Isoproply is fine. Someone in another thread was trying to recover a floppy with mould. Iso worked.

Personally I'd try another erq appropriate drive and pc.

Yeow. 1000 disks would stress even the best drives. My suggestion is to actually keep the floppy drive out of the computer case. Remove the drive's top cover, so you can easily access the reading heads.

Get your isopropyl ready with qtips. As you read your disks, stop to clean the drive heads with alcohol every 10-15 disks (depending on how dirty the disks are). This will increase your chances of successful reads.

For imaging full disks, I'd go with IMGDisk first, or VGACopy second. But in your case, the disks don't seem to need imaging if all they contain is user files. Those can be just copied out to a folder on the HDD with "copy a:\*.* C:\Disk493" . If a read error occurs, try retrying like a handful of times. If still fails, either clean the drive head or the disk surface (using alcohol, yes). You can revert to IMGDisk/VGACopy to try and read and tell exactly where on the disk surface the problem is, so you can locate it and clean it directly.

As the author of ImageDisk, I am kinda partial to that one 😀

The good thing is that if you have problems with it, you can easily
reach me here!

ImageDisk uses the PC Floppy Controller which means a few things...

1) The disks must be formatted with in a way the the PC FDC can handle,
this is a Nec765 compatible device, and HardSector disks (very rare) and
certain non-standard soft sector formats (looking at you Apple) won't work.

2) You want an AT+ .. XT does not do high-density which uses a different
data rate and spindle speed ... some fairly rare/unususal low density
formats use data encoding which need these.

3) PC floppy controllers are NOT all alike. Many (especially on newer
systems) implement only the minimum needed to read PC format disks...

I provide a program called TESTFDC which will evaluate what your machines
FDC can do - If you are planning to read disks from many different systems
it's well worth the time to find/test as many systems as you can to find
one that has more capability.

Some Add-In cards have better FDCs than some mainboards.

4) ImageDisk runs under DOS because it accesses that FDC hardware directly
and it some non-standard ways... DOS lets you "have" the machine, most
modern systens are too restrictive to this without a "world of hurt" in
special drivers and permissions - IMD also needs very stable timing, not
possible under modern OSs. I do have ways to easily move images to/from
DOS and modern systems.

As someone who has read/archiven many disks, I agree completely with the
previous poster about mounting the drive outside the machine... Not only
does this make the drive easier to clean, but much easier to change (eg:
HD for DD) .. these drive spin at differnt speeds and have different head
widths... you may find disks you can read on one type but not the other.

Re: cleaning ... some drives make it hard to reach the head, and a Q-tip
may not fit ... I've used just a strip of paper, get it under the head,
close the drive (you may have to hold it partially closed to get the heads
to touch the strip without so much pressure it tears) and move it back and
forth inder the head.

I've also made "cleaning disks" (you can buy them, but on many disks you
may go through) - sacrifice and old disk, cut very edge on non drive side
to remove diskette media ... lay the media on a piece of paper and
trace it inside and out so you can carefully cut one out of the
paper to put back in the jacket ... you'll need "the right" paper to
stand up, but you can get several cleanings for each piece, them make
another.

Good luck ... btw: if you'd like to see a way I automater the reading of
over 1000 DVDs, look at "Daves Old Computers" -> "DVD Robot" (near bottom)

DaveDDS wrote on 2025-05-04, 14:10:

I've used just a strip of paper, get it under the head,
close the drive (you may have to hold it partially closed to get the heads to touch the strip without so much pressure it tears) and move it back and forth inder the head.

Nice one! I prefer that technique over the cloth-like media that cleaning disks typically have. Cleaning disks have a potentioal to get caught by the drive head and wreak havoc. A more solid media like a paper or a thin carton is safer, as long as you don't close the heads all the way, which is possible to achieve with 5.25" drives but not 3.5" drives... (unless you know of a trick ti do that as well 🙂)

I was talking more about 5.25" disks, which I've found to be the ones
more people want to archive as least in the realm of "old computers" (and
yeah, I do realize the any type of floppy is on an "old" system these days)

For a 3.5" drive, you will at least have to take the cover off.
I've seen a few different designs, some you can manually press
the head down part way without locking, some you can make a jig to
reduce spring pressure, and in some I've had to use a bit of light
cardboard/thin paper folded/rolled-flattened to make thicker/springy
and work it between the heads in raised position.

This is all of course assuming you can get at the heads... In some
really tricky drives I've made thin plastic "disks" (think margarine
tub lid) with paper wrapped around the center (inside to outjust a few
times the width of a diskette slot and move it in/out of the "open" drive...
Trick it to make it not lay flat at the end which will go under the head.
(sgsin, roll/flatten a piece under, and tape in place.

Btw, another recommendation about IMD - I usually use set the drive to "NONE"
in BIOS - so the BIOS/OS won't see/initialize it or do anything "odd" to it.
I have run into a system where this actually disabled FDC access to the
drive, but it all others - since IMD talks to the hardware directly, it
still can use the drive.

PS: Sorry for many typo's - been travelling this past month and accessing Vogons on phone/small tablet...
back to "normal" soon!

If you want to go insane, attempt to archive 1000 floppy disks.

I've been working on this project for about 5 days now and I've successfully read less than a dozen discs.

Between multiple hardware failures, faulty discs, faulty discs fouling out the read heads damaging other discs, etc its a never ending nightmare.

Yeah... nah.

Unless it is unique, it's probably faster to download it from someone else who took the time.

myne wrote on 2025-05-11, 13:32:

Yeah... nah.

Unless it is unique, it's probably faster to download it from someone else who took the time.

Alot of this stuff seems to be unique, otherwise unpreserved software though.

And SO MANY of these disks are refusing to read at all. Most of them still throw unformatted/unreadable errors and refuse to even list the file index even tested across multiple drives. Tools that claim to be able to read disks with bad index tracks report every single sector as bad. Thats not possible, there would have to be visible physical damage to the disk. Again, this lot of disks has gone from a 90 percent survival rate, to a 10 percent survival rate just being packed in a box. They werent stored that badly.

I am borderline enraged at this point. This doesn't make any sense.

Whats worse is many times a disk will show its file index once, then refuse to read again. Disk drives I've just cleaned thoroughly are leaving visible marks on disks. I can't even tell anymore if on any given disk the disk is bad, the drivehead is contaminated, etc. How am I supposed to do my due diligence to technological archealogy and make sure each and every disk is truly unsalvageable or contains nothing of use before I discard them if I can't even get a reliable baseline? A common occurrence is:

* disk wont read
* try disk that did read fine as a test
* won't read, test disk destroyed
* clean drive
* test original disk, original disk destroyed
* drive is recontaminated by the original disk that was destroyed due to the drive being dirty.
* clean drive
* try new disk
* new disk is bad, surface disintegrates and contaminates drive

this is literally impossible to continue as a workflow

I once received a pack of ~100 5.25" floppies.. Tried a few and they all weren't reading properly and started to bust my drives. At that point I understood that all the disks were subjected to the same conditions, and thus they will mostly all fail the same. I cleaned a couple and they worked, so I could move to mass-production phase:

Over the course of a few days, every time I go into a passive meeting (working from home), I grab a few disks and clean them up. Several days later and the whole bunch was done. After that, no more damage to the drive. Most disks read fine, but a handful required a second pass. In the end the pack was fully restored, and I use them daily for various retro tasks.

Moral of the story, if one or two disks of the pack damaged your drive, don't risk the rest. Try to clean a disk or two and see if they come back to life and stop damaging the drive. If they still damage the drive after cleaning, then the entire disk set is probably disintegrated and belong to the garbage. Drives are much more expensive than disks.. at least the 5.25" drives...

I think your drive is faulty.
Again, find a period correct drive and PC with a classic floppy cable.
Plenty of LGA775 era boards still had the classic 34pin header.

Sounds like maybe they were somewhere hot or humid for a while?
I was going through some coverdisks from magazines a while back that haven't been stored too well but 95% of them read just fine. There were some bad ones and I found some odd issues like one disk the spinner had become detached from the media which caused really odd behaviour. They were stored in a cardboard box thats lived inside for several years then outside for a summer, but they were all taped together, that is just an English summer though, not all that hot but the storage they're in was hitting 45c on hot days.
Some disks got discoloured from the heat but none of the disks got warped which probably helped their survival.

Though, I've also had disks that I can't recover at all where poor storage has lead to there being permanent warping on the disk surface itself in one spot, which makes the whole disk useless. The driver disks for a parallel CD drive I was trying to use were unusable from that and I couldn't recover it. Trying to iron the disk flat just resulted in the data being lost (curie point?).
If you want to check whether perhaps that's happened to some of yours, place the disk face down and slide back the cover then use a flathead screwdriver to rotate the spinner and look for any non-uniform surface features.

Since you're seeing the disks get damaged, it might be the media or it might be the drive? Narrowing down where the problem is might help.
If you hear intermittent rubbing coming from the drive then the disks are probably warped. If the plastic housing of the floppy disk is no longer flat then that could put the disk media too close to the heads during rotation.

Though perhaps it's just the one drive? I think more than once I've consigned a laptop FDD drive to the bin where it must've been knocked and the adjustments were off so putting a good disk in, the damaged drive scraped off material and ruined the disks.
Maybe check with the secondary drive to see if you get better results or better sound indications that reads are going okay, I haven't seen good 3.5" drives mark the disk surface unless the disks are badly warped.

