24 Aug
2021
24 Aug
'21
4 p.m.
Back in the 2007 time frame, Andrew Lynch had written a utility to read
Vector Graphic hard-sectored diskettes on a Catweasel board. Called "CWVG",
does anyone have a copy of the program?
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
25 Aug
25 Aug
6:39 a.m.
I don't have that code, however a couple of points:
First of all, I think Vector Graphic actually used Micropolis drives
(and I suspect yours are because of the hard sectoring). I have
Micropolis drives on my Altair. Back in 2006 I imaged my floppies - by
reading the sectors on my Altair and transferring the data over to my PC
and saving an image file.
Note that at least for the 100 TPI Mod II model you'd have to use an
actual Micropolis drive or the tracks won't line up. I don't know if
that is also true of the Mod I drives which have a more standard tracks
per inch.
Back in 2006 I modified some existing catweasel code I got from Jim
Battle to handle Data General hard sector floppies, and if I remember
correctly - this was all back in 2006 - also changed the code so that
more stuff was in tables. You can find a link to download the code at:
https://www.computercollection.net/index.php/specialized-interfaces/
It would require source code changes to read Micropolis floppies directly.
Another option might be the open-source fluxengine, which has
dead-simple hardware that plugs into USB, and can also process catweasel
cell capture files. (Greaseweazle can't handle hard sector floppies yet)
http://cowlark.com/fluxengine/index.html
http://cowlark.com/fluxengine/doc/disk-micropolis.html
I am actually looking to re-image my Micropolis floppies (single sided,
both Mod I and Mod II) to confirm my existing images, and have the
requisite hardware on order from DigiKey which should arrive this week.
Fluxengine already has support for hard sectoring, so I will give that
a try and report back in a few days.
JRJ
On 8/24/2021 6:00 PM, Mike Loewen via cctech wrote:
?? Back in the 2007 time frame, Andrew Lynch had written a utility to
read Vector Graphic hard-sectored diskettes on a Catweasel board. Called
"CWVG", does anyone have a copy of the program?
Mike Loewen??????????????? mloewen at cpumagic.scol.pa.us
Old Technology??????????????? http://q7.neurotica.com/Oldtech/
1:57 p.m.
On Wed, 25 Aug 2021, Jay Jaeger via cctalk wrote:
I don't have that code, however a couple of
points:
First of all, I think Vector Graphic actually used Micropolis drives (and I
suspect yours are because of the hard sectoring). I have Micropolis drives
on my Altair. Back in 2006 I imaged my floppies - by reading the sectors on
my Altair and transferring the data over to my PC and saving an image file.
Note that at least for the 100 TPI Mod II model you'd have to use an actual
Micropolis drive or the tracks won't line up. I don't know if that is also
true of the Mod I drives which have a more standard tracks per inch.
Important points!
Yes, Vector Graphics was 100tpi. That was not the only unusual aspect to
their format.
Yes, Micropolis ALSO made 48tpi single sided drives. I don't remember
whether they were 35 track or 40 track; I used one for years on a TRS80
(which wass 35 track); it was an exceptionally reliable, albeit heavy and
slow-steppping drive.
For 100tpi, there also existed Tandon TM100-4M drives. The M stood for
"Micropolis". One of the TM100-4M drives that I had did not have the
'M'
on the label.
--
Grumpy Ol' Fred cisin at xenosoft.com
2:37 p.m.
On 8/25/21 1:57 PM, Fred Cisin via cctalk wrote:
For 100tpi, there also existed Tandon TM100-4M
drives.? The M stood for
"Micropolis".?? One of the TM100-4M drives that I had did not have the
'M' on the label.
The early Micropolis 48 tpi drives (1015 Mod I) were spec-ed at 35 tracks.
The 100 TPI models (mod II) were 77 tracks.
Later drives included both 100 TPI and 96 TPI (e.g. 1115-5 and -6). For
a 5.25" drive, a beast of one. No wonder they quit making floppy drives.
Tandon was pretty sloppy with their labeling. I've got a TM-100-4M
that's identified as such on the body label, but the door latch handle
has a "96 tpi" sticker. Whatthehell, a bezel's a bezel, right?
--Chuck
3:31 p.m.
> For 100tpi, there also existed Tandon TM100-4M
drives.? The M stood for
> "Micropolis".?? One of the TM100-4M drives that I had did not have the
> 'M' on the label.
On Wed, 25 Aug 2021, Chuck Guzis via cctalk wrote:
The early Micropolis 48 tpi drives (1015 Mod I) were
spec-ed at 35 tracks.
That's what I thought, but I wasn't sure. Thank you.
The 100 TPI models (mod II) were 77 tracks.
Yes. I had forgotten that. Thank you.
Later drives included both 100 TPI and 96 TPI (e.g.
1115-5 and -6).
For a 5.25" drive, a beast of one. No wonder they quit making floppy
drives.
My 35 track Micropolis drive outlasted every other drive (shugart SA400
had radial alignment issues, others had speed control, etc.) and the
machine itself. The lead screw positioner was kinda slow, but so was the
spiral "record" of the SA400.
Tandon was pretty sloppy with their labeling.
I've got a TM-100-4M
that's identified as such on the body label, but the door latch handle
has a "96 tpi" sticker. Whatthehell, a bezel's a bezel, right?
So, that's TWO mislabels within a tiny sample.
The 100 TPI models (mod II) were 77 tracks.
The
first "1.2M" drive that I got was from a computer swap and was among
projects and prototypes. It looked like a production model, with no
obvious board re-work nor jumper wires. It had a 50 pin header, instead
of a 34 pin card edge connector! I tested it, and it did go to 80 tracks,
but I heard that when IBM was developing the 1.2M drive, some at
Microsoft wondered whether IBM was re-introducing 8 inch drives!
4:23 p.m.
On 8/25/21 3:31 PM, Fred Cisin via cctalk wrote:
> Later drives included both 100 TPI and 96 TPI
(e.g. 1115-5 and -6).
> For a 5.25" drive, a beast of one.? No wonder they quit making floppy
> drives.
The 1115/16 series was Micropolis' last gasp at floppy drives, I think.
Not the steel plate construction of the earlier drives, but cast
(zamac, I think) both body and diskette carrier. The worm-drive
positioner pivots with the diskette carrier, just like the originals. I
suspect that this contributes much to the accuracy of the track positioning.
The 1115 uses a MOS MCU (6502-based perhaps?)to implement buffered
seeking, which ameliorates the slower positioning times and allows for
some measure of "ballistic seek".
We had a problem finding suitable alignment disks for 100 tpi, so a
couple of the engineers got a 3/4" aluminum plate and an external
stepper coupled through a 100:1 right-angle precision drive coupled to
the floppy drive. It took about 3 minutes to step from one end of the
disk to the other, but the setup worked.
Early on, with the 100 tpi units (about 1976-77, we had a problem with
the disk clamping. Everyone was new to the 5.25" drive with its DC
motor, so naturally, one turned on the motor only when needed (much like
the later IBM PCs). What happened when the spindle was stopped was that
the clamping was off-center frequently. The hub area of the floppy
would get wrinkled and then it became a craps shoot to get the mangled
floppy to read correctly.
Our primary vendor was Dysan for floppies. Initially, they introduced
a kit that enabled one to attach an adhesive-backed reinforcing ring.
It seemed to work, and Dysan began supplying floppies with the ring
installed. Eventually, Micropolis determined that adding a microswitch
to the drive that detected when the drive door was being closed, turning
on the motor for a couple of seconds allowed for accurate clamping and
centering of the diskette on the clamping "cone".
By the time the 1.2M 5.25" drives came out, everyone was doing this, so
rings on high-density floppies were unnecessary. Although there are
some exceptions (obviously), a good rule of thumb is if the ring is
present, the disk is DD; if absent, HD.
--Chuck
26 Aug
26 Aug
12:46 p.m.
On 8/25/2021 3:57 PM, Fred Cisin via cctalk wrote:
On Wed, 25 Aug 2021, Jay Jaeger via cctalk wrote:
35 track. The aforementioned "Mod I" models.
I don't have that code, however a couple of
points:
First of all, I think Vector Graphic actually used Micropolis drives
(and I suspect yours are because of the hard sectoring).? I have
Micropolis drives on my Altair.? Back in 2006 I imaged my floppies -
by reading the sectors on my Altair and transferring the data over to
my PC and saving an image file.
Note that at least for the 100 TPI Mod II model you'd have to use an
actual Micropolis drive or the tracks won't line up.? I don't know if
that is also true of the Mod I drives which have a more standard
tracks per inch.
Important points!
Yes, Vector Graphics was 100tpi.? That was not the only unusual aspect
to their format.
Yes, Micropolis ALSO made 48tpi single sided drives.? I don't remember
whether they were 35 track or 40 track; I used one for years on a TRS80
(which wass 35 track); it was an exceptionally reliable, albeit heavy
and slow-steppping drive.
