1 Sep
2022
1 Sep
'22
6:59 p.m.
On 9/1/22 18:43, Mike Katz wrote:
Taking my memory back to the early 1980's and the
Western Digital floppy
disk controller chip family (177X single density and 179X double
density). I wrote the 6809 drivers for Gimix Flex. The controller chip
used the index pulse for sector zero position and for timing out a
failed read or write command. I don't recall if that controller chip
family could handle hard sectoring (one hole per sector) or not.
All WD and NEC floppy controllers use the index hole for formatting as
Fred mentioned. Further, I believe that three passes of the index hole
while attempting to read or write gets you "Sector not found" if the
address ID isn't detected. See datasheet and app notes for details.
No WD or NEC floppy controller handles hard-sectored recording schemes.
HOWEVER:
A few years back, I was sent a bunch of 8" HS disks that were really
puzzling--the sector ID address headers didn't line up with the sector
hole timing. In fact, they were WAY off.
It turns out that some 8" drives can be set to separate the sector holes
from the index hole (separate output pins for index and sector). Doing
so, gives you what amounts to a soft-sectored floppy, regardless of what
the physical object is. I know that I used a Siemens FDD-200 drive
jumpered accordingly to read them.
Good times.
--Chuck
1 Sep
1 Sep
7:22 p.m.
New subject: Flipping an 8" diskette
On 9/1/22 18:43, Mike Katz wrote:
> Taking my memory back to the early 1980's and the Western Digital floppy
> disk controller chip family (177X single density and 179X double
> density). I wrote the 6809 drivers for Gimix Flex. The controller chip
> used the index pulse for sector zero position and for timing out a
> failed read or write command. I don't recall if that controller chip
> family could handle hard sectoring (one hole per sector) or not.
On Thu, 1 Sep 2022, Chuck Guzis via cctalk wrote:
All WD and NEC floppy controllers use the index hole
for formatting as
Fred mentioned. Further, I believe that three passes of the index hole
while attempting to read or write gets you "Sector not found" if the
address ID isn't detected. See datasheet and app notes for details.
OB_Obscure stuff:
At least some of the Teac 5.25" FD55-x series drives ALSO used the index
pulse for determining "drive ready". If you mask off the index hole
with a write protect tab, then those drives will give DRIVE NOT READY.
So, on THOSE drives, masking off index pulse (for disks (such as come
Cromemco) with first sector too soon after index, etc.) should be done by
interrupting that pin of the cable, or modifying the FDC controller board.
(ANY masking of the index pulse will "corrupt" the FDC's guesses of what
an error is)
On a bet, . . .
we noticed that the Central Point Option Board software could not work
without index pulse. You CAN modify a drive to index off of the spindle,
without an index hole. FLIP a hard sectored floppy, and format it with
indexing off of the spindle. THAT disk can now be read and/or written.
But, the Central Point board software can't handle it. Punching an index
hole access hole in that flippy won't help, because the disk is hard
sectored. Thus, a disk whose content canbe copied with COPY or DISKCOPY,
but not with the Option Board (with supplied software).
No WD or NEC floppy controller handles hard-sectored
recording schemes.
But, how about a WD TRACK READ with the index pulse masked?
HOWEVER:
A few years back, I was sent a bunch of 8" HS disks that were really
puzzling--the sector ID address headers didn't line up with the sector
hole timing. In fact, they were WAY off.
It turns out that some 8" drives can be set to separate the sector holes
from the index hole (separate output pins for index and sector). Doing
so, gives you what amounts to a soft-sectored floppy, regardless of what
the physical object is. I know that I used a Siemens FDD-200 drive
jumpered accordingly to read them.
Good times.
--Chuck
COOL!
8:03 p.m.
New subject: Flipping an 8" diskette
It turns out that some 8" drives can be set to
separate the sector holes
from the index hole (separate output pins for index and sector). Doing
so, gives you what amounts to a soft-sectored floppy, regardless of what
the physical object is.
The flippy 8" diskette I linked a picture of this morning was written that way. Took
me a little time to figure out. Part of a massive lot of diskettes all done like that.
I'm guessing the original owner either worked for Varityper at some point, or knew
someone who did, and was able to acquire a huge quantity of hard sector diskettes
cheaply.
I know that I used a Siemens FDD-200 drive jumpered
accordingly to read them.
Shugart SA-801 (or 800 with the LSI chip and 801 jumper closed) will also do it!
Thanks,
Jonathan
8:11 p.m.
New subject: Flipping an 8" diskette
On 9/1/22 19:22, Fred Cisin via cctalk wrote:
But, how about a WD TRACK READ with the index pulse
masked?
Well, okay, but I never found that to be very useful, because to achieve
correct byte alignment, you need one of the standard address marks to
get the data separator working right, which few HS recording schemes
employ. That's not to say that it can't be done, but it's a lot of
trouble, particularly for MFM encoding. There's also the dodge on PCs
of using two drives--one with a standard formatted soft-sector floppy to
start a "read track with a very large sector length code" operation and
changing the drive select to the second drive while the read's in progress.
The "Read Track" feature is useful for recovering data from regular
soft-sectored IBM System/3 or 3740-type floppies, however.
All in all, a cheap MCU is easier--and cheaper, I think. Consider a
little STM32F405 MicroPython board--it runs at 168MHz, timers that can
capture pulses at 84 MHz and has about 192KB of fast SRAM inside. Cost
is around $11-15 and includes MicroSD and USB interfaces.
Similarly, you can get an Option Board working on index-less or
hard-sectored floppies by using a cheap MCU to either generate or mask
index pulses. You can even use the same MCU to simulate index pulses
for equipment requiring hard-sectored floppies. A little 8-pin <$1 MCU
is more than adequate for the job.
Silicon has gotten so cheap!
--Chuck
596
days inactive
596
days old
3 comments
3 participants
participants (3)
-
Chuck Guzis -
Fred Cisin -
Jonathan Chapman