Alt-BEAM Archive
Message #00054
To: "'Noam Rudnick'" rudnick1@cwix.com, beam@corp.sgi.com
From: Wilf Rigter Wilf.Rigter@powertech.bc.ca
Date: Sun, 31 Jan 1999 10:24:25 -0800
Subject: [alt-beam] Re: Adaptability in Two Motor Walkers
Hello Noam,
If the front legs are at the usual angle they will not give much problem
after tuning the corresponding core components. Gravity will do the rest to
center the front legs. (I guess the back motor could also be slanted in the
same direction but I haven't tried this) Anyway the effect is similar to a
pendulum, even if you push it only one way it will swing around an average
position.
For the rear legs there is (generally) no gravity to set things right but
you could add a spring to each leg which would store energy (like the
pendulum) and force the legs to reciprocate around an average (rotary)
position. There is no positional feedback inherent in the electronics,
however there is the Piotter fix of the servo type motor using the pot as a
limit switch. The value of the pot is small compared to the normal >1M
resistor values in a Nv , so he cut the pot in 2 places near the end of the
carbon tracks creating in effect a rotary switch with the sliding contact
connected to ground and the two ends connected to the corresponding Nv
"bias points" . Other people have used limit switches which can be two
pieces of bare wire "free-formed" to contact the bare wire of each leg at
the end of travel. Each wire is connected to the corresponding Nv "bias
point" and the leg itself is connected to ground. I like this last fix the
best because of its simplicity but it's a little like wearing your brain on
the outside: just don't bang you head (wire). (hehe!)
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: Noam Rudnick [SMTP:rudnick1@cwix.com]
> Sent: Sunday, January 31, 1999 7:51 AM
> To: beam@corp.sgi.com
> Subject: Adaptability in Two Motor Walkers
>
> So I built my first walker using those sweet bgmicro lens motors,
> and It
> works pretty well, but due to some variations in the exact values of my
> components (they all say they are the same value, just slight errors in
> manufacturing probably), each legs goes one way slightly more than the
> other way. This behavior is tolerable for short periods of time, but
> eventually it causes the robot to flip over (quuite interesting to watch
> actually =)) Anyway, this problem could probably be easily fixed by
> adjusting the RC time constants with a potentiometer, but if I where to do
> this, then the robot would have trouble when it encountered terrain which
> required a slightly different walking gait. So this is where I thought
> feedback came in, until a few weeks ago when I was told that the 245
> chip's
> inputs are isolated from its outputs. Then I was told that by adding 47K
> resistors across the inputs and outputs of the 245 chip, one could achieve
> feedback. So I tried this, only to find virtually the exact same behavior
> in my walker as before. I really would like to give my robot the ability
> to adapt to terrain, and I was wondering if any of you could help me out.
> Currently, I am using three stacked 245 chips because I needed all of them
> with my old motors, and I never bothered to take them off. Would this
> make
> any difference? Should I choose resistances lower than 47K because of
> this? I would really appreciate any advice.
>
> Thanks a Lot,
> Noam
>
> P.S. I decoupled all of my chips from the power supply as well, but I
> really can't imagine that this would make much of a difference...would it?
>
> thanks again
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/list/alt-beam
Free Web-based e-mail groups by eGroups.com
Home