For 100tpi, there also existed Tandon TM100-4M drives.? The M stood for
"Micropolis".?? One of the TM100-4M drives that I had did not have the
'M' on the label.
--
Grumpy Ol' Fred???????????? cisin at xenosoft.com
JRJ
25 Aug
25 Aug
8:49 a.m.
Back in the 2007 time frame, Andrew Lynch had written
a utility to read
Vector Graphic hard-sectored diskettes on a Catweasel board. Called
"CWVG", does anyone have a copy of the program?
The program doesn't appear to have made it onto vector-archive.org.
I'll add to Jay's comment about 100 tpi floppies: there was a Tandon
model (100-4M?) that did 100 tpi also. Of both the M and T models are
scarce.
De
3:58 p.m.
On Wed, 25 Aug 2021, Dennis Boone via cctalk wrote:
On Wed, 24 Aug 2021, Mike Loewen via cctalk wrote:
Indeed, the two Micropolis drives on my Vector 1 are the 1043-2 (master)
and 1023-II (slave). Both are 100tpi, single-sided. The S100 controller is the
Micropolis FDC. I haven't had time to find out if they work.
As a few of the signals on the 34-pin connector are different than the
Shugart layout, I'm considering making up a custom cable and connecting it to
my Catweasel MK4+. I have a utility from Andrew Lynch called "cwns", which
is a modified version of "cw2dmk" to read Northstar hard-sector (10 SPT)
diskettes. I might be able to modify that to read the VG 16 SPT diskettes, if
the 1043-2 will work with the Catweasel.
It's been pointed out that I could write something for the Vector 1 to read
diskettes and transfer the contents out the serial port. I'd rather attack it
Back in the 2007 time frame, Andrew Lynch had written a utility to read
Vector Graphic hard-sectored diskettes on a Catweasel board. Called
"CWVG", does anyone have a copy of the program?
The program doesn't appear to have made it onto vector-archive.org.
I'll add to Jay's comment about 100 tpi floppies: there was a Tandon
model (100-4M?) that did 100 tpi also. Of both the M and T models are
scarce. from the point of my media imaging system, and there is
a literal shoebox full
of diskettes to image.
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://q7.neurotica.com/Oldtech/
4:45 p.m.
On 8/25/21 3:58 PM, Mike Loewen via cctalk wrote:
On Wed, 24 Aug 2021, Mike Loewen via cctalk wrote:
Back in the 2007 time frame, Andrew Lynch had written a utility to read
Vector Graphic hard-sectored diskettes on a Catweasel board.? Called
"CWVG", does anyone have a copy of the program?
Some years ago, I received a batch of Vector 4 memorite floppies.
Using a catweasel, I did a conversion to RTF (Memorite has a lot of
features not suited to straight ASCII conversion; e.g. footnoting). It
was a multi-step process.
First was capturing the flux transition data using a catweasel and
writing to a file. Next was interpreting the data and converting to a
diskette image. Extraction of the files using 22Disk was next and
finally conversion of the files to RTF.
This was not Andrew's program, but one rolled from scratch here. I
doubt that it will suffice for most people.
--Chuck
26 Aug
26 Aug
12:51 p.m.
On 8/25/2021 5:58 PM, Mike Loewen via cctalk wrote:
On Wed, 25 Aug 2021, Dennis Boone via cctalk wrote:
?? As a few of the signals on the 34-pin connector
are different than
the Shugart layout, I'm considering making up a custom cable and
connecting it to my Catweasel MK4+. I have a utility from Andrew Lynch
called "cwns", which is a modified version of "cw2dmk" to read
Northstar
hard-sector (10 SPT) diskettes. I might be able to modify that to read
the VG 16 SPT diskettes, if the 1043-2 will work with the Catweasel.
It will not - but it has nothing to do with the cabling. I already
tried with my catweasel. I was able to read flux, but between the hard
sectoring and (possibly) different meta markers for the floppy sectors,
the .scp file is useless. I just got my Cypress board for a fluxengine
in the mail yesterday, but haven't set it up yet - maybe tomorrow. The
website for fluxengine indicates it ought to work.
?? It's been pointed out that I could write
something for the Vector 1
to read diskettes and transfer the contents out the serial port. I'd
rather attack it from the point of my media imaging system, and there is
a literal shoebox full of diskettes to image.
That was what I did back years ago.
Mike Loewen??????????????? mloewen at cpumagic.scol.pa.us
Old Technology??????????????? http://q7.neurotica.com/Oldtech/
JRJ
27 Aug
27 Aug
8:08 p.m.
On 8/26/2021 2:51 PM, Jay Jaeger via cctech wrote:
JRJ
On 8/25/2021 5:58 PM, Mike Loewen via cctalk wrote:
The fluxengine almost works with my Micropolis drives - but there are
some problems.
1) The drive select pins are different. I am having *some* luck access
one driver of my daisy-chained pair, but not the other. The fluxengine
setup and/or the drive also seem to get confused when I try to access
the other drive. I submitted an issue on github for more flexible drive
selection / motor control capability. One could work around this with
suitable cabing / jumpering.
2) So far I have not been able to read an entire Mod II disk
successfully - lots of good sectors on two, but not all without errors
on either one. But I have not tried cleaning the heads on my drives, or
trying the Mod I with different select jumpering.
3) The Micropolis drives are slow stepping, so I added multiple commands
to seek to cylinder 0 to my script - not sure if that is actually
working right.
On Wed, 25 Aug 2021, Dennis Boone via cctalk
wrote:
??? As a few of the signals on the 34-pin
connector are different than
the Shugart layout, I'm considering making up a custom cable and
connecting it to my Catweasel MK4+. I have a utility from Andrew Lynch
called "cwns", which is a modified version of "cw2dmk" to read
Northstar hard-sector (10 SPT) diskettes. I might be able to modify
that to read the VG 16 SPT diskettes, if the 1043-2 will work with the
Catweasel.
>
It will not - but it has nothing to do with the cabling.? I already
tried with my catweasel.? I was able to read flux, but between the hard
sectoring and (possibly) different meta markers for the floppy sectors,
the .scp file is useless.? I just got my Cypress board for a fluxengine
in the mail yesterday, but haven't set it up yet - maybe tomorrow.? The
website for fluxengine indicates it ought to work.
Mike Loewen??????????????? mloewen at cpumagic.scol.pa.us
Old Technology??????????????? http://q7.neurotica.com/Oldtech/
JRJ
28 Aug
28 Aug
12:12 p.m.
On 8/27/2021 10:08 PM, Jay Jaeger via cctalk wrote:
JRJ
JRJ
On 8/26/2021 2:51 PM, Jay Jaeger via cctech wrote:
Another update. Without messing with the cable pinouts the fluxengine
will only support a Micropolis drive strapped for DS2 (as fluxengine
drive 0).
I have successfully read a couple of Mod I disks (143KB). But some have
not read very well. Sometimes it goes "off the rails" and says there
are 77 tracks, even when the input config specifies 35. My guess is
that at the least the decoder could probably need some improvement - a
task I am not up for, at present.
On 8/25/2021 5:58 PM, Mike Loewen via cctalk wrote:
The fluxengine almost works with my Micropolis drives - but there are
some problems.
1)? The drive select pins are different.? I am having *some* luck access
one driver of my daisy-chained pair, but not the other.? The fluxengine
setup and/or the drive also seem to get confused when I try to access
the other drive.? I submitted an issue on github for more flexible drive
selection / motor control capability.? One could work around this with
suitable cabing / jumpering.
2)? So far I have not been able to read an entire Mod II disk
successfully - lots of good sectors on two, but not all without errors
on either one.? But I have not tried cleaning the heads on my drives, or
trying the Mod I with different select jumpering.
3) The Micropolis drives are slow stepping, so I added multiple commands
to seek to cylinder 0 to my script - not sure if that is actually
working right.
On Wed, 25 Aug 2021, Dennis Boone via cctalk
wrote:
??? As a few of the signals on the 34-pin
connector are different
than the Shugart layout, I'm considering making up a custom cable and
connecting it to my Catweasel MK4+. I have a utility from Andrew
Lynch called "cwns", which is a modified version of "cw2dmk" to read
Northstar hard-sector (10 SPT) diskettes. I might be able to modify
that to read the VG 16 SPT diskettes, if the 1043-2 will work with
the Catweasel.
>
It will not - but it has nothing to do with the cabling.? I already
tried with my catweasel.? I was able to read flux, but between the
hard sectoring and (possibly) different meta markers for the floppy
sectors, the .scp file is useless.? I just got my Cypress board for a
fluxengine in the mail yesterday, but haven't set it up yet - maybe
tomorrow.? The website for fluxengine indicates it ought to work.
Mike Loewen??????????????? mloewen at cpumagic.scol.pa.us
Old Technology??????????????? http://q7.neurotica.com/Oldtech/
JRJ 
12:27 p.m.
On 8/28/21 12:12 PM, Jay Jaeger via cctalk wrote:
I have successfully read a couple of Mod I disks
(143KB).? But some have not read very well.
I played with one when they came out. The software wasn't ready for prime time.
I had hoped it had gotten better since then. Even something basic like ^C out of a
transfer
locked the board up.
1:06 p.m.
On 8/28/21 12:27 PM, Al Kossow via cctalk wrote:
On 8/28/21 12:12 PM, Jay Jaeger via cctalk wrote:
I believe that I used an MPI 100 tpi drive when I processed the Vector 4
images. Came off without a hitch using a Catweasel.
Raw decoded image was about 600K.
One of the first jobs was to decode the memorite disk itself.
Here's the disk image dump, if you're curious.
https://app.box.com/s/mw9pk2xmrcr8ageu2ueuutvk4b6iw721
--Chuck
I have successfully read a couple of Mod I disks
(143KB).? But some
have not read very well.
I played with one when they came out. The software wasn't ready for
prime time.
I had hoped it had gotten better since then. Even something basic like
^C out of a transfer
locked the board up.
3:01 p.m.
On 8/28/2021 3:06 PM, Chuck Guzis via cctalk wrote:
On 8/28/21 12:27 PM, Al Kossow via cctalk wrote:
Back in 2005 I imaged both my Mod I and Mod II floppies using my Altair
and a program I wrote way way wayyyyyy back when for transferring floppy
images called "XMT". But I like to image floppies two different ways,
and the Greaseweazle cannot (yet) deal with hard sectoring.
Of course, I could also use my catweasel with these drives directly -
but kind of a pain to do.
JRJ
On 8/28/21 12:12 PM, Jay Jaeger via cctalk
wrote:
I believe that I used an MPI 100 tpi drive when I processed the Vector 4
images. Came off without a hitch using a Catweasel.
Raw decoded image was about 600K.
One of the first jobs was to decode the memorite disk itself.
Here's the disk image dump, if you're curious.
https://app.box.com/s/mw9pk2xmrcr8ageu2ueuutvk4b6iw721
--Chuck
I have successfully read a couple of Mod I disks
(143KB).? But some
have not read very well.
I played with one when they came out. The software wasn't ready for
prime time.
I had hoped it had gotten better since then. Even something basic like
^C out of a transfer
locked the board up.
3:14 p.m.
On 8/28/21 3:01 PM, Jay Jaeger via cctalk wrote:
Back in 2005 I imaged both my Mod I and Mod II floppies using my Altair
and a program I wrote way way wayyyyyy back when for transferring floppy
images called "XMT".? But I like to image floppies two different ways,
and the Greaseweazle cannot (yet) deal with hard sectoring.
It can't---that's a surprise to me. I used a MK III by the way--the
image is from 2007.
Nowadays, I use an STM32F4 microcontroller. MOre than capable enough
for anything I can throw at it.
--Chuck
7:04 p.m.
On 8/28/2021 5:14 PM, Chuck Guzis via cctalk wrote:
On 8/28/21 3:01 PM, Jay Jaeger via cctalk wrote:
That is correct. It can't. I expect it is just a matter of software
and firmware. You can get a .scp file out of it by telling it to read,
oh, 60 revolutions, but the .scp file is pretty useless.
JRJ
Back in 2005 I imaged both my Mod I and Mod II floppies using my Altair
and a program I wrote way way wayyyyyy back when for transferring floppy
images called "XMT".? But I like to image floppies two different ways,
and the Greaseweazle cannot (yet) deal with hard sectoring.
It can't---that's a surprise to me. I used a MK III by the way--the
image is from 2007.
2:59 p.m.
On 8/28/2021 2:27 PM, Al Kossow via cctalk wrote:
On 8/28/21 12:12 PM, Jay Jaeger via cctalk wrote:
And, still does, unfortunately. I opened a ticket for that. ;)
I have successfully read a couple of Mod I disks
(143KB).? But some
have not read very well.
I played with one when they came out. The software wasn't ready for
prime time.
I had hoped it had gotten better since then. Even something basic like
^C out of a transfer
locked the board up.
969
days inactive
974
days old
18 comments
6 participants
participants (6)
-
aek＠bitsavers.org -
cclist＠sydex.com -
cisin＠xenosoft.com -
cube1＠charter.net -
drb＠msu.edu -
mloewen＠cpumagic.scol.pa.us