Did you tried VGA-Copy?

The most important thing in the condition of the diskette media surface - if it still looks good
(not worn or showing flecks of missing coating), chances are the diskettes are
probably still readable.

Very important to make sure whatever drives you try and read on are very clean,
it's all to easy to gouge up old media with even a very tiny pit of a spur on the head...

I can't see for sure in your pic... does it have a small ring around the center hole?
This would indicate that it's most likely a 360k diskette... lack of a ring usually means
1.2m disks. Make sure you are using the right type of drive.

and if it is 360, try both drive types -- 1.2 has thinner tracks which might see
less "edge noise" .. or if you've are already using an 80-track drive, a 40-track
might see more of the flux ...

Also try several different drives (making sure each is very clean first) slight
alignment differences can be significant when reading old/damaged media.

You can use IMDs Alignment/Test to see how readable the disk is in various places.
- if you get more errors on the inner(or outer) tracks - look closely at those for even
minor visible damage.

If you are able to perform partial reads on different drives and are getting
different "good" sectors - you can combine these into a more complete image.

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

Pieszka wrote on 2025-02-27, 18:32:

I received from a friend of mine some old floppy disks containing AutoCAD release 11 with polish language. (...) It's shame that such a rare software cannot be preserved.

So first of all, at least tape over the write protection holes. That still won't protect the floppy from badly behaving drive during power-on (or off) brownouts so do not leave the floppy in the drive. And if you tried to read those using any Win9x you've already introduced changes to the FAT structures. I really wish people would first read about data preservation before "experimenting", which usually will only degrade the data further or even kills the floppies. What did you even use for the cleaning?

Anyway, if the sector headers can't be found then you can't read the sectors, period. In fact sector data might be good but if the header is glitched then no software will help you. The only way is to do a flux copy and work with that. In fact copies, plural, as there might be weak bits that will require multiple passes. Some bits might be gone completly. So if you really care about this get yourself a Greaseweazle. Or if you don't want to spend the money, find someone nearby who already has one. Note, if you are going to send the floppies, via post or courier service (including Paczkomat), pack it well. And I mean WELL - assume people will approach the box with strong magnets. Or don't bother at all. And preferably wait for spring so that there are no high temperature variations during day and night during transit.

And seriously, if you want to preserve software start with reading, not doing.

analog_programmer wrote on 2025-02-27, 20:30:

Did you tried VGA-Copy?

No, I didn't. I read a few sources on floppies and how they work beforehand to not mess something up and found that ImageDisk and DskImage are good DOS software for this job (I'm running FreeDOS to be precise; I don't want Windows to mess something up, and besides, I don't think Windows allows such low-level access to floppy controllers as DOS does, but I might be wrong). That said, I've now tried VGA-Copy, but it's the same. I'm attaching a photo of the result.

DaveDDS wrote on 2025-02-27, 20:47:

The most important thing in the condition of the diskette media surface - if it still looks good (not worn or showing flecks of […]
Show full quote

The most important thing in the condition of the diskette media surface - if it still looks good
(not worn or showing flecks of missing coating), chances are the diskettes are
probably still readable.

Very important to make sure whatever drives you try and read on are very clean,
it's all to easy to gouge up old media with even a very tiny pit of a spur on the head...

Well... now it gets interesting. I cleaned the floppy (and the drive) with IPA and a fiber optic cleaning cloth on my first try. But now looking again at the floppy, I see some streaks—probably IPA made it invisible for a while, and I overlooked it. It's a little bit concerning, because if IPA didn't remove it, what else could? To be honest, I cleaned it very gently because I don't want to destroy or scratch the magnetic disk.

DaveDDS wrote on 2025-02-27, 20:47:

I can't see for sure in your pic... does it have a small ring around the center hole? This would indicate that it's most likely […]
Show full quote

I can't see for sure in your pic... does it have a small ring around the center hole?
This would indicate that it's most likely a 360k diskette... lack of a ring usually means
1.2m disks. Make sure you are using the right type of drive.

and if it is 360, try both drive types -- 1.2 has thinner tracks which might see
less "edge noise" .. or if you've are already using an 80-track drive, a 40-track
might see more of the flux ...

Also try several different drives (making sure each is very clean first) slight
alignment differences can be significant when reading old/damaged media.

You can use IMDs Alignment/Test to see how readable the disk is in various places.
- if you get more errors on the inner(or outer) tracks - look closely at those for even
minor visible damage.

If you are able to perform partial reads on different drives and are getting
different "good" sectors - you can combine these into a more complete image.

No, it doesn't have a small ring, so it probably isn't a 360k floppy. It's also written on the label "1.2M". I tried IMDs Alignment/Test, and it correctly detected all tracks; no issues there, just when dumping the data, it throws errors. Also looked for visible damages but found nothing (of course, besides these streaks I've mentioned earlier, but I don't think they are permanent—I hope).

I don't have any other drives, so I'm hoping it can be done on my unit (I've tested it already with a completely different set of floppies, and it dumped them just fine—both ImageDisk and DskImage), but if not, I will search for another one.

That's how it is with stuff recovered from school. It was a technical school with an electronics profile and very small soldering classrooms, so the box was in pretty bad shape and smelled like soldering flux (fortunately, it was easy to clean). I hoped that maybe the floppy disks survived, but I guess they are in pretty bad shape too (if even IPA didn't help).

Pieszka wrote on 2025-02-28, 10:23:

I see some streaks—probably IPA made it invisible for a while, and I overlooked it. It's a little bit concerning, because if IPA didn't remove it, what else could?

Looks quite like a fungi. The mold can grow on a floppy surface in damp conditions.

It's a 1.2MB disk as VGACopy has determined.
The error pattern indicates that it's not totally lost. The most terrible thing would be if an entire track is gone, which would be an indicator that the drive head scratched the disk.
In your case the errors are concentrated around a specific region of sectors on both sides of the disk media. This seems to match that mold you show in the other photo.
Mold can be cleaned with alcohol, or even soap water. I would use a Qtip, dip it in alcohol, then gently wipe the mold away, dry the surface area, then give the disk a few more tries.
If that doesn't work, try soapy water, but also make sure you dry it well.
In all cases, we're talking about a tiny spot of liquid just to wet the qtip a little. Then dry the area using the other side of the qtip stick.

Do not rub strongly or repeatedly against the surface. I found that it could cause the magnetism to get messed up.. meaning the data will be totally destroyed.

Pieszka wrote on 2025-02-28, 10:23:

No, I didn't. I read a few sources on floppies and how they work beforehand to not mess something up and found that ImageDisk and DskImage are good DOS software for this job (I'm running FreeDOS to be precise; I don't want Windows to mess something up, and besides, I don't think Windows allows such low-level access to floppy controllers as DOS does, but I might be wrong). That said, I've now tried VGA-Copy, but it's the same. I'm attaching a photo of the result.

VGA-Copy is low level floppy access DOS software. If it can't recover the errors (not caused by physical damage), than nothing will help here.

I've never used VGA copy but IMD is also "low level access", it communicates directly with the FDC and bypasses DOS and the BIOS. I've had good luck getting difficult disks to read with IMD.

Can you at least get a directory listing?
The majority of the files are likely to be identical to English.
So, if it fails on one or 2 files, find the English copy and compare the file sizes. You can probably rebuild it unless you're extremely unlucky and it just happens to be the polish file.

wbahnassi wrote on 2025-02-28, 10:52:

It's a 1.2MB disk as VGACopy has determined. The error pattern indicates that it's not totally lost. The most terrible thing wou […]
Show full quote

It's a 1.2MB disk as VGACopy has determined.
The error pattern indicates that it's not totally lost. The most terrible thing would be if an entire track is gone, which would be an indicator that the drive head scratched the disk.
In your case the errors are concentrated around a specific region of sectors on both sides of the disk media. This seems to match that mold you show in the other photo.
Mold can be cleaned with alcohol, or even soap water. I would use a Qtip, dip it in alcohol, then gently wipe the mold away, dry the surface area, then give the disk a few more tries.
If that doesn't work, try soapy water, but also make sure you dry it well.
In all cases, we're talking about a tiny spot of liquid just to wet the qtip a little. Then dry the area using the other side of the qtip stick.

