Newsgroups: comp.robotics
Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!hellgate.utah.edu!croc.utah.edu!gunderse
From: gunderse%croc.utah.edu@cs.utah.edu (Eric Gundersen)
Date: 29 Aug 90 12:38:55 MDT
Message-ID: <1990Aug29.123855.2035@hellgate.utah.edu>
Sender: gunderse@cs.utah.edu (Eric H. Gundersen)
Organization: University of Utah CS Dept

Subject: Re: PUMA 560 AND IBM PC/XT


In article <5664@quanta.eng.ohio-state.edu> CHIN@rcgl1.eng.ohio-state.edu (Pei-chieh Chin) writes:
>
>	Hi!  I am working on a project which needs to connect a PUMA 560
>robot and a supervisory computer.

  	I have done a vax 750 to puma interface. I looked at the network
method and decided it looked to complicated for the time I had to work on my
project. I decided to use the teminal serial port and just have the vax send  
straight val II commands (after issuing a 'disable interactive'- just stops 
user friendly messages from fouling things up). I'm sure it was easier but 
has it's problems (time delays being the biggest, far from real-time 
controll).
	I would also be interested in how the network system works etc. I'm
sure it has many advantages.

	To comment on your real question, your problem may likely be a 
serial line parameter (parity, stop bits etc) problem. I noticed that when
I had a parity problem, I was reading in only half the characters but it 
would write fine without any problem. Scanf's (in C) also are pretty picky 
about CR and NEWLINE. Scanf's can skip thier fields if they get a newl
or cr that a previous scanf didn't "use up". Any way I found serial line
communication to be a real issue. Try hooking up a spare terminal's input
lead (pin 2 or 3, I can't remember) to each of 2 and 3 (one at a time)of one 
of your serial line connections to see the data  being read and written. also
note that you wont see multiple CR's. (one of my problems was one of my DTE's
was sending a \n\r\r for an end of line and I didn't know there were two \r's
and so my next scanf was skipped). Anyway have fun.

Eric Gundersen
 


