Re: TI-H: TI Port Doubler & Data/Address bus conections
[Prev][Next][Index][Thread]
Re: TI-H: TI Port Doubler & Data/Address bus conections
Richard Piotter wrote:
>
> What about having the data alternate between two types, and have a
> simple multiplexer/demultiplexer switch between two devices on each
> clock signal. You would loose half the speed, but the Multiplexer could
> send a clock pulse to the recieving device so it would end up running
> half the speed, but allow two or MORE non-I2C chips or devices. Also, I
> have noticed that I2C chips don't affect the normal transfer. You could
> also use the paralel I2C 8-bit chip to select 1 of 8 dual signal
> multiplexer chanels (2x8 bidirectional multiplexer). The I2C chip
> shouldn't interfere with the other chips transmision.
I know using the TI protocol should not affect I2C chips, but how about
using the I2C protocol when devices that use the TI protocol are connected
to the same link port? Knowing both protocols I'd assume this to be
disastrous as the TI protocol device (say, a Graph Link) would acknowledge
each high-to-low transmission by pulling the other data line low.
I haven't tested it, though.
> The Memory cart on the 92 has to have the data AND address bus leading
> to it. If someone knew these, paralel devices could be connected
> directly to the TI-92's data and address bus and you could use the free
> addresses to access those add ons! It would be like all the custom
> devices that people used to build for those old 8-bit computers. One
> idea would be to add a SID chip (like in a Comodore) and a tiny speaker.
> You could set it to an unused address and be able to create wave form
> generated sound of different loudneses and with 3 voices! You could
> probably even add the Universal Bus Extender Card directly with the
> calc!. The UBEC was designed a long time ago, but lets you control 1500
> something I/O ports using only 256 bytes of unused addresses (I think).
> It has Input and output cards and DAC and ADC cards too. You can even
> buy the boards, pre-etched from somewhere.
> I suppose it's only an idea, but imagine what you could do with direct
> data/addres access!
Yeah, sure, and most people trying to do this would fry their calcs. You'd
need some sort of a buffer in between the processor's data and address pins
and the port (UBEC or whatever), otherwise you'll quickly destroy something
crucial with e.g. a burst of static.
Besides, what do we really need UBEC for? Maybe it's a standard, but does
it give any advantages?
-Ozone
--
*** Osma Suominen *** ozone@clinet.fi *** http://www.clinet.fi/~ozone ***
References: