maxtherabbit wrote on 2021-05-24, 21:39:

Does knocking the heads against the over travel stops disalign them? I know some drives like Apple ][ and some commodore units didn't even have track 0 sensors, but it seems like all that impact would not do the head assembly any favors...

I accidentally knocked the heads on a Tandon TM-100 a bit when trying to use IMD to clean my other floppy drive (that was 80 tracks)

Im not familiar with Apple or Commodore stuff other than whaI have seen on the YT. These do not have track sensor? How do they even work, is the stepper motor an absolute one with encoder for position readout? I mean one could just try to step back N times but as you say, it would cause the whole assembly to just bang agains the limiter a lot...

It shouldn't really cause any head aligment issues, unless they were either about to rip off anyway or the assembly was unscrewed and not put together properly. But it can strain the mechanism, some will take it better than others. I'd say the flex steel tape one would be mostly OK, the vibrations will be fed back to the rotor but it too usually has nice, tight fit bearings. It will die at some point but not anytime soon. The mechanism with a long screw on the other hand, like what is usually used on 3.5" drives, the parts that engage the screw are plastic. I could see that wear out and cause aligment issues. But not until some real abuse.

I think the early PC BIOSes used a seek trick to detect 360k drives, since these could only step about 43 or so times. BIOS tries more (I can tell by the sound) and then counts back to track 0 sensor signal. This is done on every boot and yet these drive worked fine for years.

Deunan wrote on 2021-05-24, 22:38:

Im not familiar with Apple or Commodore stuff other than whaI have seen on the YT. These do not have track sensor? How do they even work, is the stepper motor an absolute one with encoder for position readout? I mean one could just try to step back N times but as you say, it would cause the whole assembly to just bang agains the limiter a lot...

My understanding is that is exactly what they do.

As for my Tandon, I considered this only because I recently discovered the alignment was very slightly off. Occasionally get errors on tracks 38 and 39 when using disks from other drives, no errors on disks formatted in the Tandon. I got this system recently so it's well possible the drive was always like this. I didn't test it thoroughly before the knocking incident.

(I'm trying to track down a Dysan 5.25" analog alignment disk if anyone wants to part with one drop me a PM)

maxtherabbit wrote on 2021-05-24, 22:44:

As for my Tandon, I considered this only because I recently discovered the alignment was very slightly off. Occasionally get errors on tracks 38 and 39 when using disks from other drives, no errors on disks formatted in the Tandon. I got this system recently so it's well possible the drive was always like this. I didn't test it thoroughly before the knocking incident.

Might not be aligment issue at all. Heads need certain amount of pressure against the disk surface to work properly (be close enough). Even a tiny warp in the helper rail (the one the head rests on when drive is not closed, or closed without a floppy) can lift the head and the signal will be considerably weaker. Then theres some really funny issues like I discovered on one of my Teacs, there are usually certain points on the clamping arm that press down on the plastic sleeve to make sure it stays flat and doesn't wobble when disc spins. If those points warp or lift, or the various rubber/sponge feet get old and compressed, the sleeve can lift ever so slightly right in the head cutout area and cause the head to lift as well. This happens more on the last cylinders as the signal there is weaker, and the plastic has less space to distribute any warping away from head area.

And then there are really warped floppies (they make an easily distinguishable noise when spinning) - that can bounce the heads as well. Run IMD and try to gently press on the floppy from above (also, the head, but gently). If that causes changes in the quality of the signal you know it's the pressure, not aligment, that's the issue. Also keep in mind as floppies get older they loose some of the magnetization due to influcence of outside fields. That will eventually degrade them to the point they can't be read, but re-writing them should help.

maxtherabbit wrote on 2021-05-24, 22:44:

(I'm trying to track down a Dysan 5.25" analog alignment disk if anyone wants to part with one drop me a PM)

Good luck with that, you'll need it. If you find two let me know 😀

I need to do a bit more testing to be 100% certain, but the results so far are that I can take the exact same physical disk and:

format it in the Tandon -> Checkit test in Tandon, 100% no errors
format it in one of my known good YD-580s -> Checkit test in Tandon, occasional errors T38 T39

If the problem was in the armature it shouldn't matter which drive formatted the disk

Just reading through, just trying to respond on the situation on knocking and alignment. I've only had it happen once I can remember.

So there is a routine in the BIOS that calibrates the drive by moving the head all the way out (towards back of drive) looking for the track zero switch. There is a black screw on the back that prevents the head from going too far.

IMD also does the calibration seeking when you tell it to use the drive. If you are saying that you heard some knocking when starting up IMD or the PC, then I'd check:
1) is the track zero switch working as expected -- you can check this in IMD.
2) you can check the alignment screw for track 0 -- this is part of the alignment check
2b) you could try to step even further backwards beyond track 0 -- this should not be possible

I had this happen too once, when I cleaned one of these drives, and then started it up again. I don't know for sure, but I'd suspect that it was the alignment screw.

So I aligned my drives without an alignment disk in IMD. It's totally doable. I did figure out how to use an oscilloscope to see the signals mentioned in the tandon manual. Not sure if that really helped, but it might be possible when fine tuning is needed.

I do totally agree that you should make sure the floppy is in good shape, when you are testing the drive.

IMD made the heads knock because I accidentally told it it was an 80 track drive before starting the cleaning routine

Oh, I think I have also done that. I think your drive is going to be fine, cause my drives are still working 😀

maxtherabbit wrote on 2021-05-25, 14:10:

format it in one of my known good YD-580s -> Checkit test in Tandon, occasional errors T38 T39
If the problem was in the armature it shouldn't matter which drive formatted the disk

You might be right but the conclusion is wrong. A head that doesn't make good contact with the floppy surface might also not erase/write properly. So it would matter which drive formatted the disk, though it has to be said the write field is pretty strong and less susceptible to non-ideal conditions.

Anyway, your method of testing is not good enough to be sure what's the problem here. I thought you were using a scope and tested signal level on tracks 38/39? I mean, why would you want an alignment disk otherwise? It's not usable without a scope. You can get the head aligned better with just IMD, but unless you are pretty lucky it'll never be perfect - just better than what you have now. Still, that might be enough for your needs.

Yes I have a scope but without an analog alignment disk to use it with I didnt see the point.

I have a 8" Dysan alignment floppy so I know the drill. Never occurred to me to try to measure the signal level of a regular disk tbh

On a good quality disk, formatted correctly on a good drive, you can use IMD and a scope. The formatted correctly is key. The disk must not have been formatted on a misaligned drive previously without being demagnitized completely and then formatted on a good drive. I've done alignment successfully this way and have not needed a special alignment disk.

I use floppies with genuine software on them, these were usually copied on heavy duty drives with near perfect aligment that was often tested, and corrected if necessary. But I start with my own floppies that were written by drives that I recalibrated. Reason being if there is something wrong with the heads or write circuitry I don't want to damage or erase my originals. Once I see the drive responding well to my attempts I switch to genuine software floppy. I should point out that these are old now and the signal level, in general, will probably be lower than a fresh written floppy. These things do degrade slowly over time.

If the head assembly was taken apart (I did that a few times to clean really dirty drives or to resolve issues with bent parts) then the upper vs lower head aligment is also needed. It's important to remember that while the signal level should be nearly identical between the two, there might be some differences and it can also vary along the floppy surface. That's why it's best to align near the end but not at the very last cylinder, and double check the results at a few more key points like track 0 and the middle of the floppy.

Interesting approach.

I'm a bit skeptical of trusting any of my own drives as the progenitor. Not because they are out of alignment per se, but because they are probably not 100% perfect and I don't want to create a tolerance stacking condition. I'm more amenable to the idea of using a professionally duplicated disk.

So do you guys just move the head sled back and forth and try to estimate the point with peak signal amplitude? I've only ever done alignment using the "cats eye" pattern with a real alignment disk

My understanding is a real alignment floppy has specific patterns written to it, and not right on the track center but offset a bit, and the perfect eye is the result of getting the head position in the middle of such patterns.
With just the signal level, which is also not pure sinusoid so will "dance" a bit due to random triggering, you can get very close but never as perfect as with eye method. But you work with what you have. It's still pretty darn close.

In order to avoid stacking errors never overwrite the floppy you've picked as reference source, and/or use (preferably more than one to be sure) original software floppies. Best I can do is try to write a floppy with several drives I have and compare the signal level between them. Assuming same write power the results should be near identical - and so far I'm pretty happy with what I got. But if I could get my hands on alignment disk I could do some proper verification.

EDIT: The peak can have a bit of a dead zone to it, not much mind you, but there is one. It's less of an issue on the close to last cylinders as the signal is weaker there and adjusment is more touchy. I recon I'm introducing more error by adjusting the head with my fingers rather than set screws though. Plus the screws would be different for each drive type anyway.

Well I'm happy to report my problem wasn't alignment related. In fact it was just the stupid belt. The belt was marginal enough that it was occasionally slipping and the spindle wasn't maintaining constant speed.

maxtherabbit wrote on 2021-05-31, 03:44:

Well I'm happy to report my problem wasn't alignment related. In fact it was just the stupid belt. The belt was marginal enough that it was occasionally slipping and the spindle wasn't maintaining constant speed.

I wonder why that effects it more on the inner tracks? It's CAV, so the number of transitions passing under the read head when the belt slips would be the same, and I would have thought the drag on the read heads would cause more slip the further out they were. Also interesting that disks formatted on your drive could be read by it, even if the speed wasn't constant. But it's good to know that belt slipping doesn't cause the same errors across a disk.

maxtherabbit wrote on 2021-05-31, 03:44:

Well I'm happy to report my problem wasn't alignment related. In fact it was just the stupid belt. The belt was marginal enough that it was occasionally slipping and the spindle wasn't maintaining constant speed.

You've never mentioned it's a belt-driven spindle or I would've suggested testing the speed. To be frank I've only dealt with completly disintegrated belts personally, not slipping ones, but I have some more "belt experience" from other stuff like older tape players. If it's stretched it's no good but possibly it's only dirty, that can cause it to slip too. The wheels can be cleaned with some degreaser, don't use sprays because that stuff easily gets everywhere it shouldn't (like PCBs and bearings), a cotton swab is perfect for this job. I stick to 99.9% isopropyl alcohol but I've heard of people using brake cleaner as well.

For the belt itself use water and soap only, then dry it in ambient air. Chemicals can easily ruin it completly so beware, try that only as last resort. I've had some success restoring proper operation by just cleaning all the belts and wheels, the dust and dirt can cause them to slip (especially the older ones that are already somewhat stretched). Even if it's only temporary solution it saves money and gives you a good idea on what is wrong with the thing.

snufkin wrote on 2021-05-31, 07:01:

Also interesting that disks formatted on your drive could be read by it, even if the speed wasn't constant.

Further testing proved this to not be the case, I got errors with disks formatted on the drive. I mentioned this in an earlier post.

I got the drive running tip top by lubricating the hub spindle and by boiling the belt in water for 10 minutes to rejuvenate it.