Converting from 8 to 3.5 inch floppy diskettes
marcusb
Get ready to use 3.5 inch diskette drives instead of those big and
noisy 8 inch drives on your Cromemco (and therefore any other CPM)
genre computers:
http://majzel.blogspot.com/2009/04/converting-from-8-to-35-inch-floppy.html
(Note that this requires trivial modification of your disk controller,
in my case the Cromemco 64FDC to connect the 8 inch disk drive ready
of pin 22 to the 5 inch ready line which is pin 34. There is a link
in my article to a previous post showing how that was done already).
Regards marcus.
glen herrmannsfeldt
Pin 34 is used for either READY or DISK CHANGE, sometimes with
a jumper for selection.
But I do agree that for actual use 3.5in floppy is a better choice.
Normally, 3.5in HD floppies run 300RPM, compared to 360RPM for
8 inch. That should only be a problem if the controller times
out waiting for the index pulse.
-- glen
Herbert Johnson
> Get ready to use 3.5 inch diskette drives instead of those big and
> noisy 8 inch drives on your Cromemco (and therefore any other CPM)
> genre computers:
>
> http://majzel.blogspot.com/2009/04/converting-from-8-to-35-inch-flopp...
> Regards marcus.
Thanks for the blog note and this post. I have comment and questions,
from a first read.
If I read your note correctly, the particular brand and model of 3.5"
drive you selected, is designed so that grounding pin 2 (high/low
density) sets the drive to read and write at double density rates, AND
speeds up the rotation rate of the diskette from 300 RPM to the
standard for 8-inch floppy drives of 360 RPM. I'm explicitly saying
"read and write at double density rates" because, I believe, the 3.5-
inch drives which permit "density" changes also change the response of
their read and write circuits, to match the frequency and bandwidth of
those specific data rates.
These features of this drive, may not be available on other 3.5"
drives. To choose other drives, someone will have to read the data
sheets, maybe use test instruments to verify results (the dreaded
"oscilloscope" perhaps). "It boots OK" won't be good enough. Thanks
for calling this out as a "japanese mode" of "1.6MB".
It's also important to use the correct diskette media for this mode of
operation. You say that you use 1.2MB media, so called "high density"
media, and not 720K or "double density" media. The terms used by the
drive manufacturer were "low" and "high" density. Modern users simply
assume "high density" is normal. But "high" and "double density" are
still written on boxes of diskette media. Again, this is an issue of
"frequency", the data density as actually written on the media (bits
per inch). Writing at the "wrong" density for the media may not work
for all tracks written, and the data may not persist. Again, it may
'boot" today from track 0, but files on track 40 may fail tomorrow.
Your blog entry notes something about data surviving for "another 25
years". That remains to be seen. Survival of disks and the actual
coatings depends on quality of materials, conditions of storage
including humidity and temperature. It's no fun to see an 8-inch
diskette, or a 5.25-inch diskette, have its oxide coating scrapped off
by a drive. These coatings have lubricants which dissipate over time.
Mold will grow on disks, catch in the drive "heads", and act like a
scraper. I've seen it, this is not myth. All the more reason to use
newer media and newer drives.
One question. You wrote "Locate OPA pads and connect a wire between
them.". What is "OPA"? (I'll read the manual also, thanks for the link
and the PDF.)
Thanks for this good work, well described and documented. I don't know
how persistent "blog" entries are: will you have a copy of this info
on your Web site?
If my comments are helpful and correct, you can quote me. Let me know
where I'm wrong. I accumulate this kind of technical stuff on the
following Web page, you can link to it if you wish:
http://www.retrotechnology.com/herbs_stuff/drive.html
..and I'll add a reference to your work there, soon.
herb johnson
retrotechnology.com
glen herrmannsfeldt
(snip on 3.5in and 8in floppy drives)
> If I read your note correctly, the particular brand and model of 3.5"
> drive you selected, is designed so that grounding pin 2 (high/low
> density) sets the drive to read and write at double density rates, AND
> speeds up the rotation rate of the diskette from 300 RPM to the
> standard for 8-inch floppy drives of 360 RPM. I'm explicitly saying
> "read and write at double density rates" because, I believe, the 3.5-
> inch drives which permit "density" changes also change the response of
> their read and write circuits, to match the frequency and bandwidth of
> those specific data rates.
The write current is different, as far as I know, the read
circuitry doesn't change. Since both DD and HD 3.5in drives
run at 300RPM I would be surprised to see one switch to 360.
The tradition for format is to start on an index pulse, write
the track ending with a gap, and keep writing until the next
index pulse. It might be that the time is too long for some
controllers and/or formatting programs. Once formatted, soft
sector systems should not notice the extra gap length.
> These features of this drive, may not be available on other 3.5"
> drives. To choose other drives, someone will have to read the data
> sheets, maybe use test instruments to verify results (the dreaded
> "oscilloscope" perhaps). "It boots OK" won't be good enough. Thanks
> for calling this out as a "japanese mode" of "1.6MB".
> It's also important to use the correct diskette media for this mode of
> operation. You say that you use 1.2MB media, so called "high density"
> media, and not 720K or "double density" media. The terms used by the
> drive manufacturer were "low" and "high" density. Modern users simply
> assume "high density" is normal. But "high" and "double density" are
> still written on boxes of diskette media. Again, this is an issue of
> "frequency", the data density as actually written on the media (bits
> per inch). Writing at the "wrong" density for the media may not work
> for all tracks written, and the data may not persist. Again, it may
> 'boot" today from track 0, but files on track 40 may fail tomorrow.
For media, it is the magnetic properties that matter. HD media
requires a higher write current. Using the high current on DD
media also doesn't work so well. I believe 3.5in drives reliably
sense this with a sense hold on the disk.
