Servo magazine 12 2008

Tạp chí Servo

Let your geek shine.Meet Pete Lewis, lead vocalist for the band Storytyme Pete recently created the RS1000,

a new personal monitor system for performing musicians It was SparkFun’s tutorials, products and PCB service that enabled him to take his idea

to market in less than a year

The tools are out there Find the resources you need to let your geek shine too

Sharing Ingenuity

W W W S P A R K F U N C O M

Introducing Logic, the new Logic Analyzer with SPI, Serial, and I2C Now shipping for $149.

Problem Seen Problem Solved.

Robothon Robot Combat 2008

28 Results and Upcoming Competitions

Robot Profile

26 Micro Drive

SERVO Magazine (ISSN 1546-0592/CDN Pub Agree#40702530) is published

monthly for $24.95 per year by T & L Publications, Inc., 430 Princeland Court, Corona,

CA 92879 PERIODICALS POSTAGE PAID AT CORONA, CA AND AT

ADDITION-AL ENTRY MAILING OFFICES POSTMASTER: Send address changes to SERVO

Build a Self-Balancing Scooter

by Balakumar Balasubramaniam and Wendi Dreesen

This inexpensive system lets you monitor your (or your robot’s!) rhythm and stride pattern.

Part 3: Base Assembly.

SERVO TankBot

by Ron Hackett

This time, take a look at centering the TankBot’s servo motors and implementing a Universal TV remote for control.

PAGE 48

PAGE 74 Features & Projects

430 Princeland Ct., Corona, CA 92879-1300

(951) 371-8497

FAX (951) 371-3052 Webstore Only 1-800-783-4624

www.servomagazine.com

Subscriptions

Toll Free 1-877-525-2539 Outside US 1-818-487-4545

P.O Box 15277, N Hollywood, CA 91615

PUBLISHER

Larry Lemieux

publisher@servomagazine.com ASSOCIATE PUBLISHER/

Dan Danknick

dan@teamdelta.com CONTRIBUTING EDITORS

Jeff Eckert Tom Carroll Gordon McComb David Geer Dennis Clark R Steven Rainwater Fred Eady Kevin Berry Michael Simpson Chris Savage Wendi Dreesen Fulvio Mastrogiovanni Shane Colton Thomas Kenny Blake Hooper Rob Farrow William Smith Ron Hackett Balakumar Balasubramaniam

CIRCULATION DIRECTOR

Tracy Kerley

subscribe@servomagazine.com MARKETING COORDINATOR

WEBSTORE

Brian Kirkpatrick

sales@servomagazine.com WEB CONTENT

Michael Kaudze

website@servomagazine.com PRODUCTION/GRAPHICS

T & L Publications, Inc.

All Rights Reserved

All advertising is subject to publisher’s approval.

We are not responsible for mistakes, misprints,

or typographical errors SERVO Magazine assumes

no responsibility for the availability or condition of advertised items or for the honesty of the advertiser The publisher makes no claims for the

legality of any item advertised in SERVO.This is the

sole responsibility of the advertiser.Advertisers and their agencies agree to indemnify and protect the publisher from any and all claims, action, or expense

arising from advertising placed in SERVO Please

Mass Customization

I’m convinced that the future of

robotics — especially service robotics

— is tied to the development and

acceptance of mass customization

Consider the typical challenges faced

by an enthusiast in creating a

complete robot First, there’s the

hardware platform — from the

supporting infrastructure to the drive

train to the effectors Unless you

start from scratch — a costly and

time-consuming affair — you must

identify and purchase a crawler or

roamer or other platform suitable to

your needs However, have you ever

tried mixing and matching an arm

from one manufacturer with the

body of another? Even if it’s possible

to readily bolt the effector securely to

the main body, there remain issues

from electrical connectivity to how to

best compensate for the extra weight

and shift in the center of gravity

Then, there’s the issue of sensor

selection — from US range detectors

to IR radar Again, this is a task- and

cost-dependent selection, assuming

what you need is even available

off-the-shelf There’s also the power

system, communications system, and

all-important computational

hardware and software to consider

Typically, you must either cobble

together a unique assemblage of

components from a variety of

vendors or buy a pre-configured

robot from a single vendor that’s (at

best) a compromise On top of it all,

because your final robot is a unique

assemblage of components, sensors,

effectors, and software, the software

routines crafted by another developer

will probably require significant

modification before they’re useable

in your robot

Other industries facing the samechallenges turned to the masscustomization model, often withastounding commercial success Forexample, take Dell and Apple Whenyou order a PC from the Dell website,you scroll through pages describingoptions and prices You addwhatever features you can’t livewithout — while knowing thatregardless of your choices, themachine will boot and run yourfavorite programs The Apple websiteoffers many of the same designoptions — including drives withvarious capacities, DVD and CDplayer/recorder options, extramemory, video memory, and monitorsizes This is possible because eachcomponent manufacturer abides byindustry standards

Mass customization has beenapplied successfully to other, lessobvious industries, as well Forexample, one of my interests isbuilding, refurbishing, and playingelectric guitars However, as withrobotics, designing a guitar involvescareful consideration of myriadfactors Critical issues range fromneck size and geometry and thenumber and configuration ofelectronic pickups, to thecomposition of the body All of thesefactors and more affect the tone andplay-ability of a guitar Furthermore,

to put the pieces together requires

an expensive set of specialized toolsand the expertise to use them

The mass customizationcompanies catering to custom guitarbuilders (such as

Mind / Iron

by Bryan Bergeron, Editor Œ

Perform proportional speed, direction, and steering with

only two Radio/Control channels for vehicles using twoseparate brush-type electric motors mounted right and leftwith our mixing RDFR dual speed control Used in manysuccessful competitive robots Single joystick operation: upgoes straight ahead, down is reverse Pure right or left twirlsvehicle as motors turn opposite directions In between stickpositions completely proportional Plugs in like a servo toyour Futaba, JR, Hitec, or similar radio Compatible with gyrosteering stabilization Various volt and amp sizes available.The RDFR47E 55V 75A per motor unit pictured above.www.vantec.com

STEER WINNING ROBOTS

WITHOUT SERVOS!

Order at

www.FirstAct.com) take care of the standards issues,

allowing you to focus on appearance and geometry The

universe of possible components isn’t available to you,

but the component selection is generally rich enough to

create a custom guitar suitable for any budget As a

result, with a click of the mouse and credit transaction a

custom guitar can be at your doorstep in a matter of

weeks No need to fully equip a woodworking shop or

learn how to wire your own pickup coils, and (given the

skill) you can play music written for any other electric

So, if mass customization can work for other

industries, why has it been slow to take off in robotics?

One factor is the lack of a deep-pocketed Dell or Apple in

the world of robotics A few hardware platforms have

been widely accepted by the enthusiast community — the

Parallax BoeBot comes to mind However, most robot

models sell in the hundreds unless they make it to the

game/entertainment market

The underlying issue is, of course, the lack of an

industry standard robot configuration (akin to the PC) or

a proprietary design with a devoted following (akin to the

Apple Mac) Certainly, standards have been advanced to

assist in the development of reusable robot software

components (such as the Microsoft Robotics Studio) and

several vendors offer platforms ready to be populated

with components But there are no real robotics

standards from a body large enough to enjoin the entire

consumer robotics industry, such as the IEEE

So, what can we do, as mere enthusiasts, to nudge

mass customization of robotics along? Do what

developers in other industries do: Identify best of breed

and then integrate them into a good product line The

first hurdle is, of course, defining “best,” given each

component manufacturer has a stake in declaring their

product as such

Take a moment and imagine what you’d do, given

unlimited access to commercial robotic components and

systems What’s the best platform out there for small

service robots? Medium sized? What are the best (i.e.,

most affordable, easily programmed, compatible, etc.)

controllers? Sensors? Motors?

As a first step in developing industry recognized

standards, a consortia of companies and developers could

interface the best out there and then provide options to

suit different classes of users If you or your

group/company is up for the challenge — or currently

engaged in the challenge — then please drop me a line

SV

Stereo Vision System

Introduced

If your bot or other homebuilt

device needs 3D vision, check out the

Surveyor SVT™ : a camera,

dual-processor Wi-Fi system geared for

robotics, embedded image processing,

and Web-based remote monitoring

Surveyor (www.surveyor.com)

points to features including on-board

programmability, Wi-Fi connectivity,

easy sensor and actuator interface,

open source architecture, and a list

price of $550 as key attributes

You may note from the photo

that the SVT employs Analog Devices’

Blackfin processors, which are

designed for such things as

multi-format audio, video, voice and image

processing, etc (details at www.ana

log.com) According to Surveyor

literature, the system takes “full

advantage of the processor’s power

efficiency, optimized video

instruc-tions, high speed video interface, and

easy interface to peripheral devices.”

