SuperDrive (floppy) maintenance

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
Hi Guys,

On the infamous “zero track alignment” issue with the floppy drives…

I fixed many drives in the past with the most usual “trial and error”, luck-driven and unnerving approach.

This time I opted for something more structured, which has proven to be rather successful (read: not much time spent).

Pre condition: you need to have a “reference drive”, realistically one with factory settings and still working.
You will use it for creating the reference floppy disk.

Drive to fix: cleaned up and lubricated, as per above instructions.

Now, what you have to do is to realign the position of the optical sensor that detects when the pick up head is in the right initial position.

Unfortunately, catching the correlation between changed position of the sensor and the actual position of the pick up is impossible, as there’s not visual aid.

So, the trick is to create a reference system which can drive you in the process.

Put the drive under a top light, means that you will have a shiny line caused by the reflection of the round shaft.

With a fine tip permanent marker, draw a tiny dot, right on top of the shiny line.

Now here’s the key information:

- when you eject the disk, the stepper motor is always receiving the same number of impulses, corresponding to an integer number of revolutions. This means you can make your dot with the marker after the disk has been ejected, which is much easier. You will see when you insert it and the pick up is brought back to the initial position that the mark will get in the exact same angular position.
- If you try to format the disk, you can easily count 10 discrete steps per each revolution. Means that each rotation of the shaft of 36 deg corresponds to one track in the floppy (0.115 mm).

Procedure I have used.

- loosen up the screw that keeps the optical sensor in place.
- Take the sensor all the way back: you will start from there.
- Now the difficult part: tighten the screw just enough to keep the assembly stable, but allow some movement when gently forced, for example when you gently tap the sensor with a screw driver or similar. You want to imperceptibly move it forward, literally one (motor) step at a time.
- Of course, you have to adjust the position after ejecting the disk, then re-insert it and see if the shaft is stopping always in the same angular position or not.
- If not, you should be able to judge by how many angular steps the position of the pick up has changed.
- Proceed one step at time, until a known good floppy is read and mounted.
- In the case of the last drive, I had to allow four rotational steps from the very end for matching the zero track of the good floppy.
 

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
I recognize my English is not good enough for giving a clear explanation of the method… I will see if I can find the time for integrating some notes and some sketches. But yes, the principle is valid for 400k drives too, at least as long as the alignment of the stator assembly has not been messed up.
 
  • Like
Reactions: fehervaria

fehervaria

Tinkerer
Sep 23, 2021
153
160
43
48
North Germany
I recognize my English is not good enough for giving a clear explanation of the method… I will see if I can find the time for integrating some notes and some sketches. But yes, the principle is valid for 400k drives too, at least as long as the alignment of the stator assembly has not been messed up.
Paolo, could you just do some video from the procedure?
I am working on a project - a heavily damaged SE restoration - and the floppy drive lost (just fallen out) one of the screw from the stepper motor's two adjusting screws. The stepper motor was just "hanging" in its place when I "find" it. (But nothing was damaged or bended.)
I reached that state with the head position adjustment the disk drive reads, boots any disk (800K and 1.44MB disks, various original Apple disks, original Microsoft Word install disks, original Lemmings install disks) but can't format or write to the disks. If I try to fine tune the Track-Zero sensor or the stepper motor's shaft position (based Larry Pina's instructions) - it makes the things worst. So, this is my best what I reached "blind".
I really would like to try your way because I've spent many-many hours to make this disk drive alive and I feel I am very close to the final settings.
 
Last edited:
  • Like
Reactions: Paolo B

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
I should still have some disassembled drives around (in theory: not working, but I just hate throwing away that kind of stuff… Doctors would have a lot to say about it, but that’s another story…), so maybe over the next weekend I will try to integrate a drive and see if it can be used for taking some pics.
 
  • Like
Reactions: fehervaria

Volvo242GT

Tinkerer
Feb 7, 2022
307
167
43
Currently Duvall, WA
Here some slides I put together explaining the ”zero track alignment“ procedure.
As far as my experience goes, 100% success rate in recovering “misaligned” drives.
I hope it can help.

Zero Track Alignment
Thanks for posting this. Was able to get my SE's floppy drive, which has never read a disk in the time I've owned the computer to actually mount a floppy, albeit slowly. Still have some other adjustments to do before it reliably loads software, but it's an improvement.
 
  • Like
Reactions: Paolo B

fehervaria

Tinkerer
Sep 23, 2021
153
160
43
48
North Germany
Here some slides I put together explaining the ”zero track alignment“ procedure.
As far as my experience goes, 100% success rate in recovering “misaligned” drives.
I hope it can help.

Zero Track Alignment
Grazie mille Paolo!
Siete fantastici, lo aspettavo tanto tempo fa, visto che la descrizione di Larry Pina non basta per farlo senza esperienze.
Grazie ancora, sei grande.
 
  • Love
Reactions: Paolo B

Stephen

BetterBit
Staff member
Founder
Sep 5, 2021
204
250
63
San Francisco
Here some slides I put together explaining the ”zero track alignment“ procedure.
As far as my experience goes, 100% success rate in recovering “misaligned” drives.
I hope it can help.

Zero Track Alignment
I'm just seeing this now — it's an incredible resource, thank you for sharing! Would you consider compiling the slides into a single Resource?
 
  • Like
Reactions: Paolo B

alxlab

Active Tinkerer
Sep 23, 2021
287
312
63
www.alxlab.com
I would like the share my recent experience and observations with 3.5" 800K floppy drive head alignment.

As probably mentioned before the The Dead Mac Scrolls offers a great problem and solution list to go through with alignment fixes being the last resort.

In my case it was definitely alignment of the optical sensor. The interesting thing that I found out, that I haven't seen mentioned anywhere else, is that it seems to be easier to get the floppy drives to work using the cable from an external 3.5" drive than the ribbon cable in the Macs.

1648659899772.jpeg


The external drive cable is very well made with the foil shield only connected to shield ground and a ferrite bead. I didn't double check how they pair the wires in the cable, but maybe they pair grounds with signals as well to reduce crosstalk within the cable.

With the external drive cable I was able to get both 800K drives to read and write disks. With the internal ribbon cable the drives could almost boot from one disk and not at all with another.

It might be interesting to try to make a new rounded ribbon cable for internal drive connection and see if the same results are achieved. Maybe a parallel port cable would do the job.

Other observations I made is that it seems if the drive's optical sensor is misaligned, disks can either not initialize at all or they can initialize and then be read afterwards, but then have trouble in other drives. I'm assuming the difference is due to the level of misalignment of the optical sensor.
 
  • Like
Reactions: fehervaria

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
I would like the share my recent experience and observations with 3.5" 800K floppy drive head alignment.

As probably mentioned before the The Dead Mac Scrolls offers a great problem and solution list to go through with alignment fixes being the last resort.

In my case it was definitely alignment of the optical sensor. The interesting thing that I found out, that I haven't seen mentioned anywhere else, is that it seems to be easier to get the floppy drives to work using the cable from an external 3.5" drive than the ribbon cable in the Macs.

View attachment 3853

The external drive cable is very well made with the foil shield only connected to shield ground and a ferrite bead. I didn't double check how they pair the wires in the cable, but maybe they pair grounds with signals as well to reduce crosstalk within the cable.

With the external drive cable I was able to get both 800K drives to read and write disks. With the internal ribbon cable the drives could almost boot from one disk and not at all with another.

It might be interesting to try to make a new rounded ribbon cable for internal drive connection and see if the same results are achieved. Maybe a parallel port cable would do the job.

Other observations I made is that it seems if the drive's optical sensor is misaligned, disks can either not initialize at all or they can initialize and then be read afterwards, but then have trouble in other drives. I'm assuming the difference is due to the level of misalignment of the optical sensor.
Strange as it might sound, but as far as I remember not all the combinations of motherboard / ribbon cable / floppy drive are compatible. Don’t remember exactly the details, but there are “yellow” and “red” ribbon cables. As for the misalignment of the optical sensor, a couple of posts above you can find a document I have preparare describing the problem and a reasonable easy way to fix it.
 

fehervaria

Tinkerer
Sep 23, 2021
153
160
43
48
North Germany
Just an interesting observation: I've spent many-many hours in the last years to "try-to-realign-the-bastard-floppy-drive" and I was able to "do" one very-very interesting situation:
I had a drive (already out from this state) like you have two (!) logical disks on the same physical disk. More explanation:
I had a floppy drive (totaly out of alignment, harvested out from a very bad state SE FDHD) and I followed Larry Pina's explanation and my own experience to try to realign - make it work again. The result of the lot of tries, including the stepper motor repositioning AND the Track-Zero sensor repositioning ended up in a very strange state.
The drive was able to read, write, boot - operate normal BUT in an other drive the same disk was working different. At the first time injecting the disk the "OK" drives wanted to format the disk, and then it was working well (read, write, boot) but putting the disk back to the "repaired" drive the original content was there too like I did not format it in the well-known dirve. Modifying the content, using for boot - everything was fine. Putting over to the known-good drive - and its content was also there! But they were totaly different bt good contents.
I have a feeling I was able to align the drive (which one was from the damaged SE) to read/write its tracks BETWEEN the usual position of the tracks and the other drives were read/write their tracks to the "real" positions...
I tried with multiple disks and everything was 100% ok, except, this drive was writing between the tracks... living its own parallel universe.
This means I was able to write 800K disk content 2 times on the same disk. I tried 1.44 HD disks and normal 800K disks and they were the same OK. First I've thought it is possible because I used 1.44 High Density disks but it was OK working with "normal" 800K disks as well.
Never seen this behaviour before - and sadly I disaligned this setting - and never was able to reproduce.
 
  • Wow
  • Like
Reactions: JDW and Paolo B

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
Just an interesting observation: I've spent many-many hours in the last years to "try-to-realign-the-bastard-floppy-drive" and I was able to "do" one very-very interesting situation:
I had a drive (already out from this state) like you have two (!) logical disks on the same physical disk. More explanation:
Yes, it can happen, as the reading tolerance is just 10% of the track width. That’s why the “in between tracks” alignment, which is controlled by the angular position of the rotor wrt a defined zero position was factory calibrated and sealed. The zero track alignment is a gross adjustment window for compensating mechanical and electronic tolerances.
 

Paolo B

Tinkerer
Nov 27, 2021
258
144
43
Nagoya, Japan
How do these drives get out of alignment in the first place? Mechanical damage? Or just drifting over time?
Honestly, I have no precise idea. Based on the inferred working principles, I‘d say misalignment of the rotor is simply not possible, unless deliberately induced. As for the zero track sensor, I would be curious to see the response curve, for sure it’s not a “perfect” step between the two states. So, maybe some drift in the response curve due to aging of the sensor and of the controlling circuitry is a possibility…
 

alxlab

Active Tinkerer
Sep 23, 2021
287
312
63
www.alxlab.com
I'm not surprised that my drives were misaligned since they got shipped/kicked across the country.

Even so it looks like the problems with my drives seems to be with the alignment of the optical endstop sensor and not the motor alignment. It seems to take VERY little movement to make it not be able to read disks properly anymore. As a result it's a pain to adjust by hand. As @fehervaria I was able to get the drive to be able to initialize it's own disks but not be usable with working factory tuned drives. The optical stop must be off by a fraction of a mm. I got tired of trying to get the alignment perfect. Maybe I'll try again in the future. Maybe 3d print some spacing blocks that increase in factions of a mm. I dunno. Doing it blind by hand is just too time consuming.
 
Last edited:
  • Like
Reactions: fehervaria

fehervaria

Tinkerer
Sep 23, 2021
153
160
43
48
North Germany
I'm not surprised that my drives were misaligned since they got shipped/kicked across the country.

Even so it looks like the problems with my drives seems to be with the alignment of the optical endstop sensor and not the motor alignment. It seems to take VERY little movement to make it not be able to read disks properly anymore. As a result it's a pain to adjust by hand. As @fehervaria I was able to get the drive to be able to initialize it's own disks but not be usable with working factory tuned drives. The optical stop must be off by a fraction of a mm. I got tired of trying to get the alignment perfect. Maybe I'll try again in the future. Maybe 3d print some spacing blocks that increase in factions of a mm. I dunno. Doing it blind by hand is just too time consuming.
Same story here. That SE what I was talking about, was very badly handled in the last 20+ years. The front bezel was multiple places broken, pieces are missing from it, the main board was standing in some dirty water (the whole machine was stinky like a pub's toilette) and the hard disk was broken off from its screws (2 was broken totaly) and the floppy drive was also suffered of the physical damage: the Track Zero sensor's screw was missing and the Track Zero sensor was blocked between the top and bottom heads of the drive.
The machine works after the restoration but still mixing the floppy drives together: connect a drive to the external connector, it will display 3 floppy disks on the Desktop. If the internal drive is plugged to the LOWER DRIVE onboard connector it works well. The SCSI works, etc... except the misaligned floppy drive...
Maybe I SHOULD NOT follow Larry Pina's instruction and move out the stepper motor from its settings and get @Paolo B 's instruction earlier, than maybe would be possible to rescue the whole machine. Actually it has another floppy drive from a donor, LogicBoard dead SE.
I will try, time to time to align this poor drive - but maybe I organise before some oscilloscope for better possibilities.