-- glen
All...@localhost.net
wrote:
This is not generally hard to do, I've done this using, Compupro
DISK1A, CCS 2242 and a few others plus my own design that were
set up for 5.25" drives (designed before 3.5"). Even My KAypros have
this as they are lighter and can store as much or more than 80track
two sided doubled density 5.25" drives and three 3.5" can fit where
two half height 5.25 were. The latter being useful for packaged
systems that are cramped for disk space.
The usual problem is if the system was setup for the most common 8"
formats it will not always work for 3.5" without some mods to the
BIOS.
Makes a good improvement if you can at least take advantage
of the drives capabilities.
Allison
marcus bennett
Of course I'd expect you to ask good questions .. to which you will
see I dont have all the answers.
The selection of the right 3.5 inch diskette drive is critial IMHO.
Just like when the IBM PC forced changes in 360K 5.25" floppy design
to the point where many drives can't select between RDY or DC
operation, it seems the same evolution has happened to 3.5 inch
drives. Most are now single speed, PC format only drives.
My choice of the Samsung SFD-321B was the result of a long search.
Initially I was excited to find this document
http://maben.homeip.net/static/S100/DISK/35%20diskettes/3.5%20diskette%20jumpers.pdf
which lists some interesting jumpers, but it was much a googling
later when I found the Datasheet to the SFD-321B.
To be clear, when pin 2 is forced low, and you insert a HD media
(2MB), the drive spins at 360rpm and goes to 1.6MB mode.
Oh and I have no idea what OPA is, but if you find out then you might
also be able to tell me what OPB is also, they dont expand that
acronym in the manual either :-(
Regards marcus
Thomas "Todd" Fischer
"marcus bennett" <amo...@gmail.com> wrote in message
news:f3d4326f-a3de-4cec...@j8g2000yql.googlegroups.com...
> Herb,
>
> Of course I'd expect you to ask good questions .. to which you will
> see I dont have all the answers.
(snip)
> Oh and I have no idea what OPA is, but if you find out then you might
> also be able to tell me what OPB is also, they dont expand that
> acronym in the manual either :-(
>
> Regards marcus
Perhaps "Option A" and "Option B"?
MdntTrain
> Dear All,
>
> Get ready to use 3.5 inch diskette drives instead of those big and
There are a variety of "tri-density" 3.5" floppy drives out there --
my favs being the Teac FD-235HG and Mitsumi D359T3. These are better
known and more documented than the Samdung.
As everyone knows, 3.5" drives normally spin at 300rpm for all modes..
with density (data rate) usually being determined by holes on the
floppy carrier itself. Yet, tridensity drives are different in that
they use a low (0v) on pin 2 to signal 360rpm HD mode... which is,
unfortunately, the opposite of the 5.25HD drive.
The 5.25HD drive uses a high (+5v) on pin 2 to signal HD and (0v) to
indicate low density. If the drive is jumpered for a constant 360rpm,
then the data rate must change between 500 and 300 for high / low
density use. Some PC controllers did this, and I preferred it for the
higher performance low density.
Anyhoo, I'm kinda sceptical about 3.5" drives outlasting either my 8"
drives OR my 5.25" drives. The 8" ones are easiest to service, at
least in terms of caps replacement. It'll be interesting to see
whether media becomes the issue, or the drives themselves.
jS
Herbert Johnson
> Herbert Johnson wrote:
> > .... I'm explicitly saying
> > "read and write at double density rates" because, I believe, the 3.5-
> > inch drives which permit "density" changes also change the response of
> > their read and write circuits, to match the frequency and bandwidth of
> > those specific data rates.
>
> The write current is different, as far as I know, the read
> circuitry doesn't change. Since both DD and HD 3.5in drives
> run at 300RPM I would be surprised to see one switch to 360.
Glenn, Marcus has a link to the engineering manual for the drive. The
manual says "360RPM" for that option as Marcus describes it. One can
measure actual speed from the index pulse, if necessary. (Marcus, did
you independently verify the rotation speed?)
I believe other discussions in c.o.c on use of 3.5" floppy drives,
have discussed drives which can rotate at 360RPM. But the IBM
alternative to dual-rotation-speed drives, was to change the
"frequency" of the read/write signals - speed up the data instead of
the drive! So 3.5" drives which can physically run at 360RPM are at
least not typical. I mention this on my Web page.
Page 5 of the manual on Marcus's site, refers to "precompensation" and
"read bit shift" values for "low" (double) density versus "high
density". I was speaking from memory when I said there were circuit
differences in "frequency and bandwidth" for different data rates.
While I recall some discussion of frequency compensation at another
time, it's sufficient to note the precompensation circuit change.
You mentioned another difference, namely higher write current for high
density - thanks. I'll see if I can describe this better in the future
on my site. My point was that floppy drive read and write circuits
operate *differently* at double density from high density. So any
scheme to use a drive "differently" must avoid fooling the drive too
much.
> The tradition for format is to start on an index pulse, write
> the track ending with a gap, and keep writing until the next
> index pulse. It might be that the time is too long for some
> controllers and/or formatting programs. Once formatted, soft
> sector systems should not notice the extra gap length.
> For media, it is the magnetic properties that matter. HD media
> requires a higher write current. Using the high current on DD
> media also doesn't work so well.
It gets a little confusing, in the current discussion, to argue about
formatting. There are many, many formatting programs which covered the
technology of the 1970's and 1980's. The bottom line is simpler: if
the floppy controller "expects" to see index pulses at 360RPM, but
gets them "later" at 300RPM, there would either be confusion or more
sectors written. But as you suggest, the "wrong" rotation rate
produces the "wrong" magnetic data "rates" (density, bits per inch),
and that is a known problem.
