Walker, Spyder 1 (Spieder 1)

My First One
Two Motor WalkerSchematic
This one has a movie1.37MB MPG Movie, 320x200 resolution, 15 seconds..
Nihon Mini Motor Bug
The Gear SetGear Modifying Instructions.
Bridge Head
CyberArm revisited
Number 5 - Bugger
Symet Page
Turbot PageFirst one - 1.37MB MPG file.Second One - 1.37MB MPG file.
Biggie Page
Slider Pagewhine-whine-whine-WHINE-WAV file, 76K.
PIC-16LF84 chip, emulates a microcore.Schematic - JPEG image 925 x 683 pixels.Source file for GP16LF84.Hex File for programming into PIC-16LF84-04/P
Your're Here !Schematic - GIF image 1423 x 906 pixels.Source for Spyder1.Yes !Yes! 964KB WMV.

Shows bug on slightly downhill rough sidewalk.  This rough surface greatly helps in motivation.  Still need to tweak leg action, but almost done.

This is a second or third attempt at making an articulated 4 legged walker, which it now does. (February 27, 2002)
This one is patterned after Tilden's Spyder but has some major differences.  Nihon mini-motors were chosen to use  instead of trying to dig up OTU's and pager motors.   In existence is the frame, with all legs fully functional and the 2.5" x 1.5" PIC controller board.  All legs are symmetrical, i.e. all the same, no legs are mirror images of it's neighbor.

Top Down View Action:
Goes forward.
Also, it does not turn around in place, although it's not impossible.
Could go forward, backward, right or left, maybe someday diagonal.

Motor placements:
Each leg has a vertical "hip" motor that can rotate the whole leg, +/- 40 degrees.  This motor and leg pivot assembly seems to be tipped back about 10 degrees (or so).
Each leg has a horizontal "knee" motor that can 'lift the lower leg/extend foot', +/- 30 degrees.

Hard Part:
Setting the front legs so that they lift and extend, clearing the ground and not dragging while moving in or out.
    Harder still: Getting the Nihon Mini-Motors to "lock" in place while the rest of the 'bot moves.
    Noting that the lower strut is in a different place, looking at the difference between front and rear legs.
Also, in my opinion for this application, the Nihon motors run about 75% too fast, a slower gear motor would be preferred.  This is partially solved by running at a reduced voltage, 4.1v from unmodified cell phone battery.

Figuring out how the "gearmotor hip-pivot assembly" is angled and rotated in the body.
Measuring the length of all leg parts from the www.solarbotics.com pictures.
Making a Li-On cell phone battery work, for the ~4.1 volts (& rechargeable, nice).
Making the centering springs work correctly. (Hi Dave)
Building a reliable "motor-output-shaft spring-assembly" from the motor to the thigh.

Figuring out the leg movement sequence, this took about 7 days.  Still not satisfied on this yet.

1/16" brass rod for the foot, upper & lower struts and underbody bracing.  3/32" brass tubing for the struts pivot points.  3/32" brass rod for the thigh part of the leg.  0.032" brass plate for the motor mountings, 1" squares.  Jumbo paperclip wire for the upper "X" bracing (not shown yet).

A device removed form a fully functional auto-focusing lens.  Has IR LED for transmitter and dual sensors at the receiving end.  Looks kinda like the one on Tilden's Spyder.  Is black plastic with two lenses mounted with some blank space in between.  The circuit board is fully functional out to about 15 feet or so.  This came from a table-top projection system, probably surplus.  Circuit gives outputs to two mini-type motors to adjust the lens for focus, etc.  These outputs may be useful in determining what just walked past the 'bot or movement detection, cats coming in for attack, for instance.

Spieder 1
Tilden's Spyder 1

Leg Sequence:
Here's the leg cycles that are being used. Each Nv cycle is divided up into 5 micro steps,
this gives an easy way to figure out how to make the legs move with overlaps in leg movement.
Even though the cycles start & stop with cycle 1, it may get moved over to starting at cycle 3 where there are no overlaps.
Overlaps also occur between cycles 7 & 8, 12 & 13 and 17 & 18 (spaced out buy 5, curious).

In Tilden's patent sheet that shows the leg PDC's hooked to the microcore, there's a MUX controlling "GAIT" between legs 3 & 4 (see his patent drawing).  How about switching sequences for legs (his legs 3 & 4) 2 & 4 ?  What would this do ? Note: doesn't seem to work, not sure what Tilden was doing.
Leg Micro Cycles
cycle 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Nv 1 2 3 4
Leg1 cw cw dn dn ccw ccw up up
Leg2 up up ccw ccw dn dn cw cw
Leg3 up up ccw ccw dn dn cw cw
Leg4 cw dn dn ccw ccw up up cw

Revised Leg Micro Cycles
cycle 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 2
Nv 1 2 3 4
Leg1 dn dn ccw ccw up up cw cw
Leg2 up up ccw ccw dn dn cw cw
Leg3 up up ccw ccw dn dn cw cw
Leg4 dn dn ccw ccw up up cw cw

Turn In Place:
Rotate ClockWise (CW) by rotating all legs CounterClockWise (CCW) then re-positioning the legs.  Should be easy.  For instance if (from the Leg Sequence Chart above) the 'bot stopped or started normal walking at cycle 3, then a RotateCW would go like: rotate all ccw (this rotates the body cw), rotate leg1 cw; move leg2 up, cw, down; move leg3 up, cw, down; rotate leg4 cw.  Now the 'bot can walk forward again.

This will be tested later.  Same as going forward, but sideways.  Will be trying left & right movements.
It seems that CrabLeft actually does a RotateCW, funny, oh well more testing.
Top-Down Views of 'Walking Animations'
Forward WalkingCurrent Walking Leg Motion (from above chart).

Bug Walk 3This animation is proposed (Run, cat coming ?)

Two Legs Driving.This is also proposed, in a larger Spyder.

Top View.Top View.These two animations don't work so well in application, but they're fun to watch.

High Front View.
Shows a frontal view.  Cell phone battery laying on table in back of 'bot, plugged in.  Circuit board is mounted upside down in center of 'bot.
Low Front View.
Lower down frontal view.
Left Side Pose.
Left Side View, in a "pose".
Left Front Leg In (Up).
Right Front Leg in or up.  Rear legs just move up or down, not in or out, well maybe both..
Left Front Leg Out (Down).
Right Front leg out or down.

More Pictures:
Bottom View, JPEG image 525 x 457 pixels.
A device similar to Tilden's Spyder's eyes has been mounted in front but lower on the 'bot.  Looks like a spyder wearing glasses.
Here is the bottom view. Here you can see the controller board with the double stacked 74AC245's, the two 74HC139's and the PIC-16LF84 microcontroller. This 'bot runs at 4.1V from a li-on cell phone battery (they call it a 3.6v battery, not so).  Also has four LED's to tell what the '139 enables are doing.  When you scratch right behind the circuit board, the right rear leg starts twitching :-)
JPEG image 494 x 403 pixels.
With Optics attached but not connected.
JPEG image 626 x 405 pixels.
Front view of optics.  Note ruler in foreground.  It fits nicely in a 8.5" by 11" wooden tray.
JPEG image 768 x 512 pixels.
Close-up of Right Front Leg Mechanics.  Here you can see how the Nihon motor is mounted in the knee.  It can come out with the removal of two mounting screws and the shaft mounting screw.  All Nihon motors are mounted this way, for easy replacement.

Paul T. Barton

This page updated: April/04/2002