Mechanical
Development of L07 Robot

Figure 1 - L07 before modifications |
Figure 2 - Modified L07 Robot
|
Figure 1 shows the L07 robot completed
by the Lunartics senior design team. The L07 robot was
modified
by the Moonavators senior design team in order to make the arm more
rigid and capable of handling larger loads (Figure 2).
The
modifications included: improved base plate, gearing change in the
elbow, smoother linear actuator motion, and stiffer arm links.
The changes did give the arm more strength and held the
joints
more rigid under loading. With these improvements, the arm
was
also able to handle more torque before bending. However, in
order
to break through hard soil and chip rock, more drastic changes needed
to be performed. The arm of the L07 was completely removed
from
the frame and a new arm was installed, named M08 (Seen in Figure 3).
Base Joint
The base was
sloppy and provided inaccurate
feedback. The slop was a result of
inaccurate machining of alignment holes and a Lazy Susan used as the
support
for the arm to rotate on. The inaccurate
machining resulted in skewed washers and spacers to make the base fit
together. The Lazy Susan supported a
base plate with an internal gear attached to the bottom. The
base plate rotated due to a drive gear on
a right angle motor coming up through from below. The Lazy Susan
allowed a 1/8”
of horizontal movement causing the drive gear to occasionally
disengage. A
potentiometer, opposite the drive gear inside the internal gear,
attempted to
stabilize the setup. The potentiometer was not strong enough to keep
the
assembly centered and the movement only created stress on
axle.
The potentiometer was also only pressed into
the frame plate and held in only by friction.
Arms
The Arms allowed for torsional movement caused
by the linear actuator and from the gearing used to rotate the actuator. This movement was because
the linear actuator
did not load each of the arms equally, and the helical gear also
imparts a
torsion load. The
arms were made of flat
aluminum stock and had a large displacement under torsion loads. The helical gear also
slips under load
because the bracket holding the drive gear was insufficient.
Linear Actuator
The linear
actuator roller bearings easily bound up and provided inconsistent
motion. The
bearings do not provide enough surface
contact with the rack. The
drive motor
with a high gear ratio was inoperable.
The rack was also able to twist almost 180 degrees within
the
bearings.
Mechanical
Development of M08 Robot

Figure 3 - M08 Robot
Poker

Figure
4 - Percussive Poker
Purpose
- to break up soil and chip hard rock
using a percussive, hammering action mechanically driven by a cam
follower
Main
Components
- Twin
lobe cam
- Poker
Follower
- Spring
Design
Description
The percussive
poker uses a cam follower to produce the continuous hammering effect.
A high speed motor spins
the cam which drives
the poker shaft down.
A
spring was
placed in the shaft to return the poker shaft up when the cam came
around.
Roller
bearings were mounted on both sides
of the cam to stabilize the cam while it rotates while minimizing the
friction
between the cam and the poker housing.
Bushings were placed in the poker shaft to give the poker
a smoother linear
motion.
Finally the
poker tip was
threaded to allow for different tips to be tested.
The
team considered two percussive designs: a
three-bar slider crank mechanism (see Appendix) and a cam-follower
mechanism. The
three-bar linkage, inspired by the design
of a sewing machine, was driven by a circular cam.
Attached
to the cam was a connector linkage that could freely rotate with the
cam. Attached at
the other end of the connector
linkage was the poker (the slider crank) constrained to a linear motion
by the
shaft which moved up and down as the cam spun.
The first problem with this design is that the shaft that
constrained
the poker would experience very high stress when the force of the rock
pushed the
connector bar against the shaft at each hit.
This could lead to very rapid wear in the mechanism (see
Appendix). The next
problem was that the design,
although simple in theory, required a complicated design of the cam and
connections between the cam and connector linkage with high tolerances
work
effectively. Thus,
the necessity for
complexity to make the design work coupled with the prospect of wear
instilled
in the design that would quickly diminish the effectiveness of the
design, this
mechanism was not chosen.
Wrist
Joint

Figure 5 - Wrist rotation
Purpose
– To alternate between the scoop and the poker and provide
torque for scooping
Main
Components:
- 3x3
scoop for digging and trenching
- Rino
gear box with a 120:1 gear ratio
- Anaheim
motor
- Slip
ring for continuous rotation without tangled wires
Design
Description
The goal of this
design was simple, to provide the M-08 a degree of freedom that allowed
it to
alternate between digging and breaking rock.
Last year’s team helped set the criterion for
the design when they
realized that the required an exceptionally powerful motor for the
scoop to be
capable of performing any work. Therefore, the design includes an
Anaheim
motor that
provides an upwards of ( I need to look up the exact amount of torque
) and a powerful gear box
with 120:1 gear
ratio.
Originally,
a more powerful motor
gearbox assembly was designed for the wrist.
However, due to shipping delays, the assembly was
redesigned with the
intent of installing either the
Anaheim
motor or the combination motor gearbox when it is shipped. A slip ring
was also
recommended from last year’s project to eliminate the burden
of wires hanging
everywhere and improve the appearance of the machine.
Elbow

Figure 6 - Elbow joint
Purpose
– The purpose of the elbow joint is to increase the precision
of linear
positioning of the percussive poker and scooper mechanism
Main
Components
- Rino Mechanical
P20-60 60:1 worm and wheel gearbox
- Rino Mechanical
P20-DX Keyed output shaft coupled to the P20-60 gearbox
- Rino Mechanical
5mm x 6mm Oldham
shaft coupler
- Anaheim
Automation 12V DC 5.5oz-in rated torque
Brushed Spur Gear Motor
Design
Description
The elbow
joint
was designed to supply high levels of torque for scooping with
relatively low
weight. The design
combines an Anaheim
Automation 12V DC motor and a Rino Mechanical P20-60 with a 60:1 gear
ratio
worm and wheel gearbox. This
motor-gearbox combination gives the elbow joint a constant rated
333oz-in of
torque to the drive shaft and, at the 1.5A stall current, an upward of
2499oz-in of peak torque.
At the 9.7A stall
current, the motor driving the linear actuator design on the L07
produced
1465oz-in of peak torque. The motor was coupled to a right angle
gearbox, but
was a 1:1 zero gear reduction gearbox therefore not increasing the
torque
supplied to the linear actuator. The
previous elbow joint design for the L07 weighed in right at 2 lbs, and
currently,
the
Anaheim
motor and Rino P20 gearbox combination weighs .9lbs. The motor for the
M08 is
approximately one third of the physical size of the L07 elbow motor,
and
produces up to 10 times the rated torque using a miniature internal
worm wheel
gearbox. The L07 used a heavy and large external worm gear and shaft
combination to translate the rotational motor motion ninety degrees
into the
elbow joint to rotate the end effectors. Not only was this setup bulky,
but
also allowed for dirt and dust to enter the gearing and after time
could
produce problems with the meshing.
Interface options
for the M08 robot

Figure 7 - Labview generated interface control
The L07 robot used an LCD touch-screen operating from the Mini-ITX
board for user interface. A software file created in LabView
had
seperate controls for each joint on the arm. This allowed the
user to move only one joint at a time and was difficult to use.

Figure 8 - Laser-6 radio frequency transmitter by Hitec
In order to make it easier to operate the arm, a radio frequency remote
control (Laser-6 by Hitec) was purchased. However, the
output of this remote controller is a PWM signal operating at
about 50Hz. The microprocessor needs an analog input.
To
convert the PWM to analog would require additional electronics,
therefore the RF remote control option was not chosen for this project.

Figure 9 - Playstation Controller
Finally, a Playstation one controller was decided upon as a simple
method of controlling the robot. The Playstation controller
sends
an analog output to the Basic Stamp microprocessor. This
method
requires no additional hardware and offers real-time control of the
robot while controlling multiple joints simultaneously.
Microcontroller

Figure 10 - RCM4110 Microprocessor
The Rabbit 4110
RabbitCore (Figure 14) controls all the motor controllers on the M-08. It is the brains of the
M-08. It reads the
control signals from a serial
port input and then takes this data and performs the programmed logic
to
operate the motor controllers accordingly.
Then the motor controllers provide the high voltage and
current needed
to operate the motors under load.
It has 4 PWM
outputs, 8 analog inputs, 6 serial ports, and 29 parallel digital I/O
lines
available on board. This
number of I/O
in addition to 512kB of onboard memory allow for connection of multiple
controllers. The
processor operates at
about 60MHz, which supports fast run-time of interrupts for controlling
the
robot arm joints simultaneously.
The
rabbit 4110 was chosen for this project primarily because
of its sufficient number of inputs and outputs to interface with the
motor
controllers and the team’s familiarity with its operation. The rabbit is also
programmed in C
programming language, which is desirable.
Motor
Controllers

Figure 11 - BLH450 |

Figure 12 - HB-25 |

Figure 13 - B15A8 |
Shoulder
Motor Controller – Oriental Motor BLH450 (Figure 11)
The BLH-450 motor controller was chosen because it is designed to
control the
BH450 motor used at the shoulder joint.
The controller has many features to control the speed and
direction of
the motor.
The
speed stability reads
the value of the motor speed and adjusts the current to hold the set
speed for
different loading on the motor.
There
are four events that may trigger the alarm setting: an overload
occurring for
more than 5 seconds, the motor getting disconnected, the motor speed
exceeding
3500
RPM,
or the input voltage to
the controller reaching 25% below or 15% above 24
VDC.
These safety precautions are beneficial
because when the arm is under high loading, the shoulder joint receives
the
majority of the torque.
The control
signals are in a parallel, meaning that a control line is either pulled
high or
low to change the operation of the corresponding control line. This
greatly
increases the number of output pins needed to control the motor.
The control
signals are: START/STOP,
RUN/BRAKE,
CW/CCW,
INT.VR/EXT,
and ALARM-RESET. The input to these control lines are inverted which
means that
a low voltage on the input to the controller will result in a logical
high.
This is done
because each input
line has an internal pull up resistor.
This type of setup is commonly used because if a failure
occurs the arm
will be turned off, stopping motion.
The
needed voltage levels for a high voltage is 5V, this is a problem
because the
Rabbit can only output 3.3V for a high voltage.
To overcome this problem a PCF 8574 I
2C
expansion chip was
used.
I
2C
is a serial
protocol
that uses
pull up resistors on
the signal lines, then the master and slave devices can toggle the line
as
desired by effectively grounding the signal line, by using an open
drain
output.
This choice
was made because the
signal lines can be 5V, this is why I
2C is
commonly used to
interface two different voltage based devices.
This also reduced the number of I/O pins used from 5 to 2.