On 18 Nov 2009 at 17:25, Fred Cisin wrote: 32 sector 8" disks are common with other manufacturers as well. Other than the ancient oddball IBM 8-sector (?) format, 32 sector is pretty much the HS format in 8". The gotcha is that several 8" drives can be jumpered to divide the sector marks to 16 or 8 sectors. These are commonly the ones with separate INDEX and SECTOR outputs. I suspect external circuitry was also used to do the same with drives that didn't explicitly support hard sectoring. --Chuck

Tony Duell wrote: It should work with anything that has a 34-pin Shugart interface. It supports the pseudo-Shugart interface PCs use (including the twisted wire), too. The current "Must Have" format support list is: - 8-inch: hard and soft-sector (untested, I don't have a drive or discs) - 5.25-inch: hard and soft-sector (I have a 96tpi double-sided drive and soft-sector discs, but no hard-sector ones; a 48tpi double-sided drive might also be worth testing) - 3.5-inch: soft-sector. Do hard-sector 3.5" discs even exist? Most of the drives I've looked at use the position of the spindle motor to derive the index signal, so I'm guessing "no". Theoretically the following should also work: - Amstrad 3-inch. Don't have one of these, don't feel like gutting a CPC/PCW/ to get one. I seem to recall the drives are 80 track and Shugart interfaced, though. - Zenith Minisport 2-inch. Assumption: the drive uses a Shugart interface, and you have a junked MiniSport to "liberate" the drive from. Disc formats ============ As far as disc formats go.... pick one. Any FM or MFM based format should be readable (and probably writeable too) with the standard hardware, microcode and firmware. If you want to add support for another format, you basically just have to change the microcode (which is loaded into the FPGA every time the hardware is cold-booted). MFM hard drives should also be fine, but a buffer/converter board will be necessary (to convert from differential to single-ended signalling). RLL should also work if you've got software to decode the contents of the drive. Software ======== Open-sourced, multiplatform. Should work on any platform with a libusb port -- that's Linux (kernel 2.2+), Darwin (Mac OS X), the various BSDs, and Windows 2000 thru Vista (32 and 64 bit). That's what, 95% of the current OS market? My discussions with the Software Preservation Society have turned sour, so it's not likely you'll be able to read a disc and produce a .IPF (or the precursor format, whose extension I can't remember right now) from it. I'm not even keen on adding an "IPF writer" extension given that the libcapsimage library is a binary blob. I might look into an extension to the ImageDisk format to store raw track data, but storing the relevant metadata (sample rate, acquisition settings...) might be tricky. Either way, the output formats are going to be open and documented, in the same way as the .IMD format. I think that answers just about all the questions... -- Phil. classiccmp at philpem.me.uk http://www.philpem.me.uk/

MMFM? Amiga (MFM, without IBM style sector headers) Tandy PDD? (3.5" ""hard sector"" (index off of spindle?)) 3" (Amdek, Amstrad, etc.) 3.25" (Dysan) are hard to distinguish from 5.25" Some formats, such as Intertec Superbrain had the data bits inverted relative to the header bits Some, such as TRS80, got weird with their Data Address Marks 2.5" and 2.9" There were some with "conventional structures", and some with single track spirals Amlyn which preceded IBM's 1.2M is 5 1.2M disks in a cartridge with a disk changer mechanism Kodak, Drivetec 2.8M "ED" "vertical recording" 3.5" 360 RPM, 1000K bits per second "Floptical" (20M 3.5", SCSI interface) LS120? (IDE interface) Iomega products (only readable by Iomega drives until the Click of death) Exatron Stringy Floppy Tape drives that used an SA400 interface I saw a "memory stick"? reader that was shaped like a 3.5" floppy, and inserted into the floppy drive . . . and then, there are some obscure ones, and even some unusual ones.

Fred Cisin wrote: Yes, with appropriate software. Yep. Set the hardware to MFM-sync, using the sync-word 0x4489 (SYNC-A1). Should be possible. If you can find me a drive and a pinout, I'll tell you :) That would be a software issue, not a hardware one. As long as there's either an FM or MFM synch word to lock onto, or an index pulse, reading the data should be possible. Decoding it might take some work. That might involve increasing the buffer size on the analyser. My prototype is 512K (because that's how big the SRAM on the Altera DE1 board is), but the "plan" is to use a 256K chip (big enough to allow a track to be read index-to-index twice in one operation, making it possible to find the "start" and "end" of a track in software). I've got one of these drives somewhere.. an IBM thing with no pinout or documentation. I know it's Shugart interface with a few extra pins added (the ribbon cable provides both power and data IIRC). Not a chance. I'm not doing IDE or SCSI, just the lower-level interfaces: MFM, RLL and (possibly) ESDI. Maybe, if someone writes the software. That would probably be SmartDisk's floppy-disc memory card reader. They were talking about its potential uses in software protection dongles in the mid-90s. I seem to recall there being an article in "Electronics: The Maplin Magazine" about it, but I don't know which issue it's in (and I have six A4 box files full of them). -- Phil. classiccmp at philpem.me.uk http://www.philpem.me.uk/

On 20 Nov 2009 at 14:46, Fred Cisin wrote: Just find a Teac FD-235J. It had a pile of jumpers on it, one of which allowed for "media select" on density. Finding DSED media might be an issue. Finding *good* DSED media, harder still. --Chuck

On 20 Nov 2009 at 23:25, Philip Pemberton wrote: Not the 235J that I have--plain old 34-pin legacy connections. Here's a product manual excerpt: http://arcarc.xmission.com/archive/Test%20Equipment/Fluke/PDFs/9100_Fi xture/Floppy%20Drives/3fd0010a.pdf Are you certain that you're not thinking of the FD235HS or the FD235JS SCSI drives? (basically a non-S drive with Teac's adapter bolted on). --Chuck

I've had very good luck with Athana DSDD 3.5" media when used in my Amiga and 1581 drives. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_!

On Fri, 20 Nov 2009, Fred Cisin wrote: Did I miss something? :) Neither my Amiga nor the 1581 uses ED media. Are you possibly referring to the CMD FD4000 drive? it was for the Commodore IEC bus and was an ED mech if memory serves. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_!

On Fri, 20 Nov 2009, Chuck Guzis wrote: ...and NOW I found my mistake. I misread the original message as referring to "DSDD" and not "DSED". My apologies. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_!

Fred Cisin wrote: The acquisition clock rate is 40MHz. That means every clock tick in the acquired timing data represents 25 nanoseconds. My current torture-test is to feed it a 10MHz "read data" signal, which it's handling fine. It's actually having more trouble with the slower bit rates because there are one or two little issues with the pseudo-RLE compression. I seem to recall mentioning that it'll read MFM hard drives, which typically have a 5Mbps data rate. 1Mbps should not be an issue. You'll have more trouble soldering the adapter cable together than you will getting a valid data readback! Knowing IBM, I wouldn't be surprised. I think my 2.88 drive is buried in a box somewhere (it was "liberated" from a dead PS/2 of some description) but I don't have any discs for it. -- Phil. classiccmp at philpem.me.uk http://www.philpem.me.uk/

IIRC Western Digial made disk controller chips with normal and inverted data buses (that was, IIRC, the difference between a 1791 nad a 1793, for example). And Intertec (?) used an inverted-bus one without an external inverting bus buffer.... Only on the directory cylinder of Model 1 TRS-DOS and clones AFAIK. When they went to double density, the had to sue standard DAMs since that's all the 179x chips could do. ANd IIRC LDOS used standard DAMs even on Model 1 single-density systems. Soeheter I have an Epson drive unit. It looks a it like the well-knwon TF20, but it isn't. It's something like a BM5. The host interface is a custom parallel one (possibly to a card that pluged into a QX10 or something). The internal controller board contains a Z80A and a _hard disk_ controller chip, the interface to the 5.25" drive is a 34 wasy and a 20 way ribbon cable. Yes, it's ST412 or similar. I have no idea what sort of floppies it uses. -tony

Philip Pemberton wrote: 800K would make more sense. The motor is controlled by the onboard microprocessor and the 20 pin interface is the same for the hd20 and superdrive. They are also much more common today than 400K.

Tony Duell wrote: !REDWC, pin 2. I've made pin 2 an I/O on my analyser -- it can either be read (pullup enabled), floating, or pulled low. Thus, you can poll the disc's type (if REDWC is used as a "high density" output), or apply density selection, or use it to drive TG43, depending on what the drive requires. That's effectively the question I was asking -- the discs don't support it, so did anyone bodge a PLL onto a drive to do it? It does look like someone's had a go at something similar to what I was thinking of: http://www.lesbird.com/sebhc/storage/storage.html Look under the heading "Hard Sector Floppy Emulator". So more or less the same technology as a 5.25" disc then -- a ~5mm hole in the jacket, a ~1mm pinhole in the disc, and an optosensor in the drive. -- Phil. classiccmp at philpem.me.uk http://www.philpem.me.uk/

Being a Tandy product, the should be a service manual for it -- Tandy were very good about supplying service manuals for all their computer products, even ones they simply 'badged' (like the pocket computers). Is this manual on-line anywhere? If not, does anyone have it? -tony

The hole sizes are somwhat different (in particlar the hole in the hjacket is smaller than 5mmm). The 3" disk has a hard plasic case with an intenral shutter, mechancially superior to the 3.5" one. Openin the shutter (done by a slider on the side of the disk which catches on a hook when you put the disk in the drive) also exposes the index hole. The hole is in the magnetic media itself, not in the plastic hub or anything like that. -tony

FAIKL there's no such thing as SS 3" _media_ -- all the disks are coated (and tested) om both sides. There is 40 cylinder and 80 cylinder media. and Amstrad (the main user of 3" drives) confused the situation by using single-head 40 cylinder drives and double-head 80 cylinder drives only. So the 80 cylinder disks got known as the double-sided ones. However, 40 cylinder double-head drives exist, I've used them. The plastic shell of the 3" disk has an offset notch in the front edge. A double-head drive has a metal bracket that fits into this notch, so you can't put the disk in 'flipped'. A single-head drive either ingores this notch or has a light barrier to detect it -- the single-head drive in my Tatung Einstein has am 'in use' LED that is red if the disk is in one way up and green if it's in the other way up. This means, of course, that a double-head drive can't read the flip side of a single-head disk. You can't flip the disk to read it with head 0 (because of the bracket) and you can't read it with head 1 because the disk is effectively turning the wrong way. AFAIK this is the only type of disk that has this problem (Yes, 8" dishs have the index hole in a different place on single and doubled sided media, but every 8" double-head drive I've seen has 2 index sensrs so as to be able to handle single-sided media too) -tony