DaveDDS wrote on 2025-06-18, 20:41:

I just double-checked on "Daves Old Computers", and under Software/Images ImageDisk shows as 1.20 I also downloaded the IMD120.Z […]
Show full quote

I just double-checked on "Daves Old Computers", and under Software/Images ImageDisk shows as 1.20
I also downloaded the IMD120.ZIP it points at an when I run that IMD.COM it shows itself as 1.20

Sometimes browsers cache a LOT - that why I recommend flushing your browsers cache before downloading.
If you have an old page cached which points at IMD119.ZIP, that archive may still be on the site.

I downloaded it from here: http://dunfield.classiccmp.org/dos/index.htm

the file was called imd.zip, not imd20.zip. I wonder if I managed to find the wrong page?

DaveDDS wrote on 2025-06-18, 20:41:

You should feel some resistance - not much, but you should be able to tell the head is being pressed on the paper. If you use a […]
Show full quote

You should feel some resistance - not much, but you should be able to tell the head is being pressed on the paper.
If you use a thin strip which can fall to the side with the drive vertical and heads open - is it still that loose when the heads are
closed on it (and make sure you are getting it under the head - some drives have shielding which makes it tricky to get
it in the right place.

If there is no output from 'A'lign/tests = it would seem there's no data from the FDC.

When you 'E'rase, does it step over the whole disk?
Can you 'E'rase with no disk in the drive?

Erase did step over the whole disk. I am about to go to bed now (in UK), but will check erasing with no disk tomorrow and also try the paper test again.

edit: ok, I have found 1.20 under software/images.

DaveDDS wrote on 2025-06-18, 20:41:

I just double-checked on "Daves Old Computers", and under Software/Images ImageDisk shows as 1.20 I also downloaded the IMD120.Z […]
Show full quote

I just double-checked on "Daves Old Computers", and under Software/Images ImageDisk shows as 1.20
I also downloaded the IMD120.ZIP it points at an when I run that IMD.COM it shows itself as 1.20

Sometimes browsers cache a LOT - that why I recommend flushing your browsers cache before downloading.
If you have an old page cached which points at IMD119.ZIP, that archive may still be on the site.

You should feel some resistance - not much, but you should be able to tell the head is being pressed on the paper.
If you use a thin strip which can fall to the side with the drive vertical and heads open - is it still that loose when the heads are
closed on it (and make sure you are getting it under the head - some drives have shielding which makes it tricky to get
it in the right place.

If there is no output from 'A'lign/tests = it would seem there's no data from the FDC.

When you 'E'rase, does it step over the whole disk?
Can you 'E'rase with no disk in the drive?

If I try to erase with no disk in the drive I get this:

I have one disk now that seems to be showing something on track 0 head 0. But no other track on that disk with either head:

I don't know what the numbers mean, but at least its a sign of life.

I then used erase disk function of IMD and then the Align method. This time it was giving ??? again. So no data. I then went back to dos and tried an unconditional 1.2mb format. It reported error "Invalid Media or Track 0 bad - disk unusable". I then went back to IMD again and find the same kind of data showing on T0 H0, but nowhere else. So it would appear that the format command can write to T0 H0, and IMD can read from that track and head.

I have done a test where I put a small strip of paper between the disk and the head, then held drive vertically. The paper didn't fall out. So head does appear to be making contact with it.

A quick follow up. I tried writing a disk image again using IMD. It gave same errors as before (NoData) but I left it longer and could see it started moving the head and was going track by track. Still reporting NoData. I stopped it after 3 tracks and then did Align test again. It is showing data from both heads from the first 3 tracks.

So, clearly, the drive is able to read and write from the heads. But clearly its not doing it correctly, as IMD and Format command both complain.

My apologies, I didn't realize an older IMD.ZIP was still available under "DOS widgets"
- I have removed it from there, and added a note under "Special purpose packages"
to indicate that it is available and where to find it.
(btw, if you found/downloaded it from somewhere else on the site, please let me know)

--

I should mention that the PC FDC (Nec765) does NOT have an "erase" function,
IMD does an effective erase by formatting the disk with an unreasnably large sector
size - which actually fails (IMD does not report this) and overwrites the entire
track with an unreadable partial sector.

In other words, 'E'rase is essentially doing a 'F'ormat, which does need the index
pulse to know where to start on the track - in other words, this means that your
index sensor is indeed working!

--

Align/Test is one of the more unusual funcitons offered by ImageDisk, and kinda
requires you to know a fair but about floppy disk internals in order to make full
use if it's capabilities. The ImageDisk documentation and interactive help do describe
what the number indicate ...

Also be aware that as described above, 'E'rase intentionally makes unreadable
sectors.

It might be worth trying 'F'ormat .. this won't create a disk that an OS can read
(it won't have a filesystem), but will low-level format a diskette, which you can
then read/align/test within ImageDisk.

--

Also, have you tried giving the drive heads a good cleaning?
A common problem with older floppy drives is that the heads get crud on them which
prevents them from reading/writing sectors correctly ... given that you are seeing
some data transfer, this seems quite likely.

The best way to clean a drive is to disassemble it to the point where you can
use an aicohol swab to give a good (and gentle) cleaning... This can be fairly
difficult on some drives...

You used to be able to get "cleaning disks" which would allow you "scrub" the
disks with an alcohol solution - unfortunately OS's don't include a function to
properly perform this, so you end up just trying to access the disk (which will
fail) and only uses the first 1-2 tracks of the cleaning media.

ImageDisk has a 'C'lean function which will scrub the head over all tracks of such
cleaning media.

I've not seen such "cleaning disks" available for a long time - I've made my own 5.25
by carefully cutting the very back edge of a diskette casing, removing the original
media, and using it as a template to carefully/accurately cut out a replacement
from thick paper. Sometimes I use the center "circle" from the original media to re-
enforce that ring at the center - FWIW

DaveDDS wrote on 2025-06-19, 09:13:

My apologies, I didn't realize an older IMD.ZIP was still available under "DOS widgets" - I have removed it from there, and adde […]
Show full quote

My apologies, I didn't realize an older IMD.ZIP was still available under "DOS widgets"
- I have removed it from there, and added a note under "Special purpose packages"
to indicate that it is available and where to find it.
(btw, if you found/downloaded it from somewhere else on the site, please let me know)

No problem. Yes, that's where I got it

DaveDDS wrote on 2025-06-19, 09:13:

I should mention that the PC FDC (Nec765) does NOT have an "erase" function, IMD does an effective erase by formatting the disk […]
Show full quote

I should mention that the PC FDC (Nec765) does NOT have an "erase" function,
IMD does an effective erase by formatting the disk with an unreasnably large sector
size - which actually fails (IMD does not report this) and overwrites the entire
track with an unreadable partial sector.

In other words, 'E'rase is essentially doing a 'F'ormat, which does need the index
pulse to know where to start on the track - in other words, this means that your
index sensor is indeed working!

Interesting info, and great news that sensor is OK.

DaveDDS wrote on 2025-06-19, 09:13:

Align/Test is one of the more unusual funcitons offered by ImageDisk, and kinda requires you to know a fair but about floppy dis […]
Show full quote

Align/Test is one of the more unusual funcitons offered by ImageDisk, and kinda
requires you to know a fair but about floppy disk internals in order to make full
use if it's capabilities. The ImageDisk documentation and interactive help do describe
what the number indicate ...

Also be aware that as described above, 'E'rase intentionally makes unreadable
sectors.

OK. I will look at the help. I did look in help for other functions, but not that one

DaveDDS wrote on 2025-06-19, 09:13:

It might be worth trying 'F'ormat .. this won't create a disk that an OS can read
(it won't have a filesystem), but will low-level format a diskette, which you can
then read/align/test within ImageDisk.

OK. I did look at that, but wasn't sure what to put in for the settings it required...

DaveDDS wrote on 2025-06-19, 09:13:

Also, have you tried giving the drive heads a good cleaning? A common problem with older floppy drives is that the heads get cru […]
Show full quote

Also, have you tried giving the drive heads a good cleaning?
A common problem with older floppy drives is that the heads get crud on them which
prevents them from reading/writing sectors correctly ... given that you are seeing
some data transfer, this seems quite likely.

The best way to clean a drive is to disassemble it to the point where you can
use an aicohol swab to give a good (and gentle) cleaning... This can be fairly
difficult on some drives...

You used to be able to get "cleaning disks" which would allow you "scrub" the
disks with an alcohol solution - unfortunately OS's don't include a function to
properly perform this, so you end up just trying to access the disk (which will
fail) and only uses the first 1-2 tracks of the cleaning media.

ImageDisk has a 'C'lean function which will scrub the head over all tracks of such
cleaning media.

I've not seen such "cleaning disks" available for a long time - I've made my own 5.25
by carefully cutting the very back edge of a diskette casing, removing the original
media, and using it as a template to carefully/accurately cut out a replacement
from thick paper. Sometimes I use the center "circle" from the original media to re-
enforce that ring at the center - FWIW

I did try cleaning the heads with a cotton bud and IPA. Some stuff did come off. But I guess I can give it another go.

Just to follow up though. Rather than trying blank formatted 1.2mb disks I thought I'd try some disks with data on. I had tried them when I first got the drive, but that was before changing the MDB jumpers suggested by @horun

First I tried my Sid Meier Civilization disks. Those are 1.2Mb and, as with the bank disks, MS DOS just says General Error reading.

I then tried StarGlider. This seems to be a 360k disk. To my surprise I got a directory listing:

After some retries was able to copy it to the HDD:

And then was able to play it!

Note that the IMD Align function seems to find nothing on the 1.2mb Civilization disk. I have no idea if the disk is good though, as I only recently got it and have no other drives

RetroPCCupboard wrote on 2025-06-19, 09:46:

Note that the IMD Align function seems to find nothing on the 1.2mb Civilization disk. I have no idea if the disk is good though, as I only recently got it and have no other drives

Put an opaque write protection tab/sticker over the notch of any original media you do not want erased. If it's not there already. Mistakes happen and the write lockout is hardwired into the drive, so other than some really weird hardware failure it will protect the floppy.

Anyway, try putting a jumper over the the BX on your drive (it's right next to the HDA/HDB jumpers). Then try the 1.2M floppies and read/write/format. Could be your drive is locked to 300rpm and not not switching to 360rpm properly. It is a dual-speed drive after all. Unless somebody replaced the spindle motor PCB with a wrong one at some point.

To do the most common PC floppy low-level formats with ImageDisk:

First:
- Set Sides to: Two *1
- Set Cylinders to 10 *2
- Set Double-step to: OFF *3
- Set Interleve to: 1

Then use 'F'ormat, pick:

Sectors/Track to the values shown below
- 360k: 40Cylinders 2sides 9sectors/track 512byte/sectors = 3,68,640 total
- 1.2m: 80Cylinders 2sides 15sectors/track 512bytes/sector = 1,228,800 total

StartSector: 1

Pick data rate depending on drive and media type:
- 360k: 250kbps
- 1.2m: 500kbps
- 360k in 1.2m drive: 300kbps *4

Pick SectorSize: 512

*1 to speed up testing, you can format only 1 side.
- You can also do this to see if side1 works (even if side2 doesn't)
- Testing only side2 is tricker, you can use the 'X'clusion map

*2 10 Cylinders will speed up testing a LOT while still moving the head
enough that you can see it. Drive Cylinders are actually as shown in table above.

*3 for 360k media in 1.2Mm drive, double-step should be set ON, but
it won't prevent testing .. I'd leave it off to simplify.

*4 360k drives rotate at 300rpm. 1.2M drives rotate at 360rpm.
To get the right bit spacing for 360k media, you have to tansfer faster on
a 1.2 - some drives are "dual speed" (and I give some instructions on my page
for making a Panasonic drive switchable) - but most OSs don't change drive
speed and still use 300kbps for DD in HD drive.

Deunan wrote on 2025-06-19, 11:18:

Put an opaque write protection tab/sticker over the notch of any original media you do not want erased. If it's not there already. Mistakes happen and the write lockout is hardwired into the drive, so other than some really weird hardware failure it will protect the floppy.

Anyway, try putting a jumper over the the BX on your drive (it's right next to the HDA/HDB jumpers). Then try the 1.2M floppies and read/write/format. Could be your drive is locked to 300rpm and not not switching to 360rpm properly. It is a dual-speed drive after all. Unless somebody replaced the spindle motor PCB with a wrong one at some point.

It was worth a try, but it didn't seem to make a difference. I still get an error when trying the DOS format command on a blank disk

DaveDDS wrote on 2025-06-19, 12:32:

To do the most common PC floppy low-level formats with ImageDisk:

Many thanks for the detailed instructions. I have formatted a 1.2mb disk with 80 cylinders. There were no errors and the Align function is seeing data. But I am not really sure what I am looking for:

RetroPCCupboard wrote on 2025-06-20, 18:44:

Many thanks for the detailed instructions. I have formatted a 1.2mb disk with 80 cylinders. There were no errors and the Align function is seeing data. But I am not really sure what I am looking for:

The first number is what the sector headers report as cylinder. So you are at 4 (from the bar at the top of the screen) and the reported value is 2... Which is wrong. Or did you format the floppy to 360k with double stepping? In that case it's correct, but still IMD should be seeing more sectors. Right now it reports seeing 2 sectors, both wrong - but that is due the cylinder mismatch I just mentioned.

Dave might correct me but IIRC 0/2 means 0 good, 2 bad. See if switching to double step changes that to 2/0. In which case it's a small success but still not entirely good, as there should be 9 not 2 sectors per each track in 360k format.

During Align/Test, the numbers shown are:

Cylinder(track) #sectors-read #match #fail

Do keep in mind that this function was basically created to help align drives..
#match is the number of sectors in which the Cylinder number in header MATCHes the physical cylinder
#fail is "" DOES-NOT MATCH ""

Sectors which are not read don't show up.
so.. #sectors should be the same as the number of sectors/track you formatted.

I highly recommend NOT enabling double-step at all in this kind of low-level
testing ... this will avoid confusion about sector numbering. The resulting
format may not be entirely "PC compatible" (ie: DD in a 1.2m drive would result
in an 80-cylinder 720k diskette) - but all results will be consistant.

#match+#fail should (will) be the same as #sectors - as you adjust the head
aignment away form "perfect" on either side, you will see #fail increase as
#match decreases. The best alignment point is 1/2 way between the outer positions
where #match=#sectors, #fail=0 on either side of "perfect".

If double-step is kept OFF, and you format a disk in a drive, then to 'A'lign/test
in that same drive, it should always show #sectors=#match=sectors/track you formatted
**EVEN IF THE DRIVE IS NOT ALIGNED** (Alignment requires using an alignment disk,
or one freshly formatted on a new (presumably well aligned) drive)

If #sectors does NOT match formatted sectors/track, this indicates the drive is
having trouble reading(or writing) the sectors. Most commonly this indicates not
good magnetic coupling between the head and the media - most commonly because of
dirty heads, or possible some other physical factor (head pressure,
positioning etc.) - worst case it could be a drive noise/electronics problem (but uncommon
to see some sectors reading good in this case)

If #fail is non-zero when testing a disk you just formatted in the same drive the
most likely problem is head positioning - could be stepping hardware, or track-zero
switch.

---
Sorry - I know the 'A'lign/Test function in kinda complicated to get your head around... In working with
some older/less-reliable drives I saw a need for it, and to be honest, I wrote ImageDisk because I wanted
something like it - I decided to publish it after I realized how useful it could be - because of this, it was never
designed to be a "just use without thinking about it" kind of tool, and you really have to "get into" floppy
drive internals to make best use of it - but ... it also allows you do things that "just use..." software would
not be able to do.

Deunan wrote on 2025-06-20, 23:19:

The first number is what the sector headers report as cylinder. So you are at 4 (from the bar at the top of the screen) and the reported value is 2... Which is wrong. Or did you format the floppy to 360k with double stepping? In that case it's correct, but still IMD should be seeing more sectors. Right now it reports seeing 2 sectors, both wrong - but that is due the cylinder mismatch I just mentioned.

Dave might correct me but IIRC 0/2 means 0 good, 2 bad. See if switching to double step changes that to 2/0. In which case it's a small success but still not entirely good, as there should be 9 not 2 sectors per each track in 360k format.

Thanks. I think I used a wrong setting. I did the test again and now the reported cylinder matches the position I am at on the disk. These are the settings I used this time

Previously I used same settings apart from O used 500kbps FM, and this time I used 500kbps MFM

DaveDDS wrote on 2025-06-21, 02:06:

Sorry - I know the 'A'lign/Test function in kinda complicated to get your head around... In working with some older/less-reliabl […]
Show full quote

Sorry - I know the 'A'lign/Test function in kinda complicated to get your head around... In working with
some older/less-reliable drives I saw a need for it, and to be honest, I wrote ImageDisk because I wanted
something like it - I decided to publish it after I realized how useful it could be - because of this, it was never
designed to be a "just use without thinking about it" kind of tool, and you really have to "get into" floppy
drive internals to make best use of it - but ... it also allows you do things that "just use..." software would
not be able to do.

No need to apologise. Looks like an amazing tool once I get my head around it. I have actually been learning a bit about floppy disk data structure from Peter Norton's "Inside the IBM PC" book. So, I am finding this very interesting.

Anyhow, now I have formatted the disk with 500kbps MFM instead of 500Kbps FM I get this:

So, if I understand correctly, it's locating the requested position on disk ok, but is struggling with reading sectors. Sometimes it finds 15, sometimes less.

I am going to have another go at cleaning the heads.

Cleaning with a cotton bud again has made no difference. Do you think a deeper clean is required? Would I need to make one of those disk cleaners? I have seen some people make them from 125mm coffee filters and make a case out of a plastic filing folder...

I have had this drive a week now, so really should return it to the seller soon if this drive is looking unfixable...

I have bad exlerience with cleaning disks, and I always found the cotton buds to work better if you have proper access to the heads.

As for returning the drive, I would have requested a return immediatelly after finding it doesn't work (unless the seller had a disclaimer on it as "not tested" or "for parts"). As a seller, I expect the return to be the item in the same state when I sold it. Dismantling and cleaning the drive is a change of state that the seller may not like. That being said, many sellers just give up on the item if they're told it's not working, so then it's yours to fix and tamper with. Otherwise it would have been best to just repacked it in its original seller box and ship it back.

wbahnassi wrote on 2025-06-21, 08:30:

I have bad exlerience with cleaning disks, and I always found the cotton buds to work better if you have proper access to the heads.

As for returning the drive, I would have requested a return immediatelly after finding it doesn't work (unless the seller had a disclaimer on it as "not tested" or "for parts"). As a seller, I expect the return to be the item in the same state when I sold it. Dismantling and cleaning the drive is a change of state that the seller may not like. That being said, many sellers just give up on the item if they're told it's not working, so then it's yours to fix and tamper with. Otherwise it would have been best to just repacked it in its original seller box and ship it back.

Hmm. OK. I did tell him straight away that it didn't work. I wouldn't say that I have dismantled it. There was just two screws to undo in order to see the head and mechanism. It was a cover - nothing more.

I haven't removed anything else that could cause alignment issues. I guess I will just point him at this thread and see if he thinks I have gone too far. In my eyes I am just trying to make work something that was mis-sold to me (unintentionally , I am sure).

If I were the seller I wouldn't want to pay for the return shipping, which is what would happen if I filed an "Item not as described" return. Ideally I want the drive to work, as the colour would fit well with the case I want to use.

wbahnassi wrote on 2025-06-21, 08:30:

I have bad exlerience with cleaning disks, and I always found the cotton buds to work better if you have proper access to the heads.

Something I've done a few times ... take a long thin strip of thick paper, insert under the head cover from one site, get it to go under the head,
move it through enough so you will be able to pull the other end ... close the heads, spray cleaner on the strip (don't forget to fold it back over
enough to get the other side as well), then carefully pull it back and forth under the heads a few times...

DaveDDS wrote on 2025-06-21, 11:25:

Something I've done a few times ... take a long thin strip of thick paper, insert under the head cover from one site, get it to go under the head,
move it through enough so you will be able to pull the other end ... close the heads, spray cleaner on the strip (don't forget to fold it back over
enough to get the other side as well), then carefully pull it back and forth under the heads a few times...

I tried that. I am trying to be gentle though, so maybe more could be done.

I have just taken a photo of the bottom head. Sorry, that quality isn't great, as was just using mobile phone and a flashlight to get more illumination. It's hard to get a good angle on it

Around the head looks kinda dusty, but I think the head itself is clean? What do you think?

Actually, I managed to get a better picture with the 10x zoom and moving further away with the drive on its side rather than horizontal.

Very nice photos. The heads are clean where it counts, although that dust (if that is dust) around is not great. It might get sucked onto the floppy surface by the draft. See if you can clean it at all with a cotton swab.

More importantly though, I just went through your YT videos again I only now I noticed that at least on one of them the spring that forces the upper head down seems to be missing? Can you verify that? It's a very important spring, the head needs a small amount of force to clamp onto the floppy surface.

Deunan wrote on 2025-06-21, 13:06:

Very nice photos. The heads are clean where it counts, although that dust (if that is dust) around is not great. It might get sucked onto the floppy surface by the draft. See if you can clean it at all with a cotton swab.

More importantly though, I just went through your YT videos again I only now I noticed that at least on one of them the spring that forces the upper head down seems to be missing? Can you verify that? It's a very important spring, the head needs a small amount of force to clamp onto the floppy surface.

Hi. Yes, that spring is there now. Forgot to mention that I took it off when cleaning the top head one time. It didn't read any data at all (from either head) without that spring.

The tension on it is enough to hold a piece of paper between the heads when I hold the drive vertical:

RetroPCCupboard wrote on 2025-06-21, 12:53:

I have just taken a photo of the bottom head. Sorry, that quality isn't great, as was just using mobile phone and a flashlight to get more illumination. It's hard to get a good angle on it

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.

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.

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.