3 Aug
2005
3 Aug
'05
10:53 a.m.
> The idea is to make a small single-board computer
with a microcontroller,
> a WD2793 or similar floppy disk controller, enough memory to buffer a
> few tracks, and a high-speed serial port for communication with the PC.
> The board would have connectors for 5.25"/3" drives and 8" drives,
and
> would properly interface to all drive types.
Dave, why not just write software for a catweasel card?
I realize that the
catweasel's FPGA has been programmed with a bitmap made from closed-source RTL,
but is that a serious impedement?
Yes. - What happens when the catweasel becomes unavailable?
When the day comes that the catweasel is no longer
supported/sold, it would be
possible to at that point design another one.
Will it? Or will nobody bother as it's a tough job.
Will it be 100% compatible, or will we find that disks that were created with
this at-least partially proprietary system will be "lost" if the system cannot
be made to work?
People here have daydreamed reinventing the catweasel
with only TTL parts so
prevent lock-in; that is fine, but nobody has done anything about it (the
daydreaming is the easy part, by a factor of 1000).
This would be my point.
The catweasel is here and real, and at $100, it is
price reasonably.
The $100 cost, and the fact that you have to order and wait for it, means that a
significant number of people won't bother (I haven't) - something you can build
in a few hours in your workshop at near-$0 cost will get done by more people.
Everyone who has a vintage computer and doesn't use it, means one more set of
disks that are not preserved.
At that last statement I know one or more people are
going to say: rubbish, I
have the parts here in my junk bin that I could put together what I want for
free. Well, what you'd have would be a pile parts that took at least ten hours
of design time, and you'd still need to spend much more than that writing
software before coming up to speed with the catweasel.
The design time is a non-issue, what I am proposing is essentially the publishing
of a tested design. It would also include basic software, and be fully documented
to encourage other to add to the software base.
Is the cat software open-source, are the internal and storage formats fully
documented - one big thing about the work that I have been doing is to make
sure that people in the future can access the data from the images so that
they can find "other means" of regenerating the disk if necessary - even if
that regenerated disk is just a different format image for a simulator.
Sure, not everybody will
be able/willing to spend $100 for something like this, but that is far less than
the number of people who would be put off by having to build their own hardware.
On this point I disagree - the board I am thinking of would be very simple,
perhaps a dozen DIP chips - easy to build. I think a lot of the people on this
list would gladly take an evening or two to build one - also if it were to be
done, printed circuit boards, and even finished/tested boards could be made
available at low cost (like the cat), but unlike the cat, fully documented and
you can build it yourself if you like. It would not be "owned" or controlled by
anyone.
The one thing that I feel is missing from the catweasel
is a 50-pin header for
connecting to standard 8" floppies.
A grave omission for something thats supposed to archive non-PC-standard disks.
Another minor gripe is that each of the versions of
the catweasel (now four)
tries to be register compatible with previous versions while adding new
functionality. It has lead to some arcane programming requirements, which is
sad: a very thin API to hide version changes would have made things much simpler.
But would have tied the thing to one particular development environment (Here's
you VisualC++ library - have a nice day). The best solution is a fully open and
documented description of the hardware and lots of sample code.
Finally, is the WD2793 able to read northstar horizon
floppy disks? If not,
then a WD2793-based solution is not general enough.
As noted in the original message, it would be soft-sector only - the idea is to
make a simple/cheap alternative to the limitations of the PC controller...
Still deciding if it's worth it - I'll record your "NO" vote.
Regards,
Dave
--
dave04a (at) Dave Dunfield
dunfield (dot) Firmware development services & tools: www.dunfield.com
com Collector of vintage computing equipment:
http://www.parse.com/~ddunfield/museum/index.html
3 Aug
3 Aug
11:27 a.m.
On Wednesday 03 August 2005 09:53, Dave Dunfield wrote:
Yes. - What happens when the catweasel becomes unavailable?
What happens when (disk controller of choice) become unavailable? :)
The people making it aren't a "huge corporation", they're just regular
people, and I'm sure you could talk to them about such things.
Fine, buy one off ebay for less than $100, and get it within a week.
That's what I did. The catweasel is a great product for its price,
IMO.
I don't have enough parts in my junk bin to build a disk controller, and
there's a lot of people that don't have the skills to assemble
something like that.
I know a lot of people that wouldn't build ANY hardware. Even more
people (like myself) would rather spend the money for a catweasel than
try to find parts and build something that doesn't have the
functionality of a catweasel.... Do you think you'll get your WD chip
solution to read Apple ][ or C64 (1571-written) disks? I'm doubtful at
best. That's a huge volume of the disks I've got laying around and
eventually want to make images of.
If you're supposing the user can wire up a disk controller, why can't
they wire up a simple 34(SA400) to 50(SA800) pin adapter cable? Also,
you can buy an SA400/SA800 adaptor board that someone has made, which
properly controls the TG43 line. Unfortunately, I can't remember the
name of the board or who made it right now.
There's no reason a bit of C wrapper code has to be tied to any
particular 'development environment'.
I'd say that by ignoring (specifically) Apple ][ and C64, you're missing
a large percentage of possible software/users.
The idea is to make a small single-board computer with
a
microcontroller, a WD2793 or similar floppy disk controller,
enough memory to buffer a few tracks, and a high-speed serial port
for communication with the PC. The board would have connectors for
5.25"/3" drives and 8" drives, and would properly interface to all
drive types.
Dave, why not just write software for a catweasel card? I realize
that the catweasel's FPGA has been programmed with a bitmap made
from closed-source RTL, but is that a serious impedement? When the day
comes that the catweasel is no longer supported/sold,
it would be possible to at that point design another one.
Will it? Or will nobody bother as it's a tough job.
Will it be 100% compatible, or will we find that disks that were
created with this at-least partially proprietary system will be
"lost" if the system cannot be made to work? The catweasel
is here and real, and at $100, it is price
reasonably.
The $100 cost, and the fact that you have to order and wait for it,
means that a significant number of people won't bother (I haven't) -
something you can build in a few hours in your workshop at near-$0
cost will get done by more people. Is the cat software open-source, are the internal and
storage formats
fully documented - one big thing about the work that I have been
doing is to make sure that people in the future can access the data
from the images so that they can find "other means" of regenerating
the disk if necessary - even if that regenerated disk is just a
different format image for a simulator.
Yes, otherwise it wouldn't work with any formats. The whole point of
the catweasel is that 3rd parties write software for it to decode disk
formats.
Sure, not
everybody will
be able/willing to spend $100 for something like this, but that is
far less than the number of people who would be put off by having
to build their own hardware.
On this point I disagree - the board I am thinking of would be very
simple, perhaps a dozen DIP chips - easy to build. I think a lot of
the people on this list would gladly take an evening or two to build
one - also if it were to be done, printed circuit boards, and even
finished/tested boards could be made available at low cost (like the
cat), but unlike the cat, fully documented and you can build it
yourself if you like. It would not be "owned" or controlled by
anyone. The one thing
that I feel is missing from the catweasel is a 50-pin
header for connecting to standard 8" floppies.
A grave omission for something thats supposed to archive
non-PC-standard disks. Another minor
gripe is that each of the versions of the catweasel
(now four) tries to be register compatible with previous versions
while adding new functionality. It has lead to some arcane
programming requirements, which is sad: a very thin API to hide
version changes would have made things much simpler.
But would have tied the thing to one particular development
environment (Here's you VisualC++ library - have a nice day). The
best solution is a fully open and documented description of the
hardware and lots of sample code. Finally, is the
WD2793 able to read northstar horizon floppy disks?
If not, then a WD2793-based solution is not general enough.
As noted in the original message, it would be soft-sector only - the
idea is to make a simple/cheap alternative to the limitations of the
PC controller... Still deciding if it's worth it - I'll record
your "NO" vote.
Not to discourage you, but, I'd echo the sentiment above. I'd suggest
that if you do want to produce something, make something that can read
the disk on a flux-transition level, and store that sort of data.
Something nice and generic. That way, you can do all the processing in
software, and handle any format of disk that comes your way.
Pat
--
Purdue University ITAP/RCS --- http://www.itap.purdue.edu/rcs/
The Computer Refuge --- http://computer-refuge.org
10:55 p.m.
On Wed, 3 Aug 2005 10:27:01 -0500
Patrick Finnegan <pat at computer-refuge.org> wrote:
I know a lot of people that wouldn't build ANY
hardware. Even more
people (like myself) would rather spend the money for a catweasel than
try to find parts and build something that doesn't have the
functionality of a catweasel.... Do you think you'll get your WD chip
solution to read Apple ][ or C64 (1571-written) disks? I'm doubtful
at best. That's a huge volume of the disks I've got laying around and
eventually want to make images of.
I think the beauty of an open design is that you in particular don't
have to be capable or motivated to build the hardware yourself, but
there are multiple other people who are capable and (possibly) willing
to build one for you. A 'closed' design lives and dies at the whim of
the person who designed (and built) it.
A publised 'open' design constructed with TTL gates could, for example,
be constructed by someone who prefers CMOS or ECL (!) who is motivated
to adapt the design. (I know, as a packrat with a huge junkbox, that
when I see a logic circuit I want to build I always immediately begin
adapting it to the TTL I happen to have on hand.)
-----------
(stepping into a discussion where I am a latecomer, and don't mean to
step on feet or exceed my expertise)- Isn't a disk controller really
just a specialized logic state machine? When I studied the 1771 in tech
school I remember it as being fairly rudimentary stuff, i.e. nothing
that a general purpose Micro couldn't be coerced into doing. In fact,
the whole reason for the dedicated controllers of the past was that when
the whole system's 'brain' was just a Z80 or 6800 (or even just a 6502)
processor, a hardwired 'dedicated peripheral controller' design was the
only way to 'extend the machine' to do as complex a task as control a
floppy drive. These days, we're not so resource-starved that the
machine it's 'plugged into' only has a simple 8-bitter as it's sole CPU
(i.e. the Z80 chip isn't a $75 item that the whole rest of the machine
surrounds,) so it seems like if all the resources of a Z80, or a PIC, or
similar processor were thrown at the task, with some additional generic
(i.e TTL gates) hardware outboard, it could be a highly flexible
controller. (pardon my processor bias in the above. plug in your
favorite arch. as desired)
12:44 p.m.
Dave Dunfield wrote:
...
Most of the catweasel complexity (which is all buried in the FPGA) comes from
supporting C64 sound chips, timers, joystick ports, and other stuff. I get the
impression that for many of the people who use this, it isn't an archiving tool
-- they plug it into their amiga or PC, install the C64 software, and then just
use their machines like a native C64.
The complexity of the formats that people create for various disk images is
independent of the complexity of the hardware. Why hasn't somebody done an open
catweasel? I think part of the issue is that for the people who would be
interested in it, $100 isn't that off-putting. If the catweasel developer lost
interest (he has been supporting it through four versions for, how long? 10
years? so far), then there would be a void that makes it more attractive to come
up with another solution.
...
I still stand by my bald assertion that more people would be put off by having
to build it themselves than those who would be put off by the $100. Also, for
your $100 you get a professionally made board. Many of the home built cards,
even if only a dozen chips, might end up flakey and frustrating for the builders.
A hassle for sure, but certainly a lot less than wirewrapping an assortment of
TTL parts PLUS the 34/50 pin headers.
That is a strawman -- it is possible to write a C library that is tied to a
particular environment, but that is self-defeating. People write portable C
libraries all the time.
*some* of the limitations
>The idea is
to make a small single-board computer with a microcontroller,
>a WD2793 or similar floppy disk controller, enough memory to buffer a
>few tracks, and a high-speed serial port for communication with the PC.
>The board would have connectors for 5.25"/3" drives and 8" drives, and
>would properly interface to all drive types.
>Dave, why not just write software for a catweasel card? When the day
comes that the catweasel is no longer supported/sold, it would be
possible to at that point design another one.
Will it? Or will nobody bother as it's a tough job.
Will it be 100% compatible, or will we find that disks that were created with
this at-least partially proprietary system will be "lost" if the system cannot
be made to work? Is the cat software open-source, are the internal and
storage formats fully
documented - one big thing about the work that I have been doing is to make
sure that people in the future can access the data from the images so that
they can find "other means" of regenerating the disk if necessary - even if
that regenerated disk is just a different format image for a simulator.
The software is open source. The people (one actually) who sell don't really
develop the software. Other enthusiasts write the code to varying levels of
support. The format used for a particular format is a question independent of
the catweasel card.
Tim Mann has written code to read/write TRS-80 and RX-somethingorother format
disks to DMK format. The source is online. I've written code for PTDOS and the
source is available, although I doubt that anybody but me has ever used it (now
there is a test case that would give your disk archiver a challenge). I have
more than 100 compucolor disks that I intend to archive, and that is guaranteed
to not be compatable with your program either. I know the C64 linux drivers are
open source.
The only thing closed is the RTL used to generate the FPGA bitmap.
Sure, not
everybody will
be able/willing to spend $100 for something like this, but that is far less than
the number of people who would be put off by having to build their own hardware.
On this point I disagree - the board I am thinking of would be very simple,
perhaps a dozen DIP chips - easy to build. I think a lot of the people on this
list would gladly take an evening or two to build one - also if it were to be
done, printed circuit boards, and even finished/tested boards could be made
available at low cost (like the cat), but unlike the cat, fully documented and
you can build it yourself if you like. It would not be "owned" or controlled
by
anyone. The one thing
that I feel is missing from the catweasel is a 50-pin header for
connecting to standard 8" floppies.
A grave omission for something thats supposed to archive non-PC-standard disks. Another minor
gripe is that each of the versions of the catweasel (now four)
tries to be register compatible with previous versions while adding new
functionality. It has lead to some arcane programming requirements, which is
sad: a very thin API to hide version changes would have made things much simpler.
But would have tied the thing to one particular development environment (Here's
you VisualC++ library - have a nice day). The best solution is a fully open and
documented description of the hardware and lots of sample code. Finally, is the
WD2793 able to read northstar horizon floppy disks? If not,
then a WD2793-based solution is not general enough.
As noted in the original message, it would be soft-sector only - the idea is to
make a simple/cheap alternative to the limitations of the PC controller... Still deciding if it's worth it - I'll record
your "NO" vote.
Dave, I applaud you taking on the archiving problem. Teledisk has been used so
long simply because there is no good free alternative. Even with the
shortcomings of the 765, DiskImage is a good replacement for teledisk in that it
is a software-only solution for the most common controller on the planet. I
think the next step after that requires hardware, and so why stop at a partial
solution?
Personally, I hope to get around to getting my catweasel IV set up soon. If the
ImageDisk format is flexible enough, I'd be happy to write catweasel software
that can read/write the formats I care about to ImageDisk foramt. I strongly
suspect that it isn't flexible enough to represent the PTDOS format, which is
the format that I've already made up my own disk image foramt for since it is so
unusual.
1:13 p.m.
On Wed, 2005-08-03 at 11:44 -0500, Jim Battle wrote:
I still stand by my bald assertion that more people would be put off by having
to build it themselves than those who would be put off by the $100.
I'm not so sure - I'd expect that the majority of people who'd be
interested in this for archive use will have at least some electronics
skills. Different if you're talking about the emulator fans as that
sounds like a different scene most of the time.
That is a strawman -- it is possible to write a C library that is tied to a
particular environment, but that is self-defeating. People write portable C
libraries all the time.
That's another thing with an external device - you'd abstract out a lot
of the really low-level stuff such that the code on the host system is a
lot more portable across platforms...
Out of interest, what's actually changed between the catweasel
revisions? I mean, I thought they always buffered stuff at the track
level and did all the decoding / encoding work in software? So I'm
surprised the hardware isn't pretty simple and able to support pretty
much anything without several revisions...
cheers
Jules
The complexity of the formats that people create for
various disk images is
independent of the complexity of the hardware. Why hasn't somebody done an open
catweasel? I think part of the issue is that for the people who would be
interested in it, $100 isn't that off-putting.
I still think there's a lot of scope for something external and that
came with full schematics etc. though. I'm not even sure if I've got any
slots free inside my main PC, plus I'd have to take the board out every
time I needed it at the museum (or take the whole PC with me). Plus of
course it'd mean digging into the case every time I wanted to change
physical floppy drives. You're right in that it seems mean geared toward
the emulator crowd than it does toward archivists.
Last time I looked at the website for catweasel (such as it is) there
was very little information on there about the product, and nothing I
could see of that listed where drivers could be found for different
formats or what formats people were supporting for the card. Seems like
anyone buying one is very much on their own there.
Sure, not everybody will
be able/willing to spend $100 for something like this, but that is far less than
the number of people who would be put off by having to build their own hardware.
On this point I disagree - the board I am thinking of would be very simple,
perhaps a dozen DIP chips - easy to build. I think a lot of the people on this
list would gladly take an evening or two to build one - also if it were to be
done, printed circuit boards, and even finished/tested boards could be made
available at low cost (like the cat), but unlike the cat, fully documented and
you can build it yourself if you like. It would not be "owned" or controlled
by
anyone. Another minor gripe is that each of the versions of
the catweasel (now four)
tries to be register compatible with previous versions while adding new
functionality. It has lead to some arcane programming requirements, which is
sad: a very thin API to hide version changes would have made things much simpler.
But would have tied the thing to one particular development environment (Here's
you VisualC++ library - have a nice day). The best solution is a fully open and
documented description of the hardware and lots of sample code. 
6 Aug
6 Aug
10:45 a.m.
New subject: Another disk imaging project (long response)
...discussion of "universal floppy controller thingy"...
Sorry to join this thread late, but I think it would be a
fantastic idea/project to create this universal vintage
computer floppy controller. Of course, I WOULD think so,
since I've been working on part of the idea for the past
couple of years - the Semi-Virtual Diskette (SVD)
http://www.theSVD.com . Feel free to download the software
from the site to get an idea of what the control system
looks
like.
There seems to be at least four catagories/layers of
functionality that would be appropriate for this type of
device as detailed below (in no particular order). There's
enough overlap of these areas that a single device could be
built to fulfill all of the features.
--- Floppy Emulation
The idea behind floppy emulation is to provide a solid-state
mechanism that would allow vintage machines to
boot/read/write directly from/to diskette images without the
need for a physical floppy drive or physical floppies.
Given, particularly, that floppy media has a degrading
lifespan, this type of device makes it far easier to operate
vintage gear. Further, this type of device makes it
possible to backup/copy diskette images for archival and
later use.
For purposes of this discussion, it is assumed that the
device is NOT a full-fledged PC. So a PC would be a
necessary part of the whole solution, being used to archive
and upload/download the images to the device.
This device would probably not have an off-the-shelf floppy
controller, since they are normally meant to operate with a
particular encoding such as M/FM for the WD179x. Encodings
such as H8/H89, Apple 2, N*, etc. are NOT stock M/FM.
Instead, this device would probably have a micro-conroller
or processor that would "bit-bang" the same type of floppy
control.
There is a downside to this type of device. The vintage
machine is a necessary part of a "floppy copy" operation.
That is, since the device looks just like a floppy drive &
floppy to the vintage computer, the computer must run a
"copy" operation to make a backup copy of a floppy. If the
target diskette is copy- protected, it is quite likely that
the vintage machine would not be able to copy it.
The SVD that I've been working on for the last couple of
years is exactly this type of device. Though the current
downside to the SVD is that it must be specifically
programmed to support a paricular floppy type. The SVD II
will be generic, taking a soft description of the floppy
type. This will allow SVD with even previously unsupported
platforms. Further, it is locked into 300rpm with 4 micro-
second bit cells.
--- Stand-alone Floppy Reader/Writer
Almost the opposite of the floppy emulator, this device
would attach to a vintage floppy drive and extract/write
images to and from the floppies inserted in the drive. It
could be used to make duplicates of floppies either from
physical to image or from image to physical. Through
multiple simultaneous floppy drive connections, it could
make dynamic copies of floppies whereby the copy would be
BETTER than the original.
The nice thing about this device is that it wouldn't need
to be connected to the vintage computer to make the copy.
Further, it need not necessarily know about the encoding
mechanism for the floppy, being able to make copies based
upon the bit transitions seen on the floppy interface
lines.
Like the floppy emulator, this device would not be a
full-fledged PC, using a periphery PC to up/download images
and for configuration.
This device would be able to deal with issues such as
rotational speed variation and bad sectors probably more
effectively than a vintage computer trying to make a copy.
However, it would also need to understand and adapt to
different mechanisms such as hard-sectored, indexed, and
non-indexed floppies (as in Heathkit, WD179x, and Apple 2
floppies respectively). It would need to be configurable
to deal with different densities, rotational speeds, sides,
tracks, and other floppy physical characteristics. It
would also need to be able to generate an appropriate
write-precompensated signal. Fortunately, it could be made
smart enough to auto-configure itself for write-precomp.
--- External Floppy Controller
Both of the above devices represent relatively stand-alone
functionality, using the PC only for control, configuration,
and image storage. Another very interesting device would be
one that would act as a dynamic external "vintage floppy
controller" for a PC.
A specialized PC-based floppy driver could be created that
"speaks" to the external controller over a (pick your
favorite) parallel/RS232/USB link using sector read/write
commands...not unlike the way in which the Commodore 64 or
Model 100 interface to floppy drives. Through driver and
external device configuration software, the external device
will present a vintage floppy drive and floppy to the PC as
a simple storage device, allowing it to read/write to it.
In addition to read/write commands, other commands would be
available to format and download raw sectors and tracks.
Although this would be a very _interesting_ device, I feel
that it probably goes too far, and that the increased work
necessary to create it provides very little extra value.
The other two devices above allow you to work with "images"
of floppies, which IMHO is enough.
--- PC Software Control
All of these wonderful features require some form of robust
control, such as what can provided by your off-the-shelf PC.
Though it wouldn't be a required part of the "function" of
the device, it would be a necessary part of the control of
the device. Through the PC's extensive user interface as
well as its connectivity and storage, it would be able to
help configure and control the device as well as provide
periphery functions such as:
- floppy image analysis, verification, and translation
- file system manipulation, including insertion and
extraction of individual files
I've spent a lot of time on this area, as well, with the
SVD. The SVD software runs on both Windoze and Linux PCs,
allowing you to upload/download images to the SVD. Further,
it understands the filesystems of many of the different
vintage boxes include the H8/H89, TRS-80, and Apple,
allowing you to download individual files.
I'd be happy to collaborate on the creation of new devices in
this area. It should be possible to reuse SVD code, or to at
least steal some of the ideas.
Eric
6920
days inactive
6923
days old
5 comments
6 participants
participants (6)
-
chenmel＠earthlink.net -
dave04a＠dunfield.com -
eric＠rothfus.com -
frustum＠pacbell.net -
julesrichardsonuk＠yahoo.co.uk -
pat＠computer-refuge.org