Re: Why did Woz use the long cycle?

[Mike Willegal <mike@willegal.net>]

Hi Ferdinand,

VA already toggles every line.  It counts just as /HPE is asserted, and 
then holds it's own value as the horizontal counter is preset.  VA isn't 
preset when vertical preload occurs, but since there is an even number 
of vertical cycles, it doesn't need to be preset at that point, either.

Jim Salter's schematic in his book is mistaken when it connects /HPE to 
VA and the VA preset line.  /HPE is not connected to either.

Anyway your idea of eliminating the long cycle and doing an XOR with VA 
for the A8 chip input may work for low res graphics. Color phase lock 
with text mode is not important.  HI-RES graphics becomes even more 
confusing than it already is, though.  This might have been the deciding 
factor, as others have already pointed out.

Regards,
Mike Willegal



ferdimh@gmx.de wrote:
> Hi Mike!
>
> Because VA is on the same '161 as the high horizontal bits, it will be
> preset every line.
> If there is a long cycle, VA will always preset to 1 because COLOR REF
> is always high when preset happens.
> Now assume that there is no long cycle. The phase of COLOR REF toggles
> every line, so VA will be preset to 1 and 0 in alternating fashion.
> The counter works as before, but in sync with the COLOR REF Phase
> shift.
> If VA is high, this is a line with COLOR REF in phase (like in the
> original Apple). If it is low, COLOR REF is out of phase. This is
> useful in HIRES programming, where the compensation for the phase
> shift needs to be done in software.
> In LORES, HO sent to A8-14 selects in phase or out of phase LORES
> patterns to compensate for the phase shift that happens every video
> cycle. By XORing that with VA' (it needs to be inverted, I just found
> out) the pattern selection is inverted every second line, in sync with
> the COLOR REF phase shift.
>
> BTW, tying C2-5 high would cause that no long cycle happens.
>
>
> Ferdinand Meyer-Hermann