Re: CBM-900 floppy disk format/encoding

From: Groepaz <>
Date: Sun, 25 Dec 2011 06:53:36 +0200
Message-Id: <>
On Saturday 29 October 2011, you wrote:
> Sorry for my late answer,

same here, i kinda forgot about it, sorry :)

> But I do have the raw track files - I haven't had the time to look into
> the format but it looks pretty raw to me - and the contents looks quite
> systematic if you ask me - as if the kryoflux actually got something out
> of the drive event though that I'm pretty sure that the drive used with
> the Kryoflux is the wrong TPI (96) where should be 100 TPI.

at least the first few tracks should be readable.... and then when stepping 
further, you will again get valid data (but for different tracks than the head 
stepped to). atleast when i tried reading 8250 disks using catweasel and a 
96tpi drive, i could read a fair amount of tracks correctly

> Additional info: I did have time to fire up a CBM-8296 and CBM-8250
> drive - both in working contition. Couldn't read any dir from the
> CBM-900 disks with the CBM-8296. From the start I didn't expect the
> CBM-900 disks to be CBM-DOS GCR format. The CBM-8250 returned DOS error
> code 23 (checksum error in data block) stating that the right
> track/sector have been read into memory but that the checksum on the
> data is wrong. Right now I don't really know what to conclude from this...

this is really interesting, it actually hints on that the disk uses GCR, and 
also a very similar format. (else you should get "sync not found" or something 
like that). infact you should be very close, since syncing and also sector 
headers are working :)

so mmh.. i'd try the following: i wrote something like this for the 1541 
loooong ago, maybe it can be done for the 8250 too:

- write a small basic program that reads one sector from disk (and prints it 
to screen somehow so you can check wether it is what you expect), make sure 
that it works with a "valid" disk obviously
- now add error checking, and if a read error occurs, read the data from the 
drive memory (it will still be available in one of the drive "buffer pages") 
and print that to screen.

that way you should at least be able to check if the data you get somehow 
makes sense (maybe you can see fragments of text in programs, or the 
directory, whatever)... and if it does, then it probably takes someone who 
knows enough about the 8250 to program it to read these disks :)

as for decoding them from a kryoflux dump, i would ask whoever makes the 
(presumably existing?) decoder for d64 images - it should be fairly easy to 
modify (just needs a couple different parameters) and while doing that it 
should also be very easy to see what makes this format different from the dos 
used by the 8250.

that said, if you can part with one or two of these disks, i'd be interested 
in them (since i will pick up work on catweasel software in the near future) 


Der gesunde Menschenverstand ist eine Sammlung von Vorurteilen, die man bis 
zum achtzehnten Lebensjahr erworben hat.  
<Albert Einstein>

       Message was sent through the cbm-hackers mailing list
Received on 2011-12-25 05:00:12

Archive generated by hypermail 2.2.0.