Been messing with my Epson ActionNote 4SLC/33. The floppy had flaked out on and off, not sure why, anyway, when it's working it doesn't seem to read disks as well as a 486 board does on a desktop floppy drive. They're old disks, but desktop one has less errors and can get past most of them with a retry or two. Anyway, so I was trying various things, reading in non-turbo, reading with cache off, screaming out the window because it went "AWOL" again for a couple of boots... So I came to try two other drives, a slightly different version TEAC than the one that was in there, and 15 years younger, and a "this century" Panasonic JU-226 ... now while they were smoother and didn't seem to drop out yet, read performance might only have been a tad better but didn't approach desktop error tolerance.

Disks that I had formatted and written in the last 5 years were no problem. Guess I need to get an LS-120 machine back up for superb error recovery and rewrite all the old fading ones.

Anyway, I think it is going to go back together with the newer TEAC, but it requires a bit of vandalism, plastic removed to clear the eject button. Also oddly it's a couple mm narrower, so needs some padding to sit in the cradle right.

Next up, contemplating whether I will "ruin" the memory board that's in there by trying to solder a bunch more DRAM to it. Wondering whether I want an FPU in it to warm it up with another couple of watts and make it fast enough to be too slow to play anything that's 486DX required.

I gave my original 1983 Breadbin (250407 model) a new case. I went for the breadbin gray case, and used a gold coloured custom case badge.
Found the case brand new from Individual computers in Germany.
They still sell them untill stock runs out, now the original injection molds are at Commodore in USA.

BitWrangler wrote on Today, 03:01:

.... The floppy had flaked out on and off, not sure why ...

Happens with age of the material holding the magnetic media to the flexible back.
This will only get worse, and it is also now very susceptible to mechanical stress, even reading it can cause further damage, let along seeking back and forth during retries.

When recovering old media, I use ImageDisk which will let me control much more of what it does, can "keep" bad sectors, and most importantly can combing the .IMD produced by multiple read attempts into a single image with all the "good" sectors you were able to read in any of the attempts.

Also, when recovering really old/damaged media during the various read attempts, I still sometimes experiment with slight misalignment slight (sometimes tracks can be a bit better at one of the edges), slight changes to head pressure (usually lightening it, but there have been cases where I had to go slightly harder - never a good idea as it will accelerate wear of already "flakey" media)

The surfaces appear in good shape visually on most of the disks in question. I was also reading them better on the XT clamshell which has a full height (Officially 1/3 of full bay height for 5.25, whereas a full height 3.5 HDD is 2/3 height or 2 normal 3.5 bays.) floppy. I am fairly sure most of these disks are good, but the actual intensity of the recorded data has faded. Anyway, I am more struck by the difference between the low profile drives and the desktop size drives, revealed by these "just so" amount of faded disks. Thinking it's something about head construction, they're not as thick for sure, maybe have to use finer wire, maybe there's nano-amps lost in resistance that impede pickup strength.

edit: I should mention that some of the disks that read perfectly are actually from the same batch of disks of some that don't. They were either a couple of blank leftovers from the box of 50 I bought in 199x or I re-used disks that had nothing important on. The difference is only that I wrote stuff to them in the last few years, vs back in the 1990s... of course I could have had a crap drive back in the 1990s but the difference is also happening on PD catalog disks and software originals that I did not write.

BitWrangler wrote on Today, 16:24:

The surfaces appear in good shape visually on most of the disks in question.

That's good - Often with age deteriorating disks you can see "lines" showing where the actual tracks (where the head touches) are.

I was also reading them better on the XT clamshell which has a full height (Officially 1/3 of full bay height for 5.25, whereas a full height 3.5 HDD is 2/3 height or 2 normal 3.5 bays.) floppy. I am fairly sure most of these disks are good, but the actual intensity of the recorded data has faded. Anyway, I am more struck by the difference between the low profile drives and the desktop size drives, revealed by these "just so" amount of faded disks. Thinking it's something about head construction, they're not as thick for sure, maybe have to use finer wire, maybe there's nano-amps lost in resistance that impede pickup strength.

Yeah, I've never had a good luck with the super-thin (laptop) drives as i do with normal size desktop drives ... but YMMV with any drive!
Which being up another suggestion .. try reading on different drives - one drive may be able to read a marginal disk while another can't.
Just make sure all drives you use have spotlessly clean heads - bits of "crap" sticking to drive heads are probably the main cause of physical damage.
(and "of course" I recommend ImageDisk "clean heads" function, as it will "scrub" the heads back and forth over a cleaning disk, which IME works *MUCH* better than just DIR or any other means of accessing the drive to make it spin)

Btw, it's unlikely the magnetic domains will just "fade" enough to be a problem, most common degradation of existing recordings are storage with stray magnetic fields, use in drives which have (or are in) an unwanted magnetic field, and of course re-writing sectors (change to files, attributes etc.) while accessing from a "bad" drive.

edit: I should mention that some of the disks that read perfectly are actually from the same batch of disks of some that don't. They were either a couple of blank leftovers from the box of 50 I bought in 199x or I re-used disks that had nothing important on. The difference is only that I wrote stuff to them in the last few years, vs back in the 1990s... of course I could have had a crap drive back in the 1990s but the difference is also happening on PD catalog disks and software originals that I did not write.

Not uncommon at all, lots of things affect individual disks - amount of use, drives used and their head & write circuitry condition.