I have a couple older PCs dedicated to servicing the really archaic, fun machines. These computers are old enough to have floppy disk controllers and real serial and parallel ports so you can hook up all the various cables and floppy drives needed to get software and data, mostly in the form of disk images, to and from the various ’80s classics.
One is a Pentium 4-ish Celeron running Ubuntu that I can use for Drivewire and ADTPro. Another is an old 333MHz Slot 1 Celeron running Windows 98. A fair bit of the software used to support classic computers doesn’t work under newer operating systems that prevent mere user applications from poking their grubby little fingers into the hardware registers. Windows 98 is pretty permissive in that respect. Also, some of these programs are just old, having been written to support the first big crop of emulators back in the ’90s. The old copy of “Amiga Forever” I got back at the tail end of the 20th Century came with Amiga Explorer. I guess I could get a new version, but then I’d need yet another computer to support the classics. The Windows 98 box also runs Star Commander for writing disk images to real Commodore 64 floppies. (Maybe someday I’ll look into OpenCBM.) Another important feature of this particular computer is that its floppy drive controller will work at single density, according to testfdc, which many controllers on PC hardware won’t do. This is necessary for reading and writing floppy disks for the TRS-80 Model I, and probably some early CP/M disk formats.
So, to that end, this computer is fitted with a half-height 5.25″ 360K floppy drive. I also use it to reformat salvaged floppy disks. Among the few virtues of MS-DOS and Windows 98 is the fact that they easily format both sides and all 40 tracks of a 360K disk and report any bad sectors when finished. So after I back up anything that needs to be preserved from ancient floppy disks, I can format them in Windows 98 and sort them into “formats with no errors”, “formats with errors”, and “very bad” categories. I’ve been entering these statistics into a spreadsheet which I’d eventually like to put on a wiki that others could contribute to. Though floppy disks have a generally well-deserved reputation for failure, I’ve been surprised at the high percentage of disks that have not failed over the past three decades or more, even under less than archival conditions.
This past week I was sifting through some Apple 2 disks, imaging interesting ones with ADTPro, and reformatting/testing them on the Windows 98 machine. I got them all done, and entered in the spreadsheet. Then I dug up a box of floppies that I think I had already formatted, but which hadn’t been labeled or recorded in the spreadsheet. I decided I would sample a few disks to see if they formatted without errors, and if they did, I would conclude that I had already determined they were all good.
I picked one out and put it in the drive, inspecting it first to make sure it wasn’t dirty with mold. It sounded awful and failed to format. I was a bit surprised, because it had looked so nice and clean, and I was fairly sure these disks had checked good not too long ago. When I pulled the disk out, I noticed that the surface was scratched all to heck. Crap. The disk head must have gotten dirty from the disks I had put in it before and just ruined this one.
So I put a little alcohol on a cleaning disk and popped it in the drive. It too sounded even more awful than the cleaning disks usually sound, and when I took it out, the center hole looked a little bent out of shape. Foolishly, I tried another disk. Again, it went through the motions, but complained that it couldn’t write track 0. Sure enough, it was also scratched up.
I removed the drive and took off its shielding to find that a little metal cover over the top disk head had come loose and was running amok in the drive, methodically ruining any floppy it could come into contact with. Circled in blue you can see the cover. The red circle shows where it belonged.
The drive is a Mitsumi D503V, and if you have one like it, I would advise you to preemptively do the repair I did on it. The little shield over the disk head had been held on by three little foam pads with adhesive on the top and bottom. These pads degraded over time, and while the sticky was still sticky, the foam came apart and the cover fell off into the works. I probably could have just left the thing off. I leave the covers off of computers all the time. But I figured there was some reasonable reasoning that went on when this thing was designed that justified cutting up another little piece of sheet metal and affixing it with a little air gap above the upper disk head. It probably just keeps things out of the head assembly, until it gets tired and becomes things itself. But I thought I could stick it on a little better and maybe it would last another quarter century before falling off again, when the floppy disks it scratches up are even more scarce.
So I cut three little rectangles of egg carton paper pulp material as spacers, and mixed up some epoxy. I cleaned the remaining foam and adhesive off the cover and head assembly with my fingernail, making sure not to let it crumble into the disk head. I put three dabs of epoxy in the appropriate spots on the head assembly, set the egg carton stand-offs in their glue drops with tweezers, put a little more epoxy on the tops, and set the cover back in place. I let it sit for a few hours, then reinstalled it in the machine.
I tried formatting another disk and this time it complained that it failed to write the directory. What? It was bed time for the kids, which meant my day was over too, but I figured the drive was ruined. Maybe the loose piece of metal had wrecked one of the disk heads in a way that wasn’t immediately obvious when I had it apart. Quickly I ran up to the attic and found a box of disk drives. I located another very similar drive with the same model number but a discolored tan faceplate that would look slightly less out of place against the beige mid-tower case than the black one had, and figured I’d try it out the next day.
When I next got some time to work on the problem, I decided it would be a good idea to check out the cover on the second drive, just to avoid a repeat performance and more ruined floppies. Sure enough, a slight tug with my fingernail pulled the shield off on that one too. So I performed the epoxy egg carton surgery on it as well. While I had them both out, I decided to compare the disk heads to see if I could see any difference between them. The comparison didn’t uncover anything that would make me think the first drive had been damaged. But the heads did look a little darker than the ones from the attic.
I put a little isopropyl on a cotton swab and wiped it across the heads. Even though I had run the cleaning disk just two disks before, the swab came back with big black streaks. Maybe that was what got scraped off the disks that got scratched. But after seeing how obviously dirty the heads were now, regardless of how they got that way, I decided I’d better try it again.
Aha, it was working now! After giving it a good test, I replaced it with the beige-ish drive, and things are happier in this basement.
8 Little Bits 