Atari 1090XL Schematics?

[Bryan]

Hello guys (m/f)

 

Why ISA? The 1090 has its own semi-standard.

 

Mostly because there are cards that can be immediately used and it's a cool reuse of retro tech. To recreate all those functions in 1090 cards will likely mean using a bunch of FPGA's and whatnot. For me, I think it would be cooler to control a CGA or Hercules card.

[Lynxpro]

Build it and then worry about what cards come after.

 

Look at the Apple II enthusiasts over in their Facebook group. People are still making expansion cards for them… Ethernet, WiFi, Bluetooth, RAM, SD Cards, Multiple Joystick Ports, Multiple Mockingboard Sound Cards, you name it.

 

Should a new 1090XL retain the 5 expansion slots as originally designed or go for 8 to match the Apple II?

[evilmoo]

I think one could get a lot of mileage out of just interrupt-driven USB and I2C controllers. A form factor the size of a Raspberry Pi could support this, allow stacked cards, and not consume the desktop real estate a 1090 would.

 

Alternately, if you are ISA obsessed, there's this: http://arstech.com/install/ecom-prodshow/usb2isar.html

[Joey Z]

*BUT* getting everyone to agree on what should be done is like trying to herd kittens! For instance, I would like it if every card followed the ~IBM type standard of including the code for itself in ROM. People making the cards would probably just as soon include a handler loaded from disk so they don't have to add ROM to a cost sensitive project. I mention this not as a criticism but a tip of the hat to people that tackle it.

well, PBI compliant devices already do that. The atari OS was designed, from the very beginning, to allow handlers to be dynamically loaded. The XL/XE expanded on this concept with handler POLLing for SIO devices and PBI math-pack region ROMs for PBI devices. The atari OS is even more flexible in this regard than the option-ROM style of the IBM PC architecture because there is a fairly standard way of accessing devices through the CIO, whereas the IBM PC had INT numbers allocated pretty much at random by the manufacturers of devices (except where a standard defined by IBM already existed) without any real standard interface to access said devices.

[Dropcheck]

The more I investigate this, the more I believe some in Atari were trying for something along the lines of the Apple II with it's ability to insert daughter cards for expandibility. Maybe as grand an idea as what the Amiga 2000 came to be.

 

I found another more detailed specification. This writeup reads like a precurser to the Amiga's autoconfigure feature.

 

Apparently the PBI could support upto 8 devices, with only 1 being active at a time. Sadly, it looks like they were already limiting the number by the time the 1090XL prototype was nearing completion.

[Bryan]

Apparently the PBI could support upto 8 devices, with only 1 being active at a time. Sadly, it looks like they were already limiting the number by the time the 1090XL prototype was nearing completion.

I believe the 8-card limit was because they were addressing using 1 bit per device. Using a couple 74LS138's, you could probably have many more.

[Joey Z]

I believe the 8-card limit was because they were addressing using 1 bit per device. Using a couple 74LS138's, you could probably have many more.

Yeah, and what would you do with more than 8 anyway? Not to mention past that, you're going to be having potentially major bus loading and timing issues anyway.

[+Stephen]

The more I investigate this, the more I believe some in Atari were trying for something along the lines of the Apple II with it's ability to insert daughter cards for expandibility. Maybe as grand an idea as what the Amiga 2000 came to be.

 

I found another more detailed specification. This writeup reads like a precurser to the Amiga's autoconfigure feature.

 

Apparently the PBI could support upto 8 devices, with only 1 being active at a time. Sadly, it looks like they were already limiting the number by the time the 1090XL prototype was nearing completion.

Yes - the Atari 8-bit engineer knew at the time the Apple expandability was a great feature. They were saddled with bullshit arcane FCC restrictions, and this was their way around it.

[Dropcheck]

Build it and then worry about what cards come after.

 

Look at the Apple II enthusiasts over in their Facebook group. People are still making expansion cards for them… Ethernet, WiFi, Bluetooth, RAM, SD Cards, Multiple Joystick Ports, Multiple Mockingboard Sound Cards, you name it.

 

Should a new 1090XL retain the 5 expansion slots as originally designed or go for 8 to match the Apple II?

 

I think 8 would be the max. I agree bus loading and timing could be an issue with more. Especially with external connection between the Atari and the expansion box.

[pixelmischief]

If this could result in something like a 130XE with an 8 slot expansion bus in an ATX case, I would sell organs to get one. Err...yeah.

[Dropcheck]

If this could result in something like a 130XE with an 8 slot expansion bus in an ATX case, I would sell organs to get one. Err...yeah.

 

I'm afraid that is way beyond what I could possibly do. I'm thinking much smaller. Along the lines of an external expansion box that connects to an Atari 800XL/130XE PBI/ECI bus connector.

[kenames99]

so, does anyone even have photos of the bottom of the board? that would help.

 

Ken

[venom4728a]

I planned to build ny own expansion box, have been purchasing parts for awhile. I am housing it in a ST Mega File case so it will match the XE. I have a 120vac to 5 and 12vdc board. Coming out of the Atari I was thinking a small PCB the allowed me to plug into the ECI and change to a 50 pin SCSI for example. Then on the back of the ST Mega file case we have another 50pin scsi. just inside the case we switch to two ribbon cables with idc edge connectors to match the back of the XE. I was planning to cut the 5vdc supplied by the Atari and use the 5vdc from my power supply. The power supply also has the 12vdc needed for hdd or whatever. The case would only be able to hold a few cards, because they will be laying flat, but I only have a few cards anyways Maybe an MIO, Karen Maxidrive and am IDE+2

 

I do not know how to design a PCB to go from atari to scsi. I figured eventually I would teach myself that and be able to finish the project.

 

Robert

[Dropcheck]

Found some more documentation:

 

Parallel Bus Revealed part 1

 

Parallel Bus Revealed part 2

 

Parallel Bus Revealed part 3

 

Parallel Bus Revealed part 4

 

The Atari Parallel Bus Interface part 1

 

The Atari Parallel Bus Interface part 2

 

The Atari Parallel Bus Interface part 3

 

The Atari Parallel Bus Interface: PC Bridge

 

XL Expansion Connector

Edited May 14, 2016 by Dropcheck

[sup8pdct]

a couple of pics

 

http://s1325.photobucket.com/user/sup8pdct/media/IMGP6215_zpsyvblk1oa.jpg.html[/url]

 

http://s1325.photobucket.com/user/sup8pdct/media/IMGP6214_zpsbu6kzdzq.jpg.html[/url]

Ignore red wire. I added it to try and power it up.

James

Edited May 14, 2016 by sup8pdct

[+kheller2]

In the grand scheme of things, exactly which 1090 document/revision did Atari decide to go with in the board pictured? I ask this because I thought there were several 1090 documents explaining slightly different visions for the product.

Edited May 14, 2016 by kheller2

[Bryan]

It looks like Atari was going to power it with another AC transformer brick, but I'd just delete the whole power supply part of the board and use a multi-voltage supply. Then it becomes much simpler to build.

[Dropcheck]

In the grand scheme of things, exactly which 1090 document/revision did Atari decide to go with in the board pictured? I ask this because I thought there were several 1090 documents explaining slightly different visions for the product.

 

Kinda hard to tell since none of the documents I've found have more than general requirements. The board bottom says revision 2A.

 

Now if we could get Curt to release the official schematics he says he had or the Eagle schematic he was going to make, we might be able to see differences.

 

One thing I did notice is that in the docs they indicate that 2 RS232C and 1 Centronics type port were going to be included. No such chips or connectors on these pics.

Edited May 15, 2016 by Dropcheck

[Dropcheck]

It looks like Atari was going to power it with another AC transformer brick, but I'd just delete the whole power supply part of the board and use a multi-voltage supply. Then it becomes much simpler to build.

 

 

Maybe no...... If it is one of the standard power supply bricks, say the computer power supply for the 800XL or 130XE then that's one less power supply most people would have to buy. Many already have an extra one laying around.

[Bryan]

Maybe no...... If it is one of the standard power supply bricks, say the computer power supply for the 800XL or 130XE then that's one less power supply most people would have to buy. Many already have an extra one laying around.

 

What I'm saying is it's obvious from the board. Those 4 giant diodes (CR1-4) make up a bridge rectifier and then there's lots of caps and voltage regulators downstream. This thing runs on AC. It has test points for 5, 12, -12 and 10V (which looks to be unregulated).

 

The 800XL/130XE brick is 5V DC.

[kenames99]

hey sup8pdct, very nice of you to post these photos. thank you.

 

Ken

[Dropcheck]

 

What I'm saying is it's obvious from the board. Those 4 giant diodes (CR1-4) make up a bridge rectifier and then there's lots of caps and voltage regulators downstream. This thing runs on AC. It has test points for 5, 12, -12 and 10V (which looks to be unregulated).

 

The 800XL/130XE brick is 5V DC.

 

 

Duh...... Missed that.

Edited May 15, 2016 by Dropcheck

[sup8pdct]

The power in plug is larger so a 1050 one wont plug in, hence the red wire to try and power it up.

The 10V supply is the main supply for each card. Each card has a voltage regulator for its own use.

 

James

[sup8pdct]

did a little work

Found what each pin does on the interface slots.

 

pin function comment1 comment2

 

1 10v main power supply unregulated

2 10v main power supply unregulated

3 Extsel unbuffered

4 audio unbuffered

5 casinh

6 gnd

7 ref

8 MPD unbuffered

9 RDY unbuffered

10 irq unbuffered

11 ac signal via 220 resistor from power in not switched

12 reset

13 pin 37 pbi unbuffered -12v ?

14 Enable pin of address buff Pulled low via resistor (enable)

15 r/w dir of data transceiver from PBI

16 -12v very low current draw

17 A15

18 Gnd

19 A14

20 02 clock

21 A13

22 gnd

23 A12

24 +5v very low current draw

25 A11

26 all slots coms between slots?

27 A10

28 39 pbi unbuffered +12v ?

29 A9

30 47 PBI unbuffered (signal only) +5v 600XL only

31 A8

32 enable of data transceiver pulled high via resistor (disabled)

33 A7

34 D7

35 A6

36 D6

37 A5

38 D5

39 A4

40 D4

41 A3

42 D3

43 A2

44 D2

45 A1

46 D1

47 A0

48 D0

49 Gnd PBI only (pin 1) detects if computer connected? Other PBI GNDs to interface GND

50 +12V very low current draw

 

Unless otherwise stated, all signals are buffered by either LS244 (address plus others) or LS645 (data)

Found an older document with the +/- 12V at computer end (not on 600XL or 800XL). Would like to see what 1400/1450XLD had.

 

James

 

P.s formatted well from excel when 1st done. posting was a different look. very hurried edit. bed time for me now

Edited May 15, 2016 by sup8pdct

[Dropcheck]

Thank you sup8pdct!

 

I think that gives me enough information to start a preliminary schematic. My eyes were beginning to cross following those traces on the back where the slots are.

 

I was trying to think why when the computer can support 8 different PBI devices, the 1090XL only had 5 slots. The preliminary specs called for 2 RS232C and 1 centronics ports. But the pics you posted had no ports or circuitry for them. I wonder if they were left off as a cost cutting measure, assuming that some third party would fill the gap.

Edited May 15, 2016 by Dropcheck