Pinout for SED9421

[Tony Duell]

> >Interesting, I've never seen a floppy drive with buffered seeks, and 
> >since there's no seek-complete pin on the interface connector, I don't 
> >see how it could really work. Every floppy controller I've ever seen 
> >sends the step pulses at a constant rate, which would have to be slow 
> >enough for the worst case. 
> 
> READY goes false while seeking.   You can issue pulses as fast as you want

Ah right... That would really confuse some controllers, I think

> (I know the drive will take a pulse every 2 msec. and probably much
> faster).  In a way, this isn't all that different from the overlapped seek

That reminds me in passing of the Shugart SA4000 hard disk. That thing 
has buffered seeks, but you can't issue step pulses at any old rate. You 
either have to give it a step pulse and wait for it to give seek_complete 
(i.e. totally unbuffered seeks) or give it all the pulse at greater than 
a certain rate and then wait for the seek complete (buffered seek). If 
you give it the pulses too _slowly_, it will miscount (there is one 
up/down counter for the step pulses, if you're not careful the pulses 
from the step input get confused with those from the drive logic when it 
moves the heads by one cylinder!).

> >Sure... I remember the Sanders dot matrix printers that ramped up (and 
> >down) the step rate of the carriage feed stepper and sent the printhead 
> >pulses at just the right times to put the dots in the right places. 
> 
> Are you certain that Sanders used a stepper for carriage positioning?  If

Well, I was prtty certain, but I don't have the service manuals 
immediately to hand. I would be very certain about the 12/7, and fairly 
certain about the 700 (which is actually built on a Diablo 630 chassis 
with totally different elecronics and a dot-matrix head).

> they were like Diablo or any of the other high-speed printers, it was more
> likely a DC motor with encoder.  Steppers were just too slow, particularly
> for bidirectional "smart" printing.  IIRC, one of the engineers who worked

I think that's the reason for a lot of the Sanders electronics (the 700 
had 3 micropeocessors in it -- 2 Z8s and a Z80) - -to handle the 
acceleration of the carriage motor and put the dots down at the right time.

> on the Diablo dot matrix printer (can't recall the model number, but it
> used the Rockwell PPS-8 set) said that the whole business of accelerating
> and decellerating a mass like a printhead made for some interesting
> programming.
> 
> Paper feed, on the other hand, was almost certainly stepper-driven.

It certainly was. The 12/7 had a separate tractor feed with its own 
stepper motor that was simply wired in parallel with the platten motor in 
the printer. Turning the knob on the tractor feed caused the motor in 
that to generate enough voltage to turn the platten roller motor, even 
with the machine turned off.

-tony