2.88MB floppy disk drive

[Multibooter]

I have been toying with the idea of turning an LS-120 drive into an 2.88MB floppy drive, using special re-formatted 120MB LS-120 diskettes.

In order to get some experience with real 2.88MB floppy drives, I have obtained an old internal 2.88MB floppy disk drive, Model MF356F-822MB by Mitsubishi, manufactured May 1997 for IBM Corporation. It also has a sticker on it "IBM FRU 52G3400". The floppy drive has a regular 34-pin floppy drive connector, a power connector with 4 pins and a big blue protruding eject button. I had connected the 2.88MB floppy drive via the regular floppy cable, instead of a 1.44MB floppy drive.

Unfortunately I cannot get the 2.88MB Mitsubishi/IBM floppy drive to work. I had set the BIOS of the Asus P5PE-VM motherboard from 1.44MB to 2.88MB. I had tried the 2.88MB floppy drive on 2 different computers with this dual-core motherboard. The 2.88MB floppy drive powers up ok, is displayed in My Computer under Win98 and WinXP as A:, but it cannot read/write/format 720KB, 1.44MB or 2.88MB diskettes. The old DOS program VGA-Copy v5.30, in a command window under WinXP, could actually go thru the steps of writing/formating a 2.88MB ED diskette, with the drive activity lights correctly flashing, but VGA-Copy only displayed yellow bars for each track, i.e. error encountered.

When I tried to boot into DOS from a 2.88MB ED boot floppy, the 2.88MB floppy drive was accessed and its green activity light lit up, but after about 10 seconds the System Commander operating systems selection menu came up instead of DOS, i.e. something went wrong while trying to boot from the 2.88MB boot floppy..

What am I doing wrong? Does this 2.88MB drive require some special tricks for connecting? Or is this 2.88MB drive just bad? Is there a chance to fix this drive? Or is the BIOS of the Asus P5PE-VM not compatible with 2.88MB floppy drives, even if it has a 2.88MB selection? I don't have another 2.88MB floppy drive to compare.

Edited September 15, 2011 by Multibooter

[Multibooter]

It looks like the 2.88MB floppy drive has finally died. When I powered up the computer with the power supply cable connected to the 2.88MB floppy drive, the computer immediately shut down; when I powered on without the 2.88MB floppy drive connected, the computer started up fine.

The 2.88MB floppy drive apparently requires a custom-made floppy cable to work with my motherboard, different from the standard floppy cable. The pin-out of 2 different IBM 2.88MB drives (PC 700 Series and PS/2) is at http://web.archive.org/web/20000612082850/http://servicepac.mainz.ibm.com/eprmhtml/eprm/h266.htm

The drive which just died has a separate 4-pin power connector and was perhaps from the IBM PC 700 Series (no +5V, +12V pins in pinout) . The drive for the PS/2 apparently has no separate power connector, since the pinout indicates that it gets its power from pins 3 and 6 of the floppy cable.

Does anyone in this forum have a 2.88MB floppy disk drive under Win98 or WinXP?

Edited September 15, 2011 by Multibooter

[jds]

Well, you're probably right in suggesting that, because your drive has a 4 pin power connector, it probably follows the second (PC 700) pinout from the link you've given, rather than the third one (Aptiva 2165).

However, I do recall a distant memory. Back in the earliest days of the PS/2 series, when 3.5" diskettes were a bizarre aberration, I do recall discovering that the PS/2 used an inverted Density Select signal on pin 2 of the floppy interface, as compared to the industry (Shugart, I believe) standard. These machines & drives supported 720K and 1.44M diskettes.

So it is possible that this signal inversion persisted for future PS/2 models, and hence any floppy disk drives built to suit.

That said, I have no specific knowledge of 2.88M drives. I see however from those pinouts, that pin 33 has joined pin 2 in selecting the drive density, which makes sense, as there are now at least three densities to support. It is possible that this additional density select line is similarly inverted (in the PS/2 range, as compared to the industry standard).

So that might explain why you were unsuccessful. Alternatively, the drive may have been faulty.

BTW, you must use the correct diskettes for the density you're using. This works (or doesn't) both ways. For example, don't think you can format 1.44M diskettes to 720M and get them to work reliably, if at all.

Joe.

[jaclaz]

BTW, you must use the correct diskettes for the density you're using. This works (or doesn't) both ways. For example, don't think you can format 1.44M diskettes to 720M and get them to work reliably, if at all.

Sure you can, with no actual perceivable difference in reliability.

This is what all the users of midi keyboards (with 720 Kb floppy disk drives) have been doing in the last, say, 10 years:

1. format the 1440 disk as 720 Kb in a PC floppy disk drive (covering second hole with tape)
   
2. usethem on the keyboard
   

There is no actual difference in the internals of a floppy disk (at least between the 720 and the 1440) and I know for sure that for at least a short period around 1994/1995 ONLY 1440 floppies were produced and they were "packed" into either a "single hole" plastic case (720 Kb) or in a "two hole" one (1440 Kb).

jaclaz

[dencorso]

I agree with jaclaz. I've been formatting 3.5" diakettes to all possible formats, from 160 kiB to 1743 kiB for the last 5 years, if not more, without any unusual reliability issues. Of course, floppies can develop bad clusters, with passing time, but they do that at the same rate, regardless of being formatted to 1440 kiB or not. Formats using more than 80 tracks are slightly more prone to developing bad clusters, but then one is using an area of the floppy that the manufaturer didn't intend to be used. Sony and Maxell floppies are very reliable. However, nowadays, it's getting difficult to find new, unopened, boxes of Maxells. But Sonys and Imations are still plentiful.

[jds]

jaclaz, dencorso,

Sorry, if you think the difference between the 720K and 1440K media is just that extra hole in the shell, you are mistaken. The compatibility problem is due to the higher coercivity of the 1440K media compared to the 720K media.

Those "bad clusters" you see after a time are due to this. When using 720K density, the magnetic field strength used by the drive is less (possibly by a factor of two, I'm not sure if it's a linear relationship) than when using 1440K density. So when you use 720K density on 1440K media, this doesn't get fully magnetised.

Why is the coercivity different? Because you need to squeeze the magnetic domains closer together on 1440K media, without them affecting each other too much (also, without the drive head corrupting the adjacent domains), which means you need higher coercivity materials.

Joe.

Edited September 25, 2011 by jds

[dencorso]

With all due respect, Joe, I disagree.

I'm fully aware that:

DD: 2 µm magnetic iron oxide (Coercivity approx. 300 OE)

HD: 1.2 µm cobalt doped iron oxide (Coercivity approx. 600 OE)

ED: 3 µm Barium ferrite (Coercivity approx. 750 OE)

However, the coercivity and the strenght of the magnetic field used need not necessarily to be matched. On the contrary, using HD media to record low density formats works OK, although it's an overkill. Media with higher coercivity will hold weaker magnetizations just as well as it does higher magnetizations, but the other way round is not true. So trying to format DD media to 1440 kiB results in bad clusters galore appearing fast, in a matter of a few days at most, in the very rare cases they don't appear right away, during the format operation. But formatting HD media to 720 kiB is just as safe as formatting DD media, although perfectionists would first apply a bulk demagnetiser to preformatted floppies. Experience (not just my own) confirms this is the case, and as I said above, the appearance of bad clusters in HD media occurs at about the same rate (which is slow) for any format from 160 kiB to 1680 kiB.

I imagine ED floppies behave in the same way, accepting both HD an DD formats just as well as 2880 kiB. However, since I've no experience with ED floppies, I leave this last comment as just a speculation, that needs experimentation to confirm.

[jaclaz]

I will repeat myself:

There is no actual difference in the internals of a floppy disk (at least between the 720 and the 1440) and I know for sure that for at least a short period around 1994/1995 ONLY 1440 floppies were produced and they were "packed" into either a "single hole" plastic case (720 Kb) or in a "two hole" one (1440 Kb).

I have physically seen myself a line of production for floppies that worked as above stated.

