Floppy Drive
Floppy Drive
Floppy Drive
June 1994 Method to Allow the Sharing of I/O Port Addresses between a Floppy Disk Controller and
an IDE Controller within a PS/2 Micro Channel System
April 1994 Conversion Connector for the IBM Personal Computer
Replacing Dried Out Capacitors (to the 8580 Common Devices page)
Comments on 8580 Floppies (to the 8580 Common Devices page)
Installing a Floppy in a 95
Formatting 720K Disk on 1.44MB Floppy
Formatting 360K 3.5" Floppies
System to Floppy Drive List
Floppy Drive to Manufacturer List
95A (82077SL) Floppy Controller
Media formats and transfer rates
Floppy Planar and Drive pinout
2.88 Floppy to Clone Hack Progress
FIFO Mode
2.88MB Floppy
Perpendicular Mode
2.88MB Floppy Disk Construction
Gap2 Information
2.88MB Floppy Source (1,000s of them!!!)
* Marked 2.88MB Floppy Drives
Function of Third Foppy Connector
Error 165- Is The Floppy Working?
8580 Floppy Drives in the 95?
Disable Floppy Under Setup
OS/2 MCA Foppy ADDs
OS/2 ver 3 and DMF Workaround
Planar Floppy Pinout
34 pin Floppy Header Pinouts
Removable Media Security
Secure Media Mode
2.88MB Electronic Eject Floppy
EE Floppy Security Circuitry
Registers
EE Commands (Registers- Lock, Unlock, Eject)
Issue an Enhanced Command
The floppy controller and interface connector reside on the system board.
Snippets In particular, the 82077SL internally samples the IDENT and MFM pin level which is used to
1 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
configure the operating mode (PC-AT, Model 30, PS/2) on the falling edge of h/w reset.
82077AA Removal of DMA Request(DRQ) During an Under/Overrun Condition
82077 SL Power-on Reset Problem
82077SL: t23a TIMING CLARIFICATION
Interface Between 82077AA/SL and the Floppy Drive
5. Sony MP-F40W - 14/15 There are dash 14 and 15 are two new drives from Sony that handle 4 MB
requirements. The MP-F40W-14 has the DENSITY SELECT 1, DENSITY SELECT 0 on pins 2 and 33
respectively, whereas the MP-F40W-15 has the DENSITY SELECT 1, DENSITY SELECT 0 on pins 2
and 6 respectively. As it is obvious from the table below, daisy chaining is easily done if the
82077AA/SL is connected in the PS/2 mode (by typing IDENT high) with either type of drive, the only
difference being the location of DENSITY SELECT 0.
Replacing The 82077SL With The 82078 (44PIN)
Replacing The 82077SL WITH 82078 (64PIN)
1 Ed. All models support the 3.5" Electronic Eject floppy drive. (and 2.88!!)
2 Ed. Known to support 2.88 Note: early 8595s do not support the 2.88MB floppy
3 Ed. All N-Q support the EE floppy (and 2.88!!)
One pattern seems to emerge from the PS/2 planars: the "souped-up" or second-gen planars have the
82077SL FDC chips. A couple of surprises though. I did find two other FDC chips on my equipment.
The first on Bermuda planar #1 is a National Semiconductor 8477AV-2 chip *without* the "(C) NEC
1979", but with "(C) NSC 1991". Probably a 82077AA replacement that is reverse enginered enough to
avoid having to use the NEC copyright. That system is unchanged from the way I bought it, with an "*"
2.88 drive. The spare 2.88 I got on eBay I am unsure of the original model is a non-"*" drive.
There is a smaller surface mount Intel 82091AA in my HP NetServer that does bear the "(C) NEC'79"
& also "(C) Intel '86 '93". Just a guess again about being a replacement for the 82077AA with the end of
the part number. All the Intel 82077AA and 82077SL chips I have otherwise have "(C) NEC 1979" &
"(C) Intel '86 '91" of course. Other clone motherboards and adapter cards I have don't look like they have
2 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
a stand-alone FDC chip. Most support the EHD drives in the BIOS, so it has to be a variant of the
82077xx somehow (Even the enhanced NEC 72065B doesn't support 2.88 drives.).
FDC chips are supposed to give which level they are by a "ver" command given to the chip. By my
reference all flavors of the 82077 return the same value. I have tried a routine for the FDC ver command
that so far has *not* worked. The PS/2 35SX and 53 486SLC2 planars both give a return value for a
standard FDC that doesn't support 2.88 drives, then make the computer unable
to read the drive! Here is the (of all things, BASIC) routine anyway & I am going to keep trying to get it
to work.
OLDVAL = INP(&H3F5)
OUT &H3F5, &H10
FDCVER = INP(&H3F5)
FDC = ""
IF FDCVER = &H80 THEN FDC = "NEC 765 / Intel 8272 or compatible FDC, no 2.88 support"
IF FDCVER = &H81 THEN FDC = "Intel 82077xx or compatible FDC, 2.88 support"
IF FDCVER = &H90 THEN FDC = "NEC 72065B or compatible FDC, no 2.88 support"
IF FDC = "" THEN FDC = "Unknown FDC returning value " + HEX$(FDCVER) + "h"
PRINT FDC
OUT &H3F5, OLDVAL
Peter blearily looks up from his bowl of Fruit Loops 'n Beer and sez:
Guess I jump in here and clear some misunderstandments.
1. Older PS/2 are non-media sensing - means: whether the floppy has the right-hand "media type hole"
or not doesn't bother these machines.
2. "Klone Chop-Suey-PCs" use to have FDDs that *do* test for the media type
hole - and consequently refuse to read from a down-formatted 1.44MB floppy. You
*need* to use a piece of transparency tape around the front edge and cover the
hole from the *underside*. This does not have any effect on the older PS/2s as
explained in 1. above. The "generic" FDDs use a set of switches on the right
side to test for a) floppy presence and b) presence of a "High Density" hole.
(2.88MB drives have a third switch that tests for "eXtra Density" hole, which
sits a bit further away from the lower edge of the floppy). Some older PS/2
FDDs have the switches too - but they are used for media presence only - not
for detecting the media type, like e.g. in a Mod. 50/60, 55/65, 70/80 and the
30-286.
3. If you'd closed the media type hole on am actually 1.44MB formatted floppy
and try to format it on a "non PS/2" machine it might complain on a false
format in a first attempt. You better use a PS/2 (see 1. )
On DOS after 3.x you need to use FORMAT A: /U /F:720 to format to 720KB.
On DOS 3.x you need to use FORMAT A: /N:9 /T:80 to force a 720KB format.
The /U parameter in later DOS (and Win95 DOS box) does an "unconditional"
format and ignores all data and formats on the floppy. It does a *physical*
format across all sectors and actually writes the 720K structures at all.
If you would use the /Q parameter the drive would only try to rewrite the first
sectors with the File Allocation Table (FAT) on that floppy and leave the rest
untouched - that will not work and will result in a media error anyway.
The /N:9 parameter on older DOS is the difference between 720K and 1.44MB format. Both use 80
tracks (the /T:80 parameter), but 1.44 uses 18 sectors (would be /N:18), while 720K uses half of them -
3 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
therefore /N:9.
The *track width* is the same on 720 and 1.44 format - because both use 80 tracks and the stepper
motor does the same step-width - and the R/W-head gap does not change during the process ... :-)
Once you'd formatted a 1.44MB floppy to 720KB you might be unable to re-format
the floppy back to 720KB - even if you remove the covering tape from the media
type hole. 1.44MB floppies use a Ferro-Chrome (FeCr) base material, which
"holds" the magnetism a bit stronger than the Ferrite-Oxyd (FeO) material
usually used for 720K floppies. The R/W amplifier on generic FDDs might be
unable to fully erase the 720K format in this case.
4 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
Mounting Hardware
Model 85/90/95 Floppy Drive Slide
64F0156
33F5613
5 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
perpendicular recording mode. This mode utilizes magnetization perpendicular to the recording medium
plane. This is in contrast to the current mode of longitudinal recording which uses the magnetization
parallel to the recording plane. By making the bits stand vertical as opposed to on their side, recording
density is effectively doubled, Figure 1. The new perpendicular mode of recording not only produces
sharp magnetization transitions necessary at higher recording densities, but is also more stable.
Gap2 Differences
The implementation of 4 MB drives requires understanding the Gap2 (see Figures 2a and 2b) and VCO
timing requirements unique to these drives. These new requirements are dictated by the design of the
``combination head'' in these drives. Rewriting of disks in the 4 MB drives requires a pre-erase gap to
erase the magnetic flux on the disk preceding the writing by the read/write gap. The read/write gap in the
4 MB drive does not have sufficient penetration (as shown in Figure 4a) to overwrite the existing data.
6 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
In the conventional drives, the read/write gap had sufficient depth and could effectively overwrite the
older data as depicted in Figure 4b. It must be noted that it is necessary to write the conventional 2 MB
media in the 4 MB drive at 500 Kbps perpendicular mode. This ensures proper erasure of existing data
and reliable write of the new data. The pre-erase gap in the 4 MB floppy drives is activated only during
format and write commands. Both the pre erase gap and read/write gap are activated at the same time.
As shown in Figure 4a, the pre-erase gap precedes the read/write gap by 200mm. This distance
translated to bytes is about 38 bytes at a data rate of 1 Mbps and 19 bytes at 500 Kbps. Whenever the
read/write gap is enabled by the Write Gate signal the pre-erase gap is activated at the same time.
All of diskettes are pre-certified and formatted 100%. They offer a lifetime guaranteee..
Prices:
The ED Diskettes are priced as follows:
To order from single pack to volume quantities, email John Schattin The listed prices DO NOT include
shipping. But how heavy is a box of floppies anyways?
From Peter
The drives with the asterisk are those for 35/40, 56/57 and 76/77 - but *not* for 85/90/95. Should be a
64F4148, while the "others" use a 64F0204. They differ slightly in the pinout and can damage the
planar on earlier Mod. 90 / 95. (Ed. I'm using an asterisk 2.88 on my 9590. Note that this is a later
planar!)
Hi Al !
>9595 floppy is FRU 64F0204. Floppy I have that is mounted to the sled is FRU 64F4148. Can I use this
floppy on my 9595/8595 without fear?
That 64F4148 is the 35/40, 56/57, 76/77 FDD. If you really like your 95 you *do not* try it in there. A
team mate once did it ... and it took us some days to solder in a new FDD-controller ... (main problem
was to find one at first)
I cannot figure out *exactly* what caused the mess, but it has to do with the "security features"
7 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
available on the 95 - and the corresponding pins on the 56 - 77 being not present and set to GND. For the
older 8595 IBM published a warning, that use of the inappropriate FDD could permanently damage the
sysboard.
I whipped out my asterix marked FRU 64F4148. It's a Mitsubishi MF356F-899MF. I just pulled my
stock 2.88 from my Bermuda planar 9577- it is a 64F4148 as well, BUT the Mitsubishi model is
MF356C-799MF. First postulation of the "Law of the Asterix" (you heard it here first, folks!) is that the
MF356F is the model that is incompatible with early 90s/95s.
As noted above, I have used the " * " floppy on a 9590 with no unusual results. I figure that there must
be a more primitive floppy controller used on the 35/40 etc. systems. I do not have one of these to check.
Anyone with the answer, please tell me!
From Peter
The Mod. 56 / 57 / 76 / 77 / 85 / 90 / 95 use a somewhat different FDD-connector on the planar. They
have the Type-3 FDD-interface, which also supports 2.88MB drives ("Media-Sense Drives"). The third
connector is for a very odd ITBU Internal Tape Backup Unit, which was a slightly modified IRWIN
120MB tape. The machine supports only 2 FDDs - as usual.
Ed. Configuring The 82077 For Tape Drive Mode
The FDD-plugs are 34-pins (2 x 17), only the planar connector is a bit strange 44-pins. This type of
interface contains also lines for security control, i.e. in connection with the "Electronic Eject 2.88MB
FDD", which can be locked and password protected.
Tried to find a pinout of the connector but haven't found any at the moment ...
Peter Responds-
Not always. A disfuntional FDD may as well cause a 165. "It is configured - but does not respond". If
the heads stuck, do not pass Track-00 tests or have RDATA stuck high or such you will surely get a 600-
series error. But if the drive has a "DC leak" and simply appears as absent it is judged as "device missing
but still present in the configuration".
8 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
These pin style diskette drives can be sub-divided into two sub-groups. The original model 8580 drives
are identified by the P/N 90X6766. I have seen these drives labeled as manufactured by Mitsubishi, Alps
Electric and YE Data. Later models were produced for the 8595 and they are identified by the P/N
72X6112 or 1619618 and also sometimes accompanied by the FRU # 64F0162 , which is also the FRU #
reported in the HMM (October 1994) for both the 8580 and the 8595. HOWEVER, I have discovered
that although the 8595 drives (FRU #64F0162) will work on the 8580, the 8580 drives (P/N 90X6766)
will NOT work on the 8595!! The drives with FRU # 64F0162 have also been labeled as manufactured
by Mitsubishi, Alps Electric and YE Data.
9 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
10 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
Signal
1 Ground 2 -HD select 1 Ground 2 Data rate select
Drive Type Drive type ID 1/ Drive
3 +5v DC 4 3 +5v DC 4
ID 1 Status1
5 Ground 6 +12v DC 5 Ground 6 +12v DC
7 Ground 8 -Index 7 Ground 8 -Index
9 Ground 10 Reserved 9 Drive type ID 0 10 Reserved
11 Ground 12 -Drive select 11 Ground 12 -Drive select
13 Ground 14 Reserved 13 Ground 14 -Security Command
15 Ground 16 -Motor enable 15 Ground 16 -Motor enable
Media type 1/ Drive
17 Ground 18 -Direction in 17 18 -Directionin
status 3
19 Ground 20 -Step 19 Ground 20 -Step
21 Ground 22 -Write data 21 Ground 22 -Write data
23 Ground 24 Write enable 23 Ground 24 -Write enable
25 Ground 26 -Track 0 25 Ground 26 -Track 0
Media type ID 0/ Drive
27 Ground 28 -Write protect 27 28 -Write protect
status 2
29 Ground 30 -Read data 29 Ground 30 -Read data
31 Ground 32 -Head 1 Select 31 Ground 32 -Head 1 select
-Diskette
33 Ground 34 33 Data rate select 34 -Diskette change
change
So if BIOS support 2.88 it does not mean it will work, the I/O - chip must support it as well.
11 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
Sony Board
12 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
With the 1.44Mb PS/2 floppy drives IBM moved one ground and one unused pin to put the 5 &
12VDC power on the 34-pin connector. By my references they seem to have inverted a half-dozen
control signals too, but left them in the same relative position on the connector. Modifying the Sony
board with 3 circuit trace cuts and 3 jumpers to account for the power connections at least allows the
clone to power up. The Motor Enable signal (one of the ones on the "twist" of the cable) is not inverted,
so it spins the drive up when a read from the floppy is given. The stepper moter doesn't move though
because that is one of the inverted signals.
I am going to see which buffer chip is used to invert the signals between the Adaptec board and the
IBM planars. There is probably a riser with the correct buffer chip(s) I can assemble to do the task with a
little work. At least at the drive end the connector is pretty much the same. What I have seen is that at the
IBM planar end the Model 35, 40, & 53 have a 40-pin connector and the Model 56, 57, 76, & 77 have a
44-pin connector. I will figure out the pinouts for those too.
FIFO Mode
The diskette drive controller uses a FIFO buffer to enhance DMA transfer operations. The FIFO buffer
is used in the data transfer phase only, and its operation is transparent to programs.
FRUs I have seen for the 2.88MB enhanced diskette drive 92F0132, 92F0129, and I saw 82G1888
mentioned as well.
