Recovering a bad floppy disk?

[tt]

Any advice for methods for recovering a floppy disk that is no longer readable? I came across an old disk that says it has some files, likely from my first hard drive. It was likely written to with a Mac Plus and stored indoors in a cabinet with some cassette tapes. I tried reading the disk in a Mac Plus and an SE/30, but it is not readable (both machines can read other disks). The disk is a 2DD 720k type, so I wonder if this made the disk unreliable for long term storage in the 800k Mac format. I have some 880k disks that have been stored in much less ideal conditions that I can at least mount on the desktop and see the contents, so I thought there was a good chance of being able to see what is on this disk.

 

[LaPorta]

720k PC is 800k Mac due to formatting differences, so no issue. So, one thing you can try is Norton Utilities and see if there is any way to salvage files or repair the disk.

 

[Daniël]

It's 1.44MB HD disks that do not work reliably as 800K DD disks, much like how they won't work well as 720K DD disks either. Track layout and magnetic properties of the media are different, which is why writing them as DD disks either won't work, or even after a bulk erase won't be stored reliably (think of this as recording a high bias cassette in a recorder not capable of biasing beyond normal bias tapes).

Other than recovery utilities of that era, the only other option is to build or find someone with a raw flux floppy reader like the Greaseweazle, to obtain a magnetic flux readout of the disk that may be able to be pieced back together into a working image.

 

[s_pupp]

As a computer consultant in college, when Norton Utilities failed to rescue a disk, I used to use Fedit (File Edit) to extract word processing documents. There’s a lot of piecing together involved, but it saved several papers from oblivion.

 

[GRudolf94]

Depending on how much you want to archive that disk: a $20-30 Greaseweazle, plus a PC floppy drive, plus FluxEngine software, plus physically cleaning the media before attempting a read.

As long as the disk isn't damaged that will give you an image. I recommend you read more about the Greaseweazle beforehand though, should you want to follow the idea.

 

[tt]

Are there any native 68k apps that archive a disk similar to a Greaseweazle with the original hardware? I tried running Norton Utilities and a couple other tools, but they want the disk to be unlocked. I am hesitant to unlock the disk and potentially ruin the disk further. It would be great if an app could scan the disk and build an image file, but I am not sure if it could be repaired as an image file.

 

[GRudolf94]

No, because IWM/SWIM decodes read data inside of itself before presenting it to the system. Greaseweazle (and similar solutions) can dump a disk as raw magnetic flux.

 

[bigmessowires]

For a 68K on-computer solution, I think you'll need to explore the path that @s_pupp mentioned. Use a tool (maybe FEdit) or a low-level disk copy tool like Copy II Mac to read every sector on the disk. One or more sectors will probably fail to read, but with some manual effort you may be able to piece together the rest, especially if it's text documents. Here's what you can probably expect to find on an 800K floppy:

sectors 0 and 1: Boot info, you can ignore this
sector 2: The master boot record, there should be a duplicate copy of this at sector 1598 (decimal). You can probably resconstruct this if necessary.
sector 3: The volume bitmap, each bit indicates whether a sector is used or unused. For read-only use, I don't think this info matters.
sectors 4-15 decimal: The extents overflow B-tree. This probably isn't used, unless the files on your floppy disk are heavily fragmented.
sectors 16-27 decimal: The catalog B-tree. This stores the info about what files are on the disk, and their names, locations, and sizes. If these sectors are bad, then you're in for a rough time.
sectors 28-1597: file data
sector 1598: duplicate MBR
sector 1599: unused

The locations of the volume bitmap, extents overflow file, and catalog file can theoretically be different, and all are specified in the MBR. But this seems to be where the Mac's normal 800K floppy format routine puts them.

If you can make an image of the disk, I can look at it and see what I can figure out. I recently made some Python scripts to analyze disk images at this level and attempt to extract files.

 

[tt]

@bigmessowires Thanks for this detailed info on the possible structure of the floppy and offer to help extract files from the disk!

I'll see if Copy II Mac can read the disk. I may also end up trying a Greasweasel, but not sure if it is worth the potential rabbit hole of getting into that hardware. But I recently got a random assortment of floppy disks so who knows.