| Small
Robot Drive Trains
By Ibrahim Kamal
Last update:
4/4/08
 Overview
This Tutorial Aims to introduce to beginner the different
techniques used to build the chassis and drive trains of
relatively small robots. What
I mean by small robots, is the category of light weight
robots where aspects like aerodynamics and vibration damping
are not taken in account.
Most of the fundamental theory remain the same for
all sizes of robots, Theory such as robot stability, traction
and steering techniques. However, This tutorial is dedicated
to the construction of relatively
small robots, explaining
how to the implement your designs using material
and equipment at the reach of any average level hobbyist. |
|
1-Traction
and steering mechanisms
There are many types of mechanisms that have been developed
to move a robot from a place to another, more than you could
imagine, however, we will only study some of those techniques
which are the most adequate to relatively small robots:
1.1- Differential Drive system: This is the
most common locomotion scheme in the world of robotics. It's
easy to build, easy to control and can permit the robot to move
in all required directions. In that system, two motors are connected,
one to each of the two drive wheels at the right and at the
left of the robot's base. Those two motors are responsible of
driving the
robot back and forth as well as steering in any required
direction. This system can even allow a robot to pivot
in its place.
The word "differential" comes from the fact
that this system relies on the velocity difference between
the two wheels to drive the robot in the required path
and make it move in any required curve.
Figure 1A shows in a simplified way the principle of
operation of differential drive. When the left and right
motors are turning at the same speed, the robot moves
forward or backward in a straight line. In order to
turn right for example, the right motor is slowed down
and the whole robot steers to the right. The bigger
the difference of speed between the two wheels, the
tighter will be the steering curve.
|
Figure 1.A: Differential
drive system
|
As you can also notice in the figure 1.A, this system needs
one or more caster wheels (also called free-wheels) to support
the rest of the chassis while freely following the movement
of the robot engaged by the two main drive wheels. Some robots
use 2 caster wheels, adjacent to the drive wheels for better
stability, but again, in the category of small robots, one caster
wheel can be enough.
Many
roboticists even replace this single freewheel with
a piece of bent metal or any other low friction surface
to act as skid that slips on the ground under the action
of the driving wheels.
Figure 1B shows the chassis of one of our line follower
robots where skids are used along with the drive wheels
to support it |
Figure 1.B: Skids used on
a differential drive train |
in a horizontal position. This technique becomes
interesting when used for lightweight robots, where caster wheels
can cause even more friction than the skids, and add additional
and unnecessary weight to the robot.
You can also notice in the figure 1B that the rear skid is not
touching the ground, and thus is not a primary skid, but rather
a secondary one just in case the robot pivots around the drive
wheels to the back during a sudden acceleration for example.
1.2- Four wheel differential drive system:
Another
variation of the differential drive system, is the 4
wheel differential drive system, which resembles the
drive trains of tanks and bulldozers.
As figure 1c shows, it consists of only
two motors and 4 wheels. Each pair of wheels are mechanically
connected together, usually by the mean of a belt forcing
them to move together, driven by a single motor.
The behavior of this drive train is similar to the previous
differential drive system; If the the two motors turn
at the same speed, the robot moves forward or backward,
if there is a difference between the speed of the two
motors, the robot will steer in the direction of the
slower motor.
This drive system implicates a
great deal of friction, specially during steering, and
thus, to overcome this |
Figure 1.C: 4 wheel differential
drive system |
friction, it requires relatively powerful motors,
as compared to the previous differential drive scheme. The advantage
of this system is that it is very easy to build and is most
suitable for running on low friction and dusty surfaces where
other drive systems would suffer.
1.3- Car-type drive system: This
drive system has many names, but calling it "car-type"
make it clear to any reader, that it consists of 2 wheels coupled
on the same axe at the back and one or two wheels capable of
steering to control the displacement of the robot. While this
technique is very difficult to implement mechanically, it provides
two major advantages over other traction systems: Ease
of control and accuracy (because traction
and steering are connected to two independent motors). The best
way to build your first 'car-type' drive train is to look at
any old RC toy car. You can also use some parts of it's drive
train to build your own, or even use the
chassis
of a brand-new RC toy car for
your robot, as we did in the WFR
project. Figure 1.D shows a simplified layout showing
the principle of operation this drive system. Connecting
the front (steering) wheels to the motor can be very
complicated, but a basic rack and pinion arrangement
can be enough for very low steering angles, otherwise,
for relatively high steering angles, the rack and pinion
arrangement will have to be enhanced with a sliding
mechanism |
Figure 1.D: Basic car-type
drive train |
allowing the green axe (in the figure) linking
the two front wheels together to move freely along the axis
of the chassis.
There can be many other configurations based on the same idea,
like a 2 wheel traction / 1 wheel steering drive train, which
may be a little simpler to construct because the front wheel
can be directly connected to a gearhead DC motor.
1.4- Divided chassis drive train:
This is another locomotion scheme that we experienced, which
has the advantage of being very powerful and highly configurable.
Inneed, this system relies on 4 independent motors, each one
coupled to one of the four wheels, allowing the robot to move
forward or backward with the
accumulated
power of four motors or to turn right and left with
approximately equivalent power, but with minimum frictional
losses.
Figure 1.E shows how to implement such a configuration,
and shows how the the chassis is divided into two parts,
hinged together, allowing it to change it's shape to
turn in tight curves.
Usually, the two front motors are used only to drag
the robot, while |
Figure 1.E: Picture and layout
of the divided chassis |
the two rear motors are used either to push it
or to bend the chassis (like a toy train) to turn right or left
by running one of the two motors faster than the other.
One last special feature of this system is that it has no head
nor tail, in other words, there is no difference between the
front part of the chassis and the rear one, making it possible
for the robot to change it's direction, to adapt to any situation,
without having to execute a 180 degrees rotation.
2-Mounting
the Motors
When you are settled on a traction
and steering mechanism for your robot, you have to find a way
of fixing the motors firmly to the chassis. At this point, in
order to construct a professional drive train, you have to take
those points in consideration:
 You have to be able
to mount and dismount the motor easily. Gluing, soldering and
other permanent fixing solutions are not good ones. As a roboticist,
building a prototype, You have to able the make any modifications
easily.
The outer shell
of a motor is not designed to bear high, undistributed stresses.
whatever the method you are using to mount and hold your motors
in place, you have to increase the contact area between the
motor's body and the mechanical part linking it to the chassis
to prevent stress concentration and eventually damaging the
motor.
There are many items that you can buy at your nearest hardware
store, to use it to mount the motor on the chassis, I present
some ideas that I find to be the most adequate to this category
of robots.
2.1- Pipe Clamps:
Because
they come in any size you want, because they are cheap,
and because it can be used with any motor, this is my
favorite choice. Initially, at the name implies, pipe
clamps are used to fix water pipes on walls. Due to
the resemblance of size and shape between the outer
shell of a standard DC motor and water pipes, those
clamps are very adequate to the job of holding motors
firmly in place, while distributing any eventual stress
along the whole contact surface.
Figure 2.A shows a standard pipe clamp, where the right
bolt was replaced with a longer 4mm screw, to be used
later to fix the motor-clamp assembly to the chassis.
Note that there are many varieties of pipe clamps, so
it's up to you to pay a visit to your hardware store
and find the one that would be the most suitable according
to your design. Most Pipe clamps come with a nut welded
to it, allowing easier fixation to any adjacent plan
(that plan can be a wall as it can be the chassis of
your robot).
Figure 2.B shows how a motor fits in that pipe clamp.
Adding small parts of rubber between the surfaces of
the motor and the clamp can be a good idea, as it will
absorb any eventual vibration, and protect the motor's
body.
Figure 2.C shows how a pipe clamp can be used to mount
the motor on the chassis. As you can notice, one of
the main advantages of using a pipe clamp, is that you
can easily adjust the distance between the motor and
the base of the robot to which the motor is fixed, which
is not the case with other solutions, like the U-Bolt
idea.. |
Figure 2.A: Pipe clamp
Figure 2.B: DC motor
& pipe clamp
Figure 2.C
|
2.2- U-bolts:
If
for any reason the pipe clamp solution cannot be used,
this is the second best choice. I've never used this
technique to fix motors, but i've seen it used by many
hobbyists. The two reasons making this technique less
effective that pipe clamps are:
1- The contact area is relatively small, thus stress
concentration can be a problem. This problem can be
solved by using more than one U-bolt for each motor.
2- The U-bolt is designed to hold a circular object
by pushing it against a fixed surface. As figure 2.E
shows, in our application, the circular object is the
motor and the fixed surface is the chassis. so the height
of the motor from the ground depends on the height of
the part of the chassis where the motor is fixed. In
contrast, with Pipe clamps, you can easily adjust the
distance between the chassis and the motor. |
|
2.3- Building your own bracket:
If your motor is too small, too big or just not adequate to
any of the previous solutions, building your own bracket to
hold your motor is not that difficult. You have to use your
imagination, and cope with the components, tools and materials
that are at your reach. At this point i'll only give you hints
of materials you could use to easily attach your motors.
-
A rubber band (figure 2.F) can hold
your motor in place exactly like a U-Bolt, while giving
it some elasticity to absorb shocks and vibrations.
You can obtain really strong rubber belts from old cassette
players. |
Figure 2.F |
-
Thick copper wire can also be used
instead of the U-bolt, although it has the major disadvantage
of being a permanent solution, any modification will
be destructive and new wires will have to be installed.
Figure 2.G shows actual pictures of DC motors that's
I've mounted on one of my first robots. |
Figure 2.G |
-
Plastic strap fastener, is still a
permanent solution, but has the advantage of providing
a strong attachment, while being easily introduced and
fastened in small places where it would be harder to
install copper wires or rubber bands. |
Figure 2.H |
- Metal sheets forming is the
best alternative to U-Bolts and pipe clamps if you have the
tools, skill, and material to do the job. Then you can form
the exact required shape to firmly hold your motor in place.
I personally recommend Aluminum as a working material. It's
corrosion resistance, it's light weight and it's machinability,
make it a perfect match for small robots.
3-Transmitting
power to the wheels
There are many way of transmitting
the power from the motor's shaft to the wheels, but in any case,
friction is your biggest enemy. The frictional losses increase
when the weight of the robot increases, specially if you're
not using any ball bearings to support the shaft, which is usually
the case in small robots. I am going to categorize the various
techniques used for power transmission into two main categories:
Direct transmission and Un direct transmission.
3.1 - Direct transmission
3.1.1- Direct transmission from a DC Motor, Which is
when the wheels are directly mounted on the motor's shafts.
This technique is only acceptable when working with very light
weight robots, otherwise, those relatively small DC motors are
not mechanically designed to bear any bending stresses that
could be caused by directly mounting a wheel on it. Usually
the internal
bearing
of those motors are designed to withstand the stresses
required to keep a gear in place in a gearbox, or to
withstand the tension in a rubber belt. In Most cases,
directly mounting the output shaft of a DC motor to
a wheel is rarely the best choice because without any
speed reduction, the developed torque wont be enough
to easily move a robot around. Figure 3.1.1 shows an
example of how a small plastic wheel can be mounted
to the tiny shaft of small motors. As you can see this
is the same idea used in most toy cars, where parts
are simply forced in their place. This is more interesting
than more advanced shaft coupling techniques when you're
working with tiny motors, because it would be very difficult
to drill it's output shaft to install setscrews or pins.
|
Figure 3.1.1
|
3.1.2- Direct transmission from a gearhead DC motor
is the most common method used by roboticists, specially for
the category of robot we are talking about. That's because the
output
shaft
from the gearbox of a gearhead motor are designed withstand
higher stresses than the output shaft of the motor itself
and has a very suitable output characteristics (low
RPM and high toque). That can be noticed from the shape
of the output shaft of a gearhead DC motor in the figures
3.1.2A and 3.1.2B, which is not fully circular to be
able to firmly secure it to a shaft coupling or to a
wheel in our case.
There are many ways to fix a wheel to the output shaft
of a gearhead DC motor, one of the simplest ways is
to drill a hole in the body of the wheel, of the exact
diameter of the screw to be used. Forcing the screw
in the drilled material for the first time will form
a helix mating the thread of your screw, allowing you
to fasted the shaft in its place by simply driving the
screw.
If the wheel you are using doesn't have an extended
shaft as shown in figure 3.1.2A, you will have to find
another mean of fastening the motor's shaft to the wheel,
but virtually any wheel you will find is designed to
be mounted on a shaft, so using your imagination, you'll
end up finding a solution to that problem, depending
on the shape and size of the wheel.
Figure 3.1.2B shows the output shaft of an old PITMAN
GearHead DC motor. |
Figure 3.1.2 A
Figure 3.1.2 B
|
3.2- Undirect transmission
Undirect transmission means that the shaft holding the wheels
is on a totally different plane than the output shaft of the
motor. Usually, between those two shaft is a reduction belt/gears
arrangement. Undirect transmission is more suited to heavy robots,
where rolling bearings are installed to support the shaft holding
the wheels, to minimize any frictional losses due to the
weight
of the robot. Mechanically speaking, any wheel is to
be fixed with rolling bearings, so this
is the best method of all, but again, you have to remember
that there are some compromises that can be made with
small robots, to reduce cost, weight and complexity.
Figure 3.2A shows a classic undirect
transmission system, where the rotation of the motor's
shaft is transmitted through a reduction pulley arrangement,
lowering the RPM of the wheel, while increasing the
torque. Notice how the the full weight of the robots
is distributed only on the two rolling bearings, making
this drive train much more robust than other where the
wheels are directly mounted on the motors.
Figure 3.2B and 3.2C shows a practical implementation
of this |
Figure 3.2A |
technique on one of our early robots, where each
wheel is mounted to a 6mm steel shaft, supported at both ends
with rolling bearings. There is no obligation to put the rolling
bearing at both sides of the wheel (the two rolling bearing
can be placed at the same side of the wheel as in figure 3.2A)
but you have to put two of them, each one as distant from the
other as possible, to reduce shear stresses on the rolling bearings.
(rolling bearings are designed to withstand very high radial
stresses). While you're looking at the pictures, you can notice
how the motors are mounted with pipe clamps as explained before.
Figure 3.2B |
Figure 3.2C |
4-Diameter
of the wheels, speed and torque
With a constant output
RPM from a motor, The diameter of the wheel is the last criteria
that can affect the the balance between the torque of the robot
and it's linear velocity.
Consider the wheel in figure 4, having a radius
r turning with a constant angular velocity
(RPM) w,
P a point on the circumference of the wheel,
T the torque developed from the motor acting
on the ground plane and F the reaction force of the ground on
the wheel.
Let Tf be the
torque developed by the ground as reaction to the wheel's
movement. This is the torque resisting the robot's movement.
For the robot to move from rest (or to accelerate),
T must be bigger than Tf.
The toque Tf
= F.r .......(1)
Also note the that the linear velocity of the robot
equals the linear velocity of any given point on the
outer circumference of the wheel,
Then the velocity V = w.r
.......(2)
From the relations (1) and (2), you can easily deduce
|
Figure 4
|
that increasing the radius of the wheel will
increase the velocity of the robot, but also increase the resisting
torque that the robot have to overcome. Thus, varying the radius
of the wheel, you can make a compromise between linear velocity
and torque of the robot.
5-Mass
distribution, Moment of Inertia and robot stability
One of the most important phenomenon that can affect the behavior
of moving robots is the moment of inertia. The moment of inertia,
which is different for each shape, size, weight and trajectory
of a moving rigid body, is the ability of that rigid
body to resist a change in its velocity. The force
developed by the moment of inertia appears at the center of
gravity of the body, and acts as a mechanical moment around
the pivot point of the robot (usually a wheel) in the a direction
opposing the change of velocity.
In order to better understand the effect of mass distribution
on the moment of inertia and consequently on the stability of
a moving robot, let's imagine this scenario where a simple robot
(as shown in figure 5.A) is moving at a constant velocity in
the horizontal direction. Recalling the second law
of Newton, and noting that the acceleration
of the robot equals zero,
the force F1 moving the robot forward
equals the resistance force R1 (which is the sum of
all mechanical friction losses in the system). The
stability problem will start to appear when this moving
robot will suddenly stop.
In that situation, there is a change of velocity,
a deceleration or a negative acceleration as some
books states and the moment of inertia of the robot
will resist this change of velocity, by generating
a force acting in the previous direction of motion,
trying to keep the robot moving in the same
direction, with the same velocity.
In figure 5.B, the effect of the moment of inertia
is somehow exaggerated to clearly show the unstability
that can be caused by this problem. To solve this
problem, and design robots that have very low inertia,
you have to understand those two main factors that
affect the inertia of a body:
A- The position of center of gravity
has to be as near as possible from the ground or from
the eventual pivot point (which are the wheels in
case of our example robot). This is verified by the
mathematical relation that state that:
I = m.(r^2)
Where I is the moment of inertia,
m the mass of the body, r
the distance between the center of mass of the body
and the pivot point
To change the position of the center of mass of your
robot, you have to lower the position of the heavy
parts of the robots like the batteries and the motors.
You also have to design the chassis to be as near
as possible from the ground (as shown in figure 5.C).
If all that is not enough, or if the chassis simply
isn't heavy enough, it's common practice to add additional
weights as near as possible from the ground with the
sole purpose of lowering the overall center of mass
of the robot.
B- The torque developed by the moment
|
Figure 5.A
Figure 5.B
Figure 5.C
|
of inertial depends on the rate of change of
velocity (the acceleration). You can greatly decrease the effect
of inertia by controlling the acceleration and deceleration
of your robot, through ingenious speed control algorithms
6-Finding
motors and building material
The easiest way to acquire all the material and equipment you
need to build a robot, is to buy it! But that's not the cheapest
way, and sometimes you won't find all what you need in the stores.
That's why - if you plan to dig your way in the field of robotics
- you should start collecting all kind of devices from which
you can scavenge all kind and sizes of DC motor, mechanical
parts, gears, pulleys, etc.
From my experience, I tried to classify the kind of devices
from which you could extract useful items, at least, this is
what I do with those devices before throwing them away.
Old cassette players and walkmans: Small Motors, pulleys, belts,
gears.
Old Floppy disk and CD-ROM drives: Very small motors, worm gear,
sliding mechanism, gears, limit switches
Old Photocopier machines, printers and scanners: quality Gearhead
DC motors, sliding mechanisms, rubber wheels, gears.
As for the chassis, it can be built with a multitude of materials.
My preferred material is Copper boards (made for PCBs) as you
can solder many metals to it like screws and nuts, you can also
design the main board of the robot on the same board to reduce
size, weight and cost.
My second best choice is the Fiberglass. It can be easily formed
with a a cutter, can be easily drilled, can sustain high stresses,
and looks very neat an professional.
Another good choice for building the chassis of the robot is
the 3mm Medium Density fibers (MDF wood) which has the advantage
of being very cheap, and can be very easily cut, drilled and
grinded to any shape you want. The only disadvantage of this
material is that it can easily deteriorate with water or even
too high humidity.
Those were the material I found to be more suitable for this
category of robots.
Preview of the last 15
messages discussing this page. Messages are sorted from the newest to
the oldest. |
Posted
by:
ikalogic
on:
19 Aug 2008 |
|
Re: Small Robot Drive Trains |
|
 |
Quoting gyaanguru: hello
we are working on a project in which we trying to make a robotic arm on a moving base. . since we are all electrical students we are having a bit of difficulty in designing base and finding motors. the total weight of the arm + batteries (lead acid) +electronics = 7 kg .
could you please tell me the type of motor required in terms of torque. how do we calculate the torque requirements in general. preferably a speed of 0.6m/s (2 ft/sec) is required.
could to teach us how to go about these calculations
thank u |
I could, but the first thing you have to know is that it would be big mistake to put all that weight on the robot. you have to keep all the batteries, motors, and electronics on the fixed base, and use wire systems to transmit the mechanical power to the different parts of the ARM...
|
|
|
|
|
Posted
by:
gyaanguru
on:
18 Aug 2008 |
|
Small Robot Drive Trains |
|
|
hello
we are working on a project in which we trying to make a robotic arm on a moving base. . since we are all electrical students we are having a bit of difficulty in designing base and finding motors. the total weight of the arm + batteries (lead acid) +electronics = 7 kg .
could you please tell me the type of motor required in terms of torque. how do we calculate the torque requirements in general. preferably a speed of 0.6m/s (2 ft/sec) is required.
could to teach us how to go about these calculations
thank u
|
|
|
|
|
Posted
by:
ikalogic
on:
23 May 2008 |
|
Re: Small Robot Drive Trains |
|
|
| Quoting saif: about 15cm dia... v had decided on 20cm but since most of the college competitions here in bangalore(india) fix the maze width at abt 20-25cm max.. nd aso they fix a max limit to ur chassis width as sumthin lik 20cm.. so that is it-15 cm... u think its too less? |
15cm is more than enough, and with that size of robots, you can use the PCB material for the chassis, as in our mini line follower robot project. I think it is the most suitable for that size
|
|
|
|
Posted
by:
saif
on:
23 May 2008 |
|
Re: Small Robot Drive Trains |
|
|
about 15cm dia... v had decided on 20cm but since most of the college competitions here in bangalore(india) fix the maze width at abt 20-25cm max.. nd aso they fix a max limit to ur chassis width as sumthin lik 20cm.. so that is it-15 cm... u think its too less?
|
|
|
|
Posted
by:
ikalogic
on:
22 May 2008 |
|
Re: Small Robot Drive Trains |
|
|
Quoting saif: salam... hey vr actually designing a chassis which v decided to be cicular with just 2 wheels nd passive skids.. is a circular chassis ok wid u? v decided on it coz it can rotate freely in any direction easily... even a 360 turn... nd wot material do u recommend 4 the chassis? it shud be light as well as sturdy..
nd do u recommend a lead acid or nicad battery? |
circuilar is ok, but, as fr the material of the chassis, i can't tell you unless you tell me, what is the size of your robot?
same thing for the battery.
so, what's the size of your robot?
|
|
|
|
|
Posted
by:
saif
on:
22 May 2008 |
|
Small Robot Drive Trains |
|
|
salam... hey vr actually designing a chassis which v decided to be cicular with just 2 wheels nd passive skids.. is a circular chassis ok wid u? v decided on it coz it can rotate freely in any direction easily... even a 360 turn... nd wot material do u recommend 4 the chassis? it shud be light as well as sturdy..
nd do u recommend a lead acid or nicad battery?
|
|
|
|
|
Posted
by:
ikalogic
on:
28 Feb 2008 |
|
Re: Small Robot Drive Trains |
|
|
| Quote: how can we construct a small robort which has only 2 wheels and can turn and move in any direction |
Good question!
You can either do it the easy way, which is to use 'skids' that will act just as a wheel but with more friction..
Or you can do it the hard way, wich is to build a robot with only 2 weels in the midle, and to constantly accelerate and decelerate the wheels, acting on the moment of intertia of the robot to stabilise it.. or maybe you have to have a mean of changing the center of gravity of the robot to give it the ability to balance itself horrisontally.. there lot of ongoing researches about this subject.. you may find some information on wikipedia..
hope this answers your question.
|
|
|
|
|
Posted
by:
ikalogic
on:
28 Feb 2008 |
|
Re: Small Robot Drive Trains |
|
|
| Quote: sir , i am very thankful for all this guidance but i want to know that how aerodynamics is used to make robotics? |
Well, unless you wroking with robots that go in the air on in the sea, or that are very fast, aerodynamics usually is not a big concern. I honnestly don't have a lot of experience in this field so i can't give you more information.
good luck
|
|
|
|
Posted
by:
ikalogic
on:
28 Feb 2008 |
|
Small Robot Drive Trains |
|
|
discussion for the page: [link]
| Quote: sir i have gone through the site,i want to know which motor is prefered for the better efficiency of the robots.(what type it is?) |
Gearhead DC motors (which are normal brushed DC motors, with a gearbox attached to it's output shaft, to decrease speed and increase torque)
|
|
|
|
You have
to be a member to post replies. |
|
|
|
