Image Map Image Map
Page 4 of 5 FirstFirst 12345 LastLast
Results 31 to 40 of 49

Thread: Brother floppy disk decoding

  1. #31
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,279
    Blog Entries
    20

    Default

    I've had no interest in writing Brother floppies, so I never investigated that aspect. (I've been reading and converting them for 20 years or so).

    However, my recollection is that the Brother WPs that use the 240K/120K floppies are z80-family based and that there are ROMs in the machine. One place to start might be reading and disassembling the things.

    While track 0 doesn't appear to be used for data storage and is not in the same format as the remainder of the floppy, I wonder if there's identification information lurking there somewhere. I recall that if you initialize a DOS-formatted floppy on the WP, that track 0 gets overwritten with something.

    Just some food for thought.

  2. #32
    Join Date
    Feb 2017
    Location
    Zürich, Switzerland
    Posts
    131

    Default

    Well, thanks, and now I feel... mildly chagrined?

    So I magnetically wiped a disk (being able to direct drive a floppy drive is awesome), formatted it on the WP, formatted it again because it didn't work the first time, read the flux, then used FluxEngine to rewrite the data to the data tracks, leaving the others untouched. Worked fine.

    Then I started fiddling with the non-data tracks, which the WP formats but which just contain noise. (Literally. https://pastebin.com/vxBmKvzc) I could not make head nor tail of them. Erasing them had weird effects. Sometimes it'd work, sometimes it wouldn't.

    Then I figured out that my PC USB voltage isn't quite high enough to make the floppy drive work, and it was sometimes seeking to the wrong track. But my cheap Chromebook produces a rock solid 5.00V and it works fine there.

    *sigh*

    Anyway, my hypothesis of the way it works is: the WP seeks backwards until it hits the stop; then it looks at tracks one-by-one going forwards until it sees something that looks like data. The easiest way to use FluxEngine to write Brother disks? Forget about all the physical/logical sector nonsense and just write track 0 to track 0, and everything is fine. Brother write support is go!

    @hedehede81: do you want to give it a try, just in case yours is different? Update from github; you'll need to rebuild the firmware.

    Now I really want to get hold of the Brother demo disk so I can reverse engineer the executable file format. CP/M would run beautifully on this thing.

  3. #33
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,279
    Blog Entries
    20

    Default

    Re: the WP finding the location of the first track--I believe I said that earlier.

    However, I believe that the WP drive can seek less than a full track, based on my experiences.

  4. #34
    Join Date
    Feb 2017
    Location
    Zürich, Switzerland
    Posts
    131

    Default

    Yes, you did --- I hadn't forgotten! But I did accidentally skip a vital detail in my last post, which is that the WP seems to be incredibly broad-minded about what constitutes data. So, tracks 0..2 containing noise and track 3 containing data is fine; but erase track 0 and the disk becomes unreadable. I was expecting it to be looking for a track 0 sector header, but apparently not. It may even be deliberately looking for noise, but I'm not sure why (or how).

    It's also really strange that it can reach track 0 but is choosing to start the data on track 3. @hedehede81's chooses track 1.

    Re microstepping: as I can do that on a conventional PC drive, so I'm just ignoring the entire issue and hoping it never comes up... at least in a sample set of 2, both WPs have conventionally aligned drives. (Although @hedehede81 did find an old disk with a non-conventional track pitch. Hmm.)

    In other news, my new PX-8 produces incredibly boring IBM scheme disks, although they are at least 40 track which lends them some interest. I'm going to have to track down someone with an old Macintosh.

  5. #35
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,279
    Blog Entries
    20

    Default

    Over the years, perhaps 20% of the thousand or so Brother WP floppies I've processed have pronouced track mis-alignments. Add to that, the 40 track WPs--where, on an 80-track drive, you must select between reading the odd-numbered or even-numbered tracks--there seems to be a bias toward using 0,2,4,..., but there are disks that are readable using only 1,3,5...

    In years past, I've dealt with the misalignment issue by simply using a Brother WP to copy the entire disk to a new one. While this works, it's incredibly time-consuming as it necessitates switching source and destination floppies about six times for a complete copy. I've since gone to using intentionally mis-aligned drives--just a half-track out in either direction seems to do the trick.

  6. #36

    Default

    hjalfi, I will build the new firmware and fluxengine and give it a try again. As I said in the PM, I think we are almost there, the floppy written by fluxengine can be read in the PC just fine, it should be those first tracks making the difference.

  7. #37
    Join Date
    Feb 2017
    Location
    Zürich, Switzerland
    Posts
    131

    Default

    I've added that to the documentation --- ta. I'll just have to see if one of these floppies ever turns up.

    Re 40-track drives: my Epson PX8 has 40-track disks. Reading them on an 80-track drive shows the data on even tracks with the odd tracks mostly blank. I was expecting to see the same data on both, as on 5.25" drives the track width for a 40-track disk is twice that of an 80-track disk. However, 40 track 3.5" drives appear to be extremely rare, so I wonder whether they just cheaped out and used an 80-track drive head, or possibly even an 80-track drive with double stepping. That would explain the narrow tracks.

  8. #38

    Default

    I am happy to report that thanks to hjalfi's latest updates, I confirm that fluxengine can now write Brother 240kb disks and the WP can read it. I first tried just a disk copy, it worked. Then I modded the brother.img and the WP can also read the modded disk.

    I will later add support to Brother WPT Tools so that it can add and delete files to the image, so that we can create images from scratch with its conversion features.

  9. #39
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,279
    Blog Entries
    20

    Default

    I've never dug into that with other machines, but will say that the Brother WP 40 track disks do have wider tracks than the 80 track ones, as evidenced by a few drops of Kyread and a microscope. Compare with the 40 and 80 track 3.0" CF2 drives.

  10. #40
    Join Date
    Feb 2017
    Location
    Zürich, Switzerland
    Posts
    131

    Default

    Once again an otherwise promising theory has been deflated by mere facts. Oh, well. (I'm not up to actually investigating the media itself.)

    So, I now need to do a bunch of code cleanup, write things up a bit, do a demo video, then try and get it reading and writing PX-8 disks. Reading's easy (I just tell it it's an IBM 80-track disk and it works; FluxEngine reassembles the image from the fragments it finds and I've managed to image the utilities disk) but writing will require a little work. Then I should probably go and do something else for a while...

    Thanks for all the assistance! This would have been impossible without it, and I appreciate it.

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •