3.5" floppy drive question(s)

[Chuck Guzis]

On 11 Oct 2008 at 21:17, Tony Duell wrote:

> The HP9836 system that I've been working on has a pair of 5.25" drives 
> which are effectively permanently enabled (the 'MX' jumper is fitted) and 
> thus have spearate cables back to the contoller board. The write-protect 
> lines trigger flip-flops for a disk change function (as well, of course, 
> as being used for the write protect function) while the index lines also 
> trigger retriggerable monostables for a ready function (if index pulses 
> are happening sufficiently often, there must be a disk in the drive, it 
> must be turning). 

I've got some 5.25" MPI drives with a separate "ready" board that 
plugs into the main drive electronics board.  It's got a 96L02 dual 
one-shot and some discretes to form a ready circuit.

One of the problems with add-on ready circuits on the controller side 
of the select circuitry is that determinining ready has to involve 
enough time (3 revs usually) to make sure that spindle's up to speed. 
Unfortunately, this creates a delay every time one changes drive 
select.  Conditioning the READY off of the motor signal is no 
guarantee, as some drives disable the motor when there's no diskette 
in the drive.  So leaving drives always selected as you describe is a 
way around the problem. 

We were plagued by customers simply changing diskettes whenever they 
felt like it.  Relying on a checksum-after-select scheme such as that 
used by CP/M and early MS-DOS versions was pretty ugly (and not at 
all airtight, something that plagued MICROSO~1 no end).  While 
playing with the driver, I noted that the write protect signal was 
conditioned only on select, not on motor, head-load or ready.  
Experimenting with some of the other engineers to see how fast one 
could insert and remove a floppy gave a safe polling interval of 
about 250 msec, which was barely noticeable in normal system 
operation and performed only if there were open files on the drive.

Some of the single-sided 600 RPM Sony (0A1?) drives were also used in 
a few CP/M systems, such as the Preis.  It was a very noticeable 
improvement over 5.25" media.

Cheers,
Chuck