DaveDDS wrote on 2025-06-21, 14:00:

Ok, so you do have easy access to the head - I was thinking that like on some drives I have, the whole
assembly is covered by a much larger metal shield which you didn't want to take off....

No need for paper, you can easily reach the heads with better cleaning stuff.

I was OK with removing the cover. That's just 2 screws, and is a part that doesn't affect the usage of the drive as far as I know. What I didn't want to do is unscrew any PCBs or unplug any leads. Basically, I didn't want to do anything that could harm it. Anyway, as time has gone on, I think I've probably done too much now to return it. So, if you have any further things I can try. Let me know.

BTW, I have bought another 5.25 drive on ebay. Its not arrived yet, but was cheap as is "for parts or not working" due to being untested. I think it's a drive from a BBC Micro. CHINON FZ-506 but has been modified with a switch at the front to change it between 40 and 80 tracks. Hopefully it will work on a PC. Lol. I guess I will be getting some practice in using ImageDisk.

DaveDDS wrote on 2025-06-21, 14:00:

It looks clean, but some kinds of crud you can't easily see.... […]
Show full quote

It looks clean, but some kinds of crud you can't easily see....

When cleaning the heads, what you don't want to do is put a lot of sideways force on it,
but you can put a reasonable amount of downward pressure (not enough to bend the mounts, but enough to get good contact/cleaning).

I'm assuming you've cleaned the upper head as well.

Yes, I cleaned the upper head. Though, of course, I can't see the surface of that without some disassembly

DaveDDS wrote on 2025-06-21, 14:00:

Look closely at the diskette media that's been passing under the heads, is there significant scoring or other visible "lines" - […]
Show full quote

Look closely at the diskette media that's been passing under the heads, is there significant scoring or
other visible "lines" - is one side worse than the other?

Have you tested both sides individually with ImageDisk - many types of head physical/electrical problems will affect
only one head. It would be good to know if both heads are experiencing similar problems or not.

I have tested both heads. As far as I can tell they are acting the same as each other. Similar numbers shown on the Align test.

I can't see any lines on the disks on either side.

Is it worth putting a small weight on the head like a piece of blutac to see if that helps with reading?

RetroPCCupboard wrote on 2025-06-21, 14:24:

Is it worth putting a small weight on the head like a piece of blutac to see if that helps with reading?

Or you can gently press it with a finger, but a small weight is easier and more uniform. To be sure though you need to try reading a known good floppy or reformat and rewrite again your own with that additional pressure.

Deunan wrote on 2025-06-21, 14:34:

RetroPCCupboard wrote on 2025-06-21, 14:24:

Is it worth putting a small weight on the head like a piece of blutac to see if that helps with reading?

Or you can gently press it with a finger, but a small weight is easier and more uniform. To be sure though you need to try reading a known good floppy or reformat and rewrite again your own with that additional pressure.

Ok. Tried with a piece of blutac. I formatted the disk again in IMD, then tried align. Both heads are able to read 12-13 sectors per track out of 15. Which I think is about the same as before.

It is still the case that IMD can't see any data on 1.2mb disks that have files on. Or at least they should have files (even if corrupt). But I have no other drives to confirm that.

RetroPCCupboard wrote on 2025-06-21, 15:07:

It is still the case that IMD can't see any data on 1.2mb disks that have files on. Or at least they should have files (even if corrupt). But I have no other drives to confirm that.

Which would suggest that this drive is, if nothing else, badly misaligned. There is little you can do further without a scope.

There is a blue trim pot on the PCB, close to the heads. With a dab of red paint on it, to mark the final position and lock it in place. See if you can tell if the paint was cracked and the pot rotated. If so maybe try to guess the original position and turn it. If the paint looks intact then there is little reason to mess with it (esp. without a scope).
You can also disconnect and reconnect the heads plug, right next to it. Sometimes it can oxidize and the contact is poor, could help. Again after any of these try formatting the floppy to check if you can see more sectors. If not, and both heads are affected, there is clearly something wrong there. With good head pressure even on misaligned drive, a fresh format should always read back properly. And that's not the case here.

Deunan wrote on 2025-06-21, 15:32:

Which would suggest that this drive is, if nothing else, badly misaligned. There is little you can do further without a scope.

I have a scope. A mini DSO that I bought maybe 10yrs ago when I was mucking around with learning electronics. It is tiny.

I have completely forgotten how to use it. Lol. I think it is a bit of a toy. So maybe not accurate enough for this. Assuming it still works.

Deunan wrote on 2025-06-21, 15:32:

There is a blue trim pot on the PCB, close to the heads. With a dab of red paint on it, to mark the final position and lock it in place. See if you can tell if the paint was cracked and the pot rotated. If so maybe try to guess the original position and turn it. If the paint looks intact then there is little reason to mess with it (esp. without a scope).
You can also disconnect and reconnect the heads plug, right next to it. Sometimes it can oxidize and the contact is poor, could help. Again after any of these try formatting the floppy to check if you can see more sectors. If not, and both heads are affected, there is clearly something wrong there. With good head pressure even on misaligned drive, a fresh format should always read back properly. And that's not the case here.

The blue pot doesn't appear to have moved:

I will try the heads plug

RetroPCCupboard wrote on 2025-06-21, 16:28:

I have a scope. A mini DSO that I bought maybe 10yrs ago when I was mucking around with learning electronics. It is tiny.

Might be good enough. Near the pot there is TP5 test point, looks to be GND. Verify that it is. Then closer to the chip there are 2 more TPs, these should be the head amp outputs. You can monitor both or just one, try with head zero first (the lower one). To figure out which is which the easiest way I found is to gently lift the upper head a bit, it'll loose the signal complety - but it also affects the lower head, as the pressure is now gone. The test point that is less affected by removing pressure from the upper head should be head zero.

Start with the floppy you formatted, note the signal level from both heads at several points (cylinders). Then insert the Civ floppy, or what have you, and look at that. You should still see something, but perhaps very weak.

Note that MFM does not produce nice, uniform patterns but there are only a couple of shapes that can happen. So try the scope in one-shot mode a few times and try to find two closest peaks. Might be easier to stick to just positive or negative peaks and then double the result. You should be able to find two peaks 4us apart, which is 250kHz, so doubled (or if measuring positive to negative and getting 2us) it is 500kbps. Which is what we expect. Signal level depends on the chips and drive but in general you want at least 200mV peak to peak.

RetroPCCupboard wrote on 2025-06-21, 15:07:

Ok. Tried with a piece of blutac. I formatted the disk again in IMD, then tried align. Both heads are able to read 12-13 sectors per track out of 15. Which I think is about the same as before.

The IMD output you showed earlier was reading 2 sectors... were you getting 12-13 sometimes?

It is still the case that IMD can't see any data on 1.2mb disks that have files on. Or at least they should have files (even if corrupt). But I have no other drives to confirm that.

IMD doesn't implement any sort of filesystem.. It's for lowest level drive access only (cylinders/heads/sectors)
but .. filesystems are built on top of cylinders/heads/sectors ... so while IMD won't be able to discern individual files,
it will most certainly see the sectors/data that make up those files!

Regarding 1.2m drive - you have to have the transfer rate set correctly. For HD (1.2m) disk in 1.2m drive. this would be
500kbps - for DD(360k) disk it will be 300kbps.

When you 'R'eadDisk function, IMD will perform an analasys to determine the transfer speed, heads and
number of sectors/track.... In otherwords, doing 'R'ead to see what IMD thinks is there!

Deunan wrote on 2025-06-21, 16:57:

Might be good enough. Near the pot there is TP5 test point, looks to be GND. Verify that it is. Then closer to the chip there are 2 more TPs, these should be the head amp outputs. You can monitor both or just one, try with head zero first (the lower one). To figure out which is which the easiest way I found is to gently lift the upper head a bit, it'll loose the signal complety - but it also affects the lower head, as the pressure is now gone. The test point that is less affected by removing pressure from the upper head should be head zero.

Start with the floppy you formatted, note the signal level from both heads at several points (cylinders). Then insert the Civ floppy, or what have you, and look at that. You should still see something, but perhaps very weak.

Note that MFM does not produce nice, uniform patterns but there are only a couple of shapes that can happen. So try the scope in one-shot mode a few times and try to find two closest peaks. Might be easier to stick to just positive or negative peaks and then double the result. You should be able to find two peaks 4us apart, which is 250kHz, so doubled (or if measuring positive to negative and getting 2us) it is 500kbps. Which is what we expect. Signal level depends on the chips and drive but in general you want at least 200mV peak to peak.

Thanks for the pointers. A bit over my head at the moment, but will see if I can figure out this oscilloscope.

DaveDDS wrote on 2025-06-21, 19:37:

The IMD output you showed earlier was reading 2 sectors... were you getting 12-13 sometimes?

I am getting data like this when I format a disk with two sides, 80 tracks, single step, interleave 1, 15 sectors per track, 512b sectors,.500kbps MFM and then (A)lign:

The first test I did with formatting in IMD I screwed up and used 500kbps FM rather than MFM.

IMD doesn't implement any sort of filesystem.. It's for lowest level drive access only (cylinders/heads/sectors)
but .. filesystems are built on top of cylinders/heads/sectors ... so while IMD won't be able to discern individual files,
it will most certainly see the sectors/data that make up those files!

Exactly what I thought. But I just seem to get a screen full of question marks with 1.2mb disks that were not formatted with IMD. The only disk with data that I could read was a 360kb disk. That read OK in DOS, though I had to retry a few times to get the files off it.

Regarding 1.2m drive - you have to have the transfer rate set correctly. For HD (1.2m) disk in 1.2m drive. this would be
500kbps - for DD(360k) disk it will be 300kbps.

Yeah, 500kbps is what I chose when formatting, amd I think that's what IMD is detecting when I align on a disk formatted with IMD.

When you 'R'eadDisk function, IMD will perform an analasys to determine the transfer speed, heads and
number of sectors/track.... In otherwords, doing 'R'ead to see what IMD thinks is there!

I will try a disk with data tomorrow and see what IMD detects.

I am a bit confused with what is happening. Is it both a head alignment issue and an issue with heads reading densely packed data (but OK with lower density media)?

RetroPCCupboard wrote on 2025-06-21, 21:52:

I am a bit confused with what is happening. Is it both a head alignment issue and an issue with heads reading densely packed data (but OK with lower density media)?

I still haven't ruled out bad spindle motor speed. If it's 300 and not 360 rpm then the flux changes (the "bits") are much closer to each other on the media, and there might be just enough interference to make the floppy kinda work (some sectors are seen) but it's very random. During read the lower speed also makes the signal level lower, so on top of that noise is more problematic. All that would be much less a problem for double density format, the rotation speed is good and the tracks are twice wider so even misaligned drive should be able to pick it up (if it's not more than a cylinder off).

You know, since you have a scope you can test that - once the floppy is rotating and drive is selected, one of the pins on the interface carries the INDEX pulses. But start with IMD, it also has a test for that. See what speed you get. 360 is what we expect but dual-speed drives might report 300 if the HD signal is not driven. Or if the jumpers block speed switching.

If you get some weird result, more than 3% off from 300 or 360, it might be a problem with the spindle motor sub-PCB. These sometimes have electrolytic caps that leak and corrode stuff around them.

RetroPCCupboard wrote on 2025-06-21, 21:52:

I am a bit confused with what is happening. Is it both a head alignment issue and an issue with heads reading densely packed data (but OK with lower density media)?

If you are reading back to the drive that formatted/wrote the disk, alignment should have no effect (it would have to be so grossly misaligned that track-0 stop doesn't work)

Alignment is just how close a drive is to laying down tracks in the right place (which is where other drives should be placing them)
without moving from one drive to another, this has no real effect)

There may be an electrical problem causing problems accessing at higher data rates... but far more common with IMD is use of the wrong data
rate settings. Remember that IMD lets/requires you to tell it exactly how to access the disk... The only "automatic" detection is when reading a disk
it will attempt to see what data rate and step type work...

IMD doesn't go through BIOS and therefore doesn't "know" what type of drive is connected (I tend to leave drives I plan to use with IMD configured
as "none") It will do some trial-and-error to figure out what works - but unlike DOS/Windows/Linux etc... it doesn't have tables of what settings
are appropriate for configured drive types.