I don't think that the manager of the factory rearranged the facility expecially the day I visited it in order to hide some manufacture secret.

So, to be more exact, I have reasons (direct experience) to affirm that at least one factory in the world in a period approximately between 1994 and 1995 used the SAME magnetic media on both the 720 and the 1440 diskettes.

Apart from the above, that may have been a black swan of some kind , and due to the fact that the period in which I have visited the factory was towards the end of the actual common use of the 720 floppies, and thus some "shortcuts" may have been used by the manufacturer, please read this:

http://www.retrotechnology.com/herbs_stuff/guzis.html

As you can see, the results in real life are controversial, and we have both a statement like:

http://mdfs.net/Docs/Comp/Disk/Densities

What about storing DD data on a HD disk? The grains are small

enough for the data, in fact smaller than they need to be. However, DD

recording strength is twice HD recording strength. Recording DD data on a

HD disk will force the data into the grains, like using a pencil with too

much pressure. It will be difficult for the recording head to erase old

data to write new data over it. The data on the disk will deteriorate

every time something is written to it.

And one like:

http://www.retrotechnology.com/herbs_stuff/guzis.html

Actual experience

I've gotten plenty of conversion work where the 3.5" floppies were a mixture of HD and 2D media written in 2D drives ("720K" drives are blind to the media sense hole in a 3.5" jacket). Usually, they're just fine, with the error rate approximately the same, whether or not 2D or HD media was used.

My personal experience is actually similar with the last quoted statement but I guess that there is also a factor connected to the actual hardware (floppy disk drive maker/model).

jaclaz

[jds]

the coercivity and the strenght of the magnetic field used need not necessarily to be matched. On the contrary, using HD media to record low density formats works OK, although it's an overkill. Media with higher coercivity will hold weaker magnetizations just as well as it does higher magnetizations, but the other way round is not true. So trying to format DD media to 1440 kiB results in bad clusters galore appearing fast, in a matter of a few days at most, in the very rare cases they don't appear right away, during the format operation. But formatting HD media to 720 kiB is just as safe as formatting DD media

Sorry, Den.

Over the years, the company I work for has supplied numerous pieces of software for our customers, which at one time, was distributed on floppy media. So we have (had) experience in copying many thousands of diskettes, sourced from a variety of vendors. Because it could be read by both types of drive, our practice was to standardize on the 720K format, even after 1440K drives became the norm. However, as 720K media became less common, we started to use 1440K media, unaware this would be a problem. So we learnt "the hard way" that this isn't reliable. Depending on the particular drive, and especially the make of diskette, we had the situation where HD diskettes would even fail the verification stage immediately after being copied (in 720K format). We might have assumed that this was a quality problem with the diskettes, except that these were actually from a very reputable manufacturer. So I raised this matter with the manufacturer, which is how I learnt about this coercivity issue. And just to confirm what manufacturer had said, we retested the "bad" diskettes in 1440K format, and they worked perfectly.

I will repeat myself:

There is no actual difference in the internals of a floppy disk (at least between the 720 and the 1440) and I know for sure that for at least a short period around 1994/1995 ONLY 1440 floppies were produced and they were "packed" into either a "single hole" plastic case (720 Kb) or in a "two hole" one (1440 Kb).

I have physically seen myself a line of production for floppies that worked as above stated.

I don't think that the manager of the factory rearranged the facility expecially the day I visited it in order to hide some manufacture secret.

So, to be more exact, I have reasons (direct experience) to affirm that at least one factory in the world in a period approximately between 1994 and 1995 used the SAME magnetic media on both the 720 and the 1440 diskettes.

Apart from the above, that may have been a black swan of some kind , and due to the fact that the period in which I have visited the factory was towards the end of the actual common use of the 720 floppies, and thus some "shortcuts" may have been used by the manufacturer, please read this:

http://www.retrotechnology.com/herbs_stuff/guzis.html

As you can see, the results in real life are controversial, and we have both a statement like:

http://mdfs.net/Docs/Comp/Disk/Densities

What about storing DD data on a HD disk? The grains are small

enough for the data, in fact smaller than they need to be. However, DD

recording strength is twice HD recording strength. Recording DD data on a

HD disk will force the data into the grains, like using a pencil with too

much pressure. It will be difficult for the recording head to erase old

data to write new data over it. The data on the disk will deteriorate

every time something is written to it.

