As irritating as it is to say, it's not unlikely that all your disks are bad.

I ordered 10 3.5" 720 kB floppies from a seller on Ebay, and I checked them in a modern computer with Debian and a floppy drive. 7 of the 10 disks the drive couldn't detect at all, 1 had bad sectors, and 2 were fine. 70% total failure rate in other words (of a small selection I know) and 5.25" floppies are older and more delicate than 720s.

Could be worth getting a strong neodymium magnet and wiping it across one of your floppies. Make sure it's a strong magnet. Basically what you're trying to do is remagnetise the media (google magnetic hysteresis for more details). "Wipe" the magnet slowly over the disk, turn the disk over, wipe over other side. Repeat a few times and then try the disk again.

This is of course assuming there' s nothing on the disk you want, as this will completely destroy all data/sectors/anything else on the disks.

Vaudane wrote:

As irritating as it is to say, it's not unlikely that all your disks are bad. […]
Show full quote

As irritating as it is to say, it's not unlikely that all your disks are bad.

I ordered 10 3.5" 720 kB floppies from a seller on Ebay, and I checked them in a modern computer with Debian and a floppy drive. 7 of the 10 disks the drive couldn't detect at all, 1 had bad sectors, and 2 were fine. 70% total failure rate in other words (of a small selection I know) and 5.25" floppies are older and more delicate than 720s.

Could be worth getting a strong neodymium magnet and wiping it across one of your floppies. Make sure it's a strong magnet. Basically what you're trying to do is remagnetise the media (google magnetic hysteresis for more details). "Wipe" the magnet slowly over the disk, turn the disk over, wipe over other side. Repeat a few times and then try the disk again.

This is of course assuming there' s nothing on the disk you want, as this will completely destroy all data/sectors/anything else on the disks.

Thanks for the tip with the magnet, I will try it in my garage 😀
I got a similar situation with 3,5" floppy drive (almost 20yo). It asked repeatedly to insert the disk. I cleaned the headers, only new drive replacement solved my problem. So this is why I'm also considering that the headers/something else wore out/went bad.

mcfly wrote:

Or am I missing something?

Try another PC. Pentium machines tend to have floppy controllers that are too weak to drive a specific 5.25" drive properly.
I got one of such a "dead" drive to work once by installing an ISA floppy controller (was a Multi-I/O card).
That beeing said, a more modern 5.25" drive now works in that said PC.
So it migh be hit and miss. Some combinations work, some don't.

360Kb Floppys are more robust if stored correctly because of track density.

i have a much higher failure rate on 1,44Mb and 720Kb Floppy Discs..

5,25" drives die quite regularly, but only some of them are really worth saving. And this drive isn't one of them since it has worm mechanism for heads positioning.
I have quite a number of them dead, but it is just because the plastic worm mechanism gets so stripped it can't move precisely at all and it keeps missng the tracs. The Panasonic drives are extra prone to doing stuff like this. Don't waste your time and money on something that was designed to fail and get some direct driven drive with big, proper stepper motor with metal ribbon stripe drive, like Teac FD-55GFR or Chinon FZ-506.

I made some progress with the drive 😀. It turned out that both detectors (read/write protection and index) were failing due to broken LED diodes (kind of unexpected, perhaps the drive got hit with some overcurrent). This is why the main motor didn't spin, but it is now. Unfortunately the header positioning motor stopped working. I suspect some more failed components - I bet on capacitors, or some cold joints. Still haven't given up on this... I am pretty optimistic with it, setting header position with ImageDisk will be peace of cake if I'll have it operational. I recently fixed like this some old 3,5" alps by alligning the header motor.

Ok. I got it finally working. After applying some hot air on the upper motherboard, the motor finally started working. Some adjustment of the header was also required. It looks like the patience is a virtue - especially when it comes to retro hardware. Now's the time to score some broken 8" floppy drive 😀

Nice work 😀

And while we're on the subject... I have one Teac drive that works fine but won't flush (or detect new media insertion) of the previous floppy contents.

So when I do "DIR A:" it lists the contents. I remove this floppy with another one. Do "DIR A:" again and it lists the same contents. I can try 10 floppies and all will have the contents of the first one. If I try to copy anything from other ones, the error will show - of course. Reboot will "fix" it. Also some PC-Tools reread of A: will show the correct contents and the fastest is "FDFORMAT A:" but abort the real format.

Where is this "media inserted sensor" located on TEAC drives? I'd try to clean it first.

kixs wrote:

Nice work :) […]
Show full quote

Nice work 😀

And while we're on the subject... I have one Teac drive that works fine but won't flush (or detect new media insertion) of the previous floppy contents.

So when I do "DIR A:" it lists the contents. I remove this floppy with another one. Do "DIR A:" again and it lists the same contents. I can try 10 floppies and all will have the contents of the first one. If I try to copy anything from other ones, the error will show - of course. Reboot will "fix" it. Also some PC-Tools reread of A: will show the correct contents and the fastest is "FDFORMAT A:" but abort the real format.

Where is this "media inserted sensor" located on TEAC drives? I'd try to clean it first.

I think I wasn't quite precise as for naming. Those two sensors are responsible for read/write mode detection and index detector. Since there is no more than those two in my panasonic drive, I bet that the crossing the index led beam by a floppy disc is actually doing a media detection. Teac drives I saw so far don't use much wiring like 475-3, it may be required to unscrew the top pcb. The sensor is black shape plastic component with a little hole inside and on the opposite side (on the bottom board) you will find a LED (probably IR). Just use some isopropyl alcohol with a cotton swab to clean this. I would also check the header alignment in imagedisk118 just to be sure that it reads everything ok.

mcfly wrote:

kixs wrote:

Nice work :) […]
Show full quote

Nice work 😀

And while we're on the subject... I have one Teac drive that works fine but won't flush (or detect new media insertion) of the previous floppy contents.

So when I do "DIR A:" it lists the contents. I remove this floppy with another one. Do "DIR A:" again and it lists the same contents. I can try 10 floppies and all will have the contents of the first one. If I try to copy anything from other ones, the error will show - of course. Reboot will "fix" it. Also some PC-Tools reread of A: will show the correct contents and the fastest is "FDFORMAT A:" but abort the real format.

Where is this "media inserted sensor" located on TEAC drives? I'd try to clean it first.

I think I wasn't quite precise as for naming. Those two sensors are responsible for read/write mode detection and index detector. Since there is no more than those two in my panasonic drive, I bet that the crossing the index led beam by a floppy disc is actually doing a media detection. Teac drives I saw so far don't use much wiring like 475-3, it may be required to unscrew the top pcb. The sensor is black shape plastic component with a little hole inside and on the opposite side (on the bottom board) you will find a LED (probably IR). Just use some isopropyl alcohol with a cotton swab to clean this. I would also check the header alignment in imagedisk118 just to be sure that it reads everything ok.

In the meantime I got another Panasonic drive 475 but the 4th revision (475-4) which I dissasembled completely to clean and grease its parts, set up headers and got it working. It seems like there is another sensor which I didn't pay attention to. I found service manual for this newer drive release https://tinyurl.com/475-4-service-manual and it seems like it's a magnetic sensor specificaly for media detection, not an optical as I previously thought. It's located near the top left corner of a disk, when the disk is inserted. Just straigtening some facts in case somebody wanted to look in there.

Do old floppies really die easily? When I started building retro computers couple years ago, I checked all my old floppies from early 90s and every single one of the formatted fine without any bad sectors and worked fine. Well, I had saved branded floppies that I thought were higher quality like 3M, sony, maxell and TDK. Still either floppies survive much longer than I thought or I had extremely good luck with those 25-28 year old floppies.

Agreed. I got some 5,25" floppies from early 90's and all of them work perfectly fine, no bad sectors. But on the other end I got some 3,5"second hand disks (like new) from the end of production (2000ish) Verbatim DataLifePlus with teflon coating (arguably the best of kind), no scratches, but many bad sectors. A couple of 5,25" had a harsh life but were still working and they are worse protected than 3,5". I'm not sure if they did something in the manufacturing process later on that floppy disks from some point in time just don't last that long. It's pathetic even, that 30yo floppies are more durable and even sometimes brand new sealed floppies have so much of defunct pieces. I got a some naive idea of treating those disk with electromagnet to magnetize it better. Will it work? We'll see 😀

Formatting disks with bad sectors many times (20-30 times) worked for me on some disks.

However it happened to me that some disks just destroy themselves when you want to read them (they make a terrible noise when they spin and when you get them off they have circular notches. It can even go to the point the magnetic stuff gets completely removed on some spots ...
And what's worse is that it have the tendency to kill your drive as the heads become really nasty (and cleaning them after reading such disk sometimes isn't enough ,..)

Anybody know a guide for doing head alignment on 5.25'' drives? I have a J.P.N. Corp. DS-53A drive that doesn't want to read or format any disk. I know at least some of the disks are good because they format and read on another drive (a Chinon one) and after having lubricated the rails, cleaned the heads and the index sensor i can't think of any other options.

Nprod wrote:

Anybody know a guide for doing head alignment on 5.25'' drives? I have a J.P.N. Corp. DS-53A drive that doesn't want to read or format any disk. I know at least some of the disks are good because they format and read on another drive (a Chinon one) and after having lubricated the rails, cleaned the heads and the index sensor i can't think of any other options.

Read it all before adjustment.
1. You got to have 2 things: DOS and IMDisk from here http://www.classiccmp.org/dunfield/img/index.htm
If you don't have pure DOS, get FreeDOS. Make a bootable USB pendrive and download SystemRescueCD (it has some console text tools like memtest, mhdd, freedos) and put also unpacked IMD118 files in there. When starting FreeDOS from thumb drive, make sure you pick '0' (zero) to disable all memory managers or it will crash.
2. Make sure that your MB supports 1.2MB 5,25" drives. I tested mine on asus P5W on a quad core cpu - worked fine
3. Unscrew upper shield covering the heads module. Loosen the two big screws that hold step motor in place. You will be moving it back& forth to get a reading. Sometimes, the screws are on the bottom, eg. in NEC drivers. Check with your drive and user manual.
4. Insert a disk and start imd118. Select alignment test. What should interest you are the last two columns which mean: number of ID's that matched the physical cylinder and number of ID's that did not match. 5,25" 1.2MB (double side) are 15 sectors/track/head. There are two heads and it should report 30 0 in last columns. For 3,5" 1.44 is 18 sectors/track/head - should be 36.
5. Basically the procedure is as follows: set a motor position, tighten one screw, then press 0, then 1 to reposition header. If you have stable reading, check other tracks in increments of ten (keys 0-8), if not - repeat previous step. The last track for 1.2MB is 79 so don't be surprised if you see question marks on track 80. On every range it should be 30 0 reading on last column. First column shows current track number. If you for example press zero, first column should be zero, if it is not, then the zero sector sensor should be moved back or forth. Do this after you got stable readings on all ranges.
6. If for some reason you got stable readings on range 0-60 but afterwards it drops (to say 20 and changing), then it would mean that the spring that presses on top header may be worn out. You may try to transplant another one or bend it a little bit to make it press harder. Under no circumstances press header directly with your fingers!!! You may scratch the surface and destroy headers. Do not also change header azimuth on top header (there are usually two small screws for it). If you do it, then get yourself an alignment diskette (if you can find one cheap - let me know) and an oscilloscope 😀
7. If everything is set up, try to read image of the disk, there is an option in imd. And make sure you try several floppy disk and all of them read correctly. One of it could potentially be formatted on a drive with misaligned headers, so better check with more samples.
8. This alignment procedure is obviously no match for oscilloscope, but for home use is absolutely ok.
9. If you have bad luck, headers may already be destroyed before you even try to start this procedure and it won't read a thing, only ? is displayed. Chances are swapping header modules from other similar drive would help
Hope it helps. Also make sure that everything is well lubricated and you should be fine. Don't forget about the second screw before the final assembly.

Edit: tip for FreeDOS. There is a known bug in some versions of format.com. After successful formatting sometimes a message 'driver not found' is displayed. A workaround for it is dir b: before issuing format b: /u command. (provided that b: is your floppy, and a: is freedos ramdisk, when booting from thumb drive/cd)

😳 these are really nice informations !

Would that work with old motherboards or a socket 775 motherboard is required ?

How would that work for a 360k disk drive ?

Yeah, it should work with most old motherboards (and contrary to popular belief, not every onboard floppy controllers are alike). I just happen to have such a board among others and I was interested to check if this will work on more 'modern' hardware. For 360K floppies, you would probably need to adjust number of tracks in IMD (if it does not detect) from 80 to 40 and It would report probably 16 or 18 and 0 on successful read in alignment test, since depending on DD media it is either 8 or 9 sectors/track (according to Wikipedia chart for IBM 5 1/4 format), but I would have to check it. My both 5,25" drives are 1.2MB, however I'm pretty sure they should be a jumper to enable 360k mode only. Either way I have to check this.

Deksor wrote:

😳 these are really nice informations !

Would that work with old motherboards or a socket 775 motherboard is required ?

How would that work for a 360k disk drive ?

Unfortunately, I'm afraid that I won't be able to reproduce it on my HD drive. No matter how I set up the drive: fixed density mode (1M) jumper, 360 5,25" mode set in the Bios - IMD always let me read SD and HD disks, as if it was ignoring everything. On DOS setting 360k driver in Bios causes it to display a usual read failure errors. What's interesting, when I formatted 360k disk in HD drive with format /4 option, during allignment test only track 0 have all cylinder match with their ID's, other show 0 18 in last two columns. Hard to say if it's ok, because I was able to copy/read files from such a floppy in IMD and under DOS. The question is if such formatted floppies could be read in the real 360k drive. No such problem with 1.2M ones. Perhaps I set something wrong in the program or missed some parameter. I do not see myself as an omnipotent expert, just sharing some info possible helpful for someone.

Ok, there is one more thing. When 360K disk is formatted in 1.2M drive, then every two sectors are being written. This is because 360K drive headers are wider. When issuing allign command a 'double step' option should be selected, so it skips an extra track. Then it correctly reports 18 0 on all sectors on DD media (ignore lower values, they show only for a moment before reading stabilizes). Well.. this was one of this rtfm things 😀

To sum up the sought for values are:
for 2HD 1.2M 5,25" 30 0 (15 sectors/track * 2 headers) IMD: set track value to 80
for DS DD 360K 5,25" 18 0 (9 sectors/track * 2 headers) IMD: set track value to 40, and double step on 1.2M HD drive with 360K floppy
for 2HD 1.44M 3.5" 36 0 (18 sectors/track * 2 headers) IMD: set track value to 80
for DD 720K 3,5" 18 0 (9 sectors/track * 2 headers) IMD: set track value to 80

You may also need a translation 250<>300 kbps set up with DS in HD drive. A quote from manual:

12.1.11 Data rate translations
2
3             Data rate translation allows you to try to recreate a disk on a
4             different drive type from that on which it was originally read.
5
6             The most common/useful translations are:
7
8                250 -> 300      = Read on 5.25 DD, write on 5.25 HD **
9                300 -> 250      = Read on 5.25 HD **, write on 5.25 DD
10
11            ** Normally you would also enable Double-Step on the HD drive.
12
13             You can see the transfer  rates  in  an  image  with  the  IMDU
14             utility.