Write 800k Floppies?

[lee4hmz]

The floppy drive mechanisms in both cases are pretty dumb devices. PC floppy drives are actually dumber than Mac floppy drives -- about the only thing they can do by themselves is reset the heads to track 0 on power-up, and the bus they use is pretty much the same as the old Shugart SA400 from 1976 -- one step at a time, no direct control of phases. Apple 3.5" floppy drives have registers inside them for various things, I've forgotten what, so they're marginally smarter than PC drives. (I do know they use the stepper motor phases as a register select bus.)

The Disk II was the dumbest of all, having nothing more than the stepper motor and a head amplifier, and I think that's what Anonymous Freak had in mind.

Anyway, the real issue here is what the controller does with the data once it receives it. The IWM, just like the Disk II before it, is just a state machine that pulls raw bits off the disk and expects the OS to make sense of them. This makes it quite flexible; you could probably read an MFM disk on an IWM if you knew what bits to tweak. (I'm not sure what magic the SWIM chip uses for MFM, but I would imagine the reason it wasn't done sooner is that either the IWM or the Mac itself couldn't handle 500-kbit MFM data reliably.)

PCs, however, use a chip that traces its ancestry back to the NEC uPD765 that was in the original IBM PC/XT; specifically, they're variants of either the Intel 82072 or Western Digital 37C65 PC/AT floppy controllers, depending on whose Super I/O chip you have. The 765 is easy to program and supports DMA, but it's also incredibly inflexible compared to the IWM; it only supports FM or MFM in the standard IBM "midrange" formats, and all data appears "cooked" -- raw disk access is pretty much impossible without hacks, as is reading a disk that strays too far out of the IBM specs. It also expects to use the same data rate across the entire disk (CAV mode); Mac 800k disks are zoned CLV, making them even trickier to read.

Pretty much all USB floppy drives use this particular setup, as it's well-known and easy to implement using off-the-shelf parts, so reading anything "weird" on them is out of the question. There's also the odd LS120/240 drive floating around still; they're very fast in floppy mode, but also limited to MFM (though they don't use 765s as far as I know).

Back in the day, there was a piece of hardware, the Copy II PC Deluxe Option Board, that got around all this by cutting the 765 out of the loop and using its own custom ICs to process the raw bits coming off the disk. It could read practically anything if it knew what to look for. Unfortunately, they're kind of hard to find now, and would require an older machine with ISA to be useful. That said, I have one, and it does actually work once you find the utilities...

Failing that, your best bet is to use an older Mac as a go-between. One of the reasons I still have my 8600 is that it can read and write 800k disks for my ROM 0 IIgs. (That and my 475's floppy controller is messed up.)

 

[tomlee59]

Thanks for the very complete explanation, lee4hmz!

 

[trag]

It may be of interest that the Outbound Model 125 reads and writes the Mac 800K format using a standard Citizen brand notebook floppy drive.

Of course there's a circuit board between the Citizen and the Outbound.

The main components on the circuit card are (IIRC):

WD 37C65

WD 92C32 (Digital Data Separator for floppies)

74LS624 (VCO)

XP9103 (digital potentiometer)

28C64

GAL16V8 (the part that holds me up cloning the thing)

and a few assorted 74 series logic chips.

Oh, yeah, and right at the input (host) end is an 85C30--makes me wonder what protocol the Outbound is using to talk to the thing.

One of these days I really need to whip up a schematic on this thing. It reads and writes 400K, 800K and 1.44MB format Mac floppies.

One kind of curious thing is that if my understanding is accurate, the 37C65 already has the functionality of the 92C32 built into it, so why the separate 92C32? I suspect that the VCO and the XP9103 are in some way used to control the motor speed of the floppy disk.

Other kind of interesting facts are that the Model 125 came out around 1990, uses the ROM of a Plus or SE and uses a standard IDE notebook drive (but only certain models). Arguably (or perhaps totally) the first Mac hardware to use an IDE drive.

 

[lee4hmz]

One kind of curious thing is that if my understanding is accurate, the 37C65 already has the functionality of the 92C32 built into it, so why the separate 92C32? I suspect that the VCO and the XP9103 are in some way used to control the motor speed of the floppy disk.

Sort of. The 92C32 and the VCO are there so the card can set an arbitrary bit rate (the 37C65 only supports the standard FM/MFM rates). That way, it can read 800k disks without changing the motor speed; PC floppy drives are built like HDDs, and are locked to a certain speed (usually 300 or 360 RPM). It's also, most likely, using the XP9103 as a cheap substitute for a DAC (I don't have a have data sheet handy for it, so this is pretty much a WAG...) to control the VCO.

The Deluxe Option Board actually uses a similar trick, though they used a UMC data separator instead of a Western Digital.

-lee

 

[trag]

Thanks for the analysis lee4hmz. Have you actually examined one of the Outbound floppies or is your analysis based on experience with your floppy adapter?

I'm always happy to find another person with experience of the Outbound Laptop Model 125.

 

[lee4hmz]

Thanks for the analysis lee4hmz. Have you actually examined one of the Outbound floppies or is your analysis based on experience with your floppy adapter?

I've never actually seen the Outbound adapter, but I did find a discussion somewhere (I've forgotten where now) of how the DOB works internally, and judging from the chips that are on it, the Outbound's adapter board would seem to work the same way.

-lee