And one like:

http://www.retrotechnology.com/herbs_stuff/guzis.html

Actual experience

I've gotten plenty of conversion work where the 3.5" floppies were a mixture of HD and 2D media written in 2D drives ("720K" drives are blind to the media sense hole in a 3.5" jacket). Usually, they're just fine, with the error rate approximately the same, whether or not 2D or HD media was used.

My personal experience is actually similar with the last quoted statement but I guess that there is also a factor connected to the actual hardware (floppy disk drive maker/model).

jaclaz

Hi jaclaz,

Well, back in the day when we had both DD and HD diskettes from the same manufacturer, we could distinctly see a difference in the coating of the disks. So your manufacturer either didn't care (or was careless), or else they developed a compromise formulation that was somewhere between DD and HD and used this for both.

Did you ever stop to think, why did the manufacturers introduce that extra hole in the shell to distinguish DD from HD? Do you think it's because they could charge a higher price for HD? No, it's because of similar issues encountered, and lessons learnt, from the 5.25" era.

I can agree with your last sentence. So what's your definition of "reliable"?

Joe.

Edited September 25, 2011 by jds

[jaclaz]

So your manufacturer either didn't care (or was careless), or else they developed a compromise formulation that was somewhere between DD and HD and used this for both.

Yep, this is a logical explanation.

I can agree with your last sentence. So what's your definition of "reliable"?

You write something, it stays written, it can be read a decent number of times.

"decent" being defined as "a few" or "several" or "many" times.

I guess the key for our different experience lies in your sentence:

Depending on the particular drive, and especially the make of diskette, we had the situation where .....

and in these from the given link:

http://www.retrotechnology.com/herbs_stuff/guzis.html

The thinner coating and the smaller particle size [of 3.5" DSHD diskettes] make for better-defined domain "edges" and generally better [resolved] higher-density recording. Unfortunately, [these factors] also make for a lower recovered data [read] signal, and that, coupled with the lower write current [when writing at] 2D, is what makes for Trouble, particularly on older drives.

On a personal note, I think that the overall quality of DSHD 3.5" media isn't what it used to be, so that might contribute to the general impression that 3.5" HD diskettes used as 2D aren't reliable. I have problems enough finding reliable 3.5" DSHD floppies used as such. But then, we're not paying $40 a box for them either.

As another anecdotal experience I can tell you that when making "extra" formats like the "real" 1722 disks:

http://www.serverelements.com/forums/viewtopic.php?t=64

I found that "no name" "BULK" floppies behaved "better" than "good brand" pre-formatted ones.

And also that "good brand" "oldish" floppy disk drives (the actual hardware) behaved "better" than newish "el-cheapo" ones, and I even had one of the latter "going beserk" after having written tracks 81 and 82 .

jaclaz

[allen2]

And also that "good brand" "oldish" floppy disk drives (the actual hardware) behaved "better" than newish "el-cheapo" ones, and I even had one of the latter "going beserk" after having written tracks 81 and 82 .

jaclaz

What do you mean going berserk ? Did it ate the floppy or did it sliced it in small pieces ?

[jaclaz]

What do you mean going berserk ? Did it ate the floppy or did it sliced it in small pieces ?

Floppy disk drive was not operating properly AFTER.

Floppy media (AFAICR) was not damaged.

Compare with:

http://thestarman.narod.ru/Linux/tomsrtbt.htm

It is well possible that it was a coincidence (another black swan ) but most probably attempting to read beyond track 80 somehow "clogged" the head arm in a stuck position (due to the value of the thingy, after attempting some percussive maintenace with no result, I simply threw the floppy drive in the waste bin).

jaclaz

Edited September 25, 2011 by jaclaz

[Joseph_sw]

ah, extra tracks, commonly used for some-kind original floppy authentification technique. (anti-copying stuff)