Re: PIC Expander


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

Re: PIC Expander



Andrew Wendt wrote:

> Why not use the graphlink itself? The computer end is certainly not a
mystery,
> it sends normal rs232 info. It's pretty universal.

>From my point of view, using a $30 product in another
product does not make good business sense. The code the
original poster was talking about was probably not more
than 20 lines long, max. It's not overly difficult to
re-code this.

> BTW, does anyone know if TI's calc-to-calc protocol supports full-duplex
> operation?

Nope, not in the true sense of the word. The TI protocol
is basically a lousy protocol, compared to something like
RS-232. This is because the TI calcs have an RC clock, and they
not precise enough for a synchronous protocol.

Here is the basic description of the TI protocol, as I understand it:

Both lines are active low. When one calculator wants to send
a "0" to the other calculator, it brings one line low. The other
calculator then clocks this in, and acknowledges by pulsing the
other line low. Then, to send a "1" the process repeats, starting
with the other wire. In a sense, there is a "0" line and a "1" line.
This does not allow full duplex, or even very high speed.
However, with an RC clock it is the best that can be done,
and TI was too cheap to use a crystal. The other calculator send
and receives in an opposite fashion.

That is my basic understanding of the protocol, as described by
Mel Tsai on IRC. If I am incorrect, please correct that and fill me
in. In any case, I will be forced to figure it out soon enough.

Regards,

Bryan Rittmeyer
bryanr@flash.net


References: