macintosh disks : which format for archiving ?

What is the recommended format for archiving Macintosh disks (400k, 800k) ? I hesitate between using my SCP card or using the standard DiskCopy 4.2 on a Macintosh.
I did archive several ones using DiskCopy 4.2 and then compressing the images in one Stuffit 5.5 container. The reasoning behind is that using a variable speed floppy disk drive on a Mac to create images would be more reliable than using a standard floppy disk drive and SCP. The DiskCopy / Stuffit combination allows to recreate the physical disks on a physical machine.


  • If the disk is not copy protected, then either way will "work".

    Just be sure to write protect the disks, as Macs love to write desktop information and last used dates back to the disks. Also, make sure disks are "clean" and in good condition before inserting them in to a drive - because early Macs usually used floppy as their only storage, these disks typically got a LOT of wear.

    For the purposes of posting on Winworld, I usually convert to an "IMG" raw sector image. The main reason being we needed to choose a standard, and many of our Mac archives were already in this format. Exceptions being Lisa software and a few others that may need the normally unused sector tags that IMG files exclude. (If there is ever any doubt, I will include source SCP/KF/SIT images).

    A couple of things about archiving Macintosh disks with a SCP - always select "splice" mode with at least 3 revolutions. The reason being, Macintosh 400/800k disks are not sector aligned. 1 or 2 revolutions is not enough for some decoding tools.

    In practice, when archiving Macintosh disks using a SCP, KF or other flux dumping tool, it is a very good idea to keep multiple 3.5" floppy drives around. All PC 3.5" drives must be tested ahead of time with known good 400/800k disks to ensure their analog filtering is sufficiently compatible with GCR (Teac and Toshiba drives are usually the best). In practice, when dumping weak/damaged Macintosh disks, it is all too common to have to switch drives to find one that will just read a marginally readable sector.

    You are right that a genuine Macintosh should have better results reading a disk. In practice, however, if a Macintosh disk is damaged, most of the time a genuine Mac will have problems reading it too.

    Which leads to another issue. With a Mac, reading a disk is all or nothing. If there is an error anywhere, Apple disk tools will usually stick their nose up and not give you anything.

    This is why, despite everything, I still lean toward using flux dumping tool for these disks. If that first read attempt rips up the disk, then a Mac will leave you with nothing, but a flux dumper will still give you something you can analyze later. Personally, I find the Kryoflux easier to use for this as I can easily mix and match tracks from multiple dumps.
  • edited May 2021
    @SomeGuy thanks for your advice, I will then go for the SCP scan. I need to test the floppy drive, to have the best results.
    How do you check if the dump has no bad sectors ? HxC doesn’t seem to recognise the Macintosh format when I open the SCP archive.
  • I normally use the PCE tools included with the PCE emulator to decode Macintosh images.

    The Kryoflux software also supports decoding of Macintosh images, both from an existing flux stream dump or interactively directly from a disk. But most of the time the PCE tools does a better job.

    The commands to convert are:

    pfi "dump.scp" disk1.pfi
    pfi disk1.pfi -c 0-79 -p set-rpm-mac-500 -r 500000 -p decode pri disk1.pri
    pri disk1.pri -p auto-align-gcr -p decode gcr disk1.psi -f -v

    psi disk1.psi disk1.img
    psi disk1.psi -L > disk1.txt

    erase disk1.pfi
    erase disk1.pri

    Although, out of habit I often use a batch file with this older method that should produce 100% identical results:

    pfi "track00.0.raw" disk1.pfi

    pfi disk1.pfi -c 00-15 -R 2 -r 381310 -p decode pri diska.pri
    pri diska.pri -p auto-align-gcr -p decode gcr diska.psi -f -v

    pfi disk1.pfi -c 16-31 -R 2 -r 349510 -p decode pri diskb.pri
    pri diskb.pri -p auto-align-gcr -p decode gcr diskb.psi -f -v

    pfi disk1.pfi -c 32-47 -R 2 -r 317700 -p decode pri diskc.pri
    pri diskc.pri -p auto-align-gcr -p decode gcr diskc.psi -f -v

    pfi disk1.pfi -c 48-63 -R 2 -r 285950 -p decode pri diskd.pri
    pri diskd.pri -p auto-align-gcr -p decode gcr diskd.psi -f -v

    pfi disk1.pfi -c 64-79 -R 2 -r 254190 -p decode pri diske.pri
    pri diske.pri -p auto-align-gcr -p decode gcr diske.psi -f -v

    psi diska.psi -m diskb.psi -m diskc.psi -m diskd.psi -m diske.psi disk1.psi
    psi disk1.psi disk1.img
    psi disk1.psi -L > disk1.txt

    erase disk1.pfi
    erase disk1.pri
    erase diska.pri
    erase diskb.pri
    erase diskc.pri
    erase diskd.pri
    erase diske.pri
    erase diska.psi
    erase diskb.psi
    erase diskc.psi
    erase diskd.psi
    erase diske.psi
  • @SomeGuy Hi ! I want to start dumping my macintosh disks and I'm in the process of choosing which is the best floppy drive to use with my SCP card. I have one 720k Drive and 3 different 1.44 drives to test. I used a floppy disk from Ashton Tate (Fullpaint). I think it's a good candidate as it's a 400K disk, bootable and I can test it on my Macintosh Plus.
    I made dumps using the SCP with the 4 floppy drives (splice 3 revolutions as you suggested) and one dump with DiskCopy 4.2.

    The DiskCopy 4.2 image works fine, I can recreate a flopy disk and it boots on my Mac.

    I'm not able to do anything with the SCP dumps. When I recreate a flopp disk using any of the 4 images it's refused by the Mac. I tried to convert from SCP to IMG using your scipt but when I open the IMG fine on HxC I get a load error.

    Am I doing something wrong ? or is this disk especially difficult to copy ?

    How can I determine which is the best floppy drive using the SCP images ? Is there a way to determine the "quality" of the dump just based on the SCP file ?

    Here are the different dumps

    Thanks in advance for your help / advice.
  • Keep in mind that IBM PC drives are not designed to handle these odd bitrates or encoding at all. So there is really no single "best". It seems like every time I archive Macintosh disks, I keep having to switch back and forth between drives trying to find one that will juuuuust read some stubborn track.

    Over all, I have had the best luck with Teac and Toshiba 1.44mb drives. These will typically read all tracks of a good condition disk with few errors. Most other brands fail miserably, with the first or last 10 tracks or so being totally unreadable (the signal bit rate frequency gets filtered out by the drive's filtering electronics.)

    So, glancing at your dumps - over all, these drives look fairly OK. You labeled the drives A-E. The dump from drive "D" shows a few errors on the first few tracks, so I would lean away from that one.

    The dump from the drive labeled "B" happened to pick up some stray bits from the second side - that is not bad, that can happen any time you are dealing with single sided media. Just be aware that when that happens, you just need to tell the decoding script to ignore the second head. Everything else decoded fine, so I would say that drive is OK.

    So, at a glance, the IMG created from A and E look fine. VMAC Mini recognizes the disk and and can run the program.

    As for the other issues...

    HxC does not know much about mac disks. The HxC drive emulation hardware can not (I think) be used in a Macintosh. They added a tiny bit of support to the software tool, but the last time I checked, it was very incomplete and unusable.

    The tracks on this disk are not "index aligned". The kryoflux folks admit that a Kryoflux can not write back unaligned tracks. Jim Drew swears up and down that the SuperCard Pro can write non-aligned tracks, but I have never once gotten that to work.

    What happens is that the KF and SCP start reading and writing right at the index mark. On a non-aligned track, this is usually right in the middle of a sector, and creates a read error.

    The only way I have been able to write Macintosh disks is to create a "reconstituted" image from an .IMG file (using PCE or a little Macintosh converter program on the Kryoflux forums). This image will be index aligned. I can write these with my Kryoflux OK, but now that I think of it, the SuperCard pro would puke on those the last time I tried it. I think it was somehow misinterpreting the bitrate or RPM value.

    So, dump "A" and "E" look perfect. Decoding "B" just needs a little tweek.

    If you want to write non index aligned images with the SCP, you would have to ask Jim Drew on the cbmstuff forum, but I would not expect anything from him other than finger pointing.
  • Thanks very much ! Drive A is a Toshiba 720k and the E is Alps 1.44 model. B is a Citizen 1.44 but I will stick with the Toshiba for the 400k/800k floppies.
    Is there an easy way to tell if the dump was ok or not, so that I can be sure before I post it ? With IBM disks I was using hxc and it would tell me if there were any bad sectors.
  • I wonder how applesauce would do. I don’t own a Disk ][ drive for 5 1/4” floppies, and they don’t seem to show a index kit (or at least instructions) for the newer 5.25” drives I’ve got. They say that the 3.5” drive doesn’t require the index kit. So maybe it would do better for early Mac disks, since it was designed with GCR floppies in mind.
  • Unless there are some newer tools that I have not seen yet, all I know to do is to look at the output log from the PCE tools and see if anything is odd.

    Of course, if you think there are any issues with a dump feel free to send it to me and I'll take a look at it when I can.

    BTW, I noticed that the first command set above needs the -R 2 command on the PFI line or it will get random errors due to the first and last revolutions not always being complete on non index aligned disks.

    To limit the decoding to the first head, add -h 0 to the PRI line.

    For example:
    pfi "FullPaint 1.0 - Program [3.5] lecteur B 1.44.scp" disk1.pfi
    pfi disk1.pfi -c 0-79 -R 2 -p set-rpm-mac-500 -r 500000 -p decode pri disk1.pri
    pri disk1.pri -h 0 -p auto-align-gcr -p decode gcr disk1.psi -f -v

    psi disk1.psi disk1.img
    psi disk1.psi -L > disk1.txt

    erase disk1.pfi
    erase disk1.pri
  • Thanks for your help, I have now all the correct parameters.
  • @SomeGuy I have a package with 800k and 1.44Mb floppy disks (Apple personal Diagnostics). I'm using Apple Disk Copy to make raw image (extension .image), do I really need to post SCP dumps with corresèponding .img files also ? if yes, with 1.44Mb, splice method does not allow more than 2 revolutions, which one do I choose for the -R parameter ?
  • If you make SCP dumps, I would like to be able to take a look at them. Analysis of flux dumps can often show me if the disks were modified by the user, help confirm there are no sector errors, and helps me verify if there is or is not copy protection. When adding to the library, I may omit flux dumps if I am satisfied there is no copy protection, but I would still like to see them.

    But if this is an inconvenience, and you are sure there is no copy protection, then just IMG files will suffice.

    1.44mb disks, even on a Mac, should all be index aligned. That means specifying either -R 1 or -R 2 should yield identical results. In other words, it shouldn't matter.

    Note that the PCE tools commands for decoding Macintosh 1.44mb disks are identical to IBM PC 1.44mb disks. For example:

    pfi disk1.scp disk1.pfi
    pfi disk1.pfi -R 1 -r 1000000 -p decode pri disk1.pri
    pri disk1.pri -p decode mfm disk1.psi -f -v
    psi disk1.psi disk1.img
    psi disk1.psi -L > disk1.txt

    erase disk1.pfi
    erase disk1.pri

    Unlike the 400k/800k Macintosh disks, other tools such as HxC should have no problems with 1.44mb Macintosh disks.
Sign In or Register to comment.