[prev in list] [next in list] [prev in thread] [next in thread] 

List:       classiccmp
Subject:    Re: Introduction
From:       ard () p850ug1 ! demon ! co ! uk (Tony Duell)
Date:       2008-06-02 2:21:01
Message-ID: m1K2zfW-000IxyC () p850ug1
[Download RAW message or body]

> 
> Grumply Ol' Fred wrote:
> > Although it is prob'ly irrelevant to this particular problem, IIRC,
> > buffered seek was one of the major differences between the ST506 and ST412
> > intrerfaces.
> 
> That's my recollection as well.

And mine.

> 
> On early drives, the point of buffered seek was to allow the electronics
> on the drive to use an acceleration profile specifically tailored to
> the stepper motor and actuator mass of the drive.  Before buffered
> seek, the time to seek n tracks was exactly n times the track-to-track
> time.  Buffered seek improved the seek performance considerably for
> longer seeks.

The SA4000 (14" Winchester) has a rather odd seek specification. It 
supports buffered seeks -- sort of. You can either send it step pulses so 
slowly that it steps one track at a time, or you can send it a burst of 
step pulses so faster that it buffers the lot and then does the seek. But 
if you send step pulses at an intermediate rate, it'll miscount them 
(there is one up/down counter clocked both by the step pulse and the head 
movement logic, and the logic around that is 'minimal'...) and the heads 
won't end up on the cylinder you expect.

-tony
[prev in list] [next in list] [prev in thread] [next in thread]