometimes they use programmable logic
which is not as easy to
review as earlier designs which used discrete TTL-class logic.
I strongly, strongly advise READING THE MANUALS for any technical
questions or design issues.
That said, there are excellent technical posts in this newgroup, and a
variety of good design opinions as well. They are useful and
informative. But a board and a schematic and a design description, all
at your desk, is hard to beat for technical clarity and completeness.
That's why I've spent about two decades of effort to provide them.
Others who provide them no doubt have similar motivations.
Herb Johnson
Herbert R. Johnson, New Jersey USA
http://www.retrotechnology.com/herbs_stuff/ web site
http://www.retrotechnology.net/herbs_stuff/ domain mirror
my email address: hjohnson AAT retrotechnology DOTT com
if no reply, try in a few days: herbjohnson ATT comcast DOTT net
"Herb's Stuff": old Mac, SGI, 8-inch floppy drives
S-100 IMSAI Altair computers, docs, by "Dr. S-100"
no....@no.uce.bellatlantic.net
I thought about it more. This way if your Altair is only a front
pannel and cover it will sing again.
Requriements:
Z80 or Z180, or eZ80 a fast one no wimpy 4mhz
but have the ability to run at 2mhz.
Memory with MMU to support a 1MB or more
external memory mapable into a 64K window.
Boot system rom/flash (post, monitor, OS, loader)
Some kind of mass storage CF, MMC SD, usb thumbstick
IDE also for CDrw.
IO, lots PPI of the 8255 flavor, at least 2 serial ports and a
Parallel (compatable with PC printers). Local IO isolated
from bus IO to avoid collisions. Don't forget the 88-ACR
cassette interface.
VGA video and PS2 keyboard interfaces. So a terminal
is not needed to use it. Terminal emulation should be
DEC VT05, Hazeltine 1000 or similar 1975 glass TTY.
If you do ASR33 emulation then MP3 must play motor at 75dbA
when TTY emupation is enabled and TTY clatter when writing
to screen at 85dbA not adjustable as that how it was...
USB and Ethernet with plug and play appliation side.
After all if you use a bluetooth or wireless NIC it must be able
set it self up.
Soundblaster comptable sound card/MP3 player.
Front pannel must work with the board. Connector
must be the moldy molex.
Cables for rear pannel connector holes.
Under 199.99 assembled and tested.
ALTAIR Documents on CD.
Allison
Grant Stockly
> Z80 or Z180, or eZ80 a fast one no wimpy 4mhz
> but have the ability to run at 2mhz.
What about a 50MHz 8080 in an FPGA? : )
I've sketched it up, and it would take quite a few bus buffers and
control lines to have a completely bidirectional card. It only works
to design it one way and stick with it.
> Memory with MMU to support a 1MB or more
> external memory mapable into a 64K window.
Probably 128k-768k. I'll probably run out of landing room if I do
more. Might even leave the 512k spot out. 128k is enough for me (and
cheaper than 32kx2).
> Boot system rom/flash (post, monitor, OS, loader)
My plan has all ways been to have a POR and POJ.
> Some kind of mass storage CF, MMC SD, usb thumbstick
> IDE also for CDrw.
IDE for fun if there is space, and 3-5 SD card capability. No USB. :
P
> IO, lots PPI of the 8255 flavor, at least 2 serial ports and a
> Parallel (compatable with PC printers). Local IO isolated
> from bus IO to avoid collisions. Don't forget the 88-ACR
> cassette interface.
I might put a parallel port on it. I hadn't thought of that before.
If there is free space on the CPLD it would go best there, or my
second option would be if the AVR had some free I/O.
>
> VGA video and PS2 keyboard interfaces. So a terminal
> is not needed to use it.
I've been debating how to do that for a while. I'm not sure if my
CPLD has enough resources for the VGA hardware, and a microcontroller
can't do 80x20... I really want to stick with my CPLD and not step up
to an FPGA. My plan was to have the terminal adapter based on the $99
Spartan-3 dev kit from digilentinc and make it a separate piece. If I
can fit VGA in, I will.
> If you do ASR33 emulation then MP3 must play motor at 75dbA
> when TTY emupation is enabled and TTY clatter when writing
> to screen at 85dbA not adjustable as that how it was...
Actually the feature for the disk drive AVR chip to override the
serial bistream to the MP3 chip has been designed in. : ) I'm not
sure how it will handle a loop though...
> USB and Ethernet with plug and play appliation side.
> After all if you use a bluetooth or wireless NIC it must be able
> set it self up.
No USB! : ) Not even linux is plug and play. Well, linux is plug
and play...just not unplug...
> Under 199.99 assembled and tested.
I talked with Howard about mass producing boards and that wouldn't be
impossible, except I won't ever stop adding features.
The only things I ever thought would interest this group is:
-128k to 128k-768k...maybe 1MB SRAM
-512k FLASH
-1 CPLD accessable SD card
Everything else (except mp3 and Ethernet) goes through the AVR and
will probably not be all that great on a z80. Even a 20MIPS
microcontroller has trouble keeping up with an 8080... It will insert
wait states if it needs them, its just a 4MHz z80 will be a 2MHz when
accessing the disk drive.
Grant
Grant Stockly
> MUCH MORE technical
> description than can often be found in on-line discussion. Morrow and
> Compupro / Godbout manuals have
> good technical descriptions, but sometimes they use programmable logic
> which is not as easy to
> review as earlier designs which used discrete TTL-class logic.
>
> I strongly, strongly advise READING THE MANUALS for any technical
> questions or design issues.
I know about your website and have used it several times, but I think
you missed my point. I do not have the time to research anything
other than Altair stuff. I am only interested in hearing from other
people on what I/O needs to go into and come out of the CPLD so that
if THEY want to modify it for their S100 system, it will be possible.
They can do the research and let me know just as well as I could. It
doesn't matter what the I/O does on the bus, it will all be sorted out
later in the CPLD. I just have to make sure the appropriate buffer is
hooked to the bus.
What it comes down to is in 5 days 3 S-100 PCBs can be produced for
$99. So its not the end of the world if someone wants to make their
own S-100 card... : )
Tom Lake
> you missed my point. I do not have the time to research anything
> other than Altair stuff. I am only interested in hearing from other
> people on what I/O needs to go into and come out of the CPLD so that
> if THEY want to modify it for their S100 system, it will be possible.
Anything is fine as long as it's compatible with the original Altair
I/O ports and conventions! Don't get "Fischeritis" and try to do too much
at once! Unless you have very deep pockets, that won't work. Start with
basic Altair compatibility and use sales of those products to fund the
next round of designs. Above all, don't start the next round if the current
merchandise hasn't sold!
Tom Lake
no....@no.uce.bellatlantic.net
<gr...@stockly.com> wrote:
>It took me to the middle to figure out you were kidding... ; )
>
>> Z80 or Z180, or eZ80 a fast one no wimpy 4mhz
>> but have the ability to run at 2mhz.
>
>What about a 50MHz 8080 in an FPGA? : )
Now theres a plan!
>I've sketched it up, and it would take quite a few bus buffers and
>control lines to have a completely bidirectional card. It only works
>to design it one way and stick with it.
>
>> Memory with MMU to support a 1MB or more
>> external memory mapable into a 64K window.
>
>Probably 128k-768k. I'll probably run out of landing room if I do
>more. Might even leave the 512k spot out. 128k is enough for me (and
>cheaper than 32kx2).
Nope the full boat.
>> Boot system rom/flash (post, monitor, OS, loader)
>
>My plan has all ways been to have a POR and POJ.
That gets you there but what runs?
>> Some kind of mass storage CF, MMC SD, usb thumbstick
>> IDE also for CDrw.
>
>IDE for fun if there is space, and 3-5 SD card capability. No USB. :
>P
>
>> IO, lots PPI of the 8255 flavor, at least 2 serial ports and a
>> Parallel (compatable with PC printers). Local IO isolated
>> from bus IO to avoid collisions. Don't forget the 88-ACR
>> cassette interface.
>
>I might put a parallel port on it. I hadn't thought of that before.
>If there is free space on the CPLD it would go best there, or my
>second option would be if the AVR had some free I/O.
the problem is AVR is "remote", not programmable regisers
accesable from 8080.
Just like the old days get some speed and wham, bottleneck.
Allison
>
>Grant
Grant Stockly
>
> That gets you there but what runs?
I don't get it. Anything in memory?
> the problem is AVR is "remote", not programmable regisers
> accesable from 8080.
Why would that be a problem? I found an example for programming the
AVR over Ethernet, so it would be easy for someone without tools to
customize it. I'm looking into programming the CPLD from the AVR
too. It doesn't seem too practical to program the AVR from the 8080
but you could write a boot loader for that if you wanted.
Grant Stockly
> I/O ports and conventions! Don't get "Fischeritis" and try to do too much
> at once! Unless you have very deep pockets, that won't work. Start with
> basic Altair compatibility and use sales of those products to fund the
> next round of designs. Above all, don't start the next round if the current
> merchandise hasn't sold!
Don't worry...I've been designing this card since 6 months before the
first kit sold. : )
no....@no.uce.bellatlantic.net
<gr...@stockly.com> wrote:
>> >My plan has all ways been to have a POR and POJ.
>>
>> That gets you there but what runs?
>
>I don't get it. Anything in memory?
You have POJ, I assume POJ get you to a monitor or software front
panel?
>
>> the problem is AVR is "remote", not programmable regisers
>> accesable from 8080.
>
>Why would that be a problem? I found an example for programming the
>AVR over Ethernet, so it would be easy for someone without tools to
>customize it. I'm looking into programming the CPLD from the AVR
>too. It doesn't seem too practical to program the AVR from the 8080
>but you could write a boot loader for that if you wanted.
Missed that one. No I ment if you using the AVR from the 8080 side
it's doesn't looke like a UART or 8255 and how would the 8080 used the
AVR for IO pins? I'd hope it's not a narrow everything goes through
one register and the protocal is really fussy.
Allison
Grant Stockly
> panel?
The Dunfield one, or a MITS rom like for the disk drive.
> Missed that one. No I ment if you using the AVR from the 8080 side
> it's doesn't looke like a UART or 8255 and how would the 8080 used the
> AVR for IO pins? I'd hope it's not a narrow everything goes through
> one register and the protocal is really fussy.
Microcontrollers are not fast enough to handle a slave bus in
software. I know people do IDE and hook to ISA cards and ethernet,
but that's different. A microcontroller just doesn't have enough
time. I got a 50MHz Scenix processor doing 1MHz ISA with no wait
states but it could only have 3 registers and you had to figure out
what you wanted in them before the host asked for it.
The AVR has 8 external interrupts and a few dozen pins sharing an
interrupt.
The CPLD has registers inside of it. Each 8080 read register can be
prefilled by the AVR. When the 8080 reads the contents the AVR is
interrupted so that it can put new data in it. If new data is not put
in it, the next read will cause a wait state. Same with writes except
the 8080 fills the register, AVR is interrupted, reads the data, and
then sits. If the AVR doesn't read it then it will generate a wait
state on the next write. This allows the AVR to think during the
other 8080 cycles. For an 8080, no wait states would be generated.
The only thing I've prototyped with the card is the above operation.
Everything else I've done before but separately.
So the 8080 talks directly to the CPLD, and the CPLD talks directly to
the AVR. The emulated 2SIO or SIOB has the same registers to the 8080
as a real one. Same thing with the disk drive. My first goal is to
get the disk drive emulation working. Then I'm going to program in
delays to make it "real", then add sounds. : )
Grant
Robert J. Stevens
I am looking for a S-100 Proto-Type board in any Condition. I need to
Wire Wrap up a Test board for my S-100 Systems
TIA
Bob in Wisconsin
John Q. Public
I have an unused Vector 8800V S-100 board. Let me know if this is what you
need.
Thanks,
Michael
"Robert J. Stevens" <ztre...@execpc.com> wrote in message
news:475935B0...@execpc.com...
cavelamb himself
I have one too - Vector 8800V.
Not quite as shiny as 20 years ago, but otherwise like new.
Richard
Robert J. Stevens
is a Vector8801
It has a complete Solder pads with buss Traces around the outside
Let Me know How much and I'll get a Money Order of Post Haste
Bob in Wisconsin
Robert J. Stevens
Bob
John Q. Public
$12.95 shipped Priority Mail. Can you send me your email address and I will
give you my mailing address.
Thanks,
Michael
wac...@yahoo.com
"Robert J. Stevens" <ztre...@execpc.com> wrote in message
news:475990B4...@execpc.com...
Robert J. Stevens
Not sure the other got sent but here is my Addy
my E-Mail below is valid
treb...@execpc.com
Please ship to the following address:
Robert J. Stevens
N89 W15687 Cleveland Ave
Menomonee Falls, WI 53051
United States
(262) 251-7907
Barry Watzman
going for so much that it might be worthwhile making a "run" of some new
ones with a PC Board house. Not exactly a difficult design, and I don't
think a particularly expensive board either.