Modern floppy disk controller chips?
Timothy D. Shoppa x4256
added an IDE drive interface, and that works wonderfully. It was even
very simple - just a couple of chips on a S-100 prototype board!
And updating the BIOS to handle the IDE drive was extremely
straightforward, as well.
I'd love to be able to use some of the modern floppy disk drives
available, like the 3.5" 1.44 MByte or 2.88 MByte drives. The old
Shugart SA-800 8" drives currenty attached to my Tarbell controller
still work fine, but I'd like to have the ability to read/write some other
formats, too. My experience with floppy disk controller chips pretty
much stops with the WD1793; I'd like to use something a bit more
modern and readily available, if at all possible. But if someone will
tell me of both a source for the WD1793 and how I'll be able
to use it with a 2.88 Mbyte 3.5" drive, I'd gladly use it.
So what's available today? I've looked at a few of the $12.00
FD/IDE/serial/parallel boards available for PC-clones, but they all
seem to use chips that are made by Korean manufacturers and are
not readily available as individual parts in the U.S. Anybody want to make a
FDC recommendation? Something widely available would be greatly preferred,
as I'd like to know that spare parts will be available for this upgrade
in another 15 or 20 years when I decide to upgrade the IMSAI again.
Something with a on-chip PLL/data separator (or whatever is in vogue these
days for separating MFM data) would be nice as well. (My current Tarbell
controller is about 50% logic, 50% PLL.) The preferred package would
be DIP, but I could live with a more modern package if need be.
Tim. (sho...@altair.krl.caltech.edu)
Kellogg Radiation Lab, Caltech.
Roger Ivie
> I'm updating one of my old IMSAI 8080's at the moment. I've
> added an IDE drive interface, and that works wonderfully. It was even
> very simple - just a couple of chips on a S-100 prototype board!
> So what's available today? I've looked at a few of the $12.00
> FD/IDE/serial/parallel boards available for PC-clones, but they all
> seem to use chips that are made by Korean manufacturers and are
> not readily available as individual parts in the U.S. Anybody want to make a
> FDC recommendation? Something widely available would be greatly preferred,
On the other hand, a translation from S-100 to ISA that's good enough to
run a floppy controller should be a matter of a couple of chips, a DMA
controller, an S-100 prototype board, and some way of mechanically sticking
the card in the system...
The 1793 (or, at least, the 2793 derivative) is still widely available from
supply houses, although you won't find it at your local PC shop. No, it
won't do 2.88 MB floppies; for that you'll need a (gack!) 765 derivative.
I've used the NS8477 and, apart from _really_ hating the 765 software
interface, had success.
--
----------------+------------------------------------------------------
Roger Ivie | Never underestimate the bandwidth of a
iv...@cc.usu.edu | truckload of tapes
Allison J Parent
just started building an IDE interface for my Z80 s100 crate.
You can Email me.
Allison
Alli...@world.std.com
Tony Duell
sho...@almach.caltech.edu "Timothy D. Shoppa x4256" writes:
[...]
> I'd love to be able to use some of the modern floppy disk drives
> available, like the 3.5" 1.44 MByte or 2.88 MByte drives. The old
You do realise, I assume, that the data rate for the 1.2Mbyte 5.25" drive and
the 1.44Mbyte 3.5" drive is the same as the old 8" unit (500kbps in MFM mode).
In fact, the 1.2Mbyte drive is almost the same (logically) as the 8" unit, and
often a simple cable is all that's needed to link one up. If you need more info,
please e-mail me
[...]
> formats, too. My experience with floppy disk controller chips pretty
> much stops with the WD1793; I'd like to use something a bit more
> modern and readily available, if at all possible. But if someone will
> tell me of both a source for the WD1793 and how I'll be able
> to use it with a 2.88 Mbyte 3.5" drive, I'd gladly use it.
Most modern disk controller IC's are designed for use in PC's, and are thus
based on the Intel 8272, or NEC 765 (these 2 devices appear to be equivalents).
There are 2 main problems
1) The software interface is seriously bletcherous!. I've just recovered from
writing an alien disk reading program on a PC, and it wasn't fun (of course, the
PC DMA system didn't exactly help!).
2) The original PC disk controller could not do single density (FM) operation.
Now, there's no problem (that I know of) with using the 8272 in single density
mode, but the original PC data separator could not handle it. In some of the
integrated FDC/PLL chips, the single density mode, although documented in the
data sheet, does not work correctly. This is no problem for the intended market,
of course, but it does cause problems on (e.g.) CP/M machines
>
> So what's available today? I've looked at a few of the $12.00
> FD/IDE/serial/parallel boards available for PC-clones, but they all
> seem to use chips that are made by Korean manufacturers and are
> not readily available as individual parts in the U.S. Anybody want to make a
I am told that National Semiconductor make an integrated FDC/PLL chip (based
on the 8272 architecture). I've not used it, but data should be avaiable
There are also chips from SMC (37C65 I think), and UMC (8398?). I've got data
sheets on both of those, and have managed to buy the bare chips in the UK. I've
not managed to get the UMC chip to work correctly in FM mode, though.
> FDC recommendation? Something widely available would be greatly preferred,
> as I'd like to know that spare parts will be available for this upgrade
> in another 15 or 20 years when I decide to upgrade the IMSAI again.
That's a problem with using chips designed for PC's - they tend to go out
of production very quickly.
> Something with a on-chip PLL/data separator (or whatever is in vogue these
> days for separating MFM data) would be nice as well. (My current Tarbell
> controller is about 50% logic, 50% PLL.) The preferred package would
> be DIP, but I could live with a more modern package if need be.
All the chips I have mentioned are available in (40 or 48 pin) DIL packages.
>
> Tim. (sho...@altair.krl.caltech.edu)
> Kellogg Radiation Lab, Caltech.
>
-tony
Tony Duell
>
> As I stated above, the ability to do single density is highly desired.
> Of course, I could keep the Tarbell controller attached for that case
> if necessary. It'd also be nice to use up to 4 floppy drives on the
> new controller, but I notice that most PC drive controllers don't have
> documented support for anything past the 2nd drive. (Many of the data
> sheets on the chips claim support for 4 drives, nevertheless.) As a
> matter of fact, it appears that the cable signals for the 3rd and 4th
> drives may not exist on the 34-conductor cable.
On the PC, the drive number is specified in _3_ different positions :
1) In the actual command to the 8272, to select the right (internal) track
register, etc.
2) In the low 2 bits of the Control port, to operate the drive selects
(binary code)
3) In the top 4 bits of the same port, to operate the motor control lines
(1 out of 4 code).
Now, on every single-chip controller I have worked with, the 8272 'cell' (these
chips seem to be built from standard cell libraries) still handles 4 drives
correctly, amd generates change-of-rady-state interrupts for drives that
_can't_ be in the system. So, I assume that if you added the extra bits to the
control port (which is, after all, a simple TTL output port), and _possibly_
did soem extra decoding - I don't know if the normal 2 drive select lines are
disabled by bit 1 of the output port.
Of course, you could ignore this port all together (except for the rest and
DMA enable bits), and control th edrive selects as you choose. I guess you could
provide a TG43 signal in a similar way.
I personally would not use the UMC chip in anything bu a PC - it includes an
internal address decoder for the standard PC ports. I think the SMC chip has
separate enable lines for the '8272', control/status ports, and clock rate port
>
> Oh speaking of the bletcherous design of modern PC hardware, are there
> any modern 5.25" or 3.5" drives that have removable terminators? It
I've yet ot find any drive that has the terminators actually _integrated_ into
the custom control chip. They're normally surface-mount resistors, and can
be removed by careful use of a soldering iron.
> appears that most PC's will drive doubly-terminated floppy systems
Most modern drives have 1k, not 150R terminators. This, of course, does not
match the ribbon cable at all. I guess the signals look pretty awful.
> quite nicely, but this may be because the cable run is so short. I'll
> probably have at least 5-6 feet of ribbon cable in my final setup, and
> I'm a bit worried about termination problems. For that matter, are
It's always a good idea to termiate the cable correctly, even for short runs. I
(once) removed all the terminating resistors from a pair of 3.5" drives, and
made a little terminator module that I connected to the end of the ribbon
cable. I could have soldered 150R resistors to the last drive, but I couldn't
find SMT stuff locally.
> there any modern 5.25" or 3.5" drives that can have their drive
> number changed by jumpers? All the PC-type ones I've seen lately use
> the flipped cable trick and have no apparent jumpers.
The 'generic' 1.2Mbyte drive I used for the PERQ experiments had internal
jumpers for drive select (1-4), motor speed (300, 360, switchable RPM),
pin 34 = ready | change, etc. The agnets for the manufacturer (I forget who)
refused to tell me what they were for, though.
>
> Tim. (sho...@altair.krl.caltech.edu)
-tony
Richard Steven Walz
>Timothy D. Shoppa x4256 (sho...@almach.caltech.edu) wrote:
>: I'm updating one of my old IMSAI 8080's at the moment. I've
>: added an IDE drive interface, and that works wonderfully. It was even
>: very simple - just a couple of chips on a S-100 prototype board!
>: And updating the BIOS to handle the IDE drive was extremely
>: straightforward, as well.
>
>: I'd love to be able to use some of the modern floppy disk drives
>: available, like the 3.5" 1.44 MByte or 2.88 MByte drives. The old
>: Shugart SA-800 8" drives currenty attached to my Tarbell controller
>: still work fine, but I'd like to have the ability to read/write some other
>: formats, too. My experience with floppy disk controller chips pretty
>: much stops with the WD1793; I'd like to use something a bit more
>: modern and readily available, if at all possible. But if someone will
>: tell me of both a source for the WD1793 and how I'll be able
>: to use it with a 2.88 Mbyte 3.5" drive, I'd gladly use it.
>
>years now, waiting for the SBC to be built to fit it in. Very nice
>chip, very complete and flexible.
>
>Will
>c...@crash.cts.com
-------------------------
I believe that one is the generic '765 that the PC used and uses, and that
many makers supply. Intel's version was the 8275, I think, and the rest
have something-765 on them. The Morrow CP/M computer used one as well, so
you might wish the 8080A code for their CP/M to pick at. That way you could
figure out how to use either a standard CP/M format, if you're staying with
S-100 CP/M or equivalent. I have that stuff for Morrow and also
schematics.
-Steve Walz rst...@armory.com
roger hanscom
(Timothy D. Shoppa x4256) writes:
>Yes, I do realize that the MFM rate and format is very similar for 3.5" 1.44
>controllers only support single density (250 kbps FM). I suppose I could go
>in, change all the capacitors in all the one-shots (there are literally
>dozens!) on the board and add a MFM data separator, but then it wouldn't
>be able to read the more than 2000 8" CP/M single density floppies I own.
No need for that. It is possible to run two controllers in the same box.
Use a more modern controller for high-capacity drives, and write some
code to retrieve files from your 8" drives. Hmmmm...., 2000 8" SD floppies
would fit on roughly 300 1.44 diskettes (@ .40 each it would only cost
you $120. or so --- did I do the math right??) 1/2 @:^)
> ......... It'd also be nice to use up to 4 floppy drives on the
>new controller, but I notice that most PC drive controllers don't have
>documented support for anything past the 2nd drive. (Many of the data
>sheets on the chips claim support for 4 drives, nevertheless.) As a
>matter of fact, it appears that the cable signals for the 3rd and 4th
>drives may not exist on the 34-conductor cable.
Not so, there are select lines for 4 drives on the 34-conductor cable.
Some PCs used DS3 and DS4 for "motor on" for drives 1 and 2, but these
days there is usually no separate motor on (selecting the drive turns
the motor on).
>Oh speaking of the bletcherous design of modern PC hardware, are there
>any modern 5.25" or 3.5" drives that have removable terminators? It
>appears that most PC's will drive doubly-terminated floppy systems
>quite nicely, but this may be because the cable run is so short. I'll
>probably have at least 5-6 feet of ribbon cable in my final setup, and
>I'm a bit worried about termination problems. For that matter, are
>there any modern 5.25" or 3.5" drives that can have their drive
>number changed by jumpers? All the PC-type ones I've seen lately use
>the flipped cable trick and have no apparent jumpers.
I'm not up on the termination situation, I did run into trouble trying
to run a 1.44 drive from a modified cable. It appeared to me that
there may have been termination problems doing that.
The Mitsumi 1.2 Mb 5.25" floppy drive has a very nice selection of
"jumperable" options that includes 4 drive selects, motor speed,
etc. I have submitted an article to The Computer Journal about my
adventures replacing 8" drives on Cromemco systems with these
Mitsumi drives. They work quite well, and I can't begin to tell
you what a relief it is to begin to think about getting rid of
those noisy, power-hungry 8" drives. In this area, Mitsumi drives
are around $40 at computer shows. I saw an ad in "Nuts and Volts"
for the previous version of this Mitsumi drive, with white faceplate,
for about $33. I don't know if the advertised drive has the same
nice jumper configuration. I can get more information if anybody
is interested.
Allison J Parent
basic controller chips (or cells) in the world. The 177x single density
chips which were amoung the first to be widely used. The 179x chips which
added double density mode to the same archecture. The remaining one is the
D765 (or intel 8272) of NEC design that is double and single density and
handles all drive data rates. All of the later chips generally integrate
such niceties as data seperators and DMA channels for reduced chip counts.
My knowledge of the D765 is current but is also based on my time as a
product specialist for NEC Microcomputers back in the early '80s when I
supported said chip. FYI: the early 8272s were actually NEC die with
intel marking on the package. It was part of a cross license back then.
The tarbell controllers and there were two were either single or dual density
using the 177x or 179x chips respectively. It is very difficult to make the
single density board do do double density. The dual density card however
does not suffer the limitation.
The primary difference between 765 and 179x cotrollers is that the 765
only does standard formats (preamble, marks and data fields) where the
179x will write anything you tell it in the format mode. This is only of
significance for older systems and media that may have written odd formats
to gain storage capacity or copy security.
I generally have interfaced disks to controllers with little regard for what
they were "intended" for. My kaypro has 2 3.5" floppies using the
Advent turborom and they provide 780k of storage. Prior to that I had 96tpi
5" drives doing the same storage. I also have a system that has 8", 5" and
3.5" drives on one 765 based controller. It does most needed modes
density wise. Specifically 8" it will do single and double, same for 5",
however, for 3.5 inch only 720/780k modes as I haven't set up the software
to select the correct clock for 1.44mb. This is a minor limitation in
the CP/m world.
Terminations, it's generally ok to have two in parallel but three or more
would cause grief. Most of the 3.5" drives the terminations can be disabled
but that info documenting how is generally non-existant. The drives I've
used it was a cut to a land on the board needed to achieve it. Also some
don't even appear to have terminations! Standards Bah!
Allison Parent
Will Rose
: [...]
: >
: > As I stated above, the ability to do single density is highly desired.
: > Of course, I could keep the Tarbell controller attached for that case
: > if necessary. It'd also be nice to use up to 4 floppy drives on the
: > new controller, but I notice that most PC drive controllers don't have
: > documented support for anything past the 2nd drive. (Many of the data
: > sheets on the chips claim support for 4 drives, nevertheless.) As a
: > matter of fact, it appears that the cable signals for the 3rd and 4th
: > drives may not exist on the 34-conductor cable.
They exist - that's why you have to flip the cable on a standard PC.
It suppresses floppies C and D. My XT controller (don't know the
chip) supports 4 floppies, and I've seen other controllers that will,
tho' most controllers handle only two.
Greg Bridgewater
> >In article <11FEB199...@almach.caltech.edu>
> > sho...@almach.caltech.edu "Timothy D. Shoppa x4256" writes:
> >
> >
> >[...]
> >
> >> I'd love to be able to use some of the modern floppy disk drives
> >> available, like the 3.5" 1.44 MByte or 2.88 MByte drives. The old
> >
> Tim. (sho...@altair.krl.caltech.edu)
As you've read here by now, there's a lot of similarity in the high
density drives for the PeeCee and the good 'ol 8" drives. You may want to
take the easy way and just find an old Teletek drive controller. It used
the 765, included the source code for software of course, and is adaptable
to just about anything.
Personally, I always thought the 765 was brain dead compared to the much
greater flexibility of the WD 17xx series, but obviously Invest Big Money
(IBM) knew better [mutter, choke and then there were the IBM selected
serial port chips, barf]. Code for the WD series always seemed much
easier to write also. JMNSHO.
--
Greg Bridgewater
voice (412)363-9713
internet: greg...@telerama.lm.com
Richard Steven Walz
No, the twist does nothing of the sort. The controller CARD simply uses
those other two drive selects for other things, namely MOTOR ON, as you can
easily see from the controller to drive cable interface below! As for the
extra external drives C and D, they are simply selected by the card and
substituted in circuitry for the first two A and B which are then not
connected for THAT external interface! Lookee here!:
-Steve Walz rst...@armory.com
Internal 5.25" DSDD 48-96 TPI Diskette Adapter & Diskette Drive Pinout:
Diskette Adapter Diskette Drive
-----------------------, ,-----------------------
unused 2 |----------------| 2 unused
unused 4 |----------------| 4 unused
alignment slot -----| |----- alignment slot
unused 6 |----------------| 6 Drive Select 4
Index 8 |<---------------| 8 Index
Motor Enable A 10 |--------------->| 10 Drive Select 1 *
Drive Select B 12 |--------------->| 12 Drive Select 2 *
Drive Select A 14 |--------------->| 14 Drive Select 3 *
Motor Enable B 16 |--------------->| 16 Motor On *
Direction Select 18 |--------------->| 18 Direction Select
Step 20 |--------------->| 20 Step
Write Data 22 |--------------->| 22 Write Data
Write Gate 24 |--------------->| 24 Write Gate
Track 00 26 |<---------------| 26 Track 00
Write Protect 28 |<---------------| 28 Write Protect
Read Data 30 |<---------------| 30 Read Data
Head Select 32 |--------------->| 32 Head Select
unused 34 |----------------| 34 unused
-----------------------' `-----------------------
Underside of boards have lands 1-33 which are common ground.
Note the lines on the drive marked with an asterisk (*). These lines
and the odd numbered ground lines between them are separated as a whole
and given a half twist at the more distant, or Drive A end of the
ribbon before assembly to the Drive A edge connector. This inverts them
exactly in order for Drive A on the end of the ribbon cable, while
preserving them as above for Drive B. The reason for this odd cable
twist is as follows:
Generally, ALL non-fixed disk drives used in IBM PC/XT/AT are set for
passive drive selection as the second drive by placing a Berg jumper on
the DS2 (in case of options DS1 to DS4) or DS1 (in case of options DS0
to DS3), (merely differing label conventions). The twist in the ribbon
before the most distant drive from the adapter, (Drive A), gives them a
positional identity on the adapter cable by the following method:
Before the Twist
Drive B is selected and motor turned on by Drive Select B [12] and
Motor Enable B [16] which then are active upon Drive B's Drive Select 2
[12] and Motor On [16]. The other drive select numbers are disconnected
if not jumper selected, [10] & [14].
After the Twist
Drive A is selected and motor turned on by Drive Select A [14] and
Motor Enable A [10] which then are active upon Drive A's Drive Select 2
[12] and Motor On [16]. The other drive select numbers are disconnected
if not jumper selected, [10] & [14].
To effect this twist properly, measure out 34 pin ribbon cable to reach
the required distance and cut it off cleanly and straight. Note pin 1
on one end of the 34 pin card edge connector and place the colored edge
of the ribbon cable at the pin 1 end of the connector. Compress this
squarely with a small vise or blocks and a screw clamp. Orient the
direction the ribbon emerges from the connector conveniently to the
geometry of the equipment positions. Since the daggers which stab the
ribbon are staggered in their rows it is not possible to misplace the
pins by the ribbon emergence direction, as long as pin 1 is kept on the
colored edge of the ribbon. Perform a test fit to reach Drive B and
mark the ribbon with a pen, allowing some slack. Again aligning pin 1
with the colored edge, compress another 34 pin card edge connector onto
the ribbon. At the final distant end of the ribbon count from the
colored edge of the ribbon 9 conductors and after them make a small
slice between conductors 9 and 10 with a sharp knife. Peel this slice
back several inches. Make another cut between conductors 16 and 17
separating those next 7 conductors and peel likewise. The half twist
must invert conductors 10 through 16, the four signal lines and the odd
numbered grounds trapped with them in the ribbon. Carefully align these
three ribbon ends with the middle ribbon end half twisted and assemble
within another 34 pin card edge connector. Carefully compress this
connector as squarely as possible. It should resemble below:
----_1____________RED_________________----_1________RED__________---- 1
| | | | | |
| | | | | |
| | | |-----------------------| |
| | | | twist-> \ / | |
| | | | / \ | |
| | | |-----------------------| |
| | | | | |
| | | | | |
| | | | | |
____----------------------------------____-----------------------____
Adapter Drive B Drive A
And one last thing. Don't forget that the drive farthest down the cable
from the adapter, no matter what its name, gets the 150 ohm
terminating resistor pack, usually a dual inline pin (DIP) package
which plugs in one way, with the notch at the same end as the white
outline mask under the socket on the PC Board of the drive.
-----------------------------------------------------------------------
External Disk Drive Adapter Pin-out and Connections to Diskette Drives
Some diskette adapter cards, in addition to their internal card edge
connector for floppy drives A and B, have a DB37S connector on the
outside of the computer for external drives C and D. The typical cards
which have both are diskette adapter only circuits. Sometimes the pads
may exist on the card for such a connector but it is not installed.
Cards which have not only an internal floppy diskette adapter circuit
but also an extra DB25S LPT2 centronics parallel port and dual DB9P or
DB25P RS-232C serial ports, a real time clock, and a DB17S joystick
game port. One may disable the diskette adapter on this card and
purchase a card of the first description very inexpensively.
DB37S External Drive Connector at Rear of Computer
19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
\------------------------------------------------------------/
\ o o o o o o o o o o o o o o o o o o o /
\ o o o o o o o o o o o o o o o o o o /
\------------------------------------------------------/
37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20
In order to make a cable from the external drive adapter to a pair of
drives, follow the cabling instructions above as to the twist but at
the adapter end of the cable use a ribbon type DB37P and insert the
ribbon so as to stab the colored edge of the ribbon with the dagger of
pin 21 skipping the daggers of pins 1, 20, and 2. This is necessary
since the pinout for ribbon connection is the same despite the row
oriented pin numbering instead of even-odd numbering but offset by
three conductors, which is what we might expect connecting a 37 pin
connector to 34 pin connectors. Also we know this since the colored
ribbon wire is an odd numbered conductor and must be ground. See below:
External Diskette Adapter
-----------------------,
unused 1 | Diskette Drive
unused 2 | ,-----------------------
unused 3 |----------------| 2 unused
unused 4 |----------------| 4 unused
| |----- alignment slot
unused 5 |----------------| 6 Drive Select 4
Index 6 |<---------------| 8 Index
Motor Enable C 7 |--------------->| 10 Drive Select 1 *
Drive Select D 8 |--------------->| 12 Drive Select 2 *
Drive Select C 9 |--------------->| 14 Drive Select 3 *
Motor Enable D 10 |--------------->| 16 Motor On *
Direction Select 11 |--------------->| 18 Direction Select
Step 12 |--------------->| 20 Step
Write Data 13 |--------------->| 22 Write Data
Write Gate 14 |--------------->| 24 Write Gate
Track 00 15 |<---------------| 26 Track 00
Write Protect 16 |<---------------| 28 Write Protect
Read Data 17 |<---------------| 30 Read Data
Head Select 18 |--------------->| 32 Head Select
unused 19 |----------------| 34 unused
-----------------------' `-----------------------
Bottom Row Sockets 20 - 37 of the DB37S Connector are Ground
The enabling of the two drives C and D follow exactly that of A and B
respectively as mentioned above, utilizing the twisted conductors in
the cable. Even these external drives must be passively selected by
Berg jumper to be the second of four drives available, for DS1 - DS4 it
must be DS2, for DS0 - DS3 it must be DS1. The resistor pack goes on
the farthest drive down the cable from the adapter.
copyright (c) 1986 Richard Steven Walz free to non-profit public domain
rst...@armory.com if this notice is intact
---------------------------------------------------------------------------
Will Rose
: In article <D3vos...@crash.cts.com>, Will Rose <c...@crash.cts.com> wrote:
[...]
: >
: >Try the WD 37C65. I've had one sitting on my bench for a couple of
: >years now, waiting for the SBC to be built to fit it in. Very nice
: >chip, very complete and flexible.
: -------------------------
: I believe that one is the generic '765 that the PC used and uses, and that
: many makers supply.
No, it's not the generic clone version. I don't know how the command
sets compare, but the hardware is much more highly integrated than an 8765.
Will Rose
: No, the twist does nothing of the sort. The controller CARD simply uses
: those other two drive selects for other things, namely MOTOR ON, as you can
: easily see from the controller to drive cable interface below! As for the
: extra external drives C and D, they are simply selected by the card and
: substituted in circuitry for the first two A and B which are then not
: connected for THAT external interface! Lookee here!:
: -Steve Walz rst...@armory.com
Odd, I've had a problem with two extra drives coming up with untwisted
cable on a standard controller, and being suppressed by a normal cable.
Must not have been quite as standard as I thought.
John Wilson
[wrt WD 37C65]
>No, it's not the generic clone version. I don't know how the command
>sets compare, but the hardware is much more highly integrated than an 8765.
The command sets are 100% identical, I think that's what he meant.
Yes the pinout is totally different and much more integrated, that's what
makes the chip so great!
John Wilson
Helmut Jungkunz
[wrt WD 37C65]
>The command sets are 100% identical, I think that's what he meant.
>Yes the pinout is totally different and much more integrated, that's what
>makes the chip so great!
way round, the chip is due to roast!
Regards
--
Helmut Jungkunz \* Zacherlstr. 14 \* D-85737 Ismaning \* ++49 89 961 46 33
ZNODE 51 CP/M BBS 15:00 - 3:00 CET (MEZ) ++49 89 961 45 75 upto 14400
---------------------- FOREVER Z! (motto by Rick Charnes) -----------
Richard Steven Walz
I just told you what IBM did. Some floppy drive makers even eliminated the
extra drive select lines!! They tried to be IBM compatible!!! It's just a
matter of rewiring though, as actually a drive has only one "select" line
and it's just divided and hooked to four lines on the 34 pin card edge
connector. The jumper you put to connect one of those simply selects it by
what line is active on the cable! You CAN build IBM/clones entirely without
the twists in the ribbon and then jumper each drive for its select number,
but they were afraid people were too STUPID to do that, and so they
invented the "twist" and setting all drives to the second drive select!
If you do that remember to rewire the controller to have four drive selects
from elsewhere on the board if you want them!
-Steve Walz rst...@armory.com
Tony Duell
han...@dis.anl.gov "roger hanscom" writes:
[Modifying / multiple controllers]
>
> No need for that. It is possible to run two controllers in the same box.
Are we talking about a PC or a CP/M (S100 bus ?) machine here. You can
(easily) run 2 floppy controllers in a single PC - I'm doing it at the moment
[How many drives can we put on a PC controller ?]
> Not so, there are select lines for 4 drives on the 34-conductor cable.
