Attempting to build a DSP memory expansion

[Tito]

OK, I don't think this is that crazy of a project after looking at the Turbo
Color schematics for the DSP slot pinout. I'll get high-quality mp3s playing
back on Black if it kills me (OK, not kill me, but just because it's fun).

I got some 64-pin RAM which I will strip the chips off of, and wire in some
20ns SRAM I bought along with a 1 to 2 demux which will switch between X and
Y. What I'm unsure about is the DSPRAMEN and DSPRAMDIS lines; I think need
to tie the DSPRAMDIS to +5V so the internal 8K disables, but should DSPRAMEN
go low?

Anybody ever tried something like this, or know anything about what it would
take to do this kind of thing?

Thanks for any help,
Rob

[David Evans]

In article <ssgc3gq...@news.supernews.com>, Tito <n0...@n0where.com> wrote:
>
>What I'm unsure about is the DSPRAMEN and DSPRAMDIS lines; I think need
>to tie the DSPRAMDIS to +5V so the internal 8K disables, but should DSPRAMEN
>go low?
>

Do you have the 56k databook? Mine is at home so I can't check this stuff
right now but I think that it talks abot it.

--
David Evans (NeXTMail/MIME OK) dfe...@bbcr.uwaterloo.ca
PhD Student, Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual

[Tito]

> >What I'm unsure about is the DSPRAMEN and DSPRAMDIS lines; I think need
> >to tie the DSPRAMDIS to +5V so the internal 8K disables, but should
DSPRAMEN
> >go low?
> >
>
> Do you have the 56k databook? Mine is at home so I can't check this
stuff
> right now but I think that it talks abot it.

This looks like a NeXT hardware thing only. They're known as DSPRAMDIS and
_DSPRMEN on the Color Turbo schematic (CSlab33-09.pdf), and I can't find
info on the MCM56824 to know the difference between _E1 and E2. I'll try
raising DSPRAMDIS and leaving _DSPRMEN alone to see what happens once I get
everything wired up.

Rob

[Timm Wetzel]

"Tito" <n0...@n0where.com> writes:

> OK, I don't think this is that crazy of a project after looking at
> the Turbo Color schematics for the DSP slot pinout. I'll get
> high-quality mp3s playing back on Black if it kills me

Sure sounds like a fun project.

Does this mean you are/someone is actively working on a
DSP-accelerated mp3 player?

I heard some rumours in the past, but nothing specific.

Regards,
Timm
--
Timm Wetzel <twe...@gwdg.de>
Biomed. NMR GmbH Tel +49 551 201-1091 FAX +49 551 201-1307
Am Fassberg 11, D-37077 Goettingen, Germany

[Tito]

> > OK, I don't think this is that crazy of a project after looking at
> > the Turbo Color schematics for the DSP slot pinout. I'll get
> > high-quality mp3s playing back on Black if it kills me
>
> Sure sounds like a fun project.
>
> Does this mean you are/someone is actively working on a
> DSP-accelerated mp3 player?

Yes, I'm working on it. Once things get a little further along I'm going to
set up a webpage about it.

Regards,
Rob

[Carl Ludwigson]

Tito <n0...@n0where.com> wrote:
:> >What I'm unsure about is the DSPRAMEN and DSPRAMDIS lines; I think need

: Rob

The 56824 is an 8k x 24 fast sram, and the schematic says that dspramdis
has a pullup resistor on it. So, i'd guess that when the expansion board
is plugged in you'd want to gnd the dspramdis line to disable the internal
56824, n'est-ce pas?

Also, this is old hat news, but the original next board (# 7013) had three
mcm6206 srams and a 74f158 decoder. MOT doesn't make the 6206 any more,
but recommends the mcm6246 as a replacement for the 6206. Not sure what
the san diego state board had on it...

The dsp expansion socket pins 57-60, 62-64 on the schematic lists xy,
rd, wr, men, ps, ramdis, and gpio0...and the sram chips list w, e, and g.
The design is to decode xy, rd, wr, men, ps, ramdis, and gpio0 into
w, e, and g to provide a NNNk x 24 memory board.

Regards,
Carl

[Tito]

I'm planning on trying 3 banks of 32kb x 24bit for X, Y, and P memory
respectively. A design very similar to the one described at
http://www.heise.de/ct/ftp/projekte/klangwerk_1/art.shtml. Basically, just
have something like:

_DSPDS & DSPXY to X and Y banks' enable
Xenable = _DSPDS AND DSPXY
Yenable = _DSPDS AND !DSPXY
_DSPPS to P bank enable
_DSPRD & _DSPWR to all 3 banks
DSPRAMDIS grounded
_DSPRMEN to ??
GPIO0 to ??

I don't believe memory maps of X, Y and P need to overlap anywhere;
supposedly MusicKit and most other NeXT DSP s/w is written to avoid any
assumptions about that sort of thing.

What do you think?

Rob

"Carl Ludwigson" <crlu...@netcom12.netcom.com> wrote in message
news:8qnrdv$d5$1...@slb0.atl.mindspring.net...