Max’s Cool Beans
By Max the Magnificent
Flashing LEDs and drooling engineers – Part 30
I
am a huge fan of Douglas Adams
(RIP). In addition to The Hitch Hiker’s Guide to the Galaxy, I really like
his tall tales about Dirk Gently, who
bills himself as a ‘holistic detective’
who makes use of ‘the fundamental
interconnectedness of all things.’ The
reason I mention this here is that I’m
currently experiencing a lot of fundamental interconnectedness myself,
which is unfortunate because I don’t
have a speech prepared and I have
nothing suitable to wear.
A future replete with robots
As you will doubtless (well, hopefully) be
aware from previous Cool Beans columns,
over the past few months, your humble
narrator and his friend Steve Manley have
been working on an animatronic robot
head. In reality, Steve has done the bulk
of the grunt work while I’ve focused my
attention on offering silly suggestions (I
like to play to my strengths). As a result
of all this activity, I’ve been spending
a lot of time thinking about things like
robots and motors and sensors.
My wife (Gina the Gorgeous) often asks
me how long it will be before we will
be able to buy a robot to help her with
her household chores. I fear she’s been
watching too many science fiction films
with me because she’s thinking of a humanoid-shaped incarnation that will be
able to perform activities like crouching
down and picking things up and putting
them away, loading and emptying the
dishwasher, similarly with the washing
machine and dryer (including ironing
and folding the clothes), watering the
flowers in the pots and baskets on our
front and back porches, accompanying
her to the grocery store where she can
say things like ‘get me two large cans of
diced tomatoes and meet me at the meat
counter,’ and so on and so forth.
Not wishing to raise her hopes with
unrealistic expectations, I’ve been telling
Gina that it is going to be quite some time
before this sort of thing comes to pass.
In my heart of hearts, I’ve feared that we
may not be around to see it. At the same
time, having recently watched the postapocalyptic science fiction thriller film
Mother Android (https://bit.ly/3N3r8tz),
I’ve also been afraid that we might live
long enough to regret wanting to live
long enough to see it (you really don’t
want to peer into what I laughingly call
my mind).
For example, although they are incredibly clever, the humanoid robots from
Boston Dynamics that can do things like
dance (https://bit.ly/3y2cezp) and perform
gymnastics (https://bit.ly/3OpPkXY) are
more along the lines of uber-expensive
proof-of-concept creations affordable only
to mega-corporations and nation states.
Having said this, I’ve seen some things
over the past couple of weeks that lead
me to believe a future replete with robots
may be closer than we think.
Robots, motors and sensors,
Oh my!
When you come to think about it, we see
and hear (no pun intended) a lot of news
about the latest and greatest in visual
Fig.1. Robot hand equipped with BeBop RoboSkin picking up a
ball (Image: Bebop Sensors).
54
and audio sensors giving machines the
ability to see and hear, including things
like object detection and recognition and
the ability to recognise sounds like glass
breaking and to understand and respond
to natural human speech. There are also
developments in gustatory and olfactory sensors that will afford machines
the ability to detect and identify tastes
and smells. Less discussed, however, are
tactile sensors that will give machines
a sense of touch, but this too may be
poised to change.
For example, I just read an interesting
article on Science Alert about how scientists have developed a ‘living skin’ for
robots (https://bit.ly/3tJyXxs). Looking
really icky (if you’ll forgive my talking
technical), this ‘material’ is water repellent, self-healing, and has a texture
just like human skin, which is perhaps
not too surprising because it’s actually
made out of human skin cells.
By some strange quirk of interconnectedness fate, I was recently chatting with
Keith McMillen, who is the founder and
chief technical officer (CTO) of a company called Bebop Sensors (https://bit.
ly/3N3m4p3). Of particular interest to us
here, Keith has developed a smart sensing fabric that is used to create a skinlike covering called Bebop RoboSkin,
which can provide humanoid robots
with tactile awareness (Fig.1).
The amazing thing is that this tactile
awareness is claimed to exceed the capabilities of human beings with respect
to spatial resolution and sensitivity. In
the example shown, there are 80 taxel
Fig.2. Robot finger equipped with BeBop RoboSkin reading
Braille (Image: Bebop Sensors).
Practical Electronics | August | 2022
Fig.3. Meet EVE, the humanoid research robot (Image: Halodi Robotics)
sensors (think ‘tactile pixels’) in each
fingertip, presented in an array with a
2mm x 3mm pitch. According to Keith,
humans have about a 4mm pitch with
respect to the nerves in our fingertips.
I must admit that I was somewhat
doubtful when I first heard this. My
human-centric knee-jerk reaction is to
think that I am the end-result of billions
of years of evolution that’s left my biological sensors ‘state-of-the-art,’ as it
were. However, Keith says that there’s
a simple way to verify his claims. If you
take two plastic knitting needles with
slightly curved tips, hold them side by
side, close your eyes, and touch the ends
of the knitting needles to the tip of one
of your fingers, it’s only when the distance between the points of the needles
is 4mm or more that you can distinguish
them as being separate.
Have you ever tried to see (again, no
pun intended) if you could read Braille?
The first version of this tactile writing
system, which is used by people who
are visually impaired, was developed
in 1824 by a 15-year-old called Louis
Braille in France who lost his sight as
a result of a childhood accident.
Braille characters are formed using a
combination of six raised dots arranged
in a 3 × 2 matrix, which is called the
‘braille cell.’ This matrix offers 64 different combinations, which can be used
to represent alphanumeric characters
and punctuation marks (a cell with no
dots is equivalent to a space).
Practical Electronics | August | 2022
I once had the opportunity to run one
of my favorite fi ngers over a book in
Braille, and I simply could not distinguish the number and locations of the
raised dots under my dexterous digit.
I understand that I could learn to do
so with practice, but I remember being
surprised by how little I could sense.
This all came back to me when Keith
showed me a picture of a robot finger
equipped with BeBop RoboSkin reading Braille (Fig.2).
Robots, robots everywhere...
I have friends at a company called Immervision (https://bit.ly/3QxxOTy). They
are working on the cutting edge of machine vision, developing cameras that
combine ultra-wide-angle lenses with
high-resolution sensors and sophisticated de-warping, image-stitching, and
machine vision software based on artificial intelligence (AI), machine learning
(ML) and deep learning (DL).
My friends were telling me that one
of their current projects is with another company called Halodi Robotics (https://bit.ly/39vFEg5). You can
probably guess what Holodi is working
on. Literally yesterday as I pen these
words, I was chatting with Dr. Nicholas
Nadeau, who is Halodi’s CTO.
The folks at Holodi have created an
awesome autonomous android called
Eve, which (or should I say ‘who’?) is
capable of performing some of the most
dexterous of human tasks. Operating
with human strength and speed, Eve
is capable of crouching down, reaching up, and using ‘her’ hands to open
doors, push buttons, and manipulate
objects (Fig.3).
One of the reasons the guys and gals
at Holodi are working with the chaps
and chapesses at Immervision is that,
in the event that Eve runs into a situation she can’t handle, she can use the
internet to call out to her human companions, who can use virtual reality
goggles and haptic solutions to establish a telepresence connection – seeing
through Eve’s eyes and controlling her
hands and body, and Immervision’s ultra-wide-angle camera subsystems are
ideal for this sort of application.
It’s easy to think of all sorts of applications for Eve-type robots. With their
ability to open doors and control elevators, tasks like patrolling buildings at
night and stocking supermarket shelves
spring to mind. Also, helping nurses by
bringing drinks and meals to hospital
patients. And, in the fullness of time,
helping around the home (I’m not going
to show Gina this article because she
will never stop asking, ‘when?’).
One of the interesting things Nicholas told me is that they are doing a lot
of work designing their own specialised motors that provide the optimal
combination of speed and torque for
their particular application. In turn,
this caused me to return my attention
to the motors powering my own animatronic appendage.
Motoring along
Have you ever thought to yourself, ‘What
is a motor?’ I have. I wish I hadn’t. It
turns out that this is a topic of mindboggling complexity. As a start, although
the terms are often used interchangeably, engines and motors are not necessarily the same thing. By one definition,
engines convert chemical fuel into mechanical force by means of combustion,
while motors (a.k.a. ‘electrical motors’)
transform electrical energy into mechanical energy.
Having said this, some people define
a motor as being ‘a device that consumes energy in one form and converts
it into motion or mechanical work.’ By
this definition, engines would form a
sub-category of motors. There are also
molecular motors – both natural (biological) and artificial (molecular machines) – that are essential agents in
living organisms. If you are interested in
learning more about these little rascals,
I would strongly recommend the book
Life’s Ratchet: How Molecular Machines
Extract Order from Chaos by Peter Hoffmann (https://amzn.to/3tK6B6r).
For the purposes of our animatronic
noggin, we are interested in electrical
55
Fig.4. (above) Each animatronic eye is controlled by two 9g servos
(Image: Steve Manley)
motors, but there are so many different
ways to ‘slice and dice’ this topic that
it makes your (human and animatronic) heads spin. For example, we might
divide things into rotary motors and
linear motors, where the latter is essentially any electric motor that has been
‘unrolled’ so that, instead of producing
torque (rotation), it produces a straightline force along its length.
I must admit that, when I originally
started to contemplate this column, I was
hoping to present you with a handy-dandy hierarchical tree-structured graphic
that illustrated the relationship between
all of the different types of electric
motors, starting (perhaps) by splitting
things up into AC motors and DC motors
and branching out from there. Since that
time, I’ve grown to be an older, sadder,
and wiser person. I’ve also come to believe that this is a task beyond the ken of
mortal man. Suffice it to say that, if you
feel daring, perform a Google search on
‘Different types of electric motors’ (you
will soon wish you hadn’t).
One reference comes at things from a
different angle. In his book Motors for
Makers: A Guide to Steppers, Servos,
and Other Electrical Machines (https://
amzn.to/3N5ucoZ), Matthew Scarpino
kicks things off with a nice and easyto-comprehend Motor Selection Flowchart. This starts with a decision symbol
that asks: ‘Do you need to control/measure the precise angle?’ There are two
options: ‘Yes’ and ‘No.’ If you select
‘Yes,’ you are directed into an action
symbol that succinctly says, ‘Choose
a stepper or servo motor’ (if you select
‘No,’ you head off into a cascade of options and decisions).
To be fair to Matthew, he does note that
his diagram is of use only for making an
initial assessment and it doesn’t cover all
the possibilities. For example, universal
motors can operate on both DC and AC
power, and if any motor is connected to
an encoder or position sensor, then its
angle can be measured and controlled.
Still and all, this chart does provide a
useful starting point.
56
Fig.5. Tower Pro SG92R Micro Servo (Image: Adafruit)
Stepper and servo motors
When it comes to position control, the
two main options are stepper motors
(‘steppers’ for short) and servo motors
(‘servos’ for short).
A stepper motor divides a full rotation into a number of equal steps. Such
a motor employs a simple form of openloop control system because it can be
commanded (via a simple sequence of
discrete pulses) to move through a precise angle and hold at the designated
step without any position sensor or
feedback. Steppers fi nd myriad uses
in things like analogue clocks, 2D and
3D printers, laser cutters, and robots.
A servo motor is a rotary (or linear)
actuator that allows for precise control of angular or linear position, velocity and acceleration. In addition to
the motor, a servo includes a sensor
to determine its current position and
a relatively sophisticated controller to
provide a closed-loop control system.
My chum Rick Curl recently sent me a
link to a YouTube video showing some
rather tasty dual-axis servos used to
implement a robotic drummer called
Zenbot (https://bit.ly/3xELy6w).
For the purposes of these discussions,
we are focusing on small hobby servos
based on DC motors in which everything
is presented in a single small package.
However, we should note that servos
come in all shapes and sizes, including
whopping industrial servos (both DC
and AC) which may employ external
sensors and control modules.
First experiments
now in a position to say that I just had a
fossick in my treasure chest of bits and
pulled out a Tower Pro SG92R Micro
Servo (Fig.5), like the ones available
from Adafruit (https://bit.ly/3n0ElbY).
This little scamp can rotate approximately 180° (90° in each direction). Not
shown here is the servo horn, which is
a short arm (or pair of arms at 180° to
each other) that clip onto the gear wheel
sticking out of the top of the servo.
These servos are controlled by means
of pulse-width modulation (PWM). In
this case, there are two parameters of
which we need to be aware: the width
of the pulse and the period of the signal
(see: Fig.6).
It’s the width of the pulse that determines the position of the servo’s horn.
The 1.5ms shown here will drive the
servo to its default (central) position.
The period of the pulses (that is, the
time between pulses) is less important but, for hobby applications, it’s
common to use 20ms, which equates
to a refresh rate of 50Hz (ie, fifty cycles
per second).
Servos may be classed as being analogue or digital. Don’t panic, because
these appear to be identical from the
perspective of the controller, which
doesn’t know or care and sends out the
same pulses regardless. The difference
is inside the servo itself. An analogue
servo employs analogue circuitry to
amplify and process the pulses from
the controller and use them to drive
the motor to the specified position. By
comparison, a digital servo contains a
microcontroller, which measures and
processes the pulses using digital techniques. Each type has advantages and
disadvantages, none of which we need
delve into here.
In my previous column (PE, July 2022),
Steve shared some interesting information with respect to the servos we
ended up using on our animatronic
head, such as the two metal-geared
9g Turnigy TGY-50090 servos of
1. 5 ms
20ms
(https://bit.ly/37pEFNb) we use to
control each eye (Fig.4).
Controller
Servo
( e. g. A rduino Uno)
On the off chance you wanted to
learn something unexpected, the
word ‘fossick’ is Australian for ‘rummage.’ Telling you this means I am Fig.6. PWM used to control a hobby servo.
Practical Electronics | August | 2022
Fig.7. The Sweep Circuit (Image: Adafruit)
Fig.8. Example Sweep Circuit program.
Different servos respond to different
PWM parameters, although the 1.5ms
is always the default position. For example, according to Motors for Makers,
some servos will accept pulses from
0.7ms (full rotation left/anticlockwise)
to 2.3ms (full rotation right/clockwise)
(by ‘full’ we mean the maximum rotation supported by that particular servo).
By comparison, according to Adafruit’s
webpage, in the case of the Tower Pro
SG92R, a 1.0ms pulse width corresponds to a rotation of –90°, a 1.5ms
pulse corresponds to 0°, and a 2.0ms
pulse corresponds to a +90° rotation.
Just to keep us on our toes, different
servo manufacturers use different colour
schemes for their wires. Common combinations are brown/red/orange, black/
red/yellow, black/red/white, and black/
red/blue. Fret not! – this is easier than it
Practical Electronics | August | 2022
seems. The red wire, which
always appears in the center,
is always +ve (let’s say 5V,
although servos using other
voltages are available); the
black or brown wires are
always –ve (let’s say 0V),
and the remaining wire is
always the control signal
wire, irrespective of whether
its colour is orange, yellow,
white, or blue.
As you will see, controlling our example servo is
easy-peasy, not least that
a Servo library is included with the Arduino’s integrated development environment (https://bit.
ly/3O3rIZu). The Adafruit
website provides some nice
experimental circuits and
associated programs. For
example, consider what
they call the ‘Sweep Circuit’ (Fig.7).
There’s no real need to use the breadboard in this case, but doing so will
make it easier to migrate to more interesting test cases later. Also, if you
happen to have one lying around, it
would be a good idea to connect a reasonably large electrolytic capacitor (say
470µF to 1,000µF) between the power
and ground rails, thereby helping to
insulate the Arduino from any servoinduced power surges, and having the
breadboard makes this easy.
The control signal to the servo can use
any of the Arduino Uno’s digital outputs
that supports PWM, which would be
pins 3, 5, 6, 9, 10, and 11. We’re using
pin 9 in this example.
An example sketch (based on the one
offered by Adafruit) is shown in Fig.8.
Our first step on Line 1 is to include
the servo library. On Line 3, we create a
servo object. In this case, we’ve called
it MyServo, but any legitimate (nonkeyword) name will do.
On Line 7 we use the attach(pin)
method to attach our servo variable to
a pin (digital pin 9 in this example).
Note that there’s another way to do this,
which is to use attach(pin, min,
max), where min defines the pulse
width, specified in microseconds, corresponding to the minimum (0 degree)
angle on the servo, and max defines the
pulse width, specified in microseconds,
corresponding to the maximum (180°)
angle on the servo.
Inside the loop() function we have
two for() loops. The first sweeps the
servo from 0° to 180°, then the second
sweeps it back from 180° to 0°. The
15ms delays between each step are
there to give the servo time to respond
and move to its new position. Both of
these loops use the write() method
to specify the desired angle in degrees
from 0° to 180° (in the case of a continuous rotation type of servo, writing
a value of 0 would set the servo to rotating at full speed in one direction, a
value of 180 would set it to rotating at
full speed in the other direction, and
a value of 90 would result in no movement at all).
Make sure you attach a horn to the
servo before you set this program running or you might not even realise that
anything is happening. If you’ve never
used a servo before, then seeing the
simple back and forth sweep motion
presented by this program will bring a
little smile to your face because it will
make you realise that you’ve just flung
open the door to a world of motioncontrol possibilities.
Next time
As exciting as all this is (assuming all
this is new to you), we’ve really dipped
only the tips of our toes into the servo
control waters. Next time, we’re going
to consider some different approaches
to control a bunch of these little rascals, like the seven servos we’re using
on our animatronic noggin. Until then,
as always, I welcome your questions,
comments, and suggestions.
Cool bean Max Maxfield (Hawaiian shirt, on the right) is emperor
of all he surveys at CliveMaxfield.com – the go-to site for the
latest and greatest in technological geekdom.
Comments or questions? Email Max at: max<at>CliveMaxfield.com
57