It incorporates two BF537 32-bit

Blackfins, two Omnivision OV9655

1.3 megapixel cameras, PWM motor

control, various interfaces, and

3.3V and 5V regulation for battery

operation You also get a Lantronix

Matchport 802.11bg radio, all of

which takes up only a 2.5 x 6 x 2 inch

volume and consumes less than 2W

Programming can be via Windows,

Mac OS X, or Linux

Bee Safer on the Road

At the recent CEATEC Japan

show, Nissan (www.nissan-global.

com) unveiled its concept for

automotive safety in the form of theBiomimetic Car Robot Drive, alsoknown as BR23C According toNissan, it is based on lessons learnedfrom the “humble bumblebee,”

specifically the bee’s vision system

The insect’s compound eyes canspot obstacles in a 300°+ range, allowing

it to fly safely within its personal space

In the automotive version, Nissan’sLaser Range Finder (LRF) system covers180° and has a sensor range of 2 mtoward the front The LRF calculatesthe distance to another object andsends the data to an onboard micro-processor, which translates it into acollision avoidance strategy But a beecan bumble up, down, or sideways toavoid other bees, whereas the BR23Ccan only move in two dimensions andonly within the limitations of the wheels

Presumably, the operating range of

2 m is a prototype limitation, as it wouldappear advisable to detect a pendingcollision with a bit more time to react

Let’s say you’re driving toward a phone pole or other stationary object at

tele-60 mph (88 ft/sec) and detect an cle at a distance of 10 ft That wouldallow only about 1/10 sec reaction time

obsta-— enough time for an airbag to deploy,but not so hot for braking or swerving

Nissan says the device “only needs

to process inputs every few seconds,and act on that.” Hmmm In any event,the company hopes that the system will

cut the number of accident fatalitiesand serious injuries to Nissan drivers byhalf by 2015, as compared with 1995

And Bee Careful of Flying Objects

Another ostensible offspring ofthe genus Bombus is the KillerBeeunmanned aircraft system, which isbeing provided to the US Navy andMarine Corps by a team made up

of Swift Engineering and Raytheon

(details at www.killerbeeuas.com).

It looks more like a manta ray leapingthrough the surface of the gulf, butit’s considerably more lethal thaneither one when connected to com-bat and command control systems.According to a recent announcement,

“KillerBee has the ability to insertpersistent intelligence, surveillance,and reconnaissance (ISR) into the bat-

tle space and rapidly deliveractionable intelligence tocombatant commanders.” Late in September, aRaytheon flight operationscrew simulated a combatenvironment by deliveringthe KillerBee system to

a remote location viaHumvees In less than 45minutes, the crew set up

by Jeff Eckert

The SVT™ dual-camera, dual-processor stereo vision system.

Courtesy of Surveyor Corp.

Nissan’s BR23C uses compound “eyes” for collision avoidance Courtesy of Nissan Motor Company

Advanced Technology Center.

The KillerBee — a fully autonomous UAS — will be used by the US Navy and Marine Corps for both ground-based and ship-launched surveillance and reconnaissance Courtesy of Raytheon Corp.

the system and launched the UAS.

The team then executed the

opera-tional scenario (details not revealed)

and safely retrieved the UAS

Specs are sketchy, but we did find

out that KillerBee is fully autonomous,

with no piloting skills needed by the

operator Maximum takeoff gross

weight is 164 lb, and it’s powered by

an 8 hp engine fitted with a

two-blade prop The stall speed is 45 kt,

and loiter speed is 55 kt Mission

endurance is 15 hr, and it has an

impressive range of 972 nautical miles

AUV Hovers Underwater

Researchers at MIT (www.mit.

