Making disk images of NorthStar CP/M disks with Catweasel - CP/M

This is a discussion on Making disk images of NorthStar CP/M disks with Catweasel - CP/M ; On Sun, 16 Sep 2007 01:02:40 GMT, Dave.Dunfield@use.techsupport.link.on.my.webs ite (Dave Dunfield) wrote: >>>Apparently, some recent 5.25" drives and especially 1.2MB HD drives >>>will only work with soft sector disks and refuse to pass the multiple >>>index pulses present with a ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 40 of 40

Thread: Making disk images of NorthStar CP/M disks with Catweasel

  1. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sun, 16 Sep 2007 01:02:40 GMT,
    Dave.Dunfield@use.techsupport.link.on.my.website (Dave Dunfield)
    wrote:

    >>>Apparently, some recent 5.25" drives and especially 1.2MB HD drives
    >>>will only work with soft sector disks and refuse to pass the multiple
    >>>index pulses present with a hard sector disk. I believe it confuses
    >>>the disk into thinking it has a much higher RPM than allowable and the
    >>>electronics package thinks it has failed or not initialized properly.

    >
    >>So your about to tell me that the FD55GFV that I'm using in a NS* with
    >>MDS controller isn't working? It may be a specific drive and model
    >>has a problem but you didn't specify. Please specify.

    >
    >Nice drive and it works "most everywhere" - but I'll back up Andrew
    >in that I've encountered at least one (I think it was HD) drive that
    >would not come ready with hard-sectored disks, even when spinning
    >at 300rpm. I've also had reports in supporting ImageDisk where people
    >had drives which would not work SD (the mainboard was known to
    >support SD and it worked with a different drive). Not all drives are
    >created equal (and it seems in the PC world, some are a lot less
    >"equal" than others).
    >
    >Dave


    I understand that. However the challenge was what is that one odd
    drive? I already know the ones that do work. It would help a lot of
    people if the ID of the odd HD (generally the HD drives tend to be
    strange) were known to avoid it. It's not helpful to say some don't
    or some general group don't.

    I did say that all of the 48 tpi drives I've tried (aka 360k) did work
    and any issues were purely a matter of fit. I did post a list of
    drives known to me by testing on a NS* Horizon. However,
    I did test all of the 96tpi drives I had that would spin at 300rpm
    and tehy all worked. That only means I didn't have the offending
    drive or didn't bother trying a drive that didn't spin at the right
    speed ( HD only, not jumper selectable).

    Actually finding drives that fit in the allowable depth in the NS*
    chassis is more limiting that any problems with hard/soft
    sectoring. Many drives are too long or can only be accomodated
    by moving power supply parts.

    Allison

  2. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sep 15, 6:48 pm, no.s...@no.uce.bellatlantic.net wrote:
    > On Fri, 14 Sep 2007 18:48:43 -0700, lync...@yahoo.com wrote:
    > >On Sep 14, 5:33 pm, MikeS wrote:
    > >> On Sep 13, 9:14 pm, d...@ihatespam.msu.edu (Dennis Boone) wrote:> > I have built a prototype disk imaging tool for making disk images for
    > >> > > use in the Dave Dunfield Northstar Horizon simulator (which also runs
    > >> > > CP/M) using a generic PC with a Catweasel. The tool is now working
    > >> > > but is not ready for release. If you are interested in testing or
    > >> > > developing please contact me. Here is a sample of output produced by
    > >> > > the tool of a vintage disk from 1985:

    >
    > >> > Andrew,

    >
    > >> > I trying to collect enough tools to image some Vector disks which are
    > >> > also hard-sectored. Do you have any idea how close your tool comes to
    > >> > being able to read arbitrary hard-sector floppies?

    >
    > >> > De

    >
    > >> ----------------
    > >> I also have a couple of Vector systems and various diskettes, although
    > >> I'm not concerned enough about archiving (other than making copies on
    > >> the Vector) to get a Catweasel at this point; the discussion interests
    > >> me though.

    >
    > >> On a related note, although I do have working hard-sector drives, I've
    > >> played with replacing them with ordinary 5 1/4 drives with no success.

    >
    > >> Maybe this is a dumb question, but is there any reason why a Panasonic
    > >> (for example) 360K drive can't work with a hard-sector controller, and
    > >> if so, why not? It's not just a jumper issue by any chance?

    >
    > >> mike

    >
    > >Mike

    >
    > >Well, its not that simple. Yes, it is possible that a regular drive,
    > >especially a relatively new and sophisticated one, is not compatible
    > >with a hard sector disk. I found this out the hard way while building
    > >my CW station during the initial CWNS testing. My drive worked fine
    > >with soft sector disks but I was having problems reading the hard
    > >sector disks. A friend told me to use my oscilloscope and see if the
    > >index/sector holes were appearing on the index pin on the floppy drive
    > >connector. Sure enough, the floppy drive was "eating" hard sector
    > >pulses.

    >
    > >Apparently, some recent 5.25" drives and especially 1.2MB HD drives
    > >will only work with soft sector disks and refuse to pass the multiple
    > >index pulses present with a hard sector disk. I believe it confuses
    > >the disk into thinking it has a much higher RPM than allowable and the
    > >electronics package thinks it has failed or not initialized properly.

    >
    > >The only way to fix this problem I know of is to actually use your
    > >oscilloscope and see if the index/sector pulses are showing up on the
    > >pin at the floppy disk connector (pin 8). If your pulses are not
    > >showing up you'll have to replace the drive with a different kind. I
    > >had this exact problem where the floppy drive worked fine with soft
    > >sector disks but refused to pass hard sector index pulses. I had to
    > >replace the drive with one that did pass the hard sector pulses.

    >
    > >I swear, I would have never suspected that for happening if I hadn't
    > >seen it with my own oscilloscope and two eyes.

    >
    > >Best of luck with you project.

    >
    > >Andrew Lynch

    >
    > So your about to tell me that the FD55GFV that I'm using in a NS* with
    > MDS controller isn't working? It may be a specific drive and model
    > has a problem but you didn't specify. Please specify.
    >


    Hi Allison,
    No, of course I am not asserting your drive does not work. :-)

    I do not know about your drive since I have not tested it. I only know
    about the drive I had in my CW station and I know for certain it did
    not work with hard sector disks but did with soft sector disks.
    Apparently, there are other models which behave similarily which
    others have found as well.

    > I'll be specific. I have tried most every 48tpi and many 96 tpi drive
    > ever used in a DEC system, PC and other CP/M soft sector systems
    > on my NS* MDS (both the SD and DD controller) with 100% success.
    > The only caveat is RX50 (two spindles and one head positioner)
    > was only useful as a single drive and many drives are too LONG
    > to fit in the NS* slots but will function.
    >
    > List includes but is limited to 5.25" as thats the only size I have
    > 10sector media for.
    >
    > Panasonic (all 5.25" 48tpi and 96tpi that are NOT 1.2MB only)
    > All TEAC FD55xxx series
    > Mitsumi 48tpi drives (360k)
    > Newtronics (other than one jumperless 1.2mb _only_ drive)
    > Toshiba (all)
    > All Shugarts.
    >
    > I may even have forgotton a few.
    >
    > Allison- Hide quoted text -
    >
    > - Show quoted text -


    I did not say *all* drives did not pass the hard sector index/sector
    holes but surely *some* drives do not. Specifically, the drive which
    caused me problems is a "Canon, Model 5501". It is a HD drive of
    fairly recent vintage and I think it was relabeled as an IBM drive if
    I recall. I verified it does not pass the hard sector index/sector
    pulses with an oscilloscope on pin 8 of the connector with multiple
    before and after trials. I used the same technique to find a DSDD
    drive which *did* pass the index/sector pulses. I do not recall what
    sort of DSDD drive the replacement is but it works great whatever
    brand it is.

    It must be a known phenomenon by someone since a friend told me to
    look for it. Sure enough, they were right and my jaw fell open when I
    saw it. I'll agree it is hard to believe but I have seen it and have
    one of the misbehaving drives right here.

    I do agree with you that I should have been more specific as to which
    model did not work. My intent was not to malign all floppy drives but
    to caution the original poster that it is possible some won't work.
    It might be useful to future persons reading this thread if anyone
    else has seen this phenomenon of floppy disk drives not passing hard
    sector index/sector holes to specifically list out the model so they
    know what ones to avoid. I will start:

    Does not work with hard sector media: Canon, Model 5501

    Maybe there are others as well. Just my luck I find the one drive
    which doesn't work on the station I am building to use that very
    function! :-)

    Thanks and have a nice day!

    Andrew Lynch


  3. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sat, 15 Sep 2007 19:27:25 -0700, lynchaj@yahoo.com wrote:

    >
    >I do agree with you that I should have been more specific as to which
    >model did not work. My intent was not to malign all floppy drives but
    >to caution the original poster that it is possible some won't work.
    >It might be useful to future persons reading this thread if anyone
    >else has seen this phenomenon of floppy disk drives not passing hard
    >sector index/sector holes to specifically list out the model so they
    >know what ones to avoid. I will start:
    >
    >Does not work with hard sector media: Canon, Model 5501
    >
    >Maybe there are others as well. Just my luck I find the one drive
    >which doesn't work on the station I am building to use that very
    >function! :-)
    >
    >Thanks and have a nice day!
    >
    >Andrew Lynch


    Thanks for the info. That it behaves that way doesn't surpise me as
    many items in the PC market are suitably odd. I'd bet there are
    others that may behave badly but that now a known one.

    FYI: I looked at the drive I didn't test... CANNON 5501, Mine didn't
    do standard DD rate only HD so it I actually wrote on it "For PC 1.2MB
    ONLY". That way I knew it was mostly useless for anything else.

    Allison

  4. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sep 15, 9:02 pm, Dave.Dunfi...@use.techsupport.link.on.my.website
    (Dave Dunfield) wrote:
    > >>Apparently, some recent 5.25" drives and especially 1.2MB HD drives
    > >>will only work with soft sector disks and refuse to pass the multiple
    > >>index pulses present with a hard sector disk. I believe it confuses
    > >>the disk into thinking it has a much higher RPM than allowable and the
    > >>electronics package thinks it has failed or not initialized properly.

    > >So your about to tell me that the FD55GFV that I'm using in a NS* with
    > >MDS controller isn't working? It may be a specific drive and model
    > >has a problem but you didn't specify. Please specify.

    >
    > Nice drive and it works "most everywhere" - but I'll back up Andrew
    > in that I've encountered at least one (I think it was HD) drive that
    > would not come ready with hard-sectored disks, even when spinning
    > at 300rpm. I've also had reports in supporting ImageDisk where people
    > had drives which would not work SD (the mainboard was known to
    > support SD and it worked with a different drive). Not all drives are
    > created equal (and it seems in the PC world, some are a lot less
    > "equal" than others).
    >
    > Dave
    >
    > --
    > dave06a@ Low-cost firmware development tools: www.dunfield.com
    > dunfield. Classic computer collection: www.classiccmp.org/dunfield
    > com Some stuff I have for sale: www.dunfield.com/sale


    Hi Dave,

    Speaking of Single Density disks and Catweasel for the NorthStar, I
    have been working more on my CWNS program and since it is reading my
    two "well behaved" SSDD test disks fairly reliably, I thought to move
    on to some of the "difficult" ones left over from my conversion
    project earlier this spring. Of course, putting the disk in with
    "bad" sectors really raised havoc with the CWNS program. After a few
    days of rewrites and debugging, now CWNS can read SSDD disks with all
    "good" and some "bad" sectors. It just skips the "bad" sectors and
    replaces them with 512 byte blocks of zeros in the NSI file.

    However, here is where things get interesting. I was looking at the
    "bad" sectors and noticed it was eight sectors all in a neat little
    row, five on one track and three on another directly adjacent. Sort
    of like a file would be if it were written to a disk. :-) So I tell
    the Catweasel and cw2dmk to do a "raw bit dump" of those two tracks
    and take a closer look at the "bad" sectors. The raw data look
    suspiciously like "good" sectors but the MFM decoder can't seem to
    make sense of them... you guessed it! The "bad" sectors are really
    FM single density sectors from files randomly mixed in. ARGH! Mixed
    density disks are the bane of disk imaging!

    So, I guess now the next step is to modify CWNS so it can read FM
    single density sectors from SSSD disks. Then when the MFM decoder
    chokes on a sector try again by sending the track to the FM decoder
    and see if it can extract the data from the "bad" sectors. Then the
    next question will be how do I write mixed density sectors into the
    NorthStar Imaging format (256 bytes data + 256 bytes of zero
    padding?). However, it is going to be a while before the single
    density sectors are decoding properly since it has taken me weeks to
    get the double density ones to finally cooperate.

    Thanks and have a great weekend!

    Andrew Lynch


  5. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sep 15, 10:40 pm, no.s...@no.uce.bellatlantic.net wrote:
    > On Sat, 15 Sep 2007 19:27:25 -0700, lync...@yahoo.com wrote:
    >
    > >I do agree with you that I should have been more specific as to which
    > >model did not work. My intent was not to malign all floppy drives but
    > >to caution the original poster that it is possible some won't work.
    > >It might be useful to future persons reading this thread if anyone
    > >else has seen this phenomenon of floppy disk drives not passing hard
    > >sector index/sector holes to specifically list out the model so they
    > >know what ones to avoid. I will start:

    >
    > >Does not work with hard sector media: Canon, Model 5501

    >
    > >Maybe there are others as well. Just my luck I find the one drive
    > >which doesn't work on the station I am building to use that very
    > >function! :-)

    >
    > >Thanks and have a nice day!

    >
    > >Andrew Lynch

    >
    > Thanks for the info. That it behaves that way doesn't surpise me as
    > many items in the PC market are suitably odd. I'd bet there are
    > others that may behave badly but that now a known one.
    >
    > FYI: I looked at the drive I didn't test... CANNON 5501, Mine didn't
    > do standard DD rate only HD so it I actually wrote on it "For PC 1.2MB
    > ONLY". That way I knew it was mostly useless for anything else.
    >
    > Allison- Hide quoted text -
    >
    > - Show quoted text -


    Hi Allison,
    You have one of these Canon 5501's too? Now that is ironic! I am
    thinking I am going to toss mine in the trash since the thing has
    already cost me more of my sparsely remaining hair than it is worth.

    Your idea of labeling it is an excellent one though. I would hate to
    have to rediscover this junky drive's peculiarities again!

    Have a great weekend and thanks for all your help!

    Andrew Lynch


  6. Re: Making disk images of NorthStar CP/M disks with Catweasel

    Dave Dunfield wrote:
    (snip)

    > I've also had reports in supporting ImageDisk where people
    > had drives which would not work SD (the mainboard was known to
    > support SD and it worked with a different drive). Not all drives are
    > created equal (and it seems in the PC world, some are a lot less
    > "equal" than others).


    That is pretty strange, even more than not supporting hard sector.

    The drive is supposed to indicate flux transitions to the controller.
    MFM is more sensitive to the exact timing of transitions, so it is
    reasonable that a drive might support SD but not DD.
    The minimum distance (time) between flux transitions is the same
    for FM and MFM, and the maximum distance for MFM is equal to the maximum
    for FM. MFM includes, in addition, an intermediate distance between
    transitions.

    My guess is that the controller was a little out of tolerance, and
    was somehow more sensitive to the timing of one drive.

    -- glen


  7. Re: Making disk images of NorthStar CP/M disks with Catweasel

    >> I've also had reports in supporting ImageDisk where people
    >> had drives which would not work SD (the mainboard was known to
    >> support SD and it worked with a different drive). Not all drives are
    >> created equal (and it seems in the PC world, some are a lot less
    >> "equal" than others).


    >That is pretty strange, even more than not supporting hard sector.


    >The drive is supposed to indicate flux transitions to the controller.
    >MFM is more sensitive to the exact timing of transitions, so it is
    >reasonable that a drive might support SD but not DD.
    >The minimum distance (time) between flux transitions is the same
    >for FM and MFM, and the maximum distance for MFM is equal to the maximum
    >for FM. MFM includes, in addition, an intermediate distance between
    >transitions.


    >My guess is that the controller was a little out of tolerance, and
    >was somehow more sensitive to the timing of one drive.


    I agree, it was pretty bizarre. Unfortunately it was a drive I didn't
    have (and I have quite a few), and the person with it was too far
    away or me to borrow it for a more complete evaluation. This
    also makes the incident second hand information, so apply
    the usual "grain of salt".

    I had him do speed tests (it was pretty close to being bang
    on), and also try it on a different known-to-support SD
    mainboard - Our final conclusion what that the drive just
    wouldn't work for SD - but I could not come up with any
    reasonable explaination as to why. I agree that it should
    have worked.

    Dave

    --
    dave06a@ Low-cost firmware development tools: www.dunfield.com
    dunfield. Classic computer collection: www.classiccmp.org/dunfield
    com Some stuff I have for sale: www.dunfield.com/sale


  8. Re: Making disk images of NorthStar CP/M disks with Catweasel

    >>>>Apparently, some recent 5.25" drives and especially 1.2MB HD drives
    >>>>will only work with soft sector disks and refuse to pass the multiple
    >>>>index pulses present with a hard sector disk.


    >>>So your about to tell me that the FD55GFV that I'm using in a NS* with
    >>>MDS controller isn't working?


    >>Nice drive and it works "most everywhere" - but I'll back up Andrew
    >>in that I've encountered at least one (I think it was HD) drive that
    >>would not come ready with hard-sectored disks,


    >I understand that. However the challenge was what is that one odd
    >drive? I already know the ones that do work.


    Sorry if I offended - Andrews assertion that (some drives don't work)
    and your response (so you are saying my Teac drive that I'm using
    doesn't work) seemed to imply that you disagreed.


    >It would help a lot of
    >people if the ID of the odd HD (generally the HD drives tend to be
    >strange) were known to avoid it. It's not helpful to say some don't
    >or some general group don't.


    Agreed - perhaps I need to add listings for drive to my mainboard/FDC
    registry.


    >I did say that all of the 48 tpi drives I've tried (aka 360k) did work
    >and any issues were purely a matter of fit. I did post a list of
    >drives known to me by testing on a NS* Horizon. However,
    >I did test all of the 96tpi drives I had that would spin at 300rpm
    >and tehy all worked. That only means I didn't have the offending
    >drive or didn't bother trying a drive that didn't spin at the right
    >speed ( HD only, not jumper selectable).


    I never use "random" HD drives for disk imaging - I've got a
    couple that I've tested and know to work in the various formats
    (I've also added front panel switches for 300/360 rpm) which
    I use in a pinch, but usually I try and use the "right" drive in
    terms of physical format. HD drives are the most problematic
    and should be avoided unless you are imaging HD disks, or
    have tested the drive to meet your requirements.

    I've got notes on my site about constructing a cable
    which lets you easily swap drives external to the PC. I've
    also got notes on the motor speed select switch.


    >Actually finding drives that fit in the allowable depth in the NS*
    >chassis is more limiting that any problems with hard/soft
    >sectoring. Many drives are too long or can only be accomodated
    >by moving power supply parts.


    The Horizon has loads of space - you should try fitting drives
    into the single-bay of my Vector 1+ ... At least the N* can accomodate
    two drives. I've yet to find 1/2 height drives that are short enough to
    fit into the Vector bay.

    Dave


    --
    dave06a@ Low-cost firmware development tools: www.dunfield.com
    dunfield. Classic computer collection: www.classiccmp.org/dunfield
    com Some stuff I have for sale: www.dunfield.com/sale


  9. Re: Making disk images of NorthStar CP/M disks with Catweasel

    >Hi Dave,

    Hi Andrew,

    >Speaking of Single Density disks and Catweasel for the NorthStar, I
    >have been working more on my CWNS program and since it is reading my
    >two "well behaved" SSDD test disks fairly reliably, I thought to move
    >on to some of the "difficult" ones left over from my conversion
    >project earlier this spring. Of course, putting the disk in with
    >"bad" sectors really raised havoc with the CWNS program. After a few
    >days of rewrites and debugging, now CWNS can read SSDD disks with all
    >"good" and some "bad" sectors. It just skips the "bad" sectors and
    >replaces them with 512 byte blocks of zeros in the NSI file.


    >However, here is where things get interesting. I was looking at the
    >"bad" sectors and noticed it was eight sectors all in a neat little
    >row, five on one track and three on another directly adjacent. Sort
    >of like a file would be if it were written to a disk. :-) So I tell
    >the Catweasel and cw2dmk to do a "raw bit dump" of those two tracks
    >and take a closer look at the "bad" sectors. The raw data look
    >suspiciously like "good" sectors but the MFM decoder can't seem to
    >make sense of them... you guessed it! The "bad" sectors are really
    >FM single density sectors from files randomly mixed in. ARGH! Mixed
    >density disks are the bane of disk imaging!


    Yes, NorthStar DOS supports saving individual files in single or
    double density, and it's possible to mix density "randomly" on
    a single disk.

    In practice I've not encountered this except for disks that I made
    to see that it does in fact work. You may not have noticed, or
    forgotten it, but this statement appears in my NST documentation:

    ----------------------------------------------------------------------------------------------------
    NST does NOT currently handle mixed density disks - the disk must
    be either all single-density or all double-density. If you have data to
    backup from mixed density disks, you will have to copy it to a uniform
    density disk first.
    ---------------------------------------------------------------------------------------------------

    I don't know why someone would record individual files in single-
    density on a double-density system, but it could be done., and no
    doubt people did (as you've just proven) - but I think this is fairly
    rare. It would be neat if you modified your decoder to recognize
    this and decode the disks, but then you are going to have to come
    up with a new image file format which has a density indicator on
    each sector. Currently my NSI images do not support mixed
    density.

    Unless you are really keen to do that, a simpler solution might
    be to convert the mixed density disks to all DD - just "IN" a
    blank DD disk, then copy all the files to it from the mixed
    density disk.

    Dave

    --
    dave06a@ Low-cost firmware development tools: www.dunfield.com
    dunfield. Classic computer collection: www.classiccmp.org/dunfield
    com Some stuff I have for sale: www.dunfield.com/sale


  10. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sat, 15 Sep 2007 19:51:39 -0700, lynchaj@yahoo.com wrote:

    >Hi Allison,
    >You have one of these Canon 5501's too? Now that is ironic! I am
    >thinking I am going to toss mine in the trash since the thing has
    >already cost me more of my sparsely remaining hair than it is worth.


    Actually mine is sitting in a pile for the aluminum scrapper Alone
    with a buch of bad SA400s and other dead (mostly mechanical)
    drives and HDs.

    >Your idea of labeling it is an excellent one though. I would hate to
    >have to rediscover this junky drive's peculiarities again!


    Same here. generally I have no use for 1.2mb format even for PCs
    as I never found it common nor reliable. So I have kept away from it.
    IF I need it I have the well known FD55GFR as I use them for my DEC
    systems as well as for QD (DSDD 80track) CP/M use.

    >Have a great weekend and thanks for all your help!


    Anytime. Lists of info out there are a good thing.

    Allison

  11. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sun, 16 Sep 2007 09:47:25 GMT,
    Dave.Dunfield@use.techsupport.link.on.my.website (Dave Dunfield)
    wrote:

    >>>>>Apparently, some recent 5.25" drives and especially 1.2MB HD drives
    >>>>>will only work with soft sector disks and refuse to pass the multiple
    >>>>>index pulses present with a hard sector disk.

    >
    >>>>So your about to tell me that the FD55GFV that I'm using in a NS* with
    >>>>MDS controller isn't working?

    >
    >>>Nice drive and it works "most everywhere" - but I'll back up Andrew
    >>>in that I've encountered at least one (I think it was HD) drive that
    >>>would not come ready with hard-sectored disks,

    >
    >>I understand that. However the challenge was what is that one odd
    >>drive? I already know the ones that do work.

    >
    >Sorry if I offended - Andrews assertion that (some drives don't work)
    >and your response (so you are saying my Teac drive that I'm using
    >doesn't work) seemed to imply that you disagreed.


    Sorry if it was a bit strong. I was trying to get to what drive is
    the nasty beast. It could save others a lot of pain.

    >>It would help a lot of
    >>people if the ID of the odd HD (generally the HD drives tend to be
    >>strange) were known to avoid it. It's not helpful to say some don't
    >>or some general group don't.

    >
    >Agreed - perhaps I need to add listings for drive to my mainboard/FDC
    >registry.


    It couldn't hurt. In the floppy world there were some wild card 8",
    no shortage of strange 5.25 and then we have the 3.x" drives.
    It would be handy to know what drives that were at least common
    to PCs are unusual in someway or limiting.

    The usual PC problem is the later model drives didn't have
    Drive select switch or jumper or were board level cuts or smt shorts
    needed to affect features. The 3.5" drives are most common with that
    affliction but some of the white box branded 5.25" drives are like
    that.

    Some of the 3.5" drives used in the PS2 series and the MCA bus
    IBMs are also non standard in pinout and feature jumpers.

    >>I did say that all of the 48 tpi drives I've tried (aka 360k) did work
    >>and any issues were purely a matter of fit. I did post a list of
    >>drives known to me by testing on a NS* Horizon. However,
    >>I did test all of the 96tpi drives I had that would spin at 300rpm
    >>and tehy all worked. That only means I didn't have the offending
    >>drive or didn't bother trying a drive that didn't spin at the right
    >>speed ( HD only, not jumper selectable).

    >
    >I never use "random" HD drives for disk imaging - I've got a
    >couple that I've tested and know to work in the various formats
    >(I've also added front panel switches for 300/360 rpm) which
    >I use in a pinch, but usually I try and use the "right" drive in
    >terms of physical format. HD drives are the most problematic
    >and should be avoided unless you are imaging HD disks, or
    >have tested the drive to meet your requirements.


    Same here, I have a set of FD55B, FD55GFR on the shelf for
    those time when I need a drive with known behavour. Actually that
    lost of drives was my "pile o drives" on the shelf all marked so
    I know what they do. They are kept seperate from systems that
    have their own drives.

    >I've got notes on my site about constructing a cable
    >which lets you easily swap drives external to the PC. I've
    >also got notes on the motor speed select switch.


    Good info.

    >
    >>Actually finding drives that fit in the allowable depth in the NS*
    >>chassis is more limiting that any problems with hard/soft
    >>sectoring. Many drives are too long or can only be accomodated
    >>by moving power supply parts.

    >
    >The Horizon has loads of space - you should try fitting drives
    >into the single-bay of my Vector 1+ ... At least the N* can accomodate
    >two drives. I've yet to find 1/2 height drives that are short enough to
    >fit into the Vector bay.


    The Vector is crramped.

    Actually the NS* has room for 4 half height but watch the depth as
    if they are too long the end of the drive shorts to the two rectifiers
    back there. Many drives are too long or can only go in certain
    ways like connector up or down to make clearance. On my NS*
    systems I had to move the rectifers a bit so that the HD used
    (Quantum D540) would not sit against and short it. The other it was
    a FD55BV that would just touch the recifiers. It was a simple matter
    of moving them a quarter of an inch. All that becuase the only
    pattern that was common at the time was the SA400 which was fairly
    short.

    Allison

  12. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sep 16, 6:00 am, Dave.Dunfi...@use.techsupport.link.on.my.website
    (Dave Dunfield) wrote:
    > >Hi Dave,

    >
    > Hi Andrew,
    >
    >
    >
    >
    >
    > >Speaking of Single Density disks and Catweasel for the NorthStar, I
    > >have been working more on my CWNS program and since it is reading my
    > >two "well behaved" SSDD test disks fairly reliably, I thought to move
    > >on to some of the "difficult" ones left over from my conversion
    > >project earlier this spring. Of course, putting the disk in with
    > >"bad" sectors really raised havoc with the CWNS program. After a few
    > >days of rewrites and debugging, now CWNS can read SSDD disks with all
    > >"good" and some "bad" sectors. It just skips the "bad" sectors and
    > >replaces them with 512 byte blocks of zeros in the NSI file.
    > >However, here is where things get interesting. I was looking at the
    > >"bad" sectors and noticed it was eight sectors all in a neat little
    > >row, five on one track and three on another directly adjacent. Sort
    > >of like a file would be if it were written to a disk. :-) So I tell
    > >the Catweasel and cw2dmk to do a "raw bit dump" of those two tracks
    > >and take a closer look at the "bad" sectors. The raw data look
    > >suspiciously like "good" sectors but the MFM decoder can't seem to
    > >make sense of them... you guessed it! The "bad" sectors are really
    > >FM single density sectors from files randomly mixed in. ARGH! Mixed
    > >density disks are the bane of disk imaging!

    >
    > Yes, NorthStar DOS supports saving individual files in single or
    > double density, and it's possible to mix density "randomly" on
    > a single disk.
    >
    > In practice I've not encountered this except for disks that I made
    > to see that it does in fact work. You may not have noticed, or
    > forgotten it, but this statement appears in my NST documentation:
    >
    > ---------------------------------------------------------------------------*-------------------------
    > NST does NOT currently handle mixed density disks - the disk must
    > be either all single-density or all double-density. If you have data to
    > backup from mixed density disks, you will have to copy it to a uniform
    > density disk first.
    > ---------------------------------------------------------------------------*------------------------
    >


    Hi Dave,

    Yes, I agree. NST is great for transferring NorthStar Horizon disk
    images from working hardware. I understand it accepts only all single
    density or all double density. It makes sense to me order to keep the
    NST stub small and reduce complexity. I have no problem with NST and
    use it a lot. Thanks for writing such an excellent tool!

    > I don't know why someone would record individual files in single-
    > density on a double-density system, but it could be done., and no
    > doubt people did (as you've just proven) - but I think this is fairly
    > rare. It would be neat if you modified your decoder to recognize
    > this and decode the disks, but then you are going to have to come
    > up with a new image file format which has a density indicator on
    > each sector. Currently my NSI images do not support mixed
    > density.
    >


    Neither do I. It makes no sense to me at all. Maybe they were copying
    a file from an old single density disk? All I know is I have many
    disks with this phenomenon. These disks are the ones from Dr. Allen
    Ashley and they have his NorthStar toolsets on them. He did a lot of
    NorthStar BASIC compiler work and things of that nature. What
    possessed him to write mixed density disks, we may never know. As
    long as you are using read NorthStar hardware and NS DOS though it
    doesn't really matter. Its only when trying to make images that it
    gets complicated!

    As for CWNS, I really do want to keep using the NSI format as is to
    retain compatibility with the existing NSI image library. My thinking
    is to just write the data from the single density sectors (256 bytes)
    into the double density format (512 bytes) as the NSI format expects.
    To make up for the difference in size, either write each single sector
    data twice or with zero padding of 256 bytes. The intent would be
    retain compatibility with the NSI format while still copying the
    single density sectors. True, the file would no longer be truly
    representative of the original but at least would be modified in a
    consistent method and recoverable. The question then becomes how will
    the NorthStar Horizon simulator handle the new NSI files. I guess we
    can cross that bridge when we get there. I still have to write the
    code so this is all just speculation at this point. :-)

    I still have to make the SD decoder logic work. I have some ideas but
    it is going to be a while. When I get an NSI image of a mixed density
    disk, I will send you a copy and we can talk more about it then when
    we have some real facts and actual disk images to analyze.

    > Unless you are really keen to do that, a simpler solution might
    > be to convert the mixed density disks to all DD - just "IN" a
    > blank DD disk, then copy all the files to it from the mixed
    > density disk.
    >


    Agree but that has also been part of the problem. As far as I know
    NorthStar DOS lacks any decent file management tools like the MS-DOS
    "Midnight Commander". Making file copies of disks using CF for disks
    with large numbers of files is really not practical. I think even NS
    DOS CD chokes on mixed density disks. One or two files here or there
    is not a problem but when you have stacks of these disks to convert I
    start looking for an automated solution.

    What would be great is a NS DOS command which combines features of CO
    and CD so that you can put the mixed density disk in one drive and
    write a new all double density disk on another drive which is all nice
    and compacted. However, I do not expect we will be seeing any new NS
    DOS versions coming out any time soon! :-)

    Thanks for your help Dave! I appreciate your insight.

    Andrew Lynch

    > Dave
    >
    > --
    > dave06a@ Low-cost firmware development tools: www.dunfield.com
    > dunfield. Classic computer collection: www.classiccmp.org/dunfield
    > com Some stuff I have for sale: www.dunfield.com/sale- Hide quoted text -
    >
    > - Show quoted text -




  13. Re: Making disk images of NorthStar CP/M disks with Catweasel

    > Is the Vector format Single Density (FM) or Double Density (MFM)?

    It's MFM. Vector made a point of going for large capacity, though I
    don't recall the exact numbers, and there were at least two different
    variants. Some of the machines used 100 tpi drives, instead of 96 tpi.

    De

  14. Re: Making disk images of NorthStar CP/M disks with Catweasel

    Dave Dunfield wrote:

    (snip)

    >>However, here is where things get interesting. I was looking at the
    >>"bad" sectors and noticed it was eight sectors all in a neat little
    >>row, five on one track and three on another directly adjacent. Sort
    >>of like a file would be if it were written to a disk. :-) So I tell
    >>the Catweasel and cw2dmk to do a "raw bit dump" of those two tracks
    >>and take a closer look at the "bad" sectors. The raw data look
    >>suspiciously like "good" sectors but the MFM decoder can't seem to
    >>make sense of them... you guessed it! The "bad" sectors are really
    >>FM single density sectors from files randomly mixed in. ARGH! Mixed
    >>density disks are the bane of disk imaging!


    I remember them from tapes in OS/360 days, too. If you add a new
    file to (the end of) a tape, the system reads the tape up to the
    point where the new file is to be added, and then starts writing
    at the specified density. If the drive can read the existing density
    it will properly read up to that point, and then write as it was
    asked to do. Such a files are only readable on drives that support
    both densities.

    > Yes, NorthStar DOS supports saving individual files in single or
    > double density, and it's possible to mix density "randomly" on
    > a single disk.


