• Please review our updated Terms and Rules here

Can't figure out 360K floppy drive

Ok, a few random thoughts.

Ok, so imagedisk is showing that it is reading a few sectors during the alignment test. That is good, that suggests most of the logic and mechanical stuff is working.

A consistent pattern of 2, 5, 8, and 9 is very odd.

The first thing I would do is try a couple of additional formatted disks and see if this pattern is the same for those. (did anyone mention yet not to use high density 1.2mb 5.25" disks?)

The alignment test should show you the track and head number that the sector it has found identifies itself as. Does this match the track and head you have selected?

As you step the head to different tracks, does this pattern remain consistent?

If you select the other head, is this pattern still the same and consistent?

You might check the spindle hub. Is it gripping the hub properly? As the disk spins, does everything seem centered? Is there any garbage stuck on the hub grip?

You might also try gently pressing down on the top head while it is reading. Does that change anything?

Does the disk seem to be spinning easily? I've seen odd behaviors when a disk is not in the drive flat - such as when a front bezel is crooked or something inside the drive is not quite where it should be and pressing against the disk.

You did say the drive completely failed to format, right?
 
Chuck(G) said:
One has to wonder if the head is making consistent contact with the medium.
Click to expand...
Yes, forcing the heads to go down without a diskette in, the heads press against each other. So, with a diskette, I'd have to assume they both make contact with the diskette.

SomeGuy said:
Ok, a few random thoughts.

Ok, so imagedisk is showing that it is reading a few sectors during the alignment test. That is good, that suggests most of the logic and mechanical stuff is working.

A consistent pattern of 2, 5, 8, and 9 is very odd.

The first thing I would do is try a couple of additional formatted disks and see if this pattern is the same for those. (did anyone mention yet not to use high density 1.2mb 5.25" disks?)

The alignment test should show you the track and head number that the sector it has found identifies itself as. Does this match the track and head you have selected?

As you step the head to different tracks, does this pattern remain consistent?

If you select the other head, is this pattern still the same and consistent?

You might check the spindle hub. Is it gripping the hub properly? As the disk spins, does everything seem centered? Is there any garbage stuck on the hub grip?

You might also try gently pressing down on the top head while it is reading. Does that change anything?

Does the disk seem to be spinning easily? I've seen odd behaviors when a disk is not in the drive flat - such as when a front bezel is crooked or something inside the drive is not quite where it should be and pressing against the disk.

You did say the drive completely failed to format, right?
Click to expand...
I'll answer everything in order.

The pattern itself seems entirely random for freshly formatted and reformatted disks. Reformatting the same diskette that was outputting 2 ,5, 8, and 9 with the same settings (I am using only DSDD diskettes btw), it is now only detecting 3 sectors, 5, 8, and 9.

It's very odd, so I did 5 more trials. I erased, and then formatted for each trial. Here's them all sorted:
OG1: Sectors: 4 (2, 5, 8, 9)
OG2: Sectors: 3 (5, 8, 9)
T1: Sectors: 4 (2, 5, 8, 9)
T2: Sectors: 4 (2, 5, 8, 9)
T3: Sectors: 4 (2, 5, 8, 9)
T4: Sectors: 4 (2, 5, 8, 9)
T5: Sectors: 4 (2, 5, 8, 9)

I've waterboarded this poor diskette 😅

I'm not going to do it to another diskette, as this is a pattern between multiple as well, and I don't feel the need to shorten the life span of the rest of my limited supply.

The heads appear to match when selecting them. I don't know how to verify if they're "correct," but they're both reading stuff.

Stepping, the pattern remains the same. Attempting to Analyze the disk on another cylinder comes up with an error "Unable to determine interleave." The sectors also remain the same.

Mind you, the interleave this entire time has been the IBM standard of 6:1, single step.

Something interesting to note, though, is that Sectors 2, 5, 8, and 9 are always the last 4 sectors ImageDisk adds while formatting, with a sector order of 1, 4, 7, 3, 6, 9, 2, 5, 8. Maybe this has something to do with something?

Switching heads and attempting to Analyze again comes up with the same error, with the same sectors present.

The diskette appears to be centered around the spindle, but it is not centered in its housing, if that matters at all.

Gently laying a finger on the head seems to slow the disk down, so I don't believe I want to try that. The top head should also be touching the diskette fine, as per the test done previously for Chuck.

The spindle hub spins 100% freely when I spin it with my hand, so that shouldn't be the issue. However, inspired by the head touching experiment, I decided to disassemble the spindle support and lube it, as that didn't spin as freely. And oh boy, that sucked. Those little U shaped metal things that go on the rods and bearings to hold them in place SUCK to take on and off. Glad that's over.

And, the results: HOLY F*** IT WORKS!

When I put the previously formatted diskette in, ImageDisk did not like it, and wouldn't even read the amount of sectors. So, I erased the disk, and then formatted it like all previous attempts before it. And it just worked! No errors, no this or that, it just works. Alignment picked up all 9 sectors instantly. Test RPM shows a steady RPM of 299-300 once everything is settled.

BUT, new issue 🙃
DOS could not read the formatted disk. Upon exiting DOS mode, the computer crashed. Reloading, Windows 98 will format the disk, but it will not read it. Instead, it prompts to format the disk again, but will instead fail this time. Going back into ImageDisk, it now displays a single cylinder: 8.

In ImageDisk, erasing the disk, and then formatting it again resulted in the disk not working again. The results are now back to where they were previously; 4 sectors, 2, 5, 8, and 9.

As a last ditch effort, I shut the machine back down, took out the drive, and just sprayed the lithium grease directly into the spindle support.

Booting back up into DOS mode, I ran through the ImageDisk formatting again, and it still doesn't work.

I got to be happy for a moment 🫠

Please mind how different each part of this post was. It was written over a few hours of tinkering, so y'all got to go through my process too by reading this haha.
 
I don't know what you're using for lube, but usually, these things were not intended to use any lube on the guide rails at all. After cleaning, the most I'd use would be a pure silicone spay lube--and very little of it.

But I think you'll concede that the issue appears to be mechanical. How about the disk clamp setup?
 
I really really do not understand this.

So, everything is brokey. I went back to waterboarding the diskette. It still wouldn't work. So, in Alignment test, I tried to Write data, which made the disk completely unreadable.

And then I formatted it again. And it worked...

Why?

This makes no sense.

I inserted another floppy that was formatted with the weird 4 sector issue, and the drive couldn't read it again. Formatting makes that diskette perfectly readable as well.

Exiting DOS, shutting down Windows 98, and then starting everything back up again into ImageDisk, it still works.

Going back into DOS, it can't read the diskette again. Trying to get back into Windows initiates that error again. System crashes. I go back into DOS and then ImageDisk. The diskette is still readable.

Going straight back into Windows, Windows tries to read the diskette, and asks to format it again. I select "360Kb" "Full" "Label: Test" and then press start. It initializes for a little while, and then starts formatting. When finished, I get a results screen with 0 bad sectors. I go back into ImageDisk. Alignment test... reads the disk.

Why is it different? What changed?

Back into Windows, I try to read the diskette again. Windows thinks it isn't formatted again? Why? I format it in Windows again. Same settings, same name. It completes successfully. Again. Windows can read the disk now. WHAT?

I guess everything is good now? I don't know why everything is good now. Nothing changed?

I'm going insane.
 
Still working on the assumption that the drive electronics are good. I would agree with @Chuck(G) that this is looking like a mechanical issue. I came here to recommend that you try to wiggle the head positioning motor to see if it is loose.

It seems you've been at this a while but have you tried doing the format/read back dance with the drive on its side so that nothing is touching the spindle motor.
 
Booted the PC back up today and it can't read properly formatted disks again. Back to the same issue. Again. 😐

I really wanted to just turn the PC on and tell y'all that it just seems to work now, but we can't have nice things.

Chuck(G) said:
Too many posts down from the start, but this is a taut-band positioning system, right? Are you sure that nothing's slipping there?
Click to expand...
No, it's direct drive. There are no belts.
Edit: Idk why I thought you were talking about the spindle motor. I don't know what you mean by taut-band positioning system. Could you please explain?

jxm said:
Still working on the assumption that the drive electronics are good. I would agree with @Chuck(G) that this is looking like a mechanical issue. I came here to recommend that you try to wiggle the head positioning motor to see if it is loose.

It seems you've been at this a while but have you tried doing the format/read back dance with the drive on its side so that nothing is touching the spindle motor.
Click to expand...
The spindle motor is recessed into the housing of the drive, so it shouldn't have the ability to touch anything. It would also make an audible sound, since I've accidentally had it touch the top of the computer before.

I'm a little busy at the time of writing, so I'll get around to opening it up and poking its brain some more later tonight. Perhaps I didn't lube the head rails enough? I don't know what could be making it so temperamental. Any other ideas for things I should try later? I'll try any suggestions y'all have.
 
Wait, you were using Imagedisk to write a format? I don't believe that will write a file system (I think), so it won't be readable in DOS.

Normal DOS disks use an interleave of 1. So the sectors are ordered 1-9. With the sector order you describe, it suddenly makes sense if part of the track is getting overwritten. For example, if you have the wrong sector size, data rate, or gap size selected, that can happen.

It's either that or a drive speed issue. If you can get it to a working point for a moment, focus on using Imagedisk to measure the drive speed.

It sounds to me like the drive may be working fine but there is something wrong with the controller or system - which is why I strongly encouraged sanity testing with a 3.5" drive.

You might try a couple of other disk formatters and see what they report. Perhaps NFORMAT and FormatQM.
 
SomeGuy said:
Wait, you were using Imagedisk to write a format? I don't believe that will write a file system (I think), so it won't be readable in DOS.

Normal DOS disks use an interleave of 1. So the sectors are ordered 1-9. With the sector order you describe, it suddenly makes sense if part of the track is getting overwritten. For example, if you have the wrong sector size, data rate, or gap size selected, that can happen.

It's either that or a drive speed issue. If you can get it to a working point for a moment, focus on using Imagedisk to measure the drive speed.

It sounds to me like the drive may be working fine but there is something wrong with the controller or system - which is why I strongly encouraged sanity testing with a 3.5" drive.

You might try a couple of other disk formatters and see what they report. Perhaps NFORMAT and FormatQM.
Click to expand...
Sorry, I should have included the fact that I made DOS diskette from a .IMD file, which Windows was able to read just fine last night. ImageDisk is also unable to read the DOS disk now, displaying all "?" again.

When it did work yesterday, ImageDisk displayed a consistent 299-300 RPM both times it started miraculously working.

I do, however, have another disk drive now. I may chuck it in the system and see what it does. It's from a whole CP/M system, and has only a single head, so we'll see how that works out later. Crossing my fingers that it isn't a different format to normal 250kbps 360k MFM drives.
 
ElectroArc said:
No, it's direct drive. There are no belts.
Edit: Idk why I thought you were talking about the spindle motor. I don't know what you mean by taut-band positioning system. Could you please explain?
Click to expand...
Sure, you have the stepper positioner attached to a capstan (in the nautical sense). A thin steel fork-shaped band is wrapped around the capstan and secured in the middle to the capstan. The free ends of the band attach to the head carriage. You can see that, as the stepper rotates the capstan, the rotary motion is translated to a linear motion. Here's the seminal patent: https://www.freepatentsonline.com/4992899.pdf
 
Sorry for the radio silence, guys.

I had the idea of updating the BIOS, and that required a 3.5in diskette, which I had to wait on the shipping for. The BIOS updated just fine, surprisingly.

Aaaaaannddd... no dice.

SomeGuy said:
Wait, you were using Imagedisk to write a format? I don't believe that will write a file system (I think), so it won't be readable in DOS.

Normal DOS disks use an interleave of 1. So the sectors are ordered 1-9. With the sector order you describe, it suddenly makes sense if part of the track is getting overwritten. For example, if you have the wrong sector size, data rate, or gap size selected, that can happen.

It's either that or a drive speed issue. If you can get it to a working point for a moment, focus on using Imagedisk to measure the drive speed.

It sounds to me like the drive may be working fine but there is something wrong with the controller or system - which is why I strongly encouraged sanity testing with a 3.5" drive.

You might try a couple of other disk formatters and see what they report. Perhaps NFORMAT and FormatQM.
Click to expand...

The 5.25 FDD still doesn't function properly. However, this does pass your sanity test, right? The 3.5 FDD functioned just fine, even updating the BIOS successfully.

Chuck(G) said:
Sure, you have the stepper positioner attached to a capstan (in the nautical sense). A thin steel fork-shaped band is wrapped around the capstan and secured in the middle to the capstan. The free ends of the band attach to the head carriage. You can see that, as the stepper rotates the capstan, the rotary motion is translated to a linear motion. Here's the seminal patent: https://www.freepatentsonline.com/4992899.pdf
Click to expand...
It would appear that it's as you described. I included some pictures, but they're not very clear. I can't get a good image with it under the circuit board like that.

It doesn't appear to be slipping. The metal band is even screwed into the motor, so I don't think it can slip.

20230606_185142(1).jpg20230606_185146(1).jpg

jxm said:
Still working on the assumption that the drive electronics are good. I would agree with @Chuck(G) that this is looking like a mechanical issue. I came here to recommend that you try to wiggle the head positioning motor to see if it is loose.

It seems you've been at this a while but have you tried doing the format/read back dance with the drive on its side so that nothing is touching the spindle motor.
Click to expand...
Putting the drive on it's side, the results are the same. It still won't read the correctly formatted disk. Reformatting it brings us back to the 4 sector problem. It will still only read the last 4 sectors. I formatted it as 1:1 interleave, so it's reading 6, 7, 8, and 9 as the only sectors this time.
 
ElectroArc said:
I'll have to purchase a printer or something, because the FDD doesn't come with a strobe disc on it. I tried drawing a mark with a sharpie, but that was rather ineffective.
Click to expand...
Go to your local kinko's/mailboxes etc./print shop and make a few, because wrong spindle speed is the only thing left that points to what is going on.

If all of your lighting is high-speed LED and you have no light sources that flash at 60Hz, one trick is to look at the strobe disc through a mirrorless camera set to 1/60th shutter speed, which should show 60 images per second if you're recording to video.
 
So, I got a used printer, and here's the result.

I don't understand how this works. My s22 Ultra can't seem to flash consistently at 60hz, and my OnePlus 3 seems to flash more consistently, but it doesn't work good either.

Using my S22 Ultra's pro camera mode, included are two pictures, one at 1/60 speed and one at 1/500 speed respectively.
20230607_195158(1).jpg20230607_195209(1).jpg
Is the pattern perhaps too large? or do I need to set the camera speed higher? I don't really know what I'm doing 😅
 
That doesn't look right to me. Was expecting to see the bars holding steady, not a fully circular pattern. Then again I am more used to the other pattern shown. And I could be wrong. I am more often than I care to admit. :)
 
If the speed is right, you will see one of the dotted lines seem to almost stand still. Never had to do this to a floppy drive, but have with many turntables.
 
You must log in or register to reply here.
Back
Top