This may seem odd - but there is a reason. IMD was created to archive/restore disks for classic computer systems. These vary widely in format, and
can (and often do) be completely incompatible with more "modern" PC operating systems... I created IMD to be able to read/write/format anything
the Nev765 PC FDC and its hardware is capable of doing... So it doesn't assume "drive types".

DaveDDS wrote on 2025-06-22, 04:06:

There may be an electrical problem causing problems accessing at higher data rates... but far more common with IMD is use of the wrong data
rate settings. Remember that IMD lets/requires you to tell it exactly how to access the disk... The only "automatic" detection is when reading a disk
it will attempt to see what data rate and step type work...

I just tried Rotor.

The disk size reported by DOS seems odd to me. I Didn't think capacities went below 360k?

Anyhow I was able to copy the files from it to c: using xcopy.

I then tried IMD read to file. It seemed to experience problems when reading. Not sure if the disk image is good or not

DaveDDS wrote on 2025-06-22, 04:06:

There may be an electrical problem causing problems accessing at higher data rates... but far more common with IMD is use of the wrong data
rate settings. Remember that IMD lets/requires you to tell it exactly how to access the disk... The only "automatic" detection is when reading a disk
it will attempt to see what data rate and step type work...

I now tried a C++ disk:

MS DOS doesn't like it at all:

IMD read Disk image to file doesn't like it:

IMD align test sees nothing on either head for any track I tried. I pressed D to read data whilst analysing:

Note though that I have no idea on whether this disk is actually OK.

Deunan wrote on 2025-06-21, 22:46:

But start with IMD, it also has a test for that. See what speed you get. 360 is what we expect but dual-speed drives might report 300 if the HD signal is not driven. Or if the jumpers block speed switching.

If you get some weird result, more than 3% off from 300 or 360, it might be a problem with the spindle motor sub-PCB. These sometimes have electrolytic caps that leak and corrode stuff around them.

Well the test speed thing doesn't seem to work on either my Rotor disk (low capacity disk) or my C++ disk (1.2mb). It just gives no result (no error though):

I then tried Deluxe paint 2, which DOS won't read and it gave me this:

I'd say this isn't very close to 360RPM. Lol.

The Align test on this disk is seeing some data on the disk:

If I am reading this correctly, it is seeing data from a different track to what was selected

I am not sure on the size of this disk. But there's a disk 1 and a disk 2 for this program.

Try that Rotor disk for speed test in IMD. It seems to at least read it, might not be good enough for that purpose though.

The IMD dump seems to work, it read 720 sectors, which is 9 per track, 2 track per cylinder, 40 cylinders. 9*2*40=720. But IMD also reported data rate of 500k for that... which is very odd. Can you perhaps upload it somewhere to test it? If you don't want to due to copyright reasons, that's OK.

As for using the scope to test the drive rpm, that's easy. Here's some basic settings so you don't have to guess them:
- horizontal set to 1ms/div
- vertical set to 1V/div (adjust if the signal is small and barely visible)
- channel coupling to DC, remember to connect GND (the probe should have an alligator clip for that)
- probe pin #8 on the connector while IMD is running align function (so that the drive is selected and rotating)
- adjust trigger to lock onto the signal, so it doesn't float around

Any floppy, even unformatted, should work here as we test for the pulses generated by the hole, not sectors on media. To have a clean signal you want GND connected as close to the signal source as possible - but this is more imporant for higher frequency, weak stuff like head signals. This is TTL so if you connect the clip anywhere else (like one of the free power connectors from the PSU maybe) it'll be good enough. Just a reminder that the drive body is metal, sure, but not always connected to GND. There are jumpers for that, on some drives these might be factory soldered or missing completly. If in doubt check with ohm meter first.

If your probe is 10:1 and has no switch for 1:1 mode, you need setup the channel for that, or just keep in mind you will get 10x attenuation so the vertical should be set to 100mV/div instead. 10:1 mode is great for high frequency signals but if you find it confusing stick to 1:1 setting for the rpm test. Here you are expecting a nice, clean square signal with almost 5V swing so anything will work.

You might have to adjust the trigger level so that if falls within the signal you see, to lock it. This is usually a line you can move up/down so it overlaps with the signal. Once locked the scope built-in functions should tell you the frequency of the signal, but if not just count the bars vs dividers. Use the scope one-shot function to capture a single, frozen frame and then you can let go of the probe and count in peace. Adjust the horizontal setting as needed. Take a photo/screenshot and post here if in doubt.

You can get access to the INDEX signal right on the ege connector, the plastic should not cover it completly, there is usually enough gold exposed to touch it with the probe tip. If not, try pulling the plastic plug back a little bit.

Deunan wrote on 2025-06-22, 09:52:

Try that Rotor disk for speed test in IMD. It seems to at least read it, might not be good enough for that purpose though.

This is IMD output when running RPM test on the Rotor disk:

Thanks for the detailed instructions on how to use scope with this drive. I will see if I can remember how to use the scope. Lol. It's been 10yrs, and I didnt really use it much. Haha.

Deunan wrote on 2025-06-22, 09:52:

As for using the scope to test the drive rpm, that's easy. Here's some basic settings so you don't have to guess them: - horizon […]
Show full quote

As for using the scope to test the drive rpm, that's easy. Here's some basic settings so you don't have to guess them:
- horizontal set to 1ms/div
- vertical set to 1V/div (adjust if the signal is small and barely visible)
- channel coupling to DC, remember to connect GND (the probe should have an alligator clip for that)
- probe pin #8 on the connector while IMD is running align function (so that the drive is selected and rotating)
- adjust trigger to lock onto the signal, so it doesn't float around

I've been struggling with this scope. Very fiddly to change the settings. Especially when trying to hold a probe onto the drive..

Anyway, it was reporting 8.000hz as the frequency. Which would make it 480RPM I believe? Not sure whether or not I believe the scope. haha.

I have to be honest, I'm not sure what all of the things are on the screen. I tried 1ms/div as suggested, but could only see one pulse at a time.

RetroPCCupboard wrote on 2025-06-24, 20:16:

I have to be honest, I'm not sure what all of the things are on the screen. I tried 1ms/div as suggested, but could only see one pulse at a time.

I picked 1ms/div because I was also interested in the rising/falling slopes, and when triggering properly (and with a decent refresh rate) it's easy to spot the pulse getting out of shape, even if briefly. But yes, to see more at once and to be able to figure out the frequency you'd need to change that to slower timebase. And you did, the 20ms on the first image is good. There's 5 full divisions, half of one on the left and 2.5/10 on the right. Which is about 5*20+7.5/10*20=115ms. The scope says dT=119ms, but I'm not sure if it's for the cursors or is it able to calculate it better than what is shown. Close enough either way.

The problem is that's way too fast. Even the slower of the two times, 119ms, is about 8.4 rotations per second, or 504rpm. And it should be 360, or 300. If we calculate 504/300=1.68, and then use that factor to adjust the MFM bitrate for DD, which is 250k, we get 250*1.68=420k. Which kinda explains why the floppy controller can, somehow, lock on to it using 500k rate, but even that is with difficulty due to quite big difference. Frankly it's surprising the PLL can lock at all.

Can you make a photo of the spindle motor PCB? It's on the bottom of the drive. There's possibly some damage there from leaked capacitors that needs fixing.

Deunan wrote on 2025-06-24, 20:41:

Can you make a photo of the spindle motor PCB? It's on the bottom of the drive. There's possibly some damage there from leaked capacitors that needs fixing.

All of the components are on the other side:

Is it just a case of undoing those 4 screws, or is there more to it than that?

RetroPCCupboard wrote on 2025-06-24, 20:55:

Is it just a case of undoing those 4 screws, or is there more to it than that?

Huh. Usually the screws are from the top, 3 of them. What you have would put the stress of the clamping mechanism on the PCB.
Well, try to undo them and lift the PCB, at least rotate it to make a photo of the components side. Try to wiggle it a bit, might be properly stuck after all these years - the fit is usually pretty good but it should not be so tight as to require any force. So if you can't seem to get it off, put the screws back and try to get a better look from the top. Could be there's screws there too.

EDIT: Do not touch these 3 screws in the rotor. Those are quite likely glued and can be left-threaded. There's no need to mess with the rotor.

Just posted this which might be of interest:

Bootable diagnostics disk