Update:

I have tested the track zero switch with a multimeter and determined that it's working as it gives continuity and beeps every time when attempting to seek.
The behavior of the drive is the same with the switch disconnected with the bottom PCB. When I've reconnected the switch back, the floppy drive is out
of alignment. The cabling from the PCB has also been tested, and there is power coming to the switch, but there has been no difference with the middle
terminal when attempting to seek, as it's probably used to indicate the drive that the switch is active.

My diagnosis is that, when the operating system attempts to open a file, it cannot calibrate the drive.

tjenjic386 wrote on 2022-09-30, 09:08:

Update: […]
Show full quote

Update:

I have tested the track zero switch with a multimeter and determined that it's working as it gives continuity and beeps every time when attempting to seek.
The behavior of the drive is the same with the switch disconnected with the bottom PCB. When I've reconnected the switch back, the floppy drive is out
of alignment. The cabling from the PCB has also been tested, and there is power coming to the switch, but there has been no difference with the middle
terminal when attempting to seek, as it's probably used to indicate the drive that the switch is active.

My diagnosis is that, when the operating system attempts to open a file, it cannot calibrate the drive.

If it is of any help, Adrian's digital basement user on YT saved a drive that had a similar issue.

I've watched a few of Adrians Digital Basement video (one where he recovered the 800k apple, and a nice Okidata), but it didn't had the issue I'm having.

The switch and the bracket has been repositioned and alignment is back. With imd, you can seek to track zero easily. However, when you access the drive
and try to access the file, it simply would not seek back to track zero. When doing a format, it would seek at each track, but when it tries to return to track
zero, it does an incomplete step between tracks causing misalignment. Then I use imd to seek manually to track zero, and then alignment is back.

With the drive, only the directory listing can be done when you manually set the trive to track zero with imd.

Testfdc failes except with one test (which I think it doing wrong as it automatically seeks to the last tracks of the floppy) and it's Double Density 250k.

But I think I've found what the issue is. The drive is a single sided, single density drive according to the manual, which is incompatible with the P3 floppy controller,
and the IBM PC's controller card.

tjenjic386 wrote on 2022-10-02, 14:42:

I've watched a few of Adrians Digital Basement video (one where he recovered the 800k apple, and a nice Okidata), but it didn't […]
Show full quote

I've watched a few of Adrians Digital Basement video (one where he recovered the 800k apple, and a nice Okidata), but it didn't had the issue I'm having.

The switch and the bracket has been repositioned and alignment is back. With imd, you can seek to track zero easily. However, when you access the drive
and try to access the file, it simply would not seek back to track zero. When doing a format, it would seek at each track, but when it tries to return to track
zero, it does an incomplete step between tracks causing misalignment. Then I use imd to seek manually to track zero, and then alignment is back.

With the drive, only the directory listing can be done when you manually set the trive to track zero with imd.

Testfdc failes except with one test (which I think it doing wrong as it automatically seeks to the last tracks of the floppy) and it's Double Density 250k.

But I think I've found what the issue is. The drive is a single sided, single density drive according to the manual, which is incompatible with the P3 floppy controller,
and the IBM PC's controller card.

Basically it's a wrong controller for this unit.
Tricky

With a bit of research now, this equivalent drive is found on the Heathkit Z89 as the model and serial numbers match, which uses a 100k hard sector format
(serial L22742-A1005-B003), which might be nearly impossible to find a controller for it. I'll try to write a program to work around the issues accessing
files, and find a controller for hard sectors (if available).

tjenjic386 wrote on 2022-10-02, 14:42:

The switch and the bracket has been repositioned and alignment is back. With imd, you can seek to track zero easily. However, when you access the drive
and try to access the file, it simply would not seek back to track zero. When doing a format, it would seek at each track, but when it tries to return to track
zero, it does an incomplete step between tracks causing misalignment. Then I use imd to seek manually to track zero, and then alignment is back.

This screams "inappropriate step rate". Very early drives couldn't step as fast as the IBM PC/XT/AT BIOS tells the floppy controller to do. This means the drives miss steps when you step multiple tracks at once, but do single steps just fine. IIRC the step rate is included in the Int 1D floppy parameter table, so you should be able to swap it with a table that contains a lower step rate. But be careful: BIOSes with support for multiple densities will overwrite that pointer when a new format is established.

I've found a good article of adjusting the step rate through asm code, which is exactly what I was looking for.
I'll freely customize it the way I want it to be, experiment and put it in autoexec.bat

https://retrocmp.de/fdd/general/floppy-steprate.htm

tjenjic386 wrote on 2022-10-03, 10:06:

I've found a good article of adjusting the step rate through asm code, which is exactly what I was looking for.
I'll freely customize it the way I want it to be, experiment and put it in autoexec.bat

https://retrocmp.de/fdd/general/floppy-steprate.htm

Summing it up,

- you have the controller that goes with it
or
- you set the correct parameters via software

Correct?

I have an IBM PC with it's original controller and a P3 with it's onboard controller. The original Siemens controller for the drive (FDC100-5) is probably for a Z80 machine. I'm going to
write a program in assembly to change the settings of the step rate to 25ms (track-to-track in manual specs). Not sure if I need to change the head load time, but if only adjusting the step works,
then I won't touch.

http://bitsavers.trailing-edge.com/pdf/siemen … nce_Mar1979.pdf

Success! Adjusting the stepper rate really does work on both machines (P3, IBM PC) with a slight difference. On the P3 machine, formatting fails on the last step when
it needs to write something to boot(on IBM-PC, format works). The only thing that fails on both machines is when attempting to make a system disk, but that's not important
anyway, as I'm using both drives as drive B. I have written steprate.com using the debugger. This program without arguments resets the FDC to default settings, if supplied with
a one number argument (range 1-15), it sets the step rate to the value specified. Copying, deleting, reading files work on both machines.

Syntax:

steprate
steprate num

1
2; exclude comments when typing the program 
3
4debug
5-a 100
6push ds ; preserve the return address
7xor ax, ax
8push ax
9mov bx, 80 ;  skip spaces from command line arguments (if any)
10mov cx, [bx] ;  should of added a test and jump to avoid code falling if there is no argument, but forgot
11xor ch, ch
12mov di, bx
13inc di
14mov bx, cx
15mov byte ptr [bx+di], 24
16mov al, 20
17cld
18repz scasb
19dec di ; this sets di to the first character of a non space character of the command line argument (if any)
20xor dx, dx 
21mov si, di
22mov ax, dx
23mov cl, 2
24lodsb ; from here numeric string converts to an integer in dx (if argument is supplied)
25cmp al, 30
26jb 0138
27cmp al, 39
28ja 0138
29xor al, 30
30add dx, dx
31mov bx, dx
32shl dx, cl
33add dx, bx
34add dx, ax
35jmp 0121
36test dx, dx ; if there is no argument, code falls and initialize FDC to default
37jnz 014a 
38xor ax, ax
39mov dx, ax
40lds, bx, [0078]
41mov byte ptr [bx], cf
42int 13
43retf
44and dx, 0f ; following initializes the FDC with the user supplied argument
45xor ax, ax
46mov ds, ax
47lds bx, [0078]
48mov dh, [bx]
49mov cl, 04
50shl dl, cl
51and dh, 0f
52or dl, dh
53mov [bx], dl
54int 13
55retf
56
57-n steprate.com
58-r cx
5966
60-w

61-q
62

For the Siemens drive steprate 4 works perfectly (steprate 5 also works). The IBM tandon works slower, and not ideally, but it still works.