Date   : Sun, 08 May 2005 23:59:11 +0100
From   : jgh@... (Jonathan Graham Harston)
Subject: Re: IDE Interface for BBC

"W.Scholten" <whs@...> wrote:
> > For anyone who has been assembling the interface at home there are two
> > corrections required to the PCB: details available by private email.
> 
> Does that remove the need for extra seeks?
 
No, that is a limitation of specific *drives*.
 
> What I'd like to see on a single PCB is the IDE interface + 8K or 32K
> SRAM to enable easy implementation of E00 DFSs as in the Opus
> Challenger. I don't think I'll get round to do it myself within
> reasonable time.
 
What do you mean by "easy implementation"?
 
DFS is eminantly *un*suitable for mass (eg hard drive) storage.
The filesystem is limited to a maximum of &3FFFF bytes (256K-1)
and a single 31-entry (or 62-entry with WDFS) catalogue).
 
If you wish to patch an E00 DFS to access an IDE interface, go ahead.
 
ADFS on the BBC Master does not use memory at &E00 and can usually be
configured to have PAGE at &E00.
 
ADFS on the BBC B uses low memory and can push PAGE up past &2100.
Condequently, I recommend using HADFS on the BBC B. This takes only
a single page of private workspace, so on a DFS-only machine pushes
PAGE up from &1900 to &1A00.
 
The IDE interface is just that - an IDE interface, a specification and
set of suggested complying interfaces. Anybody is welcome to use the
design in their own board. I set off to design a working IDE interface,
not to jump six or seven steps and try a design an integrated multiple
interface board without first getting the IDE interface itself working.
 
-- 
J.G.Harston - jgh@...                - mdfs.net/User/JGH
Spectrum Emulators - http://mdfs.net/Apps/Emulators/Spectrum