Re: TI-H: SPinterface
[Prev][Next][Index][Thread]
Re: TI-H: SPinterface
<Note, this message was sent to me but I have a feeling Richard wanted
it sent to the list too, so I CC'd my reply to the list.. you should
probably read it all. ;-) >
Richard Piotter wrote:
> A lot of people want to use RF or IR links for chat or terminal use
> etc., but won't that require a driver or ASM program to run it. I got a
> sample sheet for a very simple chip designed for IR data transfer. Add a
> UART and you have a fully opperational IR tranciever. RS232 or whatever
> it is can use just a simple send and recieve line, right? couldn't you
Well, unfortunately, it isn't quite that simple. The TI calculators
use an RC pair for their oscillator, which means that the timing
is not very precise.. in fact its almost impossible to calculate
directly since it depends on humidity, part tolerance, battery
voltage, etc. For this reason, TI engineers had to use an
asynchronous protocol... this basically means that there is no
set "bit rate" as there is in RS-232(C). If you tried RS-232C
on a TI calculator... it would be a total disaster since the
calculator's oscillator is not precise enough to handle the
timing required by RS-232. And even if you did get it to work
on one calc, chances are it would not work on another.
So essentially, no, you cannot use one I/O line for RS-232 receive
and the other for RS-232 send... unless you replaced your calc's
RC pair with a crystal, but this would bring on other hassles.
> reserve one line for sending and the other for recieving? The software
> in the calc would control it, but you wouldn't need an expensive link
> cable or PIC chip that's hard to program. If the thing would be used
> mainly with an assembly language app or driver, then why stick with this
> impossible to use TI protocol. The drivers can use a simpler, more RS232
Well, the "impossible to use TI protocol" is probably the best
asynchronous protocol that can be used with a TI calculator, as far as
I know. If TI used a $0.50 more crystal everything would work... but
crystals are fairly fragile and TI knows how much we drop our calcs. :-)
So, as I said above, its probably not possible to get RS-232C (The C
means that the logic levels are TTL, not the +12 and -12 used by a PC)
from any of the TI calculators.
> Hope Someone can use this
If TI's IR chip is inexpensive enough, I could add one into the
Extender...
or better yet, design a modular extender using a system similar to the
SPinterface... in which the basic extender would have the PIC and the
memory chip and then a small bus, which could support... hmm, about 3
other chips, max. I still have 5 I/O lines free on the PIC... two could
be used for very vulgar RS-232 (though the Extender will probably use
a crystal for RS-232 support, I haven't calculated the bit rate with
any great accuracy and thus the PIC would probably not like the
RS-232 stream from your PC, for example, since the bit rates would be
off a bit) and then the other three could be used to select the chip
that the PIC is talking to. Sounds like an interesting system, I'll
have to give that some thought.
Regards,
Bryan Rittmeyer
bryanr@flash.net
Follow-Ups:
References: