Date   : Thu, 10 Aug 2006 09:06:44 +1000
From   : "Richard Wilson" <rich@...>
Subject: Re: Cursor sizes

Hi Jonathan,

Thanks very much for the info, it has clarified things quite a bit.

> No, the cursor isn't delayed after any previous delay. The cursor is
> delayed from the same place the output is delayed from:
>  
> Bitmap modes:
>                     +--+  +--+  +--+  +--+  +--+
>                   --+  +--+  +--+  +--+  +--+  +
>                     Fetch
>                     Display
> cell 1 cursor       -------
> cell 1+2 cursor     --------------
> cell 1+2+3+4 cursor ----------------------------
>  
> Teletext mode:
>                     +--+  +--+  +--+  +--+  +--+
>                   --+  +--+  +--+  +--+  +--+  +
>                     Fetch
>                           Display
> cell 1 cursor       ------
> cell 2 cursor             -------
> cell 1+2 cursor     -------------
> cell 1+2+3+4 cursor ---------------------------

>From the SAA5050 documentation:

The character output time from the SAA5050 is a huge 2.67us, and there's
an
extra latch involved in the curcuit, possibly accounting for an extra
0.33us
or so, which means the actual output of the SAA into the SAA is nearly
3us
delayed. This accounts for the 6845 cursor delay of two clock cycles
(2us),
plus the extra delay of 1 character by not using Bit 7 of the ULA
control
register. It appears the ULA doesn't use the DISPEN from the CRTC
internally
when in teletext mode, but DISPEN provides the LOSE signal to the SAA,
and
I'm presuming the SAA doesn't output RGB 2.67us after the LOSE line goes
low.

The 6845 HSYNC position is changed to suit teletext mode by only 2
character
periods. On a real BBC, this may cause the display to shift right by one
character in teletext mode. Is this actually observed?
 
> > Another question, if anyone can answer it. Is the cursor size
> > determined by the CRTC clock rate, or by the pixel clock rate
> > (Bits 2 and 3 of the ULA)?
>  
> The CRTC clock rate as set by b4. To see this, observe that 
> the ULA is set
> to:
>         7 6 5 4 3 2 1 0
> MODE 1: 1 1 0 1 1 0 0 F
> MODE 4: 1 0 0 0 1 0 0 F
>  
> The pixel clock rate is the same (40 chars per line), but the 
> MODE 1 CRTC
> rate is twice the MODE 4 CRTC clock rate and consequently, the MODE 1
> cursor width is twice the MODE 4 cursor width.

Yeah, that was a silly question really, to which I already knew the
answer
:)

Another quick question though.. How does the palette affect the cursor
display in the different modes? Can you change the colour of the cursor?
I
am getting a cursor shown in Revs, and I can't see any reason it
shouldn't
be shown, but the entire palette is set to blue and other data doesn't
display on the screen.

Is it possible to change the palette and get a red cursor (for example)
in
Mode 7?

Cheers,
Richard