Alt-BEAM Archive
Message #12163
To: beam@sgiblab.sgi.com
From: Daniel Grace lifebytes_98@yahoo.com
Date: Sun, 19 Mar 2000 15:23:27 -0800 (PST)
Subject: [alt-beam] Re: irrelevant...?(A lie)
Strange you would ask to see a show of hands of people
who haven't seen you ask them to raise their hands.
Hmmmm.....
~Daniel
> I would
> like to know how many people have deleted this
> message before getting this far,
> may I see a show of hands, don't be shy raise those
> hands.
=====
ICQ # 39402143
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
12164 Sun, 19 Mar 2000 15:45:51 -0800 (PST) [alt-beam] Yahoo! Chat BEAM List Daniel Grace Does anyone actually go to these? I'm here, but nobody
else is!
~Daniel
=====
ICQ # 39402143
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
12165 Sun, 19 Mar 2000 19:49:10 -0500 [alt-beam] Some mini geared motors on eBay beam@sgiblab.sgi.com "Arthur J. Lekstutis" Sorry for the blatant capitalism, but thought some of you might find
this useful...
I'm currently selling an extra zoom telephoto lens I have on eBay (using
the others for my own projects :-). Has two geared motors that might be
useful to those building small robots. See the pictures on eBay if
interested.
http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=287962792
Later,
Artie
12166 Sun, 19 Mar 2000 17:04:02 -0800 [alt-beam] Re: Funkie Bicore beam@sgiblab.sgi.com Bruce Robinson Max Inggal wrote:
>
> Ok today I was determined to build a bicore walker
> using a Master/slave bicore circuit. But when I
> hooked the motors the the circuit for some reason the
> legs turn more in one direction and less the other
Hi, Max.
The simple answer is, random variation. The caps and resistors in your
circuit don't have precise values ... typically they could vary by as
much as +/- 10% each. If you get really lucky, they might average out.
Chances are, however, that one Nv stays active a little longer, so the
motor turns a little bit further. That's the electronic part.
Mechanically, it isn't unusual to find a gear motor turns a little
faster in one direction than the other ... once again, allowing it to
turn just a little bit further. I have one here that turns 12 degrees
further in one direction than the other direction, for exactly the same
running time.
> anyone know why and how I can fix it.
You need some kind of force that will try to return the gearmotor shaft
to a central position. Ideally you can use a method will not exert too
much force when the legs are nearly centered, so you won't be using up
all your walker's energy just fighting resistance. The geometry you
choose will make a big difference.
The most common method is to use a spring to center a set of legs.
Here's an ASCII layout of a very practical system.
pivot
_________________
0_______ o _______0
leg | | leg
| |
|.| <- arm
|
>
< <- spring
>
|
o <- anchor point
The arm is attached firmly to the legs. When the leg starts to swing
away from center, the spring is barely stretched, so it doesn't have
much effect. The further the legs turn, the more quickly the spring is
stretched, so it begins to exert a lot of resistance when the legs get
too far away from center.
Want a bit of mechanical feedback? make your legs like this, but hook
the two arms together with one spring instead of having two springs
anchored to the body. The legs will influence each other slightly as
they turn.
A second way to get a centering action is to use gravity (this only
works when the motor is tilted, or horizontal. Whenever a leg is off
center, it is lifting part of the robot, so the robot's weight tries to
return it to a centered position. This works best with efficient
(smooth) gearboxes. It also has a slight tendency to counteract a
gearmotor that runs faster in one direction than the other.
With a gravity design, the position of the feet relative to the motor is
important. Here's another ASCII layout. Imagine the motor is between you
and the screen, and the bottom end of the motor (where the shaft
extends) is higher up the screen than the top end (closest to you).
0\ /0
\\ //
\\ pivot //
\\_____________//
\_____ o _____/
leg leg
During the first part of its motion away from center, the motor has a
bit of a mechanical advantage as it lifts the weight of the robot. The
more it turns, however, the greater the resistance (unless it goes TOO
far).
In a typical bicore configuration with one motor vertical and one motor
angled, you can't use this arrangement on both motors.
> PS=I know about attaching springs to the motor shaft
> but looks ugly.
Hide 'em :)
> Beside do you see any springs on Mr. T's walkers?
Yup, sure do. Look closely. Some like Walkman, Lobster, and Strider have
a mechanical layout that can use gravity on all 4 legs. If you see one
with a vertical motor, look for a spring. It's probably there, and it
probably doesn't look ugly :)
Bruce
12167 Sun, 19 Mar 2000 17:03:52 -0800 [alt-beam] Re: HPV Questions beam@sgiblab.sgi.com Bruce Robinson Justin wrote:
> Yep, I really like the look of that circuit, and will hopefully play
> with it i the future. I'm kind of wondering how you plan to get herding
> behavour with it?
> I'm inclined to suspect that what I think of as "herding behavour" is a
> few steps down the track, and this is the first step, so it won't
> actually give herding behaviour, but rather, allow us to achieve it
> later on.
I've had a sneak preview of a circuit that combined my (not too elegent)
sound counter with one of Craig Maynards phototropic designs ...
modified to "hear" sound. Then Wilf came along and cleaned up the sound
counter enormously, but the auraltropic (there's a new one for you)
feature got lost.
> I've only skim-read it, but it pointed out the ability to stop bots when
> they hear enough chirping (ie like a game of "sardines", they eventually
> end up in a herd, but it's a static herd).
Richard's original concept had the robots flocking together. What I
would like to see is auralphobic robots ... robots that turn away from
the last sound heard. Will they tend to distribute themselves evenly
around the park? I think so. And I think they'll tend to keep moving
too, as the sound bounces around the herd.
Want to throw a wrench in the works? Have one that periodically turns
off its sound circuit and just travels for a while in a straight line,
then fires up the sound again.
Or for survivalist characteristics, have them auralphobic when the light
level is good, so they spread out to cover the whole park. Then as the
light level falls, have them auralphilic, with the chirp rate increasing
in proportion to light intensity. Then they'll tend to flock to the
robot in the brightest light.
And for a touch of the bizarre, mount a long needle on each one,
pointing straight ahead. Have a weight sensor that triggers a continuous
high frequency beep if one gets picked up ... a frequency that attracts
all the other 'bots. That'll keep people from ripping off your robots :)
Bruce
12168 Sun, 19 Mar 2000 20:12:31 EST [alt-beam] Re: FRED circuit (beginner) beam@sgiblab.sgi.com "Corey Centen" Ok...I did some of your recommended tests. When I shorted out the FLED on
engine 1(the one with the non-working motor) the FLED stopped flashing but
the motor still did not turn on. For SE 2(the one with the working motor)
The FLED already was not lit because the motor was using all the power, but
the motor would not turn off when the FLED was shorted. Did you recommend
replacing the 4.7uf cap with a 4700uf? or did you mean to hook the 4700
directly to the battery?
Thanx.
>Yes you are bypassing some of the circuit - ALL of it! *grin* To power
>the circuit from a 9V battery, connect up a 1k resistor in series with the
>9V battery. This will keep your motors from burning out and stop your
>poor little 9V battery from having a hernia. You will still need a 4700
>uF or so storage cap.
>
>The motors will kick over pretty quickly with a 4700 uF cap - this might
>be the 'stuttering' you were talking about.
>
>To make the circuit 'fire' manually, short out the FLED. This should test
>most parts of the circuit. If this doesn't work, put a 470 Ohm
>resistor in between the PNP base and ground. If that doesn't make it
>fire, try putting the 470 Ohm between the positive terminal of your
>main storage capacitor, and the base of the NPN
>transistor. Following these tests in this order will help me (and others
>on the list) figure out what, if anything, is wrong with the circuit.
>
>hth,
>Ben
______________________________________________________
12169 Sun, 19 Mar 2000 17:38:19 -0800 [alt-beam] Re: H-Bridge Question (Newbie) beam@sgiblab.sgi.com Bruce Robinson David Simmons wrote:
>
> ... Somehow I didn't think this one was a smokeless.
>
> Kyle is looking at breadboarding the almost complete walker, again from
> Beam On-line, and assumed that the bridges would be a match to that
> circuit.
Yep, they make a good match. It's not too likely that you will turn both
sides of the bridge on with a bicore. I wouldn't get too hung up about
protecting it.
Hey, Sparky. What did you do with Lotus? Bridges hooked directly
to the "complete walker" outputs?
> I assume you recommend spending the little extra on the 2N2907 and the
> 2N2222 transistors for the Fig 6 Bridge.
About half my motors are incapable of overloading the 2N3904/2N3906
transistors. The other's will almost certainly fry them, so I just
bought a bunch of the '2907 and '2222 transistors (like about 40 of
each) and don't worry about it. If you've got an ammeter, you could
check your motor's current draw ... slow the shaft down so you can see
what it draws under load.
By the way, the 2N prefix is obsolete on the '2907 and '2222.
Everywhere I've looked they are PN2907 and PN2222.
> ... What is the engage (-ve connection.
Aha. That enables the decoder (which is what the chip is). The chip has
two decoders (one per side) and each one has an enable, or select pin
(pins 1 and 15) that turns the decoder off. The -ve indicates that
connecting the pin to negative (or ground) enables the decoder ... just
the opposite of what you would expect.
This chip would most commonly be used in a microcore or more complex Nv
net. The enable pin would allow you to turn off the chip (and therefore
the motors) when the Nv net was first going through its startup cycle.
Mine are hooked to GND most of the time.
Bruce
12170 Sun, 19 Mar 2000 21:12:58 -0800 [alt-beam] Re: Funkie Bicore beam@sgiblab.sgi.com Laura Malinowski Laura/Rob Malinowski Several other people are probably writing better answers as we speak,
but my walker did the same thing. Since the capacitor tolerances are
about 10 - 20% I thought I would change them around. When I did the gait
changed also. I hoped in the jalopie and scooted off to that place (the
evil rat place) and picked up a couple packs of caps. I put a set in
that seemed to bring the gait close to equal movement and tweaked the
pots until It walked perfectly (off the edge of a table). Try
interchanging the caps, it just might work somewhat and sortof. It's a
starting point.
Good luck
R.M.
Max Inggal wrote:
> Ok today I was determined to build a bicore walker
> using a Master/slave bicore circuit. But when I
> hooked the motors the the circuit for some reason the
> legs turn more in one direction and less the other
> anyone know why and how I can fix it.
>
> PS=I know about attaching springs to the motor shaft
> but looks ugly. Beside do you see any springs on Mr.
> T's walkers?
>
> __________________________________________________
> Do You Yahoo!?
> Talk to your friends online with Yahoo! Messenger.
> http://im.yahoo.com
Home