Personally I use my LS120 superdisk drive to read floppies since it's so fast at it. But if I know I'm testing a disk that looks suspect in that it's warped, has marks on the disk surface or made a bad noise in the LS120 - I pull out a spare Teac FD05 USB floppy drive to test further.
Both drives I know are good, but it's far easier and safer to clean the heads on a regular floppy drive than an LS120 drive.

* What are the best methods for cleaning floppy disks I suspect might be dirty? Would a q-tip an isopropyl alcohol damage the media surface?

It may be worth checking out some amiga forums for information on this, I recall them developing some tools to clean the media itself since people's game disks might get dirty / mouldy from poor storage. There are also VHS / tape recovery people that developed tools to clean mold off of tapes so recover video

TheAbandonwareGuy wrote on 2025-05-13, 02:43:

this is literally impossible to continue as a workflow

I have more experience with 5.25" floppies but I dumped quite a few 3.5" ones as well. What particular drive are you using for this? Late Samsungs (newer than 2000 or so) are poorly made, price was everything. Preferably stick to drives from mid to late '90, after cleaning of course. Pretty much all of them have similar head loading mechanism, where the upper head "snaps" onto the surface. That might be the initial damaging point that results in excessive shedding if the head tends to hit with one corner first. These are rounded but not all that much.

That being said these floppies could be all but gone if they were stored together in very poor conditions. Take one that's already damaged, try scraping the magnetic material with a piece of rounded plastic or something of the sort. Start with very light pressure and keep trying harder to see at which point it starts to go. This will tell you if the media is bad, or it's something to do with the drive.
You can try to "bake" some of these floppies a bit. Not in literal sense, do not let them go beyond 45C or so. But keeping them warm, in dry place with preferably really low humidity and some airflow might help them dry and become more stable. It's worth a try but it takes time. Quite frankly you should be taping the metal shutter open and rotating them a bit every day, for a few days, to get anywhere near uniform drying. I've seen people cut 5.25" and 8" floppies out of the envelope to try, before moving to another clean envelope for just that one read. I've never done that myself and certainly not to 3.5" media.

And again I will mention Greaseweazle, which I use for this purpose. I only do PC dumping with media that's good and the content is not rare. Because sometimes you do get only one or two attempts before things fall apart and flux dump can be processed in software afterwards. Often the software decoding is doing much better job than hardware, and you can play with PLL parameters or try some manual bit forcing or reconstruction.

I've had pretty good luck with TEAC drives... also for DD (not HD) the original Shugart SA-450 can be quite
useful - you would want to make it external - partly because it's full-height, but mostly because you can adapt
the board mount to hold it "out of the way" so you can very easily clean the heads (It also doesn't "snap" the
upper head down the way some drives do).

When I ran my company (DDS) and shipped my software on diskette, I had a "Victrory" autoloader which I
used to make dozens of diskettes at a time - something like that *could* be used to auto-read stacks of disks,
but likely not all that helpful with old media as you would have to remove the drive (easy) and clean the heads
often.

Just a thought but most magnetic materials really dislike temperature extremes, extreme cold and heat can demagnetize media like floppy disks, it would also hasten the deterioration of that material which when hit by the drive head essentially rubs off and fouls it up. (Once its lose in the drive it can remagnetize and make a royal mess and be difficult to fully remove).

I would suspect any disk that has been left sitting in such extremes let alone ones left like that for a decade or more and never get it near any drive I wanted to use later, even with cleaning that drive will still be full of disk media from the damaged disks.

All of this can be made worse if the disks are of ...suspect quality .. there were some really shitty disks produced back in the day using subpar magnetic media, 2 - 3 decades of storage would not have helped their quality at all. (3.5 disks were hit hard by shit quality and super cheap production)

You need a few good tested drives to archive lots disks, never rely on just one. Keep a few known good factory disks as test disks but read them only when needed.
Recently tried to save/image Act! v3.0 (Win3.1/95) floppies from a never opened package, not one disk was readable and do not know their history of storage. The media looked fine. Luckily they did not damage the drive.
Best guess is they got demagnetized over time...

Horun wrote on 2025-05-13, 13:54:

Best guess is they got demagnetized over time...

Some 3.5" drives can switch to 360rpm - there are odd formats, about 1.2M, that require this. Dump these floppies in flux mode at 360rpm. You see the signal level is a function of magnetic field change over time. You can't control the field but you can control the time 😀 This will raise the signal level a bit, might just be enough to decode it in software. Conversion to different rpm/bitrate is trivial, Greaseweazle tools already support some of it and if not hacking together a converter in Python would not be that difficult.

Deunan wrote on 2025-05-14, 11:01:

Some 3.5" drives can switch to 360rpm - there are odd formats...

I've not seen much in 3.5 ... but I've run into a number of older non-PC formats on 5.25
which needed to be able to change speed...

As part of the ImageDisk package, I include information on modifying a Panasonic 5.25"
HD drive (had good lock with those too) to put a switch on the front of the drive to
change to speed to 300rpm (HDs are normally 360)