On Sat, 16 Jan 1999, Pete Turnbull wrote:
On Jan 15, 19:35, Max Eskin wrote:
In fact, I've always found the limitations of
DOS and PCs curious. Does
anyone know why:
*They refuse to boot from anything but A: or C:
*DOS FDISK refuses to create more than 1 primary partition
*DOS can't find an ATAPI cd-rom without a driver, even though it's an IDE
device
Pass.
*Why we are limited to 2 floppy drives in DOS
Actually, you can have 4, if you have a controller at the secordary floppy
controller address as well as at the primaryfloppy controller address.
Each controller can only have two drives because of a silly hardware setup
chosen by IBM.
Actually, you can have a lot more than that, given the right
controllers. I have one machine with six different floppy drives operable.
With the MicroSolutions CompatiCard IV - regrettably, no longer
manufactured - you could have up to sixteen drives, four each controlled
by one of four of the cards. I don't really know why you might want that
many drives, but it is possible.
- don
On a normal SA400-style interface, there are 4 drive
selects, and one of
each other control signal; each control signal is normally gated (on the
drive) with the approriate select line. IBM chose to use an arrangement
where completely separate motor-on signals were provided for the two
drives. The way they did this was to arrange the signals in the cable such
that one drive used a particular drive select and the motor signal in the
standard place, and the other drive used a different select with the moror
signal on a different wire. Hence the infamous cable twist, such that each
drive sees a particular drive select (it's the second select line) and the
motor-on signal in the standard place. However, from the point of view of
either drive, there is another motor signal on a pin which would otherwise
be a select line for another drive. Thus you can't use all 4 drive
selects, and hence only two drives per controller cable.
Another way to look at it, is to assume that IBM wanted all drives to have
the same jumper settings, including the drive select jumper. But instead
of putting a twist in the cable which would just swap some of the select
lines, they used a twist over a wider section. Unfortunately that also
moves the motor control line to where a select line would normally be.
Thus they need another motor control line somewhere else (where a drive
select would be, obviously), and end up only being able to put half as many
drives on the bus.
--
Pete Peter Turnbull
Dept. of Computer Science
University of York