How did a PC 5.25 High Density Floppy Differentiate Disks?

[+Larry]

A 5.25" HD floppy drive spins at 360 rpm (vs 300 for DD drive). The data rate is reduced when using a DD disk. In addition, the drive must double-step to create a 40-track format, making it semi-compatible with a regular 40-track disk/drive. But the tracks are still too narrow for full compatibility (although usually it is workable).

 

But I can't find or remember how the PC "knew" what type of disk was being used. I know that to format a DD disk, a special dos command modifier had to be used with the format -- /s or something similar. But if you inserted a pre-written disk, did the PC bios try to read it in both methods to "decide" which kind of disk had been inserted?

 

-Larry

[Rybags]

I can't think of any PC variant that proactively monitored floppy state such as if a new disk was inserted like Amiga and Mac does. In the case of those it's usually an automated process where the info is read and an icon presented to the desktop.

 

On the PC in Dos or Windows, you insert a new disk but nothing happens. The controller does have a register bit that can do media change detect though. There's also status bits to enquire or set which speed mode is in effect.

 

But of course the format needs to be determined to begin with. My guess would be that when an initial command is executed against "new" media, it probably performs a read sector ID command with speed set at both settings. That would allow determining which media type is present in a fairly quick fashion.

A fail on both speeds would likely mean the disk is blank though I would suspect that the Bios or Dos itself would probably seek the head to track 0 and retry there.

[Bryan]

I don't think it knew unless the disk was formatted. It was up to you to format your disks correctly.

[+Larry]

Sure -- the /4 modifier, else it would format the disk in HD (IIRC).

 

They made a step in the right direction with the 3.5" disks, adding the extra "window."

[Bryan]

In my experience, 1.44M floppies were really unreliable. I don't know if it's the media or the drives, but it seemed like a lot of them developed bad sectors and/or refused to format cleanly. I didn't have that problem with 720K disks.

[Rybags]

I think probably both - it was pushing things maybe a bit beyond sensible limit for archiving. I imagine 2.8 Meg was even worse. But those types were a rarity and even the 1.2 Meg drives didn't get much of a foothold since by the late 80s 5.25" disks were obviously headed for the scrapheap.

[MEtalGuy66]

Actually, previous to the mid-late 90s, 1.44meg floppies were ALOT more reliable. Some time around 1996-1998ish, manufacturers started making boith the drives and the floppies ALOT SHITTIER.. I guess the thinking was that a) hardisks had come down on price and become standard on all machines, negating the possibility of anyone using floppies as a PRIMARY means of storage.. b) Optical drives had pretty much taken over as original installation media.. c) There were a wide range of archival/backup media types available that made alot more sense in that role than floppies because the scale (size in bytes) of everything had increased.

You can see what i'm talking about if you take apart a 3.5" 1.44meg floppy from the latge 80s/early 90s and compare it with one from the late 90s - on. The plastic cases are flimsier. The dacron "felt" no longer protects the entire surface of the disk, just a small piece (god knows what the thinking was here, other than utter cheapness).. The drives were reduced to absolute junk, all plastic moving parts, no internal "adjustments"..

 

It is true that the DD stuff tends to be more reliable over time than the HD stuff.. However, of all of my HD disks (1.44meg 3.5" and 1.2meg 5.25") from the 80s and early 90s, I have seen a very small percentage of failures. 1.44meg Floppies that I have bought NEW and tried to use in the last 18 years or so have had a VERY HIGH rate of failure after VARY LOW useage time, relative to the pre-cheapness-era ones..

 

Anywayze.. As usual.. All of this stuff has already been totally beaten into the ground on previous discussion threads on this forum. Do a search and you can read ALOT more specific info regarding what I mentioned above and alot more..

Edited June 8, 2016 by MEtalGuy66

[ijor]

A 5.25" HD floppy drive spins at 360 rpm (vs 300 for DD drive). The data rate is reduced when using a DD disk...

But I can't find or remember how the PC "knew" what type of disk was being used ... did the PC bios try to read it in both methods to "decide" which kind of disk had been inserted?

 

Yes, of course, the Bios (or the OS driver) tries to read in both densities. This not specific to 5.25 HD drives. It is the same for 3.5 HD ones. The extra hole in the latter is used only by the drive mechanism itself. There is no universal standard for communicating the presence of the hole to the floppy controller.

 

This isn't much difference as in our Atari, is it? The 1050 performs about the same procedure for determining the disk density.

[1050]

Actually the BIOS is set by the user, for 5.25 you only have the two choices. 360K or 1.2. If the floppy drive is misidentified in the BIOS, the boot process will halt with 'Floppy Fail 40' error which is exactly how I find my 360K drives. So it's one or the other and never both. To flesh this out further in reference to Larry's query then there actually is a hardware identification process going on but I've not been able to find much info about it. The drive needs no disk in it, this ID is done by BIOS code alone and it never fails to properly tag the drive correctly. It must be a dedicated register response from the drive itself to a special command issued from the PC then and again I have no clue about the exact procedure other than witnessing the results many, many times. I've tried to trick it with with wrong disk in right drive or other way around and you just can't fool it.

 

As to 1.44 reliability, I find the worst offender to be Windows itself -- In my opinion only, they have not done a very good job with Windows drivers for any floppy. Guaranteed to have an error and in short order is my experience with Windows and floppies, not only 1.44 either. Use Win98 Boot to DOS mode and we are rock solid just like an Atari, rock solid as in NEVER a hiccup or error. I find that quite telling about 32/64 bit protected mode floppy drivers myself. If you want to ruin a floppy let Windows have a turn at it is my take away. I cringe every time I need to do that.

 

One glaring exception is Hias' AtariDsk program running in a Windows Dos box which you should know is actually a DOS emulator. Here too we have rock solid Atari like performance all day and night long. But this is using DD 5.25 only and we can expect that there.

 

In the distant past I bought a case of super cheap 1.44 disks so I might be the problem more than the platform, you will find quite a few cheap ones that can't be formatted right out of the box. But once it's been formatted in DOS mode and it goes good, I have no more problems usually especially if I can keep the disk away from Windows. When I run out of the case of cheapos, I've often thought I should do this again but spend a little next time hoping for some quality. I just keep coming across yet another box of cheap ones so far.

[Rybags]

I'm not sure the Bios setting for drive type matters beyond using it for the boot process.

 

I generally found Windows to be OK for floppy operations - you get discrepancies for media depending on OS or command line as far as retry logic, whether the data is cached etc which can make one or the other look better or worse in temp error conditions.

[MEtalGuy66]

In the distant past I bought a case of super cheap 1.44 disks so I might be the problem more than the platform, you will find quite a few cheap ones that can't be formatted right out of the box. But once it's been formatted in DOS mode and it goes good, I have no more problems usually especially if I can keep the disk away from Windows. When I run out of the case of cheapos, I've often thought I should do this again but spend a little next time hoping for some quality. I just keep coming across yet another box of cheap ones so far.

Find some new-old-stock pre-1995 made, name-brand (Eg. Verbatim, Maxell, etc.) 1.44meg HD floppies.. See if you don't notice a huge improvement in quality over the newer cheaper ones...

[+Larry]

Find some new-old-stock pre-1995 made, name-brand (Eg. Verbatim, Maxell, etc.) 1.44meg HD floppies.. See if you don't notice a huge improvement in quality over the newer cheaper ones...

 

That's a good point about the drives and disks. I've still got a couple of Sony-branded drives that give very good success. And of course, I still have quite a bit of original software that was released on 1.44's. Every once in awhile, I load something up and it still works. So with good disks and good drives, I don't think there is anything inherently unreliable with 1.44's. Now, 1.44's used on an Atari with a Floppy Board is different. Not terribly reliable, so I gave that up fast. Wonder how reliable the HDI drive disks were? The FB used a WDC 1782-02-02 fdc, and it was "on the edge" with 36 sectors per track. HDI (IIRC the name correctly) used a different chip set.

 

-Larry

[ijor]

Actually the BIOS is set by the user, for 5.25 you only have the two choices. 360K or 1.2. If the floppy drive is misidentified in the BIOS, the boot process will halt with 'Floppy Fail 40' error which is exactly how I find my 360K drives. So it's one or the other and never both. To flesh this out further in reference to Larry's query then there actually is a hardware identification process going on but I've not been able to find much info about it.

 

We are not talking about how to identify the drive type, but about how to identify the media (the disk inserted in the drive). Larry asked how the PC knows if the disk inserted in a 5.25 HD drive is HD or DD.

 

[atarifanboi]

If I recall correctly (And I may not, my brain is a bit worn out ), the 5 1/4 disks did require you to choose when you formatted the disk, there was no media sense, only a small hole to determine actual spin speed, and a write protect notch, which you enabled/disable with a small piece of tape or a sticker.

 

However, I think it was also the case that in order for you to be able to format to the 1.2Mb capacity, the coating of the disk was different, hence why if you formatted to 1.2Mb on disks not rated for it, you ran the risk of corruption.

 

I believe the same was the case for 1.44Mb disks in terms of the coating being different, but the media sense made our lives easier.

 

I believe the reason for the different coatings was down to magnetic coercivity and the need to have tighter, cleaner tracks on the higher capacity disks.

Edited June 9, 2016 by atarifanboi

[nanochess]

I wrote a disk formatting utility many years ago. The format program tries 1.2 mb first, formatting then verifying and then tries 360k if not successful.

 

Same when reading the disk, first setting disk density for 1.2mb then for 360k (this accounts for delayed startup in post-1992 systems).

 

For some crazy reason 1.2mb floppies cannot be formatted as 360k and when sometime you force it (for example trying several disks) the resulting disk is unreliable.

[Kyle22]

IIRC, some do it something like this:

 

seek track 0

step out 42 tracks (bump stop if 40 track)

step in 40 tracks

test for track 0 sensor

[ijor]

For some crazy reason 1.2mb floppies cannot be formatted as 360k and when sometime you force it (for example trying several disks) the resulting disk is unreliable.

 

Hd disks need to be written with stronger magnetic strength. HD drives can configure the heads for writing to both types of disks. 3.5 drives, thanks to the density hole, detect automatically the media and will select the correct magnetic strength. But 5.25 HD drives don't know, they must be told by the controller. And if you write at lower data rates, the controller will configure the drive for lower magnetic strength.

 

So if you write at DD, the drive will use the lower magnetic strength at the head. And will produce a weak, unreliable signal.

 

IIRC, some do it something like this:

 

seek track 0

step out 42 tracks (bump stop if 40 track)

step in 40 tracks

test for track 0 sensor

 

This can, with some limitations, detect the type of drive, not which type of disk is inserted. Also note that this method only detects if the drive is 40 or 80 tracks, not exactly the same as detecting the density. An 80 tracks is not necessarily HD. Not very common, but there are (were) 5.25 drives, DD, 80 tracks. They were equivalent to 3.5 DD drives.

[Bryan]

I just had an idea... I wonder how easily the 1050 circuit could be modified to write HD disks reliably. You could have a switch for greater field strength and HD disks are everywhere and cheap (plus, they're not as old!).

[tschak909]

That sounds crazy enough to try.

 

-Thom

[kogden]

I just had an idea... I wonder how easily the 1050 circuit could be modified to write HD disks reliably. You could have a switch for greater field strength and HD disks are everywhere and cheap (plus, they're not as old!).

 

Would the switch be necessary? Is there any downside to using greater field strength to write to DD disks?

[Joey Z]

I just had an idea... I wonder how easily the 1050 circuit could be modified to write HD disks reliably. You could have a switch for greater field strength and HD disks are everywhere and cheap (plus, they're not as old!).

What about the difference in track width between an HD disk and a DD disk? or are you saying write a 40 track HD disk?

[Bryan]

Would the switch be necessary? Is there any downside to using greater field strength to write to DD disks?

My first thought is that it might print-through and corrupt the other side of the disk with standard media.

 

What about the difference in track width between an HD disk and a DD disk? or are you saying write a 40 track HD disk?

Same 40-track configuration with DD width. Just trying to make the disks reliable.

[Joey Z]

 

My first thought is that it might print-through and corrupt the other side of the disk with standard media.

 

Same 40-track configuration with DD width. Just trying to make the disks reliable.

 

well given that DD tracks are double the width of HD tracks (approximately) you'll probably need higher head current even.

[Bryan]

The point is to figure out what changes are necessary to make a lower coercivity HD disk (written on a 1050) read back with the same signal strength as a DD disk.

[ijor]

I just had an idea... I wonder how easily the 1050 circuit could be modified to write HD disks reliably. You could have a switch for greater field strength and HD disks are everywhere and cheap (plus, they're not as old!).

Not that I tried, but I understand it is not only a circuit issue, it requires a different head.

 

Would the switch be necessary? Is there any downside to using greater field strength to write to DD disks?

I don't remember if I ever tested it, but there are a couple of problems that make it unreliable (if it works at all).

 

well given that DD tracks are double the width of HD tracks (approximately) you'll probably need higher head current even.

No. The track pitch is unrelated to the density and the write current. There were 80 tracks DD drives, both as 3.5 (ST, Amiga, PS/2) and 5.25 (not very common). And in theory, you could have HD on 40 tracks, if you wanted to.