Re: 74ls612, clocking 6502and others...

From: Marko Mäkelä (msmakela_at_cc.hut.fi)
Date: 1998-04-22 23:24:05

On Wed, 22 Apr 1998, Andrew Vardy wrote:

> Also, the same thing occurs with the C128 in 2Mhz aka Fast mode.  At
> least, if I unplug my Swiftlink.

Unless you know the horizontal raster position, it will only work with
probability 125/130 in your case (121/126 on PAL machines).  Remember the
five refresh cycles per line during which the CPU is slowed down, and the 
video chip will read from $3FFF-$3F00.

> Without unplugging anything, I typed:
> 
> A1300
> LDY #$FF
> LDA $D720,Y
> BRK
> 
> GF1300
> 
> The result: Acc. is left with $FD.
> 
> (And the color in color RAM was all $D.)

Of course, if you are reading the color RAM ($d81f), you will get data
from there. BTW, the 4 high bits from color RAM are not random.  I've
written a program that runs in color RAM.  Most of the color RAM is filled
with zeros, and the video bank is initialized so that it will supply
opcodes like $60 most of the time.  The stack is filled with $DA I think,
so the program will be fetched from $dadb most of the time.  It took me
two weeks to write the program.  It changes the screen color when you
press space.  It is almost trivial to write a program to $de00-$dfff once
you know all of the video timing, but writing one to color memory is
extremely difficult, since in that case also the value of the program
counter matters.

	Marko

Archive generated by hypermail 2.1.1.