    I thought I remembered some system where the boot tracks are written
    as SD, the rest as DD.

    > In practice I've not encountered this except for disks that I made
    > to see that it does in fact work. You may not have noticed, or
    > forgotten it, but this statement appears in my NST documentation:


    The DD format that DEC uses for the RX02 drive uses FM headers on
    MFM data blocks. If you write a 256 byte sector, it should then take
    up the same amount of space as a 128 byte FM sector.

    -- glen


  15. Re: Making disk images of NorthStar CP/M disks with Catweasel


    >I thought I remembered some system where the boot tracks are written
    >as SD, the rest as DD.


    This is quite common, especially in CP/M where the system tracks are not
    visible in the user volume space. What Andrew is talking about is mixed
    density within the user accessable volume space (ie: some user files are
    in single-density and some are in double-density) - not many systems
    support this, but NorthStar DOS does - the main reason it is able to
    support it is because it is so simple and has no drive allocation tracking.

    Dave


    --
    dave06a@ Low-cost firmware development tools: www.dunfield.com
    dunfield. Classic computer collection: www.classiccmp.org/dunfield
    com Some stuff I have for sale: www.dunfield.com/sale


  16. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sun, 16 Sep 2007 10:11:54 -0800, glen herrmannsfeldt
    wrote:

    >Dave Dunfield wrote:
    >
    >(snip)
    >
    >> In practice I've not encountered this except for disks that I made
    >> to see that it does in fact work. You may not have noticed, or
    >> forgotten it, but this statement appears in my NST documentation:

    >
    >The DD format that DEC uses for the RX02 drive uses FM headers on
    >MFM data blocks. If you write a 256 byte sector, it should then take
    >up the same amount of space as a 128 byte FM sector.
    >
    >-- glen


    RX02 is not MFM in DD mode it's M2fm, there is a differnce and it's
    why there are no "floppy chips" with RX02 capability.

    At least one other that did m2fm was the intel 8" DD drive system used
    with the MDS800. There are others.

    Allison



  17. Re: Making disk images of NorthStar CP/M disks with Catweasel

    >> I don't know why someone would record individual files in single-
    >> density on a double-density system, but it could be done., and no
    >> doubt people did (as you've just proven) - but I think this is fairly
    >> rare. It would be neat if you modified your decoder to recognize
    >> this and decode the disks, but then you are going to have to come
    >> up with a new image file format which has a density indicator on

    >As for CWNS, I really do want to keep using the NSI format as is to
    >retain compatibility with the existing NSI image library. My thinking
    >is to just write the data from the single density sectors (256 bytes)
    >into the double density format (512 bytes) as the NSI format expects.
    >To make up for the difference in size, either write each single sector
    >data twice or with zero padding of 256 bytes. The intent would be
    >retain compatibility with the NSI format while still copying the
    >single density sectors. True, the file would no longer be truly
    >representative of the original but at least would be modified in a
    >consistent method and recoverable. The question then becomes how will
    >the NorthStar Horizon simulator handle the new NSI files. I guess we
    >can cross that bridge when we get there. I still have to write the
    >code so this is all just speculation at this point. :-)


    I don't think you can make disks that will work correctly in mixed
    density with my current simulator. The problem is that the image file
    doesn't contain any information about the type of individual sectors.
    (It's a VERY simple format) The simulator simply looks at the size
    of the image you are mounting and makes a decision as to single
    or double density, which is how it represents all sectors.

    We I would have to do is to reserve 513 bytes for each sector,
    which would include up to 512 bytes of DD data, and a flag to
    indicate the density of that particular sector. When SD data is
    written, 256 bytes of data space would be wasted - this should
    work, but is it worth doing. I've not seen mixed density disks
    among the material I've taken in.



    >Agree but that has also been part of the problem. As far as I know
    >NorthStar DOS lacks any decent file management tools like the MS-DOS
    >"Midnight Commander". Making file copies of disks using CF for disks
    >with large numbers of files is really not practical. I think even NS
    >DOS CD chokes on mixed density disks. One or two files here or there
    >is not a problem but when you have stacks of these disks to convert I
    >start looking for an automated solution.


    >What would be great is a NS DOS command which combines features of CO
    >and CD so that you can put the mixed density disk in one drive and
    >write a new all double density disk on another drive which is all nice
    >and compacted. However, I do not expect we will be seeing any new NS
    >DOS versions coming out any time soon! :-)


    It would not be hard to write a command to do that, N* DOS is pretty
    simple. One potential issue us that file sizes would change (a 3 sector
    SD file (768 bytes) becomes a 2 sector.DD file (1024 bytes), but
    hopefully this wouldn't affect too much.

    Can you determine which files are single-density? Perhaps a particular
    bit of software was used that accessed the disk directly and only worked
    in single-density - this could prevent a simple conversion... ?

    Dave


    --
    dave06a@ Low-cost firmware development tools: www.dunfield.com
    dunfield. Classic computer collection: www.classiccmp.org/dunfield
    com Some stuff I have for sale: www.dunfield.com/sale


  18. Re: Making disk images of NorthStar CP/M disks with Catweasel

    On Sep 16, 5:22 pm, Dave.Dunfi...@use.techsupport.link.on.my.website
    (Dave Dunfield) wrote:
    > >> I don't know why someone would record individual files in single-
    > >> density on a double-density system, but it could be done., and no
    > >> doubt people did (as you've just proven) - but I think this is fairly
    > >> rare. It would be neat if you modified your decoder to recognize
    > >> this and decode the disks, but then you are going to have to come
    > >> up with a new image file format which has a density indicator on

    > >As for CWNS, I really do want to keep using the NSI format as is to
    > >retain compatibility with the existing NSI image library. My thinking
    > >is to just write the data from the single density sectors (256 bytes)
    > >into the double density format (512 bytes) as the NSI format expects.
    > >To make up for the difference in size, either write each single sector
    > >data twice or with zero padding of 256 bytes. The intent would be
    > >retain compatibility with the NSI format while still copying the
    > >single density sectors. True, the file would no longer be truly
    > >representative of the original but at least would be modified in a
    > >consistent method and recoverable. The question then becomes how will
    > >the NorthStar Horizon simulator handle the new NSI files. I guess we
    > >can cross that bridge when we get there. I still have to write the
    > >code so this is all just speculation at this point. :-)

    >
    > I don't think you can make disks that will work correctly in mixed
    > density with my current simulator. The problem is that the image file
    > doesn't contain any information about the type of individual sectors.
    > (It's a VERY simple format) The simulator simply looks at the size
    > of the image you are mounting and makes a decision as to single
    > or double density, which is how it represents all sectors.
    >


    Hi Dave,

    I have been experimenting with adding an SD decoder to CWNS. I think
    this problem may help solve itself. When you read SD sectors in
    cw2dmk, unless you tell it otherwise, it sends two bytes for every
    byte of FM data it decodes. In other words, even if it is a 256 byte
    SD sector, the resulting sector read will be 512 bytes because every
    byte repeats itself. So the SD sectors should fall inline with the
    rest of the NSI format size wise.

    We do have some sector density information as well. The specific
    files and sectors are marked with single or double density attributes
    in the NS DOS file table. So we know which sectors should be read as
    double density or which are single density. As for the specific
    implementation of the NorthStar Horizon simulator, I do not know
    enough about it to speculate if the approach is possible or
    practical. We will just have to see when I complete some code for
    mixed density disks. Maybe it could be hooked into the MDS-AD
    emulation code somehow? I see there is a control bit which selects
    format density.

    http://www.sol20.org/northstar/cbios.asm

    ; Case 2: Controller Orders
    ;
    ; +----+----+----+----+----+----+----+----+
    ; | DD | SS | DP | ST | DS |
    ; +----+----+----+----+----+----+----+----+
    ;
    ; Load 8-bit order register from low order 8 address bits.
    ;
    ; DD Controls density on write.
    ; DD=1 for double density.
    ; DD=0 for single density.



    > We I would have to do is to reserve 513 bytes for each sector,
    > which would include up to 512 bytes of DD data, and a flag to
    > indicate the density of that particular sector. When SD data is
    > written, 256 bytes of data space would be wasted - this should
    > work, but is it worth doing. I've not seen mixed density disks
    > among the material I've taken in.
    >


    Maybe just 512 bytes per sector of repeated single density data and
    the density attribute from file table would be enough. Possibly when
    NS DOS accesses the MDS-AD put a "handler routine" on that specific
    bit in the controller? This is all speculation but I need to finish
    some CWNS code so we can see some real images and do some experiments.

    > >Agree but that has also been part of the problem. As far as I know
    > >NorthStar DOS lacks any decent file management tools like the MS-DOS
    > >"Midnight Commander". Making file copies of disks using CF for disks
    > >with large numbers of files is really not practical. I think even NS
    > >DOS CD chokes on mixed density disks. One or two files here or there
    > >is not a problem but when you have stacks of these disks to convert I
    > >start looking for an automated solution.
    > >What would be great is a NS DOS command which combines features of CO
    > >and CD so that you can put the mixed density disk in one drive and
    > >write a new all double density disk on another drive which is all nice
    > >and compacted. However, I do not expect we will be seeing any new NS
    > >DOS versions coming out any time soon! :-)

    >
    > It would not be hard to write a command to do that, N* DOS is pretty
    > simple. One potential issue us that file sizes would change (a 3 sector
    > SD file (768 bytes) becomes a 2 sector.DD file (1024 bytes), but
    > hopefully this wouldn't affect too much.
    >
    > Can you determine which files are single-density? Perhaps a particular
    > bit of software was used that accessed the disk directly and only worked
    > in single-density - this could prevent a simple conversion... ?
    >


    Yes, in a NS DOS directory listing it tells which files are single or
    double density. It certainly is possible it was a custom application
    which is causing the single density sectors. I just do not know since
    I haven't really had a chance to dig into all the programs and data
    buried on this stack of disks.

    > Dave
    >
    > --
    > dave06a@ Low-cost firmware development tools: www.dunfield.com
    > dunfield. Classic computer collection: www.classiccmp.org/dunfield
    > com Some stuff I have for sale: www.dunfield.com/sale- Hide quoted text -
    >
    > - Show quoted text -


    Thanks Dave. I really appreciate your help and the NorthStar Horizon
    toolset. They are incredibly helpful for archiving this vintage
    computer data.

    Andrew Lynch


  19. Re: Making disk images of NorthStar CP/M disks with Catweasel

    no.spam@no.uce.bellatlantic.net wrote:
    (snip, I wrote)

    >>The DD format that DEC uses for the RX02 drive uses FM headers on
    >>MFM data blocks. If you write a 256 byte sector, it should then take
    >>up the same amount of space as a 128 byte FM sector.


    > RX02 is not MFM in DD mode it's M2fm, there is a differnce and it's
    > why there are no "floppy chips" with RX02 capability.


    I forgot about that. In any case, the FM sector headers won't work
    with the usual floppy chips.

    -- glen


  20. Re: Making disk images of NorthStar CP/M disks with Catweasel

    Just quick update on Catweasel NorthStar: I am back from a business
    trip and tonight I fixed the SD reading function. Now CWNS can read
    and make NSI compatible disk images of SSDD and SSSD NorthStar floppy
    disk images. Next on the agenda is to fix the sector re-reading
    code. Hopefully that won't take to long but it has to work properly
    before I can tackle the mixed DD and SD floppy disks.

    Thanks!

    Andrew Lynch


+ Reply to Thread
Page 2 of 2 FirstFirst 1 2