• Please review our updated Terms and Rules here

My ultimate multi copy machine

H

Hans01

Experienced Member
Joined
Feb 17, 2014
Messages
459
Location
Netherlands
Having always trouble with copying disks I built an ultimate copy machine.

It consists of an old MSDOS computer with an opti green motherboard
with an ISA bus and award bios, a 486-DX2-66 processor and a 720 and a
1.44 floppy drive.

It is very important to use the Adaptec 1542CP controller to connect the
floppies.
It is the only controller which can handle single and double density data.
It is not possible to connect a SCSI harddisk to the adaptec as
the award bios doesn't scan the memory blocs in the right way.
Any tric to solve this problem didn't help.
The only way is to use a multi I/O adapter to connect an IDE harddisk
I installed MSDOS 6.22 and later Windows 98SE too, to support some extra tools.

The adaptec controller is connected with a flatcable to the option board and
passes trough the board to the both floppys.
I brought this cable outside the computer and split the cable in two parts with
a 34 pin male and female connector.
In the BIOS I can swap the A: and B: drive in case of a
floppy boot with different type of floppy
If I connect the two parts together I can use both floppys.
If I open the connectors I can connect to one part of the cable to one or
two 5.25 drives or my new FDADAP adapter.
(This is realy a great tool to support painless 8" drives, a really must to
have device for Model II users)

With this machine I can Read, Write and Edit all Tandy 720, 1.44, 5.25 and 8"
SS and DS disks in Single and Double density.
I can backup all my disks to my harddisk now and rebuild them in a moment.

As software a use COPYIIPC, TELEDISK, IMD (Image) and many other tools.

It is a lot of work to build this machine but it is worth the costs and the time
a spent for it.
 
I should post a photo of my old copy system--6 floppy drives (and yes, every single one of them can handle FM disks). All six in operation, if I want, with 3 active at any given time.
 
I have 3 floppy drives (3.5" 720/1.44M, 5.25" 1.2M and 5.25" 360k) in my XT. And all of them are bootable. No need to switch anything. Unfortunatelly no 8" and no FM:)
 
Hans01 said:
It is very important to use the Adaptec 1542CP controller to connect the
floppies.
It is the only controller which can handle single and double density data.
Click to expand...
Not the only ones, but they are the easiest to find. Just make sure to get one with a National Semiconductor FDC, not the Intel FDC if you want compatiblity with 128-byte MFM sectors.
 
@Hans01, how about some pictures? Would love to take a look at your setup!
 
Has anyone thought about making a Shughart floppy interface card for PC with, say, the WD1771 series of controllers? The latest of those, the WD1772PH02-02, supports 125/250/500 kbps, at both FM and MFM, and does not even require any external components on the floppy-side (except maybe a buffer or two). The only needed external logic would be an IO-address decoder.

Still, while such a card would be able to read 8" floppies, it's limited to sector-images. Then something like the KryoFlux would problably be a better option for more or less the same price anyways.
 
I've got a wire-wrapped controller using one of those (or the WD1770, I don't remember). ISTR that the instruction set does include a "read track" and, of course, as with the 17xx series of controllers, the format is what you make it.
 
Chuck(G) said:
I've got a wire-wrapped controller using one of those (or the WD1770, I don't remember). ISTR that the instruction set does include a "read track" and, of course, as with the 17xx series of controllers, the format is what you make it.
Click to expand...

Yes.

In a Z80 machine I have, using the WD1797, the read track instruction might sometimes swap the clock and data signals somewhere in the middle of the data sector. The datasheet explains the command as more of a format debugging feature, and I don't think it was indended as an alternative way to getting sector-data off the disk.
 
Precisely--and you're warned about it. Since the address mark detector is always on, a write splice can cause false triggering of the AM detector, causing it to synchronize on the wrong phase.

I think the AM detector in the WD chips looks at only 16 bits; that is, 8 clock and 8 data bits interleaved, so it's easy to fool. If you're designing your own FDC from an MCU, you could extend that to any number of bits you care to. A 32-bit MCU would be ideal for that.

I guess the point I've been trying to make all along is that FDC chips are ancient technology and just about any modern MCU could easily perform the task of an FDC. Mostly, it's just a matter of software.
 
You must log in or register to reply here.
Back
Top