To determine whether the mode and commands are supported for a specific drive:
1. With the enhanced-command bit set to 1, test the state of the drive type (1,0) signals by reading the
Drive Status register.
Note For info on enhanced-command bit, refer to System Control Port C (Hex 007C).
2. With the enhanced-command bit set to 0, retest the state of the signals. If the signals change to a
binary 11, the mode and commands are supported for that drive.
An optional 2.88 MB diskette drive with security features is available on some IBM PC Server systems.
13 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
The diskette drive is a 3.5-inch, one-inch high drive with media sense capability for the standard diskette
capacities of 720 KB, 1.44 MB, and 2.88 MB. It can read and write data up to a formatted capacity of
2.88 MB, while maintaining read and write capability with 720 KB and 1.44 MB diskette drives.
A control signal has been added to the diskette interface that supports LOCK, UNLOCK, and EJECT
commands issued by the operating system. If the privileged-access password is not set, the diskette is
unlocked during POST. If the password is set, the boot process does not unlock the diskette drive unless
it is the designated IPL source. an operating system utility. For SCSI devices, there is a proposed
standard UNLOCK command. In this case, the operating system will control the LOCK command if the
privileged-access password is set. Access to the unlocking function with specific user authorization can
be controlled by secured system software.
In the event of power loss, the system retains its state (secured or unsecured) independent of the state
of the battery. A diskette can be inserted in the drive, but it cannot be removed if the power is off. When
the drive is turned on and locked, the media cannot be inserted or removed.
Tim Clarke
Hi Louis,
I think what the announcement (and Peter et al.) talked about was software to *cause* the floppy to
eject *not* prevent/disable the feature, although I won't swear to this. I would think the physical switch
is your only option for that. I have the 'electronic eject' 5.25" slimline floppy in one of my 95s and none
of the PC-DOS 7 DRVLOCK /on, DRVLOCK /off or EJECT utilities accept it (B:) as a supported drive.
And, having thought about it ***there ain't no signal lines to control this on the floppy interface***. So,
perhaps you should get a microswitch and patch that into the circuit, then mount the microswitch
somewhere devious. Or, even sneakier, figure out if any of the drive select 2/3 or motor enable 2/3
signals are actually passed through the ribbon but not connected on the floppy drive's PCB circuitry and
use one as 'gating control' signal for the eject signal.
>The *electronic* eject drive is like the standard push button drive. SONY made both of them. The EE
drive allows you to lock the floppy disk in the drive by flipping the eject button disable switch on the
side of the drive. Unfortantly the program to software eject the disk via the OS is missing in action ...
The DISKETTE.DGS diagnostics "overlay" tests the "security features" of the electronic diskette
drives when possible. A bit of reverse engineering might reveal what is needed for recreation of these
utilities. However, doesn't PC-DOS v7/2000 have the DRVLOCK and EJECT commands? More fodder
for the dedicated hacker.
Peter
It has - but it fails with the EE-FDD as far as I can tell. At least when using it in a Lacuna. As far as I
14 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
can tell they have been intended for CD-ROMs and MODs. DRVLOCK de-activates the eject button and
EJECT forces a media ject on them. *That* is known to work under PC-DOS 7.0 with a CD-ROM
installed ... I use EJECT recently on my last PC-DOS survivor.
Lock Drive Disables the load-and-eject mechanism. The drive will not eject a loaded diskette, nor
will it load a diskette. (Depending on the characteristics of the drive, it may load the diskette and
immediately try to eject it).
Note Allow 500 mS after an Eject Media command before issuing Lock Drive cmd.
Unlock Drive Enables the load-and-eject mechanism, which allows diskettes to be removed from and
inserted into the selected drive.
Eject Media Same as pressing the eject button on the front of the drive; it causes the drive to eject a
diskette. This command is ignored if the drive is locked.
NoteIf the drive is deselected before the enhanced-command bit is reset to 1, the drive does not perform
the command.
15 of 16 15/09/2023, 09:15
floppy http://www.walshcomptech.com/ohlandl/floppy/floppy.html#Floppy...
16 of 16 15/09/2023, 09:15