From Steven.D.Ligett@Dartmouth.EDU  Ukn Jun 29 14:22:13 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA00757; Tue, 29 Jun 93 14:22:07 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA14219; Tue, 29 Jun 93 11:17:47 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA19499; Tue, 29 Jun 93 11:17:43 -0700
Received: from dartvax.dartmouth.edu 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA14215; Tue, 29 Jun 93 11:17:31 -0700
Received: from prancer.dartmouth.edu by dartvax.dartmouth.edu (5.65+D5/4.5HUB)
	id AA19814; Tue, 29 Jun 93 14:16:09 -0400
Message-Id: <5304101@prancer.Dartmouth.EDU>
Date: 29 Jun 93 14:16:03 EDT
From: Steven.D.Ligett@Dartmouth.EDU (Steven D. Ligett)
Subject: Re: Radical video idea
To: riscy@pyramid.com
Status: RO
X-Status: 

First -- there are too many ideas zooming by for me to sort through them all. 
For now, I'll concentrate on video.  Other facets later.

I think that the idea of having a 3041 generate the video syncs IS too
radical, but is the kind of thinking needed to come up with clever designs. 
Let me expand a little on the video system I was thinking about.  Here is the
list of parts needed for the video system:

2 Mac-compatible 512K VRAM SIMMS  $30/ea
1 TI tlc34076 RAMDAC              $32/ea
1 NSC LM1882 Video sync generator  $8/ea
                                 ----
total                            $100

VRAM appears to the processor as DRAM.  No differences, really.  However, you
can't easily find 60 ns VRAM, and I can get 8MB 60 ns DRAM SIMMs for <$300
any day of the week.  Which is a long way of saying that the VRAM bank(s)
will have different timing from the DRAM bank(s).

The 34076 handles 1, 2, 4, 8, 16, and 24 bits per pixel.  1, 2, 4, and 8 are
translated by a table into 24-bit pixels.  The 34076 completely handles the
serial (video) side of the VRAMs, having only to be set up once to tell it
exactly what to do.  The 34076 does not generate hsync, vsync, blank.

The 1882 generates hsync, vsync, and blank.  Just tell it once what the
timing is to be.

Once the registers in the 34076 and 1882 are set up, the processor(s) just
read and write the VRAM.

For rock-steady video, drive it from a crystal oscillator.  I expect that
there will be a 80 mhz oscillator on the board, so use that.  Giving 1024 x
768 pixels at 72 fps.  Leave space at the 34076 for other oscillators so
folks can use other resolutions - the 34076 can handle 5.

 
From SMACKINLA@cc.curtin.edu.au  Ukn Jun 29 01:31:30 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA28019; Tue, 29 Jun 93 01:31:21 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA05196; Mon, 28 Jun 93 22:29:12 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA19342; Mon, 28 Jun 93 22:28:41 -0700
Received: from cc.curtin.edu.au 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA04730; Mon, 28 Jun 93 22:28:23 -0700
Received: from cc.curtin.edu.au by cc.curtin.edu.au (PMDF #3256 ) id
 <01GZYA7K3RMOFUMLVJ@cc.curtin.edu.au>; Tue, 29 Jun 1993 13:27:20 +0800
Date: 29 Jun 1993 13:27:19 +0800
From: Pat Mackinlay <SMACKINLA@cc.curtin.edu.au>
Subject: Re: Radical video idea
To: caret@pyramid.com
Cc: riscy@pyramid.com
Message-Id: <01GZYA7K419UFUMLVJ@cc.curtin.edu.au>
X-Envelope-To: caret@pyramid.com, riscy@pyramid.com
X-Vms-To: IN%"caret@pyramid.com"
X-Vms-Cc: R3KPC,SMACKINLA
Mime-Version: 1.0
Content-Transfer-Encoding: 7BIT
Status: RO
X-Status: 


>If we add a 3041 as Steve has been suggesting, why not have it
>generate all the timing signals for the video itself?  This CPU

I have only one thing to say about this idea: yuk! <grin> I would also have 
to say that personally, I wouldn't want to write the various interrupt 
handlers. Can you imagine what would happen when the thing crashed?

>For that matter, why not have it deal with the keyboard directly
>also?  The hard thing about the IBM style keybaords is that they

I think it might be better to use (if absolutely necessary) a PIC to handle 
the keyboard. It would be better to use an 8041, as in the AT, but I have 
no idea about prices.

>If the software is going to be so real time as we suggest, then
>maybe it could also do the necessary cycles to refresh DRAM.

This is getting worse and worse <grin>. Again, what happens when the 3041 
crashes? I shudder to think... Interrupt latency takes on a whole new 
meaning...

Pat -- "There's only one thing left to do Mama, I got to ding a ding dang
	my dang a long ling long" (Jesus Built My Hotrod -- Ministry)
GCS d* -p+ c++ l++ m--- s+/- !g w- t- r

 