Do not rub strongly or repeatedly against the surface. I found that it could cause the magnetism to get messed up.. meaning the data will be totally destroyed.

And... it worked. 😁

IPA couldn't clean the mold off, so I used soapy water (then IPA to be safe), and it came off very easily. After cleaning every single bit of mold, VGACopy dumped the image just like that. Many thanks for your help. 😁

Now, I need to repeat it like 8 times for every floppy (I hope not), but knowing what to do will make it easier. Are there any other things I need to keep in mind while repeating the whole process?

myne wrote on 2025-02-28, 13:58:

Can you at least get a directory listing?
The majority of the files are likely to be identical to English.
So, if it fails on one or 2 files, find the English copy and compare the file sizes. You can probably rebuild it unless you're extremely unlucky and it just happens to be the polish file.

ImageDisk also has an option to "keep bad sectors" - which means the file will be corrupted, but
it may only be a byte or two - for text files you can often tell what it's "supposed" to be,
executables are trickyer, but if you manage to get the same program in another language, a bit of
work with a disassembler could give you the information you need to resurrect it

(ok - it would be a lot of work with a disassembler - it really depends on how much you want to
recover those exact files)

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

Excellent work!
My routine for floppy disks I acquire nowadays is to get the disk under sunlight and visually inspect the surface on both side via the access hole, I turn the disk media using two fingers pushing against the ring from the inside.

If visual inspection shows any irregularities in reflectivity then a clean up is due. If all is consistent, I smell the surface. If it smells moldy/wet, then a clean up is due also.

The above is important before you insert the disk in any drive, otherwise you are risking losing both the drive and the disk data for good.

The initial cleanup is done by IPA. This usually makes it safe to insert into the drive, but as you saw.. it doesn't remove all mold. Then I use IMGDisk to make an initial read. If there's an identifiable pattern of errors, then it's usually a moldy spot that requires further cleaning. Sometimes it's worthwhile to retry the read several times. The head can sometimes do some of the remaining cleaning for you if the remnants aren't too stubborn.

Pieszka wrote on 2025-02-28, 14:10:

And... it worked. 😁

The attachment IMG20250228145321.jpg is no longer available

IPA couldn't clean the mold off, so I used soapy water (then IPA to be safe), and it came off very easily. After cleaning every single bit of mold, VGACopy dumped the image just like that. Many thanks for your help. 😁

From my practice VGA-Copy have no match in magnetically damaged floppy disks restoration, but I didn't expect to work with this mold damaged diskette. Glad to see it helped you to save your diskette image with the rare old software.

And here we are. Successfully dumped all 9 floppy disks. Now this ancient piece of software can be preserved (although without the hardware key it probably won't work). Nevertheless, many thanks for your help again. It was a very long and painful journey, but totally worth it. 😁

I'll probably upload it somewhere for anyone to use; for now, I need some rest. :>

Pieszka wrote on 2025-02-28, 20:21:

And here we are. Successfully dumped all 9 floppy disks. ...

Awesome and congrats. Glad we could help!

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

Pieszka wrote on 2025-02-28, 20:21:

And here we are. Successfully dumped all 9 floppy disks. Now this ancient piece of software can be preserved (although without the hardware key it probably won't work). Nevertheless, many thanks for your help again. It was a very long and painful journey, but totally worth it. 😁

The attachment Screenshot 2025-02-28 210515.png is no longer available

I'll probably upload it somewhere for anyone to use; for now, I need some rest. :>

A hundred years ago I sat for about 3 days learning just enough assembly and softice to crack xp's terminal services to server.
Essentially it boiled down to one bit of one byte changing jump equal to jump not equal. 74h to 75h or vice versa. Plus a bunch of reg keys, but that was easy.

It's likely anyone with real skill will achieve the same result with this software in minutes.