Re: TI-H: Multiple device control


[Prev][Next][Index][Thread]

Re: TI-H: Multiple device control



Christopher Kalos wrote:
> 
> Having absorbed all the data I could on the SPinTerface, the Expander
> SF, and just about every other insane add-on devised, I decided that we
> need to make them all work at once.
> The Expander can be about the size of two camera batteries
> The speaker is about the size of five stacked quarters
> Then we need a bypass for a link cable, and then a fourth peripheral.
> Use a binary counter (2 bits) so that is counts from 0 to 3, overflows
> and goes back to zero.  assign each device a number, write up an
> assembly language driver for this single device, and that's it.  A few
> sleepless nights of hardware interface design ought to do the trick.
> I acknowledge that I'm just about a total neophyte to TI hardware, but
> I think my design is at least halfway feasible.  In fact, use device #4
> as a PCMCIA port on the back of the calc, use Rich's battery expander,
> and then toss on the TI-Glow device, and you have yourself a huge device
> (about the size of the calc itself) that can do almost anything!
> My next idea is to tear apart my 386 and use its parts to build a TI
> NetServer.  Of course, I want to wait for a working RT link before I do
> this.

I'm usually sceptical about every new TI hardware idea, especially the
ones presented by people with little knowledge in electronics, but I
think this idea about multiple device control can actually become
something. It's not very smart to use a counter like you suggested
because all the software would have to be rewritten, but you could use a
PCF8574 I2C chip, which is a 8-bit I/O expander. This chip would be
connected to the link port (or SPinTerface port) as usual, and all the
other peripherals would be connected in parallel with it. Then you would
use the 8 I/O pins on the PCF8574 as a sort of Chip Select pins, so that
one pin would enable/disable one device. This would allow 8 devices to
be connected, but one of the devices could be the I2C bus that can
itself hold many more devices. You would use a separate asm program to
choose the active device (or devices). The PCF8574 would probably not
interfere with any non-I2C traffic, and in I2C it would only cause
trouble if another I2C device used its address. The only potentially
troublesome situation would be the event of disabling a non-I2C device;
it might get confused by the I2C data that is sent to the PCF8574 to
switch the active device, e.g. a speaker would crackle for a very short
time.

I haven't yet figured out how the chip select function would actually be
implemented; you could, for example, use AND gates between the 8 devices
and the link port, but there probably exists a better way of doing this,
the many logic chips would take up lots of space.

-Ozone

> Brought to you by the guy who wants to write the TI-8x TCP/IP stack,

I think there are more important projects than this, but if you really
want to, then go for it...just don't try writing it in TI-BASIC ;)

> Christopher Kalos

-Ozone

-- 
*** Osma Suominen *** ozone@clinet.fi *** http://www.clinet.fi/~ozone
***


References: