Alt-BEAM Archive

Message #09694



To: beam@sgiblab.sgi.com
From: iphotope@canada.com
Date: 1 Feb 2000 18:22:05 -0800
Subject: [alt-beam] Alternate locomotion



------------=_949458125-1473-1
content-disposition: inline

Hello all.
Over the months, I have accumulated many ideas
for alternate forms of BEAM robot locomotion. Most
of the following would probably be terrible energy
wasters, but they all look interesting to me. A
couple even look like they could improve a robot's
adaptability.

1. My first idea has been "discovered" many times
on this list, and I am only adding to what has been
written. I am talking about jumping.
What I have to add follows;
Instead of repelling a magnet, my prototype uses a
solenoid with a shaft out one end. At the end of the
shaft is a foot with a little grip on it. The solenoid is
encased in a cylinder with only the shaft and 2 leads
protruding. The top of the cylinder is attached to a
motor shaft. When there is no power activating the
solenoid, the 'leg' shaft is retracted. To 'hop', the
solenoid is quickly fired (camera flash capacitor?) to
make the 'leg' shaft fire out and retract quickly.

Prior to firing, all 4 legs are pointing backwards. The
legs 'hop' and the two front legs point forward so that
the robot does not flip over itself. Having landed, the
robot's to back legs 'hop' and the front legs point
backwards again to come to the orginal resting position.
To go backwards, do the same thing, but point the legs
the opposite way.

To turn left, the robot does a normal 'hop', but instead
of moving it's front legs forward, it moves those on it's
right forward. It then fires all the legs again and returns
to the original resting position, yet having turned slightly.

2. This is an actual system patented by Lockheed but
never used, the Tri-Star wheel arangement:

_
(_)
_ / \ _
(_)----(_)

All three wheels have gears are powered gy the same drive
shaft, their own shafts supported by the triangular metal
plate. Only two wheels touch the ground at the same time,
yet they all turn. If the friont wheel gets stuck, trying to go
over something it can't, for example, the whole wheel gang
flips so that what once was the top wheel is now over the
front wheel trying to climb the object the front wheel got
stuck on. This could be very useful in making photovores
more all-terrain.
This will be discussed in more depth on my web page when
I put it up.

3. "Hover" locomotion.

The basics are rather simple, one big fan aiming down, and
two smaller fans aiming back and you have a hovering 'vore.
Theoretically, this thing could hover over water, and I
believe it with all my heart, yet if it runs out of sun before it
crosses the water, the results could be devastating. I suggest
you put a strong baloon of some sort in a ring around the robot,
and it could land on that when it runs out of power. If you
implement this design, I suggest using a circuit which runs both
(directional) motors at the same time so that the 'bot does more
than pivot.

4. Magnet gun..

http://www.mitedu.freeserve.co.uk/schematics.htm

Go to "miscellaneous", "magnetic gun" and look around. This is
probably far-fetched, but here goes anyway!

If you had a larger bore (probably longer too), you could fire
out a 'probe' tethered by a string which would then rewind
the string so that it drags the bore to it, then loads itself into it.

5. Attached is a picture of a leg which could be run from a
photovore circuit, but would still walk. On the picture;

1= The robot's frame
2= a rectangular plate
3= a slot in the plate (2) Must be able to move freely
4= a nail or screw going through the slot (3) and into the frame (1)
5= Where the plate (2) meets the leg (10)
6= A circular plate with a hole in the top (7)
7= The point where the plate (2) meets the plate (6)
This spot must be able to pivot
8= the motor
9= The motor shaft
10= a bent piece of wire used as the leg


The magnet gun and hop ideas would need extra circuitry, but
would amaze people alot more then a simple wheeled robot.

Hope this helps someone,
Iphotope


__________________________________________________________
Get your FREE personalized e-mail at http://www.canada.com

------------=_949458125-1473-1

Attachment: new gen leg.bmp

------------=_949458125-1473-1--



9695 Tue, 1 Feb 2000 19:45:42 -0800 [alt-beam] Re: The 240 microcore - was Bicore question "'beam@sgiblab.sgi.com'" Wilf Rigter

And they said it couldn't be done!

Here is a tested 74HC240 microcore (no not a bicore) circuit which behaves
just like a 74HC14 type microcore and motor driver combined. I have bench
tested the circuit with small lens motors and VCR ejector motors. Stalling
the motors by hand produces a slightly shorter pulse duration but otherwise
the circuit just keeps on ticking. The motion of the motors is the
characteristic non-overlapping microcore sequence. The circuit is similar
to the "quasi microcore" design I reported some time ago.

The new 240 uCore circuit is similar to a microcore but uses normal HC240
inverters instead of Schmitt triggers. Such a linear microcore circuit
normally would not work but this problem is overcome by adding positive
feedback from the non-inverted motor driver outputs using a capacitor. This
positive feedback speeds up the transition through the linear region using
what I call AC hysteresis. Assuming a switching threshold of 1/2Vdd, the
ratio of C1/C2 must be greater than 2 and the time constant will be
proportional to R*(C1+C2) since the bias point input voltage will clamp at
Vdd on the positive edge of the input signal. Note the location of the C1
caps, some of which are mounted between pins on the solder side of the chip.
The two diodes (which can be any type 1N914/1N4448 etc) provide an automatic
PNC function, far simpler and easier to use than the old Nu type PNC
circuit. The layout of this circuit would appear to lend itself very well to
freeforming. If possible, I will try to incorporate a reverser function to
this circuit without adding (many) more parts unless someone beats me to it
of course.

enjoy

wilf
<<240UCOR1.gif>> <<240UCORE.gif>>
> -----Original Message-----
> From: Phillip A. Ryals [SMTP:phillip@ryals.com]
> Sent: Sunday, January 30, 2000 9:21 AM
> To: beam@sgiblab.sgi.com
> Subject: Re: Bicore question
>
> > Ok I'm feed up with trying to make a Microcore walker.
> > Ao I've turned to the bicore.
>
> Thanks what I'm screaming! I've made two m/s bicore walkers, and decided
> to
> go to the microcore. I thought it would act more along the lines of what
> I
> wanted it to do.
>
> I breadboard it, and it works fine. I solder it all together, and it
> refuses to work. I've checked all my solder joints, and made sure
> everything's touching where it should, but I still can't manage a working
> model. All I'm left with now, is perhaps that I ruined the chip with
> heat.
> Are the 14's more sensitive than the 240's? I'm using one of those
> universal PCB's, and I soldered all the pins of the 14 at once. Is that a
> problem? I haven't had a problem doing it that way with 240's and 245's,
> so
> I figured I could get away with it.
>
> Any suggestions?
>
> -phillip



Attachment: 240UCOR1.gif


Attachment: 240UCORE.gif

Home