It's simpler to operate disks and drives "correctly", than to play
with marginal conditions. Marcus has demonstrated that the brand and
model drive (and disk media) he's described can, within its design,
operate like an 8-inch drive and media.
> I believe 3.5in drives reliably sense [HD or DD media] with a sense [hole] on the disk.
It's not uncommon for users to cover the HD sense hole, to write on HD
disks in DD format. They report "success" many times at doing this. As
you say, the magnetic media matters, and I think it's a bad idea.
Differences in drives and media by brand, and age and time, may
produce success now but failures later. Again, it's informative to
review old c.o.c posts (and posts in other newsgroups) on this subject
through the years.
Another note on disk lifetime. The manual cited above says "disk life;
3 X 10**6 passes per track". 3 million sounds like a lot, but at
360RPM, that's under 10,000 minutes while in rotation. Ten minutes a
day every day, that's three years use; 30 minutes, one year's use.
Plus the media dries out, over decades.
Herb Johnson
retrotechnology.com
Axel Berger
>AND speeds up the rotation rate of the diskette from 300 RPM to the
>standard for 8-inch floppy drives of 360 RPM.
Before seeing that data sheet, I'd have bet money that such a thing did
not exist. That said, I don't really see the need. As far as I'm aware
there should be no problem whatever reading and writing from and two 15-
sector tracks on a standard 3.5" HD-floppy. The one thing that may well
not work, thought it might do all the same, is formatting.
glen herrmannsfeldt
(snip, I wrote)
>> The write current is different, as far as I know, the read
>> circuitry doesn't change. ?Since both DD and HD 3.5in drives
>> run at 300RPM I would be surprised to see one switch to 360.
> Glenn, Marcus has a link to the engineering manual for the drive. The
> manual says "360RPM" for that option as Marcus describes it. One can
> measure actual speed from the index pulse, if necessary. (Marcus, did
> you independently verify the rotation speed?)
OK, not an ordinary 3.5in drive.
> I believe other discussions in c.o.c on use of 3.5" floppy drives,
> have discussed drives which can rotate at 360RPM. But the IBM
> alternative to dual-rotation-speed drives, was to change the
> "frequency" of the read/write signals - speed up the data instead of
> the drive! So 3.5" drives which can physically run at 360RPM are at
> least not typical. I mention this on my Web page.
That is what IBM did for 5.25in. For 5.25in, HD drives run at 360RPM,
(and 500kb/s data rate), DD drives at 300RPM (and 250kb/s data rate).
IBM used HD drives to read/write DD data at 360RPM and 300kb/s.
For 3.5in normally (that is, for DOS/Windows, Apple, Sun, HP-UX
machines) both DD and HD run at 300RPM. This change is why HD
5.25in disks are 1.2MB formatted, while 3.5in are 1.44MB.
> Page 5 of the manual on Marcus's site, refers to "precompensation" and
> "read bit shift" values for "low" (double) density versus "high
> density". I was speaking from memory when I said there were circuit
> differences in "frequency and bandwidth" for different data rates.
> While I recall some discussion of frequency compensation at another
> time, it's sufficient to note the precompensation circuit change.
In the usual case, precomp is in the controller, not the drive,
but yes that changes. It is also only for writing.
> You mentioned another difference, namely higher write current for high
> density - thanks. I'll see if I can describe this better in the future
> on my site. My point was that floppy drive read and write circuits
> operate *differently* at double density from high density. So any
> scheme to use a drive "differently" must avoid fooling the drive too
> much.
The difference between HD and DD 3.5in is much smaller than
for HD and DD 5.25in. Close enough that it will likely work.
>> The tradition for format is to start on an index pulse, write
>> the track ending with a gap, and keep writing until the next
>> index pulse. ?It might be that the time is too long for some
>> controllers and/or formatting programs. ?Once formatted, soft
>> sector systems should not notice the extra gap length.
>> For media, it is the magnetic properties that matter. HD media
>> requires a higher write current. Using the high current on DD
>> media also doesn't work so well.
> It gets a little confusing, in the current discussion, to argue about
> formatting. There are many, many formatting programs which covered the
> technology of the 1970's and 1980's. The bottom line is simpler: if
> the floppy controller "expects" to see index pulses at 360RPM, but
> gets them "later" at 300RPM, there would either be confusion or more
> sectors written. But as you suggest, the "wrong" rotation rate
> produces the "wrong" magnetic data "rates" (density, bits per inch),
> and that is a known problem.
When I got my first computer with a floppy drive, I ordered
the (free) data books from Western Digital. They include a lot
of details about the controllers, index pulses, and formatting.
If I remember right, for the 1793 at least, for normal read/write
operations it counts index pulses but doesn't time them. If five
pulses go by before it finds the desired sector then it reports
an error. Formatting is different.
> It's simpler to operate disks and drives "correctly", than to play
> with marginal conditions. Marcus has demonstrated that the brand and
> model drive (and disk media) he's described can, within its design,
> operate like an 8-inch drive and media.
>> I believe 3.5in drives reliably sense [HD or DD media]
>> with a sense [hole] on the disk. ?
> It's not uncommon for users to cover the HD sense hole, to write on HD
> disks in DD format. They report "success" many times at doing this.
(snip)
> Another note on disk lifetime. The manual cited above says "disk life;
> 3 X 10**6 passes per track". 3 million sounds like a lot, but at
> 360RPM, that's under 10,000 minutes while in rotation. Ten minutes a
> day every day, that's three years use; 30 minutes, one year's use.
> Plus the media dries out, over decades.
For 8 in drives, traditions was to run the motor (usually on 120VAC)
all the time, and use a head load solenoid such that the disk
only contacts the head when needed. My first TEAC drive, an FD55F,
has (I still have it) a head load solenoid. When DOS machines
got popular, they removed it from the design, replacing it with
a device that keeps the heads apart unless a disk is inserted.
Another possibility for CP/M machines, I have an LS-120 drive
with IDE interface that can read/write DD and HD floppies.
-- glen
glen herrmannsfeldt
> On Apr 16, 8:48?am, marcusb <amou...@gmail.com> wrote:
>> Get ready to use 3.5 inch diskette drives instead of those big and
>> noisy 8 inch drives on your Cromemco.
> There are a variety of "tri-density" 3.5" floppy drives out there --
> my favs being the Teac FD-235HG and Mitsumi D359T3. These are better
> known and more documented than the Samdung.
If you are counting density, then you should also include ED
(2.88MB formatted) drives/disks. They use a completely different
media (I believe ferrite), and perpendicular recording.
IBM PS/2 and NeXT used them. Otherwise, they didn't catch on.
> As everyone knows, 3.5" drives normally spin at 300rpm for all modes..
> with density (data rate) usually being determined by holes on the
> floppy carrier itself. Yet, tridensity drives are different in that
> they use a low (0v) on pin 2 to signal 360rpm HD mode... which is,
> unfortunately, the opposite of the 5.25HD drive.
Reminds me of the first time I connected a 5.25in HD drive
to an LSI-11. I connected all the pins to the appropriate
cable line, including the Write Current pin. I should have
thought it through before, but anyway the first format operation
worked fine until it got to track 44. Eight inch tradition
is to reduce the write current on inner tracks (the bits
get closer together), and that is what the Write Current
pin does on 8 inch drives. So my quick fix was to ground
that pin such that it would work.
-- glen
Barry Watzman
leave the data rate alone, with the disk turning slower, you get more
bytes per track.
There are no standards issues here, and normally the formatting program
has to be custom written for every CP/M system anyway. So leave it at
300 rpm, in fact use the sector size that will give the largest capacity
(might be 1k sectors, or 512 byte sectors) and configure the formatting
and BIOS appropriately. If you have a system with some type of bank
memory, you can even do full track reading/writing, and you will get
hard drive speeds from a floppy (I did that with 1K sectors, 9 of them
per track, on the 8" floppy drivers for the OS' I wrote for the Heathkit
/ Zenith Z-100 series (although, of course, I still also supported
standard SSSD and also all of the formats supported by Heath/Zenith)).
If you can do full track disk I/O it is actually pretty shocking just
how fast a floppy system can get.
bud
Group: comp.os.cpm Date: Fri, Apr 17, 2009, 11:09am (CDT-2) From:
herbrj...@gmail.com (Herbert Johnson)
script:
>You mentioned another difference,
>namely higher write current for high
>density - thanks.
I submit (with neither evidence nor credentials) that the write current
is _lower_ at high densities than at low densities.
________________________________________
Group: comp.os.cpm Date: Fri, Apr 17, 2009, 6:55pm (CDT+5) From:
g...@ugcs.caltech.edu (glen herrmannsfeldt)
script:
>… I should have thought it through
>before, but anyway the first format
>operation worked fine until it got to
>track 44. Eight inch tradition is to reduce
>the write current on inner tracks (the bits
>get closer together), and that is what the
>Write Current pin does on 8 inch drives.
Co-incidentally interesting. ((-;
salaam,
dowcom
To e-mail me, add the character zero to "dowcom". i.e.:
dowcom(zero)(at)webtv(dot)net.
The fact that 'conventional wisdom' is indeed 'conventional',
does not, in any way, imply that it is wise.
glen herrmannsfeldt
> I submit (with neither evidence nor credentials) that the write current
> is _lower_ at high densities than at low densities.
Everything else equal, yes. 8 inch disks use lower write
current on the inner tracks.
But high density disks use higher coercivity materials
http://en.wikipedia.org/wiki/Coercivity
requiring higher current.
-- glen
Greegor
Is there a good online reference that explains
the different coercivities ( in Oersteds ) used
for the ferric oxide coatings on various
formats of various sizes of floppy?
Single Density
Double Density PC/XT
Quad Density (96 TPI)
High Density (100 TPI)
? Density (135 TPI?)
glen herrmannsfeldt
(snip)
> Is there a good online reference that explains
> the different coercivities ( in Oersteds ) used
> for the ferric oxide coatings on various
> formats of various sizes of floppy?
> Single Density
> Double Density PC/XT
> Quad Density (96 TPI)
> High Density (100 TPI)
> ? Density (135 TPI?)
http://en.wikipedia.org/wiki/List_of_floppy_disk_formats#Physical_composition
Quad density is somewhat of a misnomer, as the bit rate is the
same, but the tracks are narrowed. That requires more sensitive
amplifiers, but not a media change.
Double density (MFM) has the same minimum bit spacing (not counting
precompensation) as single density (FM).
-- glen
Herbert Johnson
> Is there a good online reference that explains
> the different coercivities ( in Oersteds ) used
> for the ferric oxide coatings on various
> formats of various sizes of floppy?
http://www.retrotechnology.com/herbs_stuff/drive.html#data
"Data about floppy drives and media" is a list by drives and media
(density). What you request, is there. (HD media is not ferric oxide,
it's cobalt). Whether it's "good" or "explanatory" is a matter of
opinion. The page represents several years of accumulation of
technical content in response to questions like yours, and examples of
people's work on floppy drives and the reading and writing diskettes.
It's not great literature or an encyclopedia, but even Wikipedia links
to it.
