Re: TI Graph Link Errors


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

Re: TI Graph Link Errors



In article <01bb7da1$e9245660$740780d0@miller>, "Rodger D. Miller"
<miller@theshop.net> says:
>
>I am currently running a Cyrix p-120+ system with 16 MB of ram.  I own a
>TI-92 and a TI-95, and recently purchased a TI-Graph Link.  When I run the
>software for the TI-92 under Windows 95, the mouse freezes on the screen,
>but works fine when I exit the program.  Also, I get "Communication
>timeout" and "Transmission Error" messages in both the TI-92 and the TI-85
>software when I try to send or receive to the appropriate calculator.  I am
>using my only free com port, which is com port 3. Anyone have any ideas
>why?
>


In addition to the need to have your serial port(s) addresses
and IRQ's correct, i.e. no conflicts as another followup to
your post points out, there is another possibility. It could be
that your serial port(s) are unable to supply the 5V current
necessary to power the RS-232 driver/receiver IC that is
embedded in the link cable. Im my case I have *always* been able
to run my link from my pc here at work no problem (once I got the
IRQ set right). From my home pc it has never worked from com3
(my only free com port). Once I got the addresses and IRQ's straightened
out I confirmed that com3 did work by connecting it to an external modem
and sending a file. Still no link however. At this point I booted DOS
only and removed the mouse from com1 which is on a *different* card
from com3/com4 and connected the link there. Bingo! Works fine.
(I'm using the DOS version of the link program which came with the
link from TI.) This is unacceptable of course but I don't know how I'm
going to get around it yet. The link85 program must drive one of the
logic lines in the com port high to power the RS-232 chip in the
cable. I don't happen to have the schematic of the home-brew serial
link with me so I don't know which pin that might be. Note that there
are no 'power' pins in the RS-232 standard, only signal lines as
implemented in pc's. According to my TTL data book a typical NAND gate
can only source -.4 miliamp or so for output high. A HEX driver (7407)
can deliver about .25 mA. The MAX220 RS-232 driver/receiver requires
0.5 mA typical @ 5V. The MAX232 device needs 15 mA. This problem must
not be too common however or we'd have heard more complaints here.
The parallel port link is sounding more attractive now as there
is no power required as I understand it. Just a couple of diodes.
This study is by no means definative (only two pc's and three comports
involved) but I've discovered the problem to my satisfaction
at least. Your milage may vary.


Bill


References: