Re: TI-H: TI-Modem one big reply

To: ti-hardware@lists.ticalc.org
Subject: Re: TI-H: TI-Modem one big reply
From: Bryan Rittmeyer <bryanr@flash.net>
Date: Thu, 07 Aug 1997 09:34:42 -0400
References: <199708070246.TAA14551@f51.hotmail.com> <33E96318.684D59C8@tir.com>
Reply-To: ti-hardware@lists.ticalc.org
Oh joy, here we go again...

Damien wrote:

> Maybe I'm wrong on this, but, doesnt the internals of a computer run on
> 5 volts (internal modem). It would be about 6-9 volts for an external

Most PCs run on +5, -5, +12, and -12 VDC. But yes, essentially, +5V is
the most used.

> one. The only problem you may have is with the amperage needed for the
> modem. Also, about trying to get 9 pins down to 3 on a serial port.

Uhhh, a modem is basically a minicomputer, and its not sucking up any
kind of vast amounts of power. Of course, it probably wouldn't last too
long on batteries.... But there are PCMCIA (sp?) modems that have pretty
low draws.

> Easy. It's called multiplexing (I think). Mel Tsai used a form of it on
> his previous version of his expander when he used older serial EEPROMS.

What an absolute waste of electronics that would be... First of all,
that
"9" pin connector only contains 3 useful pins (for our purposes anyway)
-
Send, receive, and ground. This is essentially why you can use
a graph link, null modem and gender changer, which have 25 pins, to
connect
a modem to a calculator with 3 "pins," in which the graph link converts
between TI's (screwed up) asynchronous protocol to normal RS-232(C).

Believe me, the graph link does not have some huge multiplexing IC
inside
of it to allow the calc to control 25 I/O pins.

There is no NEED to multiplex (nor was their in the original expander,
as Mel fortunately figured out ;-)

> And it was fast (for a TI calculator). I think he said he estimated
> about 10k a sec thoughput, but I could be wrong. Still, thats faster

Well, you have to remember... the output of a multiplexing IC is still
not RS-232. If we did want to multiplex -- and we certainly do not --
you would still need some kind of device to convert the multiplexed
output into RS-232. And I can almost guarentee when you were done
you would have sliced the bitrate to pieces (as the graph link and
its bloated PIC (?) code, in all its glory, manages to do very nicely)
This is why the parallel link is so much faster. Its just not wise
from a speed point of view to convert an asynchronous protocol to
a synchronous protocol.

> Maybe you should research things a bit more carefully before shooting
> your mouth off. There are many people on this list who are extreemly
> well versed in electronics and have been working on projects like this
> for some time. I think they know what they are doing.

<soapbox opinion="true" ifoffended="go_to_next_message">
My personal view is that extra input is always beneficial, even if
incorrect. It keeps the "experts" (a group to which I certainly do
not belong) on their toes, and lets the new people learn, however
slowly, the tricks of the trade. We could certainly do with more
intelligence and less flaming around here, and the two are
connected.

The original poster's idea was certainly implausible. However,
one only has to look through a history book to find great
scientists who were laughed out of the scientific community
for an revelation that later had a huge impact on their field.
</soapbox>

Regards,

Bryan Rittmeyer
bryanr@flash.net