OSI-300 Replica

[ChristopherB]

Over the summer I worked on a smaller replica of the OSI-300 computer trainer. Using images provided by others, I was able to draw out the schematic, bread board it, and then have boards made. I have put details on my website, including the schematics.

http://randomvariations.com/category/osi-300-trainer/

I have a number of full kits, partials (need chips,) board + switch kits, or fully assembled systems for shipment within the US for those who are interested, just send me a private message.

[dave]

Hi Christopher,

This is a nifty project--I just noticed it, and am surprised nobody commented on it. The next step is to get Zork running on it

Dave

[ChristopherB]

Dave,

While the 300 might be a bit too difficult without lots of additional circuitry, I think that the 400 series would be a fairly straightforward operation with a 6502 processor, although it would need ram added to the base 400, and some access to mass storage, such as the 400 series disk controller. Adding an SD or Compact Flash card is fairly straightforward with existing examples in 6502 code as well.

If I had a 400 series system, I would definitely work on it as a project. The Apple I version would make for a good starting point.

[dave]

I was kidding about the 300, but only partly--I did see your demo of Zork on the Apple 1 at VCF East 9.1, and was very impressed. It would be awesome to get a v3 z-machine up and running for the OSI C1/2/4 platforms.

The 400 series CPU boards are pretty rare, but the 500 systems (C2/4/8) are not too rare, and C1/Superboard II/600 systems are not hard to come by at all, especially with Grant's 600/610 clone PCBs.

I think it could run on an 8" (or 3.5") floppy system. With a bit more work, it could probably be made to run on a dual-floppy 5.25" system.

[ChristopherB]

Dave,

A direct copy of the Apple II version interpreter might be possible if the memory map on the OSI systems allows for programs to run in the 0x0800 address space, and includes enough contiguous memory to load the static memory portion of the game file, the interpreter, a small buffer for the dynamic memory portion and the disk routines.

The Apple II disassembly at https://www.brouhaha.com/~eric/if/zip/ indicates that the Apple II version starts at 0x0800 and uses up to 0x1C00 bytes for the interpreter and virtual memory maps. An additional 512 bytes are used by the disk routines. The rest of memory up to 0xC000 is used by the game files. The interpreter uses the zero page and stack space.

My Apple I version consists of a pre-loader which patches the small number of locations which call the I/O routines, my I/O (including disk) routines, and then the core interpreter. This provides for a bit more memory for the loaded game. If the disk routines can be kept in the space between the stack and 0x0800, a 32k system should have enough RAM to run.

Due to the paging system used by the interpreter, the game really needs to fit on one disk, although the saved games could be placed onto a second disk if that feature was implemented.