Newsgroups: comp.robotics
Path: brunix!uunet!gatech!hubcap!dawill
From: dawill@hubcap.clemson.edu (david williams)
Subject: Re: Mini Board serial line
Message-ID: <1992Aug28.220606.5435@hubcap.clemson.edu>
Keywords: Miniboard
Organization: Clemson University
References: <1992Aug27.142429.16553@hubcap.clemson.edu> <1992Aug27.205711.23050@news.media.mit.edu> <1992Aug28.135419.3725@hubcap.clemson.edu> <1992Aug28.184714.17895@news.media.mit.edu>
Date: Fri, 28 Aug 1992 22:06:06 GMT
Lines: 62

fredm@media.mit.edu (Fred G Martin) writes:

>You are absolutely right, the Mini Board serial circuit will only be
>reliable in half-duplex protocols, where each side knows whether it's
>its turn to transmit or receive.

>This isn't so horrible, though.  For most applications, this works
>just fine.  See for example the protocol used by the DLM downloader,
>or the HEXMON program supplied with the Mini Board.

   True, but it still bugs me.  Hey, the RS-232 standard calls for
bidirectional transfer.  We've taken away that possibility.

>>   Even worse, what if we try to hook the MiniBoard up to another 
>>serial port that works the same way? (sending side supplies the negative
>>voltage) - No communications will be possible! 

>Believe it or not, two Mini Boards can communicate to each other.
>Think about it:  the Mini Board's own circuit doesn't need a negative
>voltage for serial line true.  The "negative" voltage will simply be
>zero volts (there is a 10K resistor pulling the RxD input to ground).
>When either board needs to transmit +5v, it can do so, and the other
>board will interpret it properly.

Duh.  [Slaps head with fist]  Oops.  You're right, I hadn't thought
about that.  But here's a question for you:  Since both minboards
are going to be echoing the signal, will we not have problems with 
data corruption from the bits that have made the round-trip colliding
with the newly transmitted bits?  This will only happen on a really
long cable at high transfer rates, so it may be a moot point.  Anybody
have an idea what sort of delay a standard (if there is such a thing) 
RS-232 cable has?  
   Since most miniboard applications will probably have a cable that
is 3 feet or less, or maybe even just a few inches long in a robot, 
this round-trip echo shouldn't be a problem.

>>   yep.  Oh well is right.  Get out the x-acto knives when you get 
>>the miniboard.

>I think this comment is a bit harsh.  The circuit works just fine for
>most protocols and most computers, and has been tested on a variety of
>clone and true-blue PCs, Macintosh models, and Unix boxes.  The main
>situation in which the circuit will be problematic is if your host
>computer can't deal with +/-5 volts instead of the standard +/-12
>volts.  And as I mentioned in a previous message, computers with this
>limitation are fairly unusual.

   Ok, I deserved that.  Of *course* you guys have tested it out.
Motorola told me while I was working on my thesis that their 1488/1489
RS-232 level shifters (which just about every serial application I've
ever seen uses) can handle RS-232 levels of +5v and -0.5 (!) just fine.
   Still, I can envision some situations where I would want
bidirectional high-speed transfer between the controlling computer and
the robot.  Just call it a matter of style, I suppose.  I can't wait
untill my 3 miniboards get here.

>	- Fred
-- 
      Dave Williams
        dawill@hubcap.clemson.edu
           "Huh?  What?  Could you repeat the question?"

