CPM/Z-80 card for the 1090XL -- Calling anyone with Z-80 experience!

[reifsnyderb]

Since there are a lot of Z-80 card specific details, I decided to open a separate thread for this card.

 

First a little background and explanation…

 

There may have been two 1090XL Z80 cards. There is a Z80 card schematic and a couple pictures of a Z-80 card. The schematic shows a crystal but the pictures do not.

 

I only recently re-examined the existing Z80 card schematic and I don’t think it would have worked because the system clock, Phi2, would have been so delayed the correct data would never have been written to the registers. The design is such that some communications, with the Atari, would have been through bit 7. I don’t know if there are other reasons it would have worked or not and there aren’t any available pictures of a card with this schematic.

 

The two available pictures, on the other hand, show the layout of the entire card. Last week, I started carefully examining the pictures for clues as to how it worked. The Z-80 chip is obvious, of course. The numbers, on some of the chips, can be read. Other chips can be determined by carefully looking at the blurred text, the number of pins, and the pin connections. There are also two chips for which the writing is impossible to guess at and the pin connections don’t make sense. I designated those chips U14 and U23. Due to the connections that disappear under the chips, I believe those chips are related to timing and control of the Z-80. All of the other chips I am confident have been determined.

 

Another oddity is that the board pictures have the traces on the wrong side of the board. This can be confirmed by the 1090XL edge connector being connected wrong and by looking at the few pictures of other 1090XL boards. For example, the “waffle pattern” around the voltage regulator should be on the back and not on the front. Because of this error, the board in the pictures would have never worked. My personal belief is that the board, in the pictures, ended up as a marketing prop. Though I do find it odd that all of the flux was never cleaned off of it for marketing purposes. Even so, it was fortunate that the traces were on the wrong side of the board because fewer connections were buried under the chips.

 

There are a few other connections that have to be guessed at. For example, the final register addresses will forever be a mystery due to those connections being obscured. A couple connections, under the Z-80, are also in question. I took some time to read sections of the Z-80 data sheet and study Z-80 connections in an attempt to ensure the Z-80 is connected in a way that will work. I’ve also examined a couple Z-80 schematics to get a better understanding of what connections are needed and how they are to be used.

 

Now for the good news…

 

While I eventually abandoned attempts to determine what U14 and U23 really are supposed to be, I believe I was able to create a working schematic from the pictures.  In place of U14 and U23, I installed a quadruple OR gate chip (74LS32) and a 74LS74, respectively. A few other traces had to change but I think what I’ve came up with will work.  (If somebody wants to take a look at U14 and U23, please let me know and I'll post the details.)

 

I also worked on the board layout in parallel with the schematic so as to keep track of where the traces should go, what chips were determined, etc.  (The board rendering shown in this post is a work in progress, but all of the chips are connected.)

 

I am looking for some members with Z-80 experience that can take a look at the schematic and let me know if they believe it will work or not before I get a board made for testing.

 

Here is the schematic:  CPM.pdf

 

Under construction board rendering: 

 

 

Original board images:

 

 

 

[reifsnyderb]

New rendering...

 

[sup8pdct]

I think Modsel signal is wrong. It is controlled from the Z80 card so the z80 can use a 1066 ram card. I believe the atari controls said line so the z80 can access ram or the 6502 can access same ram, not both at once.

I tend to believe schematic that is around then that card. It could be a very early prototype.

Next problem is the bios, was there one written?

[reifsnyderb]

6 hours ago, sup8pdct said:

I think Modsel signal is wrong. It is controlled from the Z80 card so the z80 can use a 1066 ram card. I believe the atari controls said line so the z80 can access ram or the 6502 can access same ram, not both at once.

I tend to believe schematic that is around then that card. It could be a very early prototype.

Next problem is the bios, was there one written?

!modsel is tied to the Z-80 set/reset toggle via signal 10-1.  It's not too clear in the schematic. 

 

You have brought my attention to the clock signal, though, as the clock signal would currently be going to the card no matter whether the Z-80 was in control of the RAM or not.  That isn't right.  I'll find a way to fix it.  The clock signal, to the RAM, is one of the lines that came out from under U-14 as well.

 

Here's the schematics of the CPM card, with the clock, and the 64k RAM card:

 

CPM_version with clock.pdf

 

64k RAM card.pdf

 

 

 

Edit to add:  Looking further, I also need an open collector output to the !modsel pin. 

2nd edit:  R/W and the clock to the 64k card all need to either have open collector or tri-state outputs as well.

 

 

 

Edited March 4 by reifsnyderb

[reifsnyderb]

10 hours ago, sup8pdct said:

Next problem is the bios, was there one written?

There isn't any bios that has been found.  I believe the Atari would load it by bank switching in a bank, on the 64k RAM card, and load it then.

 

After doing a little research into CPM, it appears it should be easy to port.  My thought is that if some sort of 8 bit protocol were to be created to go between the Z-80's I/O and the Atari while leveraging the Atari's CIO system that getting CP/M to work shouldn't be that bad.

Edited March 4 by reifsnyderb

[reifsnyderb]

Here's an updated schematic.  I think the open-collector issues are all resolved.   🙂

 

CPM.pdf

[+kheller2]

Remember that there are two CPM source disks that were released by the Museum, that I think were for the stand alone device, but might be useful for the 1090.

 

[reifsnyderb]

1 hour ago, kheller2 said:

Remember that there are two CPM source disks that were released by the Museum, that I think were for the stand alone device, but might be useful for the 1090.

 

I just checked the internet archive.  There is a character ROM, BIOS ROM, and a document.  No mention of source disks, unfortunately.

 

The document is of interest to give some clues as to the intentions.

 

sweetpea_prodspec.pdf

Edited March 5 by reifsnyderb

[reifsnyderb]

Here's a video about porting CP/M:

 

 

Here's a couple links to Z-80 schematic examples I was using:

 

http://searle.x10host.com/z80/SimpleZ80.html

 

http://www.z80.info/z80test0.htm

 

 

 

 

 

Edited March 5 by reifsnyderb

[scitari]

Yes, please

[+mytek]

22 hours ago, reifsnyderb said:

New rendering...

 

Any reason not to have the ground plane extend over the entire front side of the board? It would seem to provide both a better heat sink (although what you have is likely more than sufficient) and it certainly would greatly minimize any ground bounce and/or noise. Just a thought

[reifsnyderb]

7 minutes ago, mytek said:

Any reason not to have the ground plane extend over the entire front side of the board? It would seem to provide both a better heat sink (although what you have is likely more than sufficient) and it certainly would greatly minimize any ground bounce and/or noise. Just a thought

It's mainly because I am trying to re-create the original board.  Unfortunately, it appears the original board in the pictures would have never worked, though.  But there is enough information in the pictures to try to re-create what was the intent of the board.

Edited March 5 by reifsnyderb

[reifsnyderb]

Front rendering....

 

 

 

Aren't there any Z-80 Guru's available?

 

 

[+kheller2]

I'm not sure what this is then.  It was from the 1090 section.

 

AtariCPM-disks.zip

[_The Doctor__]

35 minutes ago, kheller2 said:

I'm not sure what this is then.  It was from the 1090 section.

 

AtariCPM-disks.zip 528.23 kB · 2 downloads

Reading the Text Files at disk 2 at the end describes quite a bit, also giving the names of who wrote what utility etc.

There are now names of people to try and contact if still living and are amenable.

[reifsnyderb]

1 hour ago, kheller2 said:

I'm not sure what this is then.  It was from the 1090 section.

 

AtariCPM-disks.zip 528.23 kB · 3 downloads

Thanks!  Where did you get those from?  Could you post a link or instructions to get there?  Is there anything else there?  Thanks!

Edited March 5 by reifsnyderb

[+kheller2]

The MacOS info says it was downloaded via a direct link to the atarimuseum from a post on facebook that Curt did.  Probably when he was trying to get SweetPea working.

 

Update, found the post on FB. BTW, FB also had pics of the 1060 sweat pea system from 1986 with Atari CORP markings.

 

 

Atari Museum

Curt Vendel  · November 8, 2019  · 

 

Hi.. I'm looking for some help. I have Data General MV4000 Restores from Atari HCD mainframe from 1984. There are two Atari CP/M disk directories.

I need someone to help me make some Atari format bootable diskettes. Judging by the dates of May/June 1984, these would most likely be the boot disks for the Atari 1066 CP/M cards for the Atari 1090 Expansion box.

I'm posting up a link to the ZIP file with the 2 directories, if someone is willing to delve into these and make me some ATR's that I could put onto some physical disks to attempt to get the 1066 CP/M card to boot, that would be fantastic and I will post a credit to you on the Atari 1066 page. thanks.

[reifsnyderb]

16 minutes ago, kheller2 said:

The MacOS info says it was downloaded via a direct link to the atarimuseum from a post on facebook that Curt did.  Probably when he was trying to get SweetPea working.

I've been all through the internet archive of the atarimuseum and can't find them.

[reifsnyderb]

Found a video, on facebook, with the CP/M card and other info.  This may be the CP/M card that is in the pictures as it appears to have the traces backwards as well. 

 

https://www.facebook.com/curt.vendel/videos/10219952385062686?idorvanity=105586892805903

 

 

[+kheller2]

This was the link: http://www.atarimuseum.com/computers/8bits/xl/xlperipherals/1090/Z80 Board/AtariCPM-disks.zip 

[reifsnyderb]

5 minutes ago, kheller2 said:

This was the link: http://www.atarimuseum.com/computers/8bits/xl/xlperipherals/1090/Z80 Board/AtariCPM-disks.zip 

Thanks!  I just checked the page and there was no link on the page to the .zip file.  That's probably why I never got it.  I wonder how many other files are in the same situation.

[Geister]

If the Z-80 card is so broken, why not look into a working Z-80 kit that can run CPM and see if it can be adapted.  

MY impression is that the 1066 was designed to work with the 80 column card for output and a memory card for ram. There's none on the 1066 card.  Why not start with a self-contained Z-80 system that has static ram and a video out circuit.  Something like this http://cpuville.com/Code/CPM.html for CPM and this for the base system. http://cpuville.com/Kits/Z80-kits-home.html

 

Edited March 6 by Geister

[+kheller2]

Or just use an Indus GT drive for CPM. 
 

I’d rather like to see the SPC card done. But that’s another dead “useless” device.  

[reifsnyderb]

1 hour ago, Geister said:

If the Z-80 card is so broken, why not look into a working Z-80 kit that can run CPM and see if it can be adapted.  

MY impression is that the 1066 was designed to work with the 80 column card for output and a memory card for ram. There's none on the 1066 card.  Why not start with a self-contained Z-80 system that has static ram and a video out circuit.  Something like this http://cpuville.com/Code/CPM.html for CPM and this for the base system. http://cpuville.com/Kits/Z80-kits-home.html

 

I was thinking of getting this working first....more for reasons of having a replica.  Then, I could use what I've learned and port everything over to a Z-80 card with SRAM built in.  The current 80 column video cards can be used either with or without a Z-80 card.

[reifsnyderb]

32 minutes ago, kheller2 said:

Or just use an Indus GT drive for CPM. 
 

I’d rather like to see the SPC card done. But that’s another dead “useless” device.  

I think there's enough information available to create an SPC card.  The biggest issue I saw was the need for RIOT chips.