Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!qt.cs.utexas.edu!yale.edu!newsserver.jvnc.net!darwin.sura.net!paladin.american.edu!howland.reston.ans.net!usc!sol.ctr.columbia.edu!eff!world!gkulosa
From: gkulosa@world.std.com (Greg A Kulosa)
Subject: Re: robot parts: info and question
Message-ID: <C1sq2I.DLC@world.std.com>
Organization: The World Public Access UNIX, Brookline, MA
References: <1993Jan31.155957.7778@sbcs.sunysb.edu> <1993Feb1.184033.19691@hpcvca.cv.hp.com>
Date: Tue, 2 Feb 1993 00:44:41 GMT
Lines: 109

abrown@hpcvcec.cv.hp.com (Allen Brown) writes:

>: I sure wish there was an equivalent volume on the genghis/attila/hannibal
>: series.

>I agree.

>I have a few magazine articles

>The C Users Journal, November 1991 p. 42
>  "A Multithreading Library In C For Subsumption Architecture"
>  by Richard F. Man

I believe that this paper is also available by FTP (I know I have it
at home).

>: Is there a reason you didn't consider servos? Price?

>They were the first thing I looked at.
>  positives:
>    Easy to drive electrically
>    Easy to mount and already
>    Geared to the right range
>  negatives: cost (the cheapest I have seen was $20 on sale),

I recently got Tower Hobbies catalog where their 'standard' servos
were on sale for $11.99.  I believe that regular price is ~$13.  Of
course, you still need at least 12 of them for a six legged robot!

>    Not very powerful.  The motor we selected is probably four times
>      as powerful as the largest hobby servo I have seen.  And the
>      non-hobby servos are astronomical in price.

I believe Genghis used standard R/C Hobby servos, but I don't know how
much he weighed.  

>    No feedback.  There is no way to to determine if the motor has
>      stalled before reaching its target position.  This was what
>      killed servos as far as I am concerned.

True, and that also kills one of the main benefits of servos; that you
don't need any feedback to move them to a specific place.  Of course,
you could put a current sensor on the power supply to the servo;  When
the motor stalls, the current draw will go way up.  Standard R/C
(Radio Control) Hobby servos use 3 wires;  5V, GND, and Signal.

>    Cannot run the motor beyond spec.  In a servo the power supply is
>      for the control electronics is the same as the motor.  This is
>      not good for noise.  It is also not good if you want to run the
>      motor at a voltage beyond its rating.  I am considering this to
>      allow a greater power to weight ratio.  Of course doing this
>      means the robot cannot run continuous duty.

>: Is the miniboard too expensive a solution? It does the same thing, and
>: has a pretty nice development environment :-).

>Our main programmer (and my main partner) has lots of experience with
>the Pic.  I think he has a development system.  So that should not be
>an issue.

>I don't know much about the miniboard.  But I got the impression that
>it has no ADCs.  Is that true?

No, the 68HC11 CPU has 8 A/D inputs on board.  It also has an RS-485
serial network type connection which would be a good way to
daisy-chain several boards.  Perhaps one Miniboard per 2 legs?  With
one master controlling them all?  Also, the 68HC11 has 8 timing
outputs which can be set-up to control the aforementioned R/C servos.

>:: We have also not worked out how to measure foot force.  But we know
>:: two of the constraints: its gotta be small and cheap.

>: There seems to be a deafening silence in c.r about this issue.
>: Shane Bouslough       | #include <stddisc.h>

>OK.  Ideas:
>  Mount piezoelectric plastic on four sides of the leg strut to
>    measure bending.
>  Use conductive foam under the foot.  Its resistance decreases as it
>    gets compressed.

This is my favorite idea, and the one that I will probably use.  The
foam is the conductive stuff that CMOS (and other) chips come packed
in.

>  Place a rubber pad near the bottom of the foot.  Put metal plates on
>    top and bottom of the pad.  Measure the capacitance between the
>    plates.  This is actually easier than it sounds.  Use the
>    capacitor as the timing element in an oscillator and measure the
>    output frequency.

This sounds good too.  

Another question;  Will you be using one of the PIC's for the 'master'
of all the others?  Or will you be using a traditional CPU?

>Thanks for the feedback.
>--
>  Allen C. Brown  abrown@cv.hp.com or hplabs!hpcvca!abrown or "Hey you!"
>      Not representing my employer.
>    The Constitution only guarantees the American people the right to
>    pursue happiness.  You have to catch it yourself. --- Benjamin Franklin


-- 
Greg A. Kulosa		|  "If we are to be damned, let's be damned for what
Systems Administrator	|   we really are"    - Jean-Luc Picard
Oberon Software		|___________________________________________________
gkulosa@oberon.com	 One Memorial Drive, Cambridge, MA 02142
