Weird drive issue

[llopis]

I have an internal 6128 drive that has a puzzling behavior.


It works fine, can read lots of different disks without a problem, etc. Then, after some time (sometimes just an hour or two) goes by, or by screwing the computer case, it will give a Read Fail error when doing a CAT. This is before even showing "Drive A". This is exactly the same behavior as if the 0-track seek sensor was in the wrong position. I can tweak the sensor position a bit, or even just move the head forward (by rotating the ring around the motor shaft), and then it will start reading disks correctly again.


Any idea what might be going on? The drive is fully lubricated, so it's not an issue of gunk hardening and making it hard to move. I'm guessing there's something in that track 0 detection that's causing the problem, but I don't understand what or why.

[llopis]

Never mind. Figured it out. Writing it down here for posterity and search feature.


The drive shaft has a ring with a screw. I'm guessing it uses that ring along with the sensor to detect track 0. The screw was loose (remember? really lubricated drive?), so as it moved a bit back and forth, it made the detection of track 0 incorrect. Finding the right spot and tightening it fixed it.

[gerald]

Never mind. Figured it out. Writing it down here for posterity and search feature.


The drive shaft has a ring with a screw. I'm guessing it uses that ring along with the sensor to detect track 0. The screw was loose (remember? really lubricated drive?), so as it moved a bit back and forth, it made the detection of track 0 incorrect. Finding the right spot and tightening it fixed it.

Not sure this is your problem.
Track 0 detection is done with the head carriage mechanism.
This bit is only used to block the motor if you try to go beyond track 0.

I am more thinking of a  faulty head motor.
There is one @Bryce  topic on that exact problem.
Edit : found http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/fixing-a-floppy-stepper-motor/

[llopis]

Not sure this is your problem.
Track 0 detection is done with the head carriage mechanism.
This bit is only used to block the motor if you try to go beyond track 0.

I am more thinking of a  faulty head motor.
There is one @Bryce  topic on that exact problem.
Edit : found http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/fixing-a-floppy-stepper-motor/

Bummer! That makes sense what you're saying about the ring, so that might not be it. I guess I should keep testing that drive for a few more days before calling it fixed.


What makes you think it might be related to Bryce's post? The fact that it doesn't stay calibrated?

[gerald]

What makes you think it might be related to Bryce's post? The fact that it doesn't stay calibrated?

Yes !

[Bryce]

It could still be something else. The comment "or by screwing the computer case" reminded me of another repair I had once. It was dry joints on the drive PCB. Flat on the desk or loose in the computer the drive worked perfectly. As soon as the case screws were tightened, it stressed the drive PCB enough to cause the dry joints to seperate and the drive failed to work.

Bryce.

[llopis]

This morning it's still working perfectly. And that's with the case securely screwed in and everything. It had never lasted this long before. I'll continue monitoring it, but I'm inclined to think that tightening the screw in that ring fixed it somehow.


I wonder if that ring being loose would slowly move forward, so that when the drive would move to find track 0, it would bump against it and not go to the right place.


That would also explain why if I moved the head forward manually, or if I turned off the computer right as the head was tracking forward, when I tried again it would work (it would start significantly forward, so it would be enough to bump into the ring and move the head to the right place). I actually remember reading a couple of threads in the forum about people who said their drive would work if they stopped at the right time and tried it again. Probably the same issue.


No idea of course. It's all a hypothesis based on what I'm seeing.