Herb Johnson
retrotechnology.com
glen herrmannsfeldt
(snip)
> "Data about floppy drives and media" is a list by drives and media
> (density). What you request, is there. (HD media is not ferric oxide,
> it's cobalt).
I believe it is cobalt doped ferric oxide. Mostly ferric
oxide with a little bit of cobalt added.
-- glen
MikeS
<snip>
>
> There are no standards issues here, and normally the formatting program
> has to be custom written for every CP/M system anyway. So leave it at
> 300 rpm, in fact use the sector size that will give the largest capacity
---------
There are indeed standards issues, at least in the context of the
original post, since the main idea there is to transparently replace
an 8" drive without the OS (CDOS or Cromix) being aware of it;
different speed or SPT would certainly be noticed.
Myself, I prefer to use 5 1/4 HD drives, since the 8" speed etc. is in
fact completely compatible with the normal 1.2HD configuration in a
PC, so that you can (re)-create 8" images and read them with anadisk/
uniform/pc-alien etc. on a completely stock PC without the need for a
special cable/adapter or 24V supply or worrying about the speed issue.
On the other side (Cromemco in this case) most 5 1/4" HD drives will
work instead of an 8" drive, whereas tri-mode (360RPM-capable) 3.5"
drives are not nearly as common. And with the addition of one jumper
on a Cromemco the drive just plugs into the 5 1/4" cable, so there is
no 50/34 adapter/cable required there either, and of course a "real"
8" drive can still remain connected as well for when it's needed.
And I just happen to have a pile of 5 1/4" diskettes, so availability
of media is not the issue for me that it might be for other folks.
But either way, 5.25 or 3.5 inches, it's good to have alternatives so
that we can minimize wear on our 8" drives and disks.
mike
glen herrmannsfeldt
>> There are no standards issues here, and normally the formatting program
>> has to be custom written for every CP/M system anyway.
(snip)
> There are indeed standards issues, at least in the context of the
> original post, since the main idea there is to transparently replace
> an 8" drive without the OS (CDOS or Cromix) being aware of it;
> different speed or SPT would certainly be noticed.
I am not sure if formatting is part of the OS or not.
For soft sector disks, it is normal to ignore the index pulse,
except possible to count them, for normal read/write operations.
(Count them to time out if the sector doesn't exist.)
Formatting does need at least to write the final gap long enough
to reach the index. It might be that for some controllers and
format programs that will fail before reaching the index on
a 300RPM drive.
Then again, the 8 inch floppy convention was that disks came
low level formatted.
-- glen
MdntTrain
It should be noted that different variations of this drive exist.. and
that possibly, the newer variations are significantly less fun and
desireable to work with than the version depicted in MarcusB's
website.
I purchased two from Newegg just to compare to the Mitsumi D359T3 and
Teac FD-235HG series.
The two I received have a very different PCB than MB's and do not
match the available .pdf manual. The manual appears to reference Rev
6, whereas these are Rev. A.. (counting in hex). Drive select pads
are missing, the other pads are much harder to discern and make sense
of, and the overall build quality of the drive while not the worst
I've seen from the Chinese just doesn't compare to Mitsumi or Teac.
I'm able to see OPA and OPB markings, but compared to MB's, these are
barely distinguishable and it's not clear which of the several pads
nearby they are referring to.
Page 25 of the 31 page SFD-321B .pdf MB provides tells how to read the
serial number. Mine is FBTA S1AU5xxxxx Rev. TA. The PCB is Rev
WT-10S 2.1.
I can't recommend this particular rev of these drives unless someone
is just desperate and can't find Teacs or Mitsumis.
I'd like to know if anyone else got this lousy rev, and if they got it
going, and if so, which pads were OPA?!
jS
MikeS
> MikeS <dm...@torfree.net> wrote:
> > On Apr 17, 6:18?pm, Barry Watzman <WatzmanNOS...@neo.rr.com> wrote:
> > <snip>
> >> There are no standards issues here, and normally the formatting program
> >> has to be custom written for every CP/M system anyway.
>
> (snip)
>
> > There are indeed standards issues, at least in the context of the
> > original post, since the main idea there is to transparently replace
> > an 8" drive without the OS (CDOS or Cromix) being aware of it;
> > different speed or SPT would certainly be noticed.
>
> I am not sure if formatting is part of the OS or not.
Well, every OS distribution I've seen and worked with included at
least one format utility, and I've found them quite useful...
> For soft sector disks, it is normal to ignore the index pulse,
> except possible to count them, for normal read/write operations.
> (Count them to time out if the sector doesn't exist.)
I wonder why most systems with a few exceptions like CBM and Apple's
GCR disks bother to have an index sensor at all then; I've just kind
of assumed that when the Cromemco format utility checks for an
inserted disk and reports the drive speed it uses the index pulse...
> Formatting does need at least to write the final gap long enough
> to reach the index. It might be that for some controllers and
> format programs that will fail before reaching the index on
> a 300RPM drive.
>
> Then again, the 8 inch floppy convention was that disks came
> low level formatted.
>
> -- glen
And whose format was that exactly, once they went to DS and DD? I
don't recall 3M or Dysan selling Cromemco formatted diskettes, and
I've just looked through several boxes of DS/DD disks of various
brands and they were all initially unformatted, most of them
explicitly so.
But hey, if you and Barry prefer to go off on your own with non-
standard RPMs and SPTs, by all means; I just prefer to be able to
download and exchange disk images that conform to the relevant specs
for the system in question, especially when that happens to be
fundamentally compatible with the ubiquitous PC.
m
Mr Emmanuel Roche, France
I don't remember seeing you before. A quick search told me that you
are Mike H. Stein, a former Cromemco dealer, living in Toronto,
Canada.
You mentioned "standards". There is, indeed, a standard disk format
for CP/M: it is the 8" disk format listed in the CP/M documentation
since 1976... (I used to call it "IBM 3740" standard.)
I don't know if you remember but, a few years ago, one English person
was surprised to discover that some 5 3/4" floppies were formatted as
8" floppies! This person only knew that the capacity of the floppies
was 243KB. I instantly reacted and, indeed, those 5 3/4" floppies were
just using the standard 8" format. No more problem transferring them.
I also noticed that later DRI OSes (like CP/M Plus, DR-DOS, etc) all
use 720KB. This is the one I use under CP/M-86 Plus. The only problem
is that FX-86 does not work in a "DOS box" under Windows. But my main
computer is running DR-DOS, so this is not a big problem for me.
Yours Sincerely,
Mr. Emmanuel Roche, France
glen herrmannsfeldt
(snip, I wrote)
>> For soft sector disks, it is normal to ignore the index pulse,
>> except possible to count them, for normal read/write operations.
>> (Count them to time out if the sector doesn't exist.)
> I wonder why most systems with a few exceptions like CBM and Apple's
> GCR disks bother to have an index sensor at all then; I've just kind
> of assumed that when the Cromemco format utility checks for an
> inserted disk and reports the drive speed it uses the index pulse...
You need it for formatting. Well, as Apple shows you don't,
but it makes it a lot easier. For the IBM standard, you start
writing (during formatting) at the index pulse, and stop at
the next index pulse. The end should be a gap (gaps actually
contain bit transitions) from the last sector to the index.
The beginning and end are gaps such that the splice point
is in a gap.
At some point in another newsgroup I figured out how it would
have to work for Apple to do it without index pulses, and
it turns out that is pretty much the way they do it. The problem
is that there is some uncertainty in the rotation speed, and you
have to at least be sure to write the whole track. (There
could be left over bits with valid sector headers.) What
Apple does is write track one with slightly more that should
be enough for a track, and then verify that all the sectors
(likely just the first and last) are readable. Then rewrite
the track increasing the gap lengths slightly. Repeat until
one of the sector headers is overwritten, then reduce the gaps
by the appropriate amount, and write all the other tracks with
the now appropriate gap size. That assumes that speed changes
slowly with time, usually a good guess. (Maybe they do
a binary search for the appropriate gap size.)
It is much easier with an index pulse. The final gap is long
enough to allow for the fastest (within tolerance) speed.
The gaps between sectors have to be big enough to allow
for the variability of speed between the formatting drive
and subsequent sector writes.
The IBM standard formats are fairly well described in
the Western Digital floppy controller books. Partly that is
because the formatting program has to do more of the work than
for the Intel (and compatible) controllers.
>> Formatting does need at least to write the final gap long enough
>> to reach the index. ?It might be that for some controllers and
>> format programs that will fail before reaching the index on
>> a 300RPM drive. ?
>> Then again, the 8 inch floppy convention was that disks came
>> low level formatted. ?
> And whose format was that exactly, once they went to DS and DD? I
> don't recall 3M or Dysan selling Cromemco formatted diskettes, and
> I've just looked through several boxes of DS/DD disks of various
> brands and they were all initially unformatted, most of them
> explicitly so.
> But hey, if you and Barry prefer to go off on your own with non-
> standard RPMs and SPTs, by all means; I just prefer to be able to
> download and exchange disk images that conform to the relevant specs
> for the system in question, especially when that happens to be
> fundamentally compatible with the ubiquitous PC.
And especially with easy to find and replace 3.5in drives.
-- glen
Barry Watzman
of the OS, meaning that it was not written or supplied by Digital Research.
CP/M has no "standard" disk format, with the exception of SSSD 8".
Beyond that, every other format, and all 5.25" formats, were developed
by the party implementing the BIOS. That same party also had to write
the format program (in particular because, even ignoring disk formats,
there is also no standard disk controller or even disk controller chip;
there were hundreds of disk controllers, some using {any of about a
half-dozen) Western Digital chips, some using NEC chips, and many (North
Star, Processor Technology, some Heath/Zenith, IMS, etc.) not using a
chip at all but using discreet logic in a totally proprietary fashion.
Each of these required a unique format program, even if the disk format
was the same (although usually, the disk format was not the same).
The index pulse is used only during the format operation. It is ignored
for all other operations (although in some cases it is the basis for the
"time out" duration, but not for diskette rotational position).
Herbert Johnson
Glen, I was going to reply to your posts about "standards" and
"conventions" and so forth, but it would take a lot of explaining to
say why there were MANY "standards" and "conventions" in use, to read
write and format 8-inch disks, during the 1970's and 80's when they
were in use on CP/M systems. You've gone on to post about other
standards, yourself. I think talking about "standards", as though
there was one way to support 8-inch floppy drives, is confusing. This
may not be your intention, Glen.
It's funny to have to argue the point, as it was an obvious issue in
the era, and anyone who worked with such systems back then ALWAYS had
to be aware of that. Any transfer of data between two systems by
diskette involved figuring out the format and schemes in use. Everyone
had DU, or programs by Sybex or others, to do "disk conversions" or
"disk analysis". Much the same occurred in the 5.25-inch drive era.
The greatest difficulty was transferring between "hard sectored"
systems and "soft sectored" systems. I think some of the work Mike S
is referring to, is about systems with hard sectored floppy
controllers - even less "standard" although common enough then and
still available now.
Anyway, it's more confusing to argue about the lack of standards and
to describe a decade's plus history of floppy disk controllers. It's
simpler to say there were many standards, which is to say that amounts
to NO standards. For specific systems, it's all in the details - which
"standards" those systems IN PARTICULAR followed. Glen, at that point
your clear knowledge of methods and designs is very useful.
This goes to the other point, raised by Barry Watzman. He seemed to
suggest that if you could get some kind of diskette scheme working on
a system, in some format however unique, that's fine. I disagree, for
the same circumstances as I've mentioned above. How do you read your
"unique" disks on someone else's system? And you *will*, or someone
else will, someday. How many posts have there been in comp.os.cpm
which amount to "...so I have these disks from X years ago, that I
can't read. Please help me"? Again, a common enough issue in the 80's
and later, still an issue today in this newsgroup.
A technical reason for some kind of "standards", is to avoid writing
data under marginal magnetic media conditions. To be brief about it:
if you cram too many data bits onto a small circle of a track on a
diskette, they may not be reliable. Or they may be OK today, or not
tomorrow. Or one drive may read them, but another may not. So, any
scheme of drive conversion has to work within standards of disk data
density, bits per inch. There are people who claim they can "beat"
such standards - but why risk losing data? 'Nuff said?
In any event, general discussions miss a key point of this discussion
thread. By finding 3.5" drives which can rotate at 360RPM, the same
rotation rate as 8-inch drives, you can AVOID MANY PROBLEMS. Use high-
density diskettes, and the 8-inch data rates will be OK for that media
- briefly put. And, the old computer's hardware and software will not
"see" a drive which is rotating too "slowly", wont' see a track that
is too "long", an index pulse rate too slow, etc. etc. That means you
don't have to change controller hardware or software.
That's great!
General discussion of general issues and general "standards", just
obscures this point. In my opinion of course, but I think general
practices and experiences of the era, and the technical points already
discussed, support my opinion.
Herb Johnson
retrotechnology.com
glen herrmannsfeldt
> Glen, I was going to reply to your posts about "standards" and
> "conventions" and so forth, but it would take a lot of explaining to
> say why there were MANY "standards" and "conventions" in use, to read
> write and format 8-inch disks, during the 1970's and 80's when they
> were in use on CP/M systems.
Well, there are two separate questions. One is the low-level
format of the blocks on the disk. For that there are the IBM
SD and DD standards, used by the Intel and WD floppy controller
chips among others. Apple and Commodore used non-IBM formats.
DEC used some IBM and some non-IBM low level formats.
> You've gone on to post about other
> standards, yourself. I think talking about "standards", as though
> there was one way to support 8-inch floppy drives, is confusing. This
> may not be your intention, Glen.
Now, separate from the low level format is the high level,
or file system format. I include in that sector interleave,
sector size, sectors/track, tracks/cylinder, track ordering.
I do like the DEC system of using separate programs for low
level and high level formatting.
> It's funny to have to argue the point, as it was an obvious issue in
> the era, and anyone who worked with such systems back then ALWAYS had
> to be aware of that. Any transfer of data between two systems by
> diskette involved figuring out the format and schemes in use. Everyone
> had DU, or programs by Sybex or others, to do "disk conversions" or
> "disk analysis". Much the same occurred in the 5.25-inch drive era.
> The greatest difficulty was transferring between "hard sectored"
> systems and "soft sectored" systems. I think some of the work Mike S
> is referring to, is about systems with hard sectored floppy
> controllers - even less "standard" although common enough then and
> still available now.
As far as this discussion goes, 8 inch drives and disks are
pretty rare, and expensive when you find them. For every day
use, 3.5in disks and drives are much easier and cheaper to find.
> Anyway, it's more confusing to argue about the lack of standards and
> to describe a decade's plus history of floppy disk controllers. It's
> simpler to say there were many standards, which is to say that amounts
> to NO standards. For specific systems, it's all in the details - which
> "standards" those systems IN PARTICULAR followed. Glen, at that point
> your clear knowledge of methods and designs is very useful.
IBM did a pretty good job of documenting their standards.
Apple and Commodore did not.
> This goes to the other point, raised by Barry Watzman. He seemed to
> suggest that if you could get some kind of diskette scheme working on
> a system, in some format however unique, that's fine. I disagree, for
> the same circumstances as I've mentioned above. How do you read your
> "unique" disks on someone else's system? And you *will*, or someone
> else will, someday. How many posts have there been in comp.os.cpm
> which amount to "...so I have these disks from X years ago, that I
> can't read. Please help me"? Again, a common enough issue in the 80's
> and later, still an issue today in this newsgroup.
Yes, good point. To me, if you use 3.5in or 5.25in disks as
direct replacements for 8in disks, with a similar low level
and high level format, you are as close as you can be to the
old formats, yet the hardware to read/write will be easier to
find in the future.
> A technical reason for some kind of "standards", is to avoid writing
> data under marginal magnetic media conditions. To be brief about it:
> if you cram too many data bits onto a small circle of a track on a
> diskette, they may not be reliable. Or they may be OK today, or not
> tomorrow. Or one drive may read them, but another may not. So, any
> scheme of drive conversion has to work within standards of disk data
> density, bits per inch. There are people who claim they can "beat"
> such standards - but why risk losing data? 'Nuff said?
Well, marginal magnetics is one. The appropriate gap sizes
are needed to allow for speed tolerance on the drives. IBM
was pretty conservative, others shrunk the gaps to get more
data on the track. That is separate from marginal magnetics.
> In any event, general discussions miss a key point of this discussion
> thread. By finding 3.5" drives which can rotate at 360RPM, the same
> rotation rate as 8-inch drives, you can AVOID MANY PROBLEMS.
But it doesn't solve the most important problem, which is
availability of drives. There are enough 3.5in drives in
the world now that a reasonable number of them should be around
for a long time. That doesn't apply to 3.5in 360RPM drives.
(My 5.25in drives always had a speed adjustment and strobe
pattern to be viewed under 60Hz arc lamps. I don't know that
3.5in drives have that.)
> Use high-
> density diskettes, and the 8-inch data rates will be OK for that media
> - briefly put. And, the old computer's hardware and software will not
> "see" a drive which is rotating too "slowly", wont' see a track that
> is too "long", an index pulse rate too slow, etc. etc. That means you
> don't have to change controller hardware or software.
With 300 RPM drives, the only difference is the time between
index pulses. If that is filled with unused (by the high level
format) sectors or gaps the disk is low level compatible with
other hardware. The bit timing and gap lengths of a 360RPM
disk are likely way too far off to be read/written (especially
written) at 300RPM.
The index pulse is used on formatting to know when to start/stop
writing format data, and in read/write to time out access to
a sector that doesn't exist. I believe that 300RPM is a better
choice for 3.5in disks emulating 8in.
-- glen
marcus bennett
I am shocked to find that other purchasers of the Samsung SFD-321B
have had less favourable experiences as using them as substitute 8"
format drives. Seems not all Samsung SFD-321B are created equal :-)
Overall I think what one needs is
a) a trimode 3.5" drive, i.e. it can simulate 77 tracks, DSDD, 360 RPM
b) that we can buy NOS (new old stock)
c) does not vary with revision, i.e. we can just go and buy one
without having to play russian roulette
d) available in Europe or USA mail order without hassle!
If we are saying that
Mitsumi D359T3
Teac FD-235HG
always fit these requirements e.g. I can go for any revision of Teac
FD-235HG, then I'd be happy to go for that drive instead. Though from
my research (d) is not satisfied unless someone can send me a European
source of the Mitsi and Teac drives?
regards marcus
Axel Berger
>Now, separate from the low level format is the high level, or file
>system format.
One might agree with that, though the distinction was never made at the
time and the term "low level format" is an IBM invention. But this
>I include in that sector interleave, sector size, sectors/track,
>tracks/cylinder, track ordering.
All that is fixed when you've initialized the raw empty sectors and
tracks, you can't get any lower level than that.
Greegor
Thanks Herb that page is awesome!
Slightly OT in a CP/M discussion perhaps except
in the sense of format cross conversion and hardware
adaptations, but where in all of that would
the short lived mutant used at first on the
Apple Lisa fit in?
It was really a nightmare I recall, the
blanks had to come ONLY from Apple!
I got the impression that later models
of Lisa even dumped it and earlier ones
got converted. Nasty!
How about a mention of how cookie substrate
actually has a ""grain"" to it and humidity and
temperature can deform the substrate "cookie"
making it slightly oval shaped?
I remember that while there were (at one point)
dozens of diskette brand names there were
only like 3 makers of the oxide coated cookies.
Dysan's standout quality reputation was based on
their reprocessing (burnishing/polishing) and QC
before they put the cookie into the diskette carrier.
Roger Ivie
> Slightly OT in a CP/M discussion perhaps except
> in the sense of format cross conversion and hardware
> adaptations, but where in all of that would
> the short lived mutant used at first on the
> Apple Lisa fit in?
You mean the Twiggy drive? According to wikipedia
http://www.retrotechnology.com/herbs_stuff/drive.html#data
"FileWare drives used 62.5 tracks per inch rather than the standard 48
or 96 TPI, and used high flux density (comparable to the later IBM 1.2MB
format introduced with the PC/AT). This required custom high-density
media. The coercivity required is similar to that of the 1.2MB format,
so it is possible to modify the jacket of 1.2MB diskettes for use in a
FileWare drive."
> It was really a nightmare I recall, the
> blanks had to come ONLY from Apple!
> I got the impression that later models
> of Lisa even dumped it and earlier ones
> got converted. Nasty!
Again, from Wikipedia:
"In the field, the FileWare drives proved to be somewhat unreliable. In
early 1984, Apple introduced the Lisa 2, which used a single 3.5 inch
Sony floppy drive in place of the two FileWare drives of the original
Lisa. A free upgrade was offered to Lisa 1 owners. The Macintosh was
also introduced using the Sony floppy drive."
--
roger ivie
ri...@ridgenet.net
Greegor
> On 2009-04-27, Greegor <Greego...@gmail.com> wrote:
>
> > Slightly OT in a CP/M discussion perhaps except
> > in the sense of format cross conversion and hardware
> > adaptations, but where in all of that would
> > the short lived mutant used at first on the
> > Apple Lisa fit in?
>
>
> "FileWare drives used 62.5 tracks per inch rather than the standard 48
> or 96 TPI, and used high flux density (comparable to the later IBM 1.2MB
> format introduced with the PC/AT). This required custom high-density
> media. The coercivity required is similar to that of the 1.2MB format,
> so it is possible to modify the jacket of 1.2MB diskettes for use in a
> FileWare drive."
>
> > It was really a nightmare I recall, the
> > blanks had to come ONLY from Apple!
> > I got the impression that later models
> > of Lisa even dumped it and earlier ones
> > got converted. Nasty!
>
> Again, from Wikipedia:
>
> "In the field, the FileWare drives proved to be somewhat unreliable. In
> early 1984, Apple introduced the Lisa 2, which used a single 3.5 inch
> Sony floppy drive in place of the two FileWare drives of the original
> Lisa. A free upgrade was offered to Lisa 1 owners. The Macintosh was
> also introduced using the Sony floppy drive."
> --
> roger ivie
> ri...@ridgenet.net
Yeah, I think that's it.
Twiggy??