Floppy drives once again

[JoeMuc2006]

Hello everyone,

got a ton of defective 3.5" floppy drives here from different Amiga and external disk drive models. I think most of them can be salvaged by readjusting them. However, I have little experience with that and currently trying to learn more about the topic, despite the fact that floppy disk drives are largely outdated and all afficionados have replaced them with USB or flash-based emulators. Amiga drives are no longer made which imho makes every drive in existence well worth saving.
Unfortunately, the Shugart protocol is very poorly documented on the internet in terms of signal sequence diagrams and the nature of the signal that a drive sends while it is reading, or the signal it needs to receive for writing. I'll be grateful for any advice where to find better information on this, even if it is a book I would have to purchase. I'm just so curious now.
From what I could find out, this is how to make a floppy disk drive read:

all odd pins: GND
pin 2 (/REDWC): HIGH (=low density)
pin 10 (/MOTEA): LOW (=enable motor of drive A)
pin 12 (/DRVSB): HIGH (=disable drive B)
pin 14 (/DRVSA): LOW (=enable drive A)
pin 16 (/MOTEB): HIGH (=disable motor of drive B)
pin 18 (/DIR): needed during track positioning
pin 20 (/STEP): LOW pulse to move head, otherwise HIGH
pin 24 (/WGATE): HIGH (=read ~ do not write)
pin 26 (/TRK00): indicates if the head is over track 0 (LOW) or somewhere else (HIGH)
pin 30 (/RDATA): serial data stream?
pin 32 (/SIDE1): select head, I think HIGH means head 0 and LOW means head 1. However, I have no idea if head 0 is the top or the bottom side of the actual disk

So if all pins are connected as shown, the motor spins and I would expect the drive to output a signal on pin 30, endlessly repeating data from the track that it is positioned on and from the head selected via pin 32. I'm a total beginner so this is just an assumption. I cannot see any other pin that could have the purpose of returning data. But the oscilloscope shows as well as nothing on that pin. The signal amplitude is 100mV at best and looks like a sawtooth where I would expect that it should be a near-5V amplitude and close to a square wave. Eventually that signal has to pass through a conductor that might be up to 50cm long, so it ought to be a lot stronger than 100mV, right? Furthermore, the signal I am getting is the same on both heads and different tracks. From what I see I cannot distinguish any of it, there is no pattern, it just looks like garbage all the time.
I think all floppy disk drives contain an internal signal amplifier for the head signals that condition the signal to a level high enough to withstand conductor issues such as length and inductivity. How come my measurements are that far off from reality apparently? It's the same with all drives I have tried so far.

TrackDiskSync is a good help for adjustments, however I didn't succeed with it yet, for instance I completely failed yesterday over an Epson SMD-400 (internal drive from A500). Even though I get a very clear and undistorted signal reading the adjustment disk (which of course I created with a drive known good), the A500 cannot use the drive with any disk. It just starts and stops spinning the disk, moves the head to track 0, then starts and stops some more times, but that's all. Not even the typical double seek the A500 usually does when it cannot read a disk. The reason may be something other than misalignment. Maybe the electronics and motor to spin the disk are broken so it does not reach the required RPM, but nothing about it sounds broken, and it either rotates at a given speed, or not at all. There is no sign of slow acceleration at startup or unintended differences in the RPM once it has spun up. It just appears to restart a few times so that specific drive may have other problems as well.
I think each drive may have its own issues. Maybe there are some people here knowing more about the SMD-400 and whether it's worth spending any more time. Maybe these drives were notorious for failing anyway and nobody wants them anymore.

But there are some more fundamental beginner questions I have.
So the signal I get from TrackDiskSync is perfect, and fiddling with the adjustment screws does change the signal so it basically works and I did what I could to align the drive properly. What I am wondering about is why only the top head can be adjusted at all whereas the bottom head is hard to reach and usually does not offer any means of adjustment.
I am missing info on what side(s) of the disk TrackDiskSync performs on. Probably just the top head because moving it changes the audio signal for sure. But what about the other head? It is not mentioned anywhere and my impression is that it just uses the top side of the disk. What if the bottom head is misaligned? Or does that never happen with any floppy disk drive?
TrackDiskSync does not offer any parameters as to which disk side to work on. There is no output at all, and the consistency of the signal suggests that no switching of disk sides happens. My best idea to include support for the bottom head is swapping the flat flex cables coming from both heads so the drive thinks it uses the top head when it actually uses the bottom one. A new alignment disk would have to be created on a working drive with swapped heads for this purpose of course.
There are some drives where the head cables cannot be swapped because the flat flex cables use one common connector instead of two separate connectors, so that won't solve it for all drives anyway. I would prefer a TrackDiskSync version that offers full support for both heads. As the source code cannot be found anywhere for this tool, I can only guess what it is currently doing, much less create a version that would offer what I want.
So is anyone out there who can explain why bottom head assigment in floppy disk drives seems to be no concern at all?

There are many people around me who still enjoy Amiga, some even for work! The players have floppy trouble from time to time, and the usual resolution is buying a new drive from eBay and tossing the old one. The drives are so rare that new ones are hopelessly overpriced. I would hope there is a better way to deal with the issue.

Have a nice day!
Joe

[guest11527]

What I am wondering about is why only the top head can be adjusted at all whereas the bottom head is hard to reach and usually does not offer any means of adjustment.

The drives I fiddled with had top and bottom head on a single slide, so adjusting one also means adjusting the other. They are both typically aligned to each other, but not necessarily to the start of track #0.

I am missing info on what side(s) of the disk TrackDiskSync performs on.

I'll check.... Ok, it writes data on both sides, but receives data only from the upper head.

Probably just the top head because moving it changes the audio signal for sure. But what about the other head? It is not mentioned anywhere and my impression is that it just uses the top side of the disk. What if the bottom head is misaligned?

That's - as I said - quite unusual because I have not yet seen a drive where you can align the heads independent of each other.

I would prefer a TrackDiskSync version that offers full support for both heads. As the source code cannot be found anywhere for this tool, I can only guess what it is currently doing, much less create a version that would offer what I want.

Well, you can always contact the author, which, as a matter of fact, you have just done.

[Hattig]

Just stick Goteks in the cases and sell them on eBay.

[JoeMuc2006]

You're right, mechanical floppy drives are pretty pointless nowadays. Well, I love them anyway, somehow I find these things appealing. Maybe because it is usually possible to fix them with reasonable effort, not like today's flash devices. For the hardcore fan the floppy sound is irreplaceable, too. And I have not seen any hardware floppy emulator so far that does this job nearly as good as WinUAE. The HxC emulator that I own has a buzzer and kind of imitates the stepper sound at least but its sound is far from being like an actual drive.
It is a little crazy but I like to have my Amigas as close as possible to the original. And that includes the pesky floppy drive.

[paul1981]

It is a little crazy but I like to have my Amigas as close as possible to the original. And that includes the pesky floppy drive.

Not crazy at all. I'd never dream of ditching the floppy in any of my Amiga's. I'm sure we're not alone too.

[JoeMuc2006]

Thanks Paul, that's what I am hoping, too. It's kind of amazing how little knowledge can be found about floppy disk drives in the technical depth required to service the damn things.
Well, I grew up with them, long before the handy 3.5-inch models were on the market, but my IT career started when disk drives were cheap enough to toss and replace, so I never learnt anything about how they actually work.
Anyway, life was based on disks those days B) I still like that vintage tech gear, including 5.25" double-height MFM hard disk drives and that sort of jazz. It's still amazing to see how well all that hardware works now that it's beyond 20 years old. And if it's broken, there is usually a way to fix it rather easily. Trying to get closer to that matter. I don't think that much of today's hardware will work half as good, if at all, 20 years from now.

[Fats]

Well, I grew up with them, long before the handy 3.5-inch models were on the market, but my IT career started when disk drives were cheap enough to toss and replace, so I never learnt anything about how they actually work.

I remember the days when your floppy drive was a computer as powerful as your main machine...

[Oldsmobile_Mike]

I remember the days when your floppy drive was a computer as powerful as your main machine...

Haha, and viruses that could infect your 1541 floppy drive.

[guest11527]

I am missing info on what side(s) of the disk TrackDiskSync performs on. Probably just the top head because moving it changes the audio signal for sure. But what about the other head? It is not mentioned anywhere and my impression is that it just uses the top side of the disk. What if the bottom head is misaligned?

Since you'd asked: I quickly added a "READBOTTOM" option to trackdisksync. Unfortunately, I don't have an Amiga here to test, so please consider this a beta release. It would still be great if you could test whether this works for you and allows you to also read the alignment data from the bottom head.

As said, I believe it is usually not needed, but maybe it helps.

Feedback welcome.

[scuzzb494]

You're right, mechanical floppy drives are pretty pointless nowadays. Well, I love them anyway, somehow I find these things appealing. Maybe because it is usually possible to fix them with reasonable effort, not like today's flash devices. For the hardcore fan the floppy sound is irreplaceable, too. And I have not seen any hardware floppy emulator so far that does this job nearly as good as WinUAE. The HxC emulator that I own has a buzzer and kind of imitates the stepper sound at least but its sound is far from being like an actual drive.
It is a little crazy but I like to have my Amigas as close as possible to the original. And that includes the pesky floppy drive.

In many ways all these later gadgets and gismos do not make for a real Amiga. I real Amiga is one fitted with a floppy drive, hard drive as an addition. An emulator is not a real Amiga and neither are these goteck thingy whatsits or flash drives. They corrupt the original base machines and in many ways destroy the beauty of the original designs. Being a purest I live with my Amigas as they were designed and only plug stuff in that was essentially part of the original game plan. I kinda am happy with some of the later expansions but sorry... camera memory and anything that cuts the case is just destructive. Worst I guess was the breaking up say a 1200 to put in what is essentially a PC tin box. Never made any sense to me.

And it is not hard at all running the original Amigas as intended. They are a joy to treasure and a day never goes by when I am not working on at least three or four machines. Like now as I decrunch dms demos on my 4000, test them on my 500 with GVP and play Settlers on my 1200 that sits behind me. And all the time I am swapping floppy disks and writing and testing stuff. It really just wouldn't be the same if the monitor didn't click between screen formats, or the drive wasn't gently clicking or there wasn't that noise of the final writing to the floppy disk. Pure magic. Its about the whole sensation and that includes the noises, sounds and experience.

[JoeMuc2006]

yeah, quite so

@Thomas: thanks, great work. I tested it and I think you are right. While the Epson SMD-400 actually has an adjustable top head, its bottom head is "fixed". I did not find any means of adjustment for it. Anyway, XCopy proved that the bottom head works as it should whereas the top head was off a bit. And this time I succeeded adjusting it with TrackDiskSync easily! Really saved my day! Thanks again!

[guest11527]

And this time I succeeded adjusting it with TrackDiskSync easily! Really saved my day! Thanks again!

You are welcome. BTW, the typical adjustment for the heads is to rotate the spindle motor which moves the heads back and forth. This is *usually* out of alignment, not the individual heads.  

Anyhow, my question is rather whether the modified version did work. As said, I've no chance to test it here.

[paul1981]

A bit off topic, but yesterday I orderded a replacement drive belt for the floppy disk motor in my dad's Korg X3R sequencer. I'd never seen a drive with a drive belt before as all the floppy drives I'd seen up until now were direct drive.

This will be another floppy drive rescued anyway.

[guest11527]

A bit off topic, but yesterday I orderded a replacement drive belt for the floppy disk motor in my dad's Korg X3R sequencer. I'd never seen a drive with a drive belt before as all the floppy drives I'd seen up until now were direct drive.

All the 5 1/4" drives I've seen where driven by a belt. In particular, the Atari 8-bit floppies (the 1050's) are also belt-driven. One reason why mine no longer work.

[JoeMuc2006]

@Thomas: My impression was that it did work. However, it was hard to differentiate between the top and bottom heads. I am not even sure the new switch caused the tool to read the bottom head actually. It would be easier to test this reliably if different frequencies could be picked for each track and side.
That is one general problem I still have with TrackDiskSync. I know the tool writes the first few tracks and applies different frequencies. But I think I have never seen (or rather heard) the drive anywhere else than on track zero. Pushing Fire on the joystick will relocate the head, still the only frequency I get is the one from track zero. Adjustment may require to check tracks further away from the disk center. How to I get there?
Honestly I don't know anything about the magic you use to write an audible frequency to a floppy disk, but if possible, it might be a great help if more command switches were available, such as:

TRACK [n] - to select a distinct track between 0 to 81 to read from (in READ mode) or write to (in INIT mode)
HEAD [n] - to select the head 0 or 1 where reading or writing is done (in both modes)
FREQUENCY [n] - to select a given frequency or one out of a set of predefined frequencies to write to the track (only applicable in INIT mode)

That way I could generate a disk where it is possible to recognize the track the head currently covers with a scope, and see if that is a match or not. If both sides have different frequency offsets, it's even possible to know what side is currently checked. Flipping between both sides would reveal if both heads are on the same track or one is on a different track from the other.
There is another idea: how about using the joystick a little more to select the active head on the fly while reading (UP = upper head, DOWN = lower head), and move the heads (LEFT = one step down to track 0, RIGHT = one step up to track 81)? I would also love to see the status being printed to the console so I know what track and disk side the program thinks it is on.
Your tool is already invaluable for diagnosis as it is. Adding these capabilities would make it a Swiss pocket knife for me, and probably many others. I can't tell you how happy I am that the tool is not discontinued

About those belt drives: Amstrad used them, too, in their darned 3" floppy drives. Back in the early days I had a CPC 6128 which had one of these built in. Almost 100% of their drives failed because the belt lost tension over the years. Good proof that nobody thought long-term. Belts may be a wise choice technically because they ensure the drive motor won't be overloaded. Instead the belt would slip when the disk refused to turn. But it's not really an everyday use case (how many disks were ever stuck so they could not be turned?) and the long term stability of the drives should have weighed more during their development. The direct-drive motors that came with 3.5" drives eventually fixed these concerns for good.