> >Unless there's some case when the 2600's data bus would not contain $0f?
>
> I like the idea. But it is possible for the 2600 to get rubbish or '0's
> floating on the databus too.
> It depends very much on the hardware pluged in and is not very reliable.
> That the last byte still floats on the databus has something to do with
how
> fast a rom can change its output state and how much load there is on
databus
Also not all 7800s will zero out the undefined bits on the TIA read
registers. This is just a side effect of a hardware fix that Atari
build into later models of the 7800. This helped with some problems
that the games with cartridge RAM had, but broke compatibility with
Activision's FE bankswitching and the Supercharger. The origial
7800's would react just like a 2600 on a TIA read. Also some people
disabled the extra hardware (only requires removing a certain
capacitor) to make their 7800 compatible with the Supercharger
and the Activision games again. So this is not a very reliable
method of detecting 7800s.
BTW, if pause is such an important feature, couldn't you use
some other switch to enable it? Or are they all used for
something else?
> I've also been looking on the schematic for a possible 7800 hardware
> detection. But found nothing concrete yet. I'm wondering though:
>
> The RIOT is addressed using the MARIA chip would there be any change on
> small incompatibility in the memorymap mirroring ?
No. In 2600 mode the memory map is exactly like on a real VCS.
> Memory configuration is externally locked. But MARIA seems not locked be
it
> will some MARIA fuctions still be avialable?
> NMI line is controlled by MARIA chip. Is there a possibility to trigger a
> NMI ?
The only way to trigger an interrupt on the MARIA is to have
the DMA find an "interrrupt bit" in the display data for a
certain scanline region. I think DMA is disabled in 2600 mode,
so you shouldn't be able to trigger a NMI. And since the DMA
is the only way to make the MARIA do anything, you shouldn't be
able to get a reaction from the MARIA, even if some of it's
write registers are still active in 2600 mode.
BTW,
How does Video Pinball play on your EPROM cart? That game
intentionally relies on the state of the undefiened bits
in the TIA read registers. If they don't stay like the
address of the register, then the ball shouldn't bounce
off the flippers anymore. We had this problem in the emulators
untill we simulated the TIA read registers properly.
Ciao, Eckhard Stolberg
