Vendest wrote on 2023-10-17, 13:41:

After booting into DOS, it gives an error "General failure reading drive B"

Did the drive spin up and try to read the floppy at all? I'm assuming that with not ready error it happens pretty much right away?

If so that would mean the drive is indeed locked to 360rpm and will work as 1.2M but not as 360k. It will read and write (with the usual caveat of narrower head) 360k floppies set as 1.2M but this is done via a software trick that all PC floppy controllers implement (by using 300kbps signalling rate instead of 250kbps). But that trick will not be used if the drive is set to 360k since it should then natively run at 300rpm and not need it.

It would also mean the RDY is output on pin 34 instead a DC signal and it trips DOS. This can perhaps be tested. You can't read or write an actual 360k floppy but you should be able to format and then read a floppy on that particular drive:
- set the drive to 360k in BIOS, no seek on boot
- use 1.2M HD floppy for this experiment as the drive will be single-stepping and you need media capable of high density 80 tracks
- use format /u command
- if DOS refuses to format the floppy (usually saying track 0 is bad) then erase the floppy first with a permanent magnet, especially the outer tracks
- if there are problems don't give up right away, try a few times, weak magnet or too strong one can leave behind magnetic domains that will not easily overwrite
- ignore bad sectors if there's only a few, this is just for testing

Note such floppy will be ureadable in other drives, you can only test it on this one, and if you want to reformat it back to PC specs you should use the magnet again.

Deunan wrote on 2023-10-17, 16:33:

Did the drive spin up and try to read the floppy at all? I'm assuming that with not ready error it happens pretty much right awa […]
Show full quote

Did the drive spin up and try to read the floppy at all? I'm assuming that with not ready error it happens pretty much right away?

If so that would mean the drive is indeed locked to 360rpm and will work as 1.2M but not as 360k. It will read and write (with the usual caveat of narrower head) 360k floppies set as 1.2M but this is done via a software trick that all PC floppy controllers implement (by using 300kbps signalling rate instead of 250kbps). But that trick will not be used if the drive is set to 360k since it should then natively run at 300rpm and not need it.

It would also mean the RDY is output on pin 34 instead a DC signal and it trips DOS. This can perhaps be tested. You can't read or write an actual 360k floppy but you should be able to format and then read a floppy on that particular drive:
- set the drive to 360k in BIOS, no seek on boot
- use 1.2M HD floppy for this experiment as the drive will be single-stepping and you need media capable of high density 80 tracks
- use format /u command
- if DOS refuses to format the floppy (usually saying track 0 is bad) then erase the floppy first with a permanent magnet, especially the outer tracks
- if there are problems don't give up right away, try a few times, weak magnet or too strong one can leave behind magnetic domains that will not easily overwrite
- ignore bad sectors if there's only a few, this is just for testing

Note such floppy will be ureadable in other drives, you can only test it on this one, and if you want to reformat it back to PC specs you should use the magnet again.

I confirm the drive spins up and tries to read the disk.
Due to some issues with the test motherboard I had to switch the DOS unit for the next test.

I tried the "format B: /U" command several times in 360k mode using a 1.2M but the drive stays "not ready"
Unfortunately I do not have a magnet available yet.

Something interesting using TESTFDC, the format / write / verify test was successful in 500kbps double-density.

Vendest wrote on 2023-10-18, 17:40:

Unfortunately I do not have a magnet available yet.

Perhaps another approach would be easier then - you can try and put some tape on the edge connector to mask pin #34. You want a thin tape to reduce any risk of damage to the contact spring in the plug on flat cable. That being said for short-term testing even standard electrical insulation tape should not be a problem.

Set the drive back to 1.2M. With the pin masked I would expect the not ready error to go away but if you swap floppies the DIR command will still show previous content. Do not use any original floppies with data you care about, this has some potential to trash the FAT structures (which will require a reformat to fix). Use your own floppies and/or write-protect them.

Deunan wrote on 2023-10-19, 08:42:

Perhaps another approach would be easier then - you can try and put some tape on the edge connector to mask pin #34. You want a thin tape to reduce any risk of damage to the contact spring in the plug on flat cable. That being said for short-term testing even standard electrical insulation tape should not be a problem.

Set the drive back to 1.2M. With the pin masked I would expect the not ready error to go away but if you swap floppies the DIR command will still show previous content. Do not use any original floppies with data you care about, this has some potential to trash the FAT structures (which will require a reformat to fix). Use your own floppies and/or write-protect them.

Thanks for your help, it works after masking up pin34 with electrical tape.
The 1.2M disk could be read and formatted in DOS.
While I could cut the trace/jumper R permanently, it seems that there is no DC signal available on the circuit board.

Vendest wrote on 2023-10-20, 05:35:

The 1.2M disk could be read and formatted in DOS.
While I could cut the trace/jumper R permanently, it seems that there is no DC signal available on the circuit board.

If the schematic is correct and this signal indeed conditions the output of INDEX via an AND gate then sadly yes, it cannot be switched to DC. Some ICs offer such selection (between RDY and DC) via jumpers next to the IC, not always directly on the output line itself. But it's not possible in this case.

Well, the good news is you can work around this problem to some extent. In DOS 5+ there is CONFIG.SYS setting called DRIVPARM that can be used to override some floppy drive settings. It can be used to disable DC and force DOS to re-read the floppy directory structure on each access - obviously this will have performance penalty but it works (this is how 360k drives operate, that's why I wanted to test this). It should work even without cutting the signal trace but will not allow you booting from this drive (as the setting is only applied once CONFIG.SYS is loaded and parsed).

DRIVPARM should in theory also work in DOS 3.3 (and perhaps even in DOS 3.0) but I never got it to work in my copy of 3.31. There are some tricks that might be used - apparently this is an internal/hidden DOS command and it might be invoked with clever abuse of special control characters in CONFIG.SYS but that pretty much makes the file uneditable via standard text editors. But on DOS 5+ it can be used like any other internal command.

Deunan wrote on 2023-10-20, 10:42:

DRIVPARM should in theory also work in DOS 3.3 (and perhaps even in DOS 3.0) but I never got it to work in my copy of 3.31. There are some tricks that might be used - apparently this is an internal/hidden DOS command and it might be invoked with clever abuse of special control characters in CONFIG.SYS but that pretty much makes the file uneditable via standard text editors. But on DOS 5+ it can be used like any other internal command.

Any source on that? I clearly remember using drivparm in config.sys without any special characters needed.

Can't remember now, but I've read somewhere that in order for DOS to recognize the command properly it had to be preceded by 4 (if memory serves) backspace characters, not sure if this was DOS bug or just a clever workaround to access this disabled command.
I do remember that I had problems with BIOS and/or certain FDC cards that I wanted to resolve using DRIVPARM on DOS 3.31, and it was not recognized as valid command.

Deunan wrote on 2023-10-20, 10:42:

If the schematic is correct and this signal indeed conditions the output of INDEX via an AND gate then sadly yes, it cannot be switched to DC. Some ICs offer such selection (between RDY and DC) via jumpers next to the IC, not always directly on the output line itself. But it's not possible in this case.

Well, the good news is you can work around this problem to some extent. In DOS 5+ there is CONFIG.SYS setting called DRIVPARM that can be used to override some floppy drive settings. It can be used to disable DC and force DOS to re-read the floppy directory structure on each access - obviously this will have performance penalty but it works (this is how 360k drives operate, that's why I wanted to test this). It should work even without cutting the signal trace but will not allow you booting from this drive (as the setting is only applied once CONFIG.SYS is loaded and parsed).

The schematic is for a different model of Y-E DATA drive but shows the same ICs and pinout.

DRIVPARM works great only after cutting the RDY signal located on the permanent jumper. I have added a pin header in case I ever need to use the RDY signal.
Posting back and front pictures of the control board for reference.

Thanks a lot for your help, the drive can now be used in both DOS mode and Win98.

weedeewee wrote on 2023-10-20, 12:16:

Deunan wrote on 2023-10-20, 10:42:

DRIVPARM should in theory also work in DOS 3.3 (and perhaps even in DOS 3.0) but I never got it to work in my copy of 3.31. There are some tricks that might be used - apparently this is an internal/hidden DOS command and it might be invoked with clever abuse of special control characters in CONFIG.SYS but that pretty much makes the file uneditable via standard text editors. But on DOS 5+ it can be used like any other internal command.

Any source on that? I clearly remember using drivparm in config.sys without any special characters needed.

https://forum.vcfed.org/index.php?threads/dri … -pcdos-7.41573/

From my own personal testing with PC-DOS, in 3.2 it works as expected without the control char, but requires it in 3.3

maxtherabbit wrote on 2023-10-22, 16:00:

From my own personal testing with PC-DOS, in 3.2 it works as expected without the control char, but requires it in 3.3

Cheers, I must've had one of the patched oem versions. I recall using the DRIVPARM command without any hack in ms-dos 3.3

Vendest wrote on 2023-10-22, 14:06:

The schematic is for a different model of Y-E DATA drive but shows the same ICs and pinout.

You know, with a drive that old and lacking DC signal it makes me wonder if it's not QD instead of HD. If you ever notice that it seems to have problems with writes (this includes format) and data read back is glitched (CRC error) then I would assume it's QD drive. Which BTW might be worth some money to some ATARI ST or perhaps Amiga crowd. These drives are rare, and the correct media is even more so.

weedeewee wrote on 2023-10-22, 16:39:

maxtherabbit wrote on 2023-10-22, 16:00:

From my own personal testing with PC-DOS, in 3.2 it works as expected without the control char, but requires it in 3.3

Cheers, I must've had one of the patched oem versions. I recall using the DRIVPARM command without any hack in ms-dos 3.3

My only experience is with IBM PC-DOS, the MS branded versions could be different.

Deunan wrote on 2023-10-22, 21:19:

Vendest wrote on 2023-10-22, 14:06:

The schematic is for a different model of Y-E DATA drive but shows the same ICs and pinout.

You know, with a drive that old and lacking DC signal it makes me wonder if it's not QD instead of HD. If you ever notice that it seems to have problems with writes (this includes format) and data read back is glitched (CRC error) then I would assume it's QD drive. Which BTW might be worth some money to some ATARI ST or perhaps Amiga crowd. These drives are rare, and the correct media is even more so.

Can't you generally just use any old DD disk as a QD one?

maxtherabbit wrote on 2023-10-22, 23:07:

Can't you generally just use any old DD disk as a QD one?

Not really. It would accept writes better but at the same time the actual magnetic domains would became "smeared" and in general considerably weaker once the erase head creates gaps. Such floppies would degrade way faster with time.
You might get better/acceptable results with a particular brand(s) of floppies, but I would still prefer to find a HD floppy that works rather than DD.

On 3.5" drives there isn't that much difference between DD and HD media but even there a DD drive would have some issues with HD media. More modern HD 3.5" drives could in theory be smart enough (thanks to being all-LSI ICs) to detect the media type and reduce write currents to improve signal integrity on DD flopies. But 5.25" drives are pretty old and "stupid", there's no way to tell what media is loaded and most don't even have media presence sensor at all.

Deunan wrote on 2023-10-20, 10:42:

Well, the good news is you can work around this problem to some extent. In DOS 5+ there is CONFIG.SYS setting called DRIVPARM that can be used to override some floppy drive settings. It can be used to disable DC and force DOS to re-read the floppy directory structure on each access - obviously this will have performance penalty but it works (this is how 360k drives operate, that's why I wanted to test this). It should work even without cutting the signal trace but will not allow you booting from this drive (as the setting is only applied once CONFIG.SYS is loaded and parsed).

Seems that I need to correct my statement after testing the drive again.
The drive works in DOS mode only with pin 34 not connected to RDY signal.
The drive works in Win98 only with pin 34 connected to RDY signal.
DRIVPARM does not seem to help if pin 34 connects to RDY signal .
Setting in Config.sys DRIVPARM=/D:1 /F:1

So I decided to leave Pin 34 connected since its primary use is with Win98.

Deunan wrote on 2023-10-23, 09:08:

Not really. It would accept writes better but at the same time the actual magnetic domains would became "smeared" and in general […]
Show full quote

maxtherabbit wrote on 2023-10-22, 23:07:

Can't you generally just use any old DD disk as a QD one?

Not really. It would accept writes better but at the same time the actual magnetic domains would became "smeared" and in general considerably weaker once the erase head creates gaps. Such floppies would degrade way faster with time.
You might get better/acceptable results with a particular brand(s) of floppies, but I would still prefer to find a HD floppy that works rather than DD.

On 3.5" drives there isn't that much difference between DD and HD media but even there a DD drive would have some issues with HD media. More modern HD 3.5" drives could in theory be smart enough (thanks to being all-LSI ICs) to detect the media type and reduce write currents to improve signal integrity on DD flopies. But 5.25" drives are pretty old and "stupid", there's no way to tell what media is loaded and most don't even have media presence sensor at all.

I said QD, not HD. To the best of my knowledge the magnetic coercivity of QD and DD media is the exact same

Vendest wrote on 2023-11-01, 22:25:

Setting in Config.sys DRIVPARM=/D:1 /F:1

with those parameters, /D:1 /F:1, on DRIVPARM, you only select drive 1 to be type 1.
You might want to try adding /C

weedeewee wrote on 2023-11-02, 07:04:

with those parameters, /D:1 /F:1, on DRIVPARM, you only select drive 1 to be type 1.
You might want to try adding /C

Thanks but adding /C did not work.

maxtherabbit wrote on 2023-11-02, 01:26:

I said QD, not HD. To the best of my knowledge the magnetic coercivity of QD and DD media is the exact same

I read you right but for some reason I misremembered coercivity for QD media. I can't find the source now (I don't trust Wiki) and my claim that HD media would be a better fit was also based on some experiments that I've run. However I don't actually have a QD drive so I used a HD one, plus not so great quality media (Verbatim Verex DD) and that might have produced worse results.

So I've investigated this again. Service manual for my SA455 drive also includes SA465 which is QD drive. Schematic and parts list differ mostly in the read path, the write chip can be the same in both models. Heads will be different so even the same write current can give different results though, and the manual calls for different media for both drives: SA154 for DD and SA164 for the QD.

But today I found this Maxell document: https://archive.org/details/1983-maxell-catal … age/n3/mode/2up
It offers MD2-D as a replacement for SA154 and MD2-DD for SA164. And both are DD class media, so I need to correct my earlier suggestion - HD media will most likely not work properly after all. That being said the DD media must be high quality, and possibly (even though coercivity might be the same) the coating thickness is different - lower for QD media.

Sorry for the derail but perhaps this will be useful to somebody.