• Please review our updated Terms and Rules here

How to recover data from unrecognized floppies?

I'll add that it's sometimes possible to recover from dropouts (they do occur) on floppies by manually editing the timing intervals between pulses. For example, an MFM disk is encoded with magnetization "flip" intervals of t, 1.5t and 2t So, if you sample (using a reasonably good MCU or say, Catweasel), you can insert a missing "flip" and come up with some readable data. Clearly, this is easiest if you're working with FM data, but it's possible to do this in MFM as well.

Not for the faint of heart, however.
 
A couple other tricks I've used.
If the data is ASCII, there is often a lot of redundancy,
even in code.
Another thing is often people don't build from a clean disk.
The fellow has made a master disk from one that has been
used for several iterations. Many of the unused sectors can
contain the bits and pieces lost.
Easy enough in ASCII text but a little tougher in binary.
I know what Chuck is talking about by flipping bits. I was once
recovering tapes done in Polyphase format ( for a Poly88 box ).
The tape had periodic folds because of poor rewinding.
I'd usually start by flipping bits near the end and work back
until it didn't make sense.
Dwight
 
SomeGuy said:
If we are determined to read the disk, I'd start by physically inspecting it to see if there are any visible scratches, or any junk on the disk. Clean with a q-tip if needed and if possible. But 3.5" disks don't usually get much junk in them.

If the disk looks perfect, then it very well may be one of those cases where re-reading a bunch may eventually increase readability. For standard formatted disks, I often just use WinImage, try to make a full image, and just mash "R" to keep retrying. If you get a good read, be sure to save it. But on disks like this, it would not surprise me too much if the readability actually improves.

I would not use an LSL-120 for that since those have a different kind of head. Since this seems to be a standard 1.44mb disk, probably the USB drive should be sufficient, although a real FDC and drive would be preferable.

If disk images are downloadable from somewhere, you can also use Winimage to compare the disk. Or other tools to compare extracted files. If they are identical up to the read error, you are probably safe just re-writing the image. (Note that Windows sometimes messes with boot sectors, so those may be different).
Click to expand...

It seems the disk did have some kind of build-up in it, I cleaned both sides with a q-tip and alcohol, and WinImage got up to 44% on a real floppy drive before refusing to go any further regardless of how long I held the R button.

Looks like I'll have to download the game since a Kryoflux or Catweasel isn't in my budget.

I'm amazed that a factory sealed game would have build-up in it.
 
It's not a buld-up, but deteriorating binder. I generally treat those disks with a drop or two of cyclomethicone before reading. Isopropanol will just further degrade the binder. It's also a good idea to "bake" floppies at about 55-57C for a few hours before reading-that tends to ameliorate the shedding problem.

Probably more than you wanted to know--but those of us doing media recovery do know some tricks.
 
Fire-Flare said:
I'm amazed that a factory sealed game would have build-up in it.
Click to expand...
It happens all the time. Shrink wrap is not air-tight, so air, pollutants, mildew, and such can slowly get inside. I suspect in some cases even the materials that make the disks themselves (adhesives, decaying plastics, and such) can also leave residue. It is always worse on poorly stored disks, but even well stored disks can suffer from this.
 
It's been well-documented that many (if not eventually, all) binder formulations are unstable over a long time. The audio tape people were the first to run into this and have various approaches; eventually it's permeated well into computer media.

Even shrink-wrapped, new media can do this. If you doubt me, go find a NOS sealed box of Wabash 5.25" disks and give them a try.

All your floppies will eventually be doomed... But then, I suspect, so will most storage, though punched card is pretty good if the stock is acid-free.
 
Bad disks in shrink wrap even happened when the disks and boxes were brand new. I have a copy of MS Office where one disk was defective from the beginning. There was a replacement disk included in the box which took a while to realize when I tried reinstalling it a few years ago. Looks like one duplicating machine broke down and corrupted all its disks for an entire lot.
 
One issue that I've run into that I can't adequately explain is the "evaporation" of data on some white-box floppies after about 30 years. I've never seen this on name-brand floppies.

Note that I don't mean "corruption"--I mean that the flux transitions are simply gone, as evidenced by checking the disk with Kyread (magnetic developer).

About all that I can imagine is that the basic oxide was mis-formulated and never was fully (magnetically) saturated when recorded and simply decayed with time.
 
Chuck(G) said:
All your floppies will eventually be doomed... But then, I suspect, so will most storage, though punched card is pretty good if the stock is acid-free.
Click to expand...

I'm finding with some of the archival work that I do (analog video in various formats, including broadcast), it's not necessarily the medium that will go bad, but rather the machines that can read them. We all have 40-yr-old 5.25" disks that read just fine, but our drives are in varying states of operation...

Chuck(G) said:
Note that I don't mean "corruption"--I mean that the flux transitions are simply gone, as evidenced by checking the disk with Kyread (magnetic developer).
Click to expand...

That happened to me once with a commercial shrinkwrapped item I bought specifically to archive. On reading it, I thought it was bad; I was then utterly shocked to find there was simply nothing there.

Thinking about this some more, I guess there's three possibilities in that case:

  1. The disk had become degaussed at some point by accident
  2. The disk had never been written to in the first place
  3. The disk somehow lost the flux transitions
Occam's Razor points to #1 or #2.
 
No, #3 is most likely. These were from a backup set. I could read perhaps 20% of the sectors there, but nothing on the rest. Normally, I'd suspect storage issues, but these were part of a set stored in a diskette box and all of the name-brand ones were readable. It was just the white-box ones. Floor sweepings being sold as real media?

I have a good stock of disk drives of various formats, so it's not an issue with me. For other media, I work to keep things in good order.

For example, here's a QIC drive capstan refurb after it's gone to goo:

W2dOdTM.jpg
 
What did you rebuild the capstan with?
Dwight

On 5.25 disk, I've had the goo that was used to hold the liner
bleed through and made sticky stuff on the disk.
It made blotches that would drag.
I had to remove then from their envelops to read them.
When using older drive that were belt driven, make sure the
belt is tight and no lumps of belt rubber stuck to the pulleys.
Slipping belts and the lumps changing speed will cause bad
reads.
Dwight
 
4 mm silicone vacuum hose. Like this

I tried a bunch of various rubber hoses, but found them not to be very "round"--i.e., the wall thickness wasn't constant. I had some of this stuff left over from replacing the brittle PVC vacuum hose on my truck. It can be stretched for a tight fit (spray some silicone lube on the inside of the tubing and you can force it onto the bare capstan) and it doesn't slip. Just compliant enough to work right and should last decades.

I've had binder bleeds on 1/2" tape. No amount of running through a tape-cleaning machine will get rid of that. There, cyclomethicone applied with a wiping felt to the tape will give you a few minutes of stick-free operation to get the data.

We live in a perishable world, I'm afraid.
 
Last edited:
I tried a number of different things. Some was too soft and some would take a set.
I didn't try silicon rubber.
The stuff that took the set looked to work but the roller would clamp on the tape
and then 10 or so seconds later when the tape started, there would be a dip
that would trash the data.
Dwight
 
I'll have to fix the machine that it runs on first. I'd wanted to to a backup
on tape but the thing is no longer booting. It doesn't look to be impossible
to fix as the disk drive doesn't spin up. It loos to be a supply issue.
It is a Sparcbook 1. It still comes up to openboot but and reports the disk
not working.
I think the disk is OK and just a supply problem. Still hard to work on
in the laptop machines with no schematic.
Dwight
 
You must log in or register to reply here.
Back
Top