> Some PCs used DS3 and DS4 for "motor on" for drives 1 and 2, but these
> days there is usually no separate motor on (selecting the drive turns
> the motor on).
Are you sure about that? I've yet to see a PC (as in IBM compatable) drive or
controller that didn't use a separate motor-on signal. In fact, the original
IBM PC and PC/XT controller, which supported 4 drives, had 2 separate cables,
each connected to 2 drives. One fitted on an internal edge connector, and the
other on a D-37 socket on the backplate.
The data sheets I have on PC disk controller chips all imply that they only
support 2 drives (but see my earlier posts for more details).
> The Mitsumi 1.2 Mb 5.25" floppy drive has a very nice selection of
> "jumperable" options that includes 4 drive selects, motor speed,
> etc. I have submitted an article to The Computer Journal about my
That's interesting. I _think_ the 1.2Mb drive I used to test my idea of
replacing an 8" drive was a Mitsumi. It certainly had a lot of links in it,
none of them documented. If anyone has managed to obtain any technical info
on these drives, I'm interested.
> you what a relief it is to begin to think about getting rid of
> those noisy, power-hungry 8" drives. In this area, Mitsumi drives
I _like_ 8" drives - they're quite fast, reliable, and easy to repair.
-tony
John Wilson
>: > It'd also be nice to use up to 4 floppy drives on the
>: > new controller, but I notice that most PC drive controllers don't have
>: > documented support for anything past the 2nd drive. (Many of the data
>: > sheets on the chips claim support for 4 drives, nevertheless.) As a
>: > matter of fact, it appears that the cable signals for the 3rd and 4th
>: > drives may not exist on the 34-conductor cable.
>
>They exist - that's why you have to flip the cable on a standard PC.
>It suppresses floppies C and D. My XT controller (don't know the
>chip) supports 4 floppies, and I've seen other controllers that will,
>tho' most controllers handle only two.
Well, no. The PC version of the 34-pin cable only has select lines for
two drives (and separates out the "motor on" pins for each). The original
IBM PC floppy controller supports 4 drives by using two separate cables
with two drives on each (the internal 34-pin HD connector and the external
DB37), and uses two halves of a 74LS240 to decide which cable's input
signals to feed to the FDC chip at any given time.
Support for 4 floppies was dropped with the AT, there's only space in the
low-memory state machine variables and CMOS RAM drive type variables for two.
John Wilson
Allison J Parent
I happen to use S100 and my S100 crate is z80 based and has 4 <FOUR>
controllers! One for the hard disks, One for the 8inch drives, one for
the 5inch drives and one for the 3.5inch drives. That does not include
the controller that simulates the Romdisk/Ramdisk. It's a metter of
writing a bios that supports them, the bus does not limit you.
Running more than one in a PC, I've done three but the third required
setting up an oddball interupt and address for the controller. PC bus
software conventions have standardized the addresses and number of
controllers somewhat. The physical bus does not limit the nuber of
controllers however.
Allison Parent
David I. Baldwin
Compaticard IV disk controller card. It uses the Intel 82077 controller
chip and some external buffers and decoders and it comes with the driver
software. The Uniform and 22disk disk translation programs use the
Compaticard to read virtually all FM/MFM formats. The 82077 is a
superset of the 765 and handles every thing from single density to 2.88
Meg drives. It does have 4 drive select outputs and 4 motor enable
outputs, but it does not have the TG43 output. It is designed to be able
to connect directly to the drive cable, but the Compaticard buffers
several signals.
I haven't had to program floppy controllers, so I can't tell you about
that part. The 82077 does have some special modes that aren't on the
765, so I suppose it's a little more difficult.
Anyway, I thought I'd mention it because it seems to do what people were
asking for. The 82077 comes in a 68-pin PLCC.
--
Dave Baldwin Internet : dib...@netcom.com
DIBs BBS : (916) 722-5799 Compuserve: 70403,2444
DIBs FAX : (916) 722-3877 Genie : D.BALDWIN3
-=-=-=- @#$%^&* I can't even quote myself! Oh,well. -=-=-=-