Hi again,
Jon De Schrijder <[EMAIL PROTECTED]> wrote:
> HARDWARE PROPOSAL:
> ------------------
> In stead of building an internal 3.58MHz clock in all SCC's, Audio's and
> Musics it would be convenient to have a fixed 3.58MHz clocksignal
> available on the cartridgeconnector and also a 'BUSCLOCK' which reflects
> the processor's speed. My proposal is to use pin16 for this (currently
> 'Reserved')
About setting a standard for this: I meant you can't change standards
that have already been set (by having in the order of a million MSX
computers based upon it). I didn't mean you can't set NEW standards,
that's okay, and this mailing-list is a good medium for discussions
about such new standards.
What I really meant: if you have this many machines based upon such a
standard, and you change 'the rules of the game' after that, it would
mean that your new standard either wouldn't be one, because there
would be a million or so computers not obeing it, OR that new
standard would make these million or so machines suddenly
non-standard. This is just a principle, no in-between.
> > If you do, I suggest you make it switchable, such that in one
> > state, it works like non-modified (like with turbo-circuits:
> > switch it off, and it works like it wasn't there).
> No need for that, because there is a 3.58MHz clock available at
> pin42 like the MSX standard requires.
The MSX standard requires "CPU clock, 3.58 MHz." on pin 42 (which
would it be?). You might consider this 2 signals in one, usually the
same. The idea discussed here, comes down to making this 2 separate
signals, CPU clock, and 3.58 MHz.
What's next, is to consider what should be on pin 42: CPU clock, or
3.58 MHz.? You suggested the latter, and that could well be easier in
practice.
Interesting detail: according to my Turbo-R service manual, the
cartridge-slot clock signal in this machine comes from the S1990
Turbo-R engine, through pin 123 (SLTCLK, can anyone verify this?).
The fact that it comes out of the engine, I think might mean it's
PROBABLY switched, or switch-able in frequency (it would be nice if
anyone could shed some light on that), otherwise it would more likely
have been hard-wired to a 3.58 MHz. signal. That would suggest you
could consider the signal on pin 42 as 'CPU clock'.
Then, think of this: suppose fixing the signal on pin 42 to either
CPU clock, or 3.58 MHz., would make the 'other' signal unavailable.
What could you do with a cartridge that requires this signal anyway:
-If it needs 3.58 MHz.: put a 3.58 MHz. oscillator inside
-If it needs CPU clock: tough luck, you can't reproduce this from the
other signals on the cartridgeslot, nor generate it.
That would also hint towards fixing the pin 42 signal to "CPU clock".
> Not so negative! 'Reserved' signals are intended for future
> expansion! Don't Fight the Future ;-)
Yes! Great! Butteh...if you want to use/define these pins, then why
not put on there:
-A signal allowing the use of Direct Memory Access (for fast
harddisk-interfaces etc.)
-A signal allowing the INPUT of a RESET-signal (so that a cartridge
might reset the computer, if for instance, it detects a 'dip' in the
power supply)
-A signal allowing a cartridge to generate non-maskable interrupts
-A signal that might be used to replace a ROM built in the computer,
with one in the cartridge
Just a few ideas (you can all come up with more like these)
Using one of only 2 reserved pins for a CPU clock or 3.58 MHz.
signal, sounds more a less like a 'waste' to me.
> AFAIK the only cartridges that will need modification(one wire
> changed) are the external memorymappers if you want to use them on a
> modified MSX (or in other words on a 7MHz MSX)
Not the Mega-mappers, by Digital KC, they don't use the clock signal
(see schematic in MSX Computer Magazine 65).
Just some other cartridges that MIGHT give trouble:
-modems/RS-232/MIDI-interfaces, if they use 3.58 MHz. to generate
baud-rates
-harddisk-interfaces; will probably have their own clock,
or don't need one, but maybe....(take your guess)
-???
> I only can see the problems today: you can't enjoy your 7MHz
> expansion because your music is screwed up.
Possible solutions:
You probably don't like the need for another switch on your computer,
certainly not for something like this, and you'd rather make as
little changes as possible to both computer, and cartridges (and no,
preferably no extra jumpers in there either). How about:
-Fix the cartridgeslot clock signal(s?) to 3.58 MHz., and modify
cartridges using "CPU clock", so that they don't need this signal
anymore. For memory mappers, for instance, you could:
-use an internal mapper (duuhhh...)
-use a Digital KC megamapper
-modify your mapper, to use an other way of generating DRAM control
signals (several ways to do this, but might require some research). I
have a 'home-built' mapper myself, with 2 DRAMs of 256K x
4 bit each (256 K mapper, thus) which I changed such, that it doesn't
need CPU clock either, and works fine on both turbo or 3.58
MHz.
-construct a mapper using SRAM, not needing "CPU clock" for any
purpose either
-Why not fix one of the cartridgeslot clock signals to 3.58 MHz., and
the signal on the other cartridgeslot (most MSX's do have 2, don't
they?) to "CPU clock". With turbo on, you can put cartridges
(un-modified) using "CPU clock" in the one slot, and cartridges
(un-modified) using 3.58 MHz. in the other slot. Just try
possible combinations of cartridges you have, then decide
what signal to put on each cartridgeslot. Cartridges not using either
one can be put in both, and with turbo off, it's all the same, and
anything should work.
-Or use a combination of the above
Greetings,
Alwin Henseler ([EMAIL PROTECTED])
http://huizen.dds.nl/~alwinh/msx (MSX Tech Doc page)
http://www.twente.nl/~cce/index.htm (computerclub Enschede)
****
MSX Mailinglist. To unsubscribe, send an email to [EMAIL PROTECTED] and put
in the body (not subject) "unsubscribe msx [EMAIL PROTECTED]" (without the
quotes :-) Problems? contact [EMAIL PROTECTED] (www.stack.nl/~wiebe/mailinglist/)
****
