Making bootable disks for the Canon Cat

[ClassicHasClass]

I was able to make, and then hack, disk images for the Canon Cat, Jef Raskin's 1987 "work processor." Despite what Canon sold it as, though, Raskin always intended it as a fully working general purpose computer. It has a 68000 CPU and its ROMs are primarily programmed in tForth, a token-threaded dialect of Forth.

With this hack, disks can now boot and take over the system like other machines (or return to the editor, if you like). It does it by finding the stored register images in the disk image file (the format is Amstrad CPC Enhanced DSK images, which accurately reproduce the disks but are also straightforward to patch and is a well documented format that Greaseweazles can read and write directly) and patching A7 to point to a trampoline routine which runs a tForth payload during the loading process.

Three demos are included:
- A Jef Raskin picture disk, which displays a photo as the preview image, waits for a key, and then lets you read the Wikipedia biography on disk.
- A simple asynchronous terminal program which autoboots or can be run from the Forth prompt.
- A two-disk slideshow. It boots from the first disk, flashes the question mark waiting for the second, displays a Happy Mac (heheh), and then runs through a full image slideshow stored as tracks.

If you don't have a Cat, there are videos as proof of hackery. Everything is on Github, including ready-to-write disk images.

Pretty pictures, bootable floppy disks, and the first Canon Cat demo?

Now that our 1987 Canon Cat is refurbished and ready to go another nine innings or so, it's time to get into the operating system and pull ...

oldvcr.blogspot.com

GitHub - classilla/catbox: Hey, Canon, look what I did to your "work processor."

Hey, Canon, look what I did to your "work processor." - classilla/catbox

github.com

 

[PL212]

This looks neat! I will try on my Cat — as soon as I repair the broken plastic guide on the floppy disk track, which I think your recent blog post detailed how to fix with a piece of metal available at a hardware store…

 

[ClassicHasClass]

Ouch! Hopefully your drive head didn't get caught on anything, which seems to be the way that kills drives.

I've got to try to build Dwight Elvey's floppy converter and plug a Gotek into it one of these days. The Canon drives suck.

 

[eric]

Always a fun read when I see an 'oldvcr' link - thanks for sharing and all you do!

 

[ClassicHasClass]

Aw, thanks! It was fun. There is more to explore on this system.

 

[PL212]

OK I think we are ready to try this...

(That's the right one reinforced; the one on the left had broken off completely and required a different bend in the metal).

^ The green disk light on the right, and "DISK" notification at the bottom of the screen, mean we're back in business...

 

[ClassicHasClass]

Excellent! Please let me know if the examples work. You may be only the second Cat system to try them!

I'll also be interested to hear if your drive has a similar issue with the track 0 switch.

 

[PL212]

I would say we are very nearly there:

The garbledness is different each boot -- not predictable. All the programs (Raskin, Catterm and Slideshow) show the same behavior of both a garbled load screen and then dumping you to the standard word processor (as opposed to the custom code). The behavior seems consistent across cold boots and USE FRONT-DISK, but a "cold" command from Forth skips any attempt at a loading screen as far as I can tell.

Here's the situation on my end:

1) I'm using (new old stock) 400k, single-sided double-density disks. I've verified one of these disks can save and restore text on the Cat normally.
2) My Greaseweazle is connected to a 2.88Mb drive (it's a long story). Is it possible this drive is writing the tracks too "thinly" for a 400k drive?
3) I have not yet recapped the Canon disk drive -- got an order on the way from Mouser to do so.

 

[ClassicHasClass]

I don't think it's a caps problem with the Canon drive (though if you're able it probably wouldn't hurt), but I am suspecting the drive you mastered the floppy with. Can that drive write 720K? I used an off-the-shelf cheap Teac PC drive which can.

Similarly, since you can write your own disks in the Cat and the Cat will read them, image one of those that your Cat will read with the 2.88MB drive and then use that image to make a copy of it (something like "gw read --format=ibm-scan --tracks=c=0-79:h=0 test.edsk" and then "gw write test.edsk"). If the copy fails in the same way, it's almost certainly the drive you're making the disks with.

The dumping into the editor is probably because the divert code embedded in the image never ran. Given that the registers are monkeyed with, I'm surprised it didn't crash, actually.

 

[PL212]

OK, I'm now trying on a new (old stock) TEAC FD-235HF (C829-US), which I believe is similar to you have. I can successfully write 20 tracks to slideshow.edsk and 78 to slideshow_2.edsk. (Does that sound correct?)

From cold boot we go right to the editor, however, and with USE FRONT + DISK we see a brief corrupted snippet of the Macintosh disk screen before dropping into the editor again.

Same pattern with raskin.edsk -- nothing from cold boot, brief distorted image from disk load.

I wonder if perhaps our two drives are differently-aligned, or if I have the infamous Zero-Track problem? I'd be glad to do any diagnostics/adjustments if there are any to try...

 

[ClassicHasClass]

Yes, that's the same drive and that sounds correct for the track counts. Disappointing that it still doesn't work.

Are you able to use that drive to copy the disk the Cat will read and write (using the commands above)? Does the Cat read the copy?

 

[PL212]

I will try that — and another idea I just had, what if I fill a disk with text on my Cat, image it using the Greaseweazle, and send the image to you? If you write it and successfully read it, that would suggest that our two drive are in alignment (at least for general-purpose data I guess.)

 

[ClassicHasClass]

Sure, by all means. I can also see if the tracks line up where they should.