edu) have revealed the new Odyssey

IV — a small, inexpensive robo sub that

can hover like a helicopter, which

should be really handy for deep-water

oil explorers, marine archaeologists,

oceanographers, and other folks who

need to loiter undersea for long periods

of time Whereas previous Odyssey

models could operate only while

moving forward continuously, the IV

can stop anywhere in a water column

and constantly correct for currents

and obstacles This makes it useful

for things like making detailed

inspections of an offshore oil

plat-form and photographing flora and

fauna around an undersea vent In a

sea trial completed off Woods Hole,

MA, it performed at depths of up to

6,000 m It can also be fitted with amechanical arm, allowing it to grabsamples and perform manipulations such

as twisting a valve or clonking a lobster

on the head Odyssey IV is pretty swift,too, with a top speed of 2 m/sec

The ultimate goal is to develop aversion that can stay underwater for

up to a year, collect and transmit data,and ultimately return to its home basewithout the need to surface at all

RoboVox: All Talk

For journalistic purposes, it is sionally necessary to be flexible with thedefinition of a robot, and in the case ofRoboVox we sag from flexible to flaccid

occa-A creation of artist Martin Bricelj, the 8

m colossus doesn’t move or do anythingrobotic, but at least it looks like anautomaton Lest you leap to the conclu-sion that this is merely a pitiful cry fornotoriety from a starving artist, notethat it is presented as a socially relevantvehicle for freedom and self-expression

According to the creator, “Its purpose is to serve as a tool for anindividual whose voice usually getslost in the sounds of the mass, the

society An individual can send a textmessage using his mobile phone tothe dedicated RoboVox’s number.Upon receiving the SMS, RoboVoxsays out loud the statement, theprotest, the declaration of love, orwhatever the message may read, thuslending its voice to the anonymousindividual.” If you’re impressed, youcan purchase RoboVox merchandisesuch as airline bags, T-shirts, and talk-

ing soft toys Just log onto www.

robovox.co.uk and take your pick It

may be the perfect Christmas gift forthat relative who has everything SV

R o b y t e s

Schematic shows Odyssey IV — a

small, inexpensive robotic submarine

that can hover in place Courtesy

of MIT Sea Grant AUV Lab.

RoboVox appears in downtown Maribor, Slovenia.

The basic EPAM is a rubber sheet

made of a custom-formulated

elastomer, explains Ilya Polyakov,

a senior mechanical engineer at

Artificial Muscle, Inc., creators of the

EPAM The polymer is applied so that

electrodes plate the rubber sheeting

on both top and bottom As voltage

is applied positively and negatively

through these electrodes, the

dielectric rubber — which doesn’t

permit the flow of electricity —

produces an electrostatic force

This force pulls the electrodes

together, pushing the EPAM out

side-to-side at tenss of

pounds-per-square-inch This actuation is then

harnessed for practical uses, including

robot muscles

“Because our muscle membrane

squeezes out proportionally to the

electrical field applied, we can simplycontrol the amount of movement,”

says Polyakov

Harnessing EPAM Properties

One of the most common building blocks using the EPAM is the universal muscle actuator (UMA) configuration, which is a jumping-offpoint for more complex applications

This configuration attaches two EPAMsheets/films at their centers, with alightweight spacer between them

“Each stack of film is attached to

a frameso it can be mounted Eachstack is also attached to an outputdisk which transmits the force andstroke of the expanding film to someload in order to perform work,”

explains Polyakov One practical usefor EPAMs and UMAs is in valves.Existing valve systems convert highspeed rotary motors running at thousands of RPMs to low speed linear motion for valve parts thatmove a few feet per second

The EPAM actuator eliminates themechanical steps necessary for thespeed conversion by offering lowspeed linear motion in the initialinstance This technology is called aroll configuration actuator, whichemploys a large stretch of artificialmuscle film/membrane wrappedaround a compressed spring

The spring pushes the cylinder ofwound film into tension, Polyakovexplains “Once in tension, the application of the electric field to the film causes it to relax and produce a force imbalance betweenthe spring and the connected films,

Contact the author at geercom@alltel.net

by David Geer

Artificial Muscle (EPAMs)

Today, Robotic Applications to Follow

Electroactive Polymer Artificial Muscles (EPAMs) are a new actuating/motion technology based on polymers that react to electricity The new actuators are useful across applications where motors are not as efficient or are simply not feasible

The valve on the left is an OTS proportional solenoid valve The one

on the right is EPAM powered with identical flow specs, and a directly attached power supply board It weighs 1/7th the original unit and operates

at 1/10th the power.

This larger solenoid coil at left is used

in commercial pneumatic valves

The coil at right — a 9.5 mm actuator stack based on EPAM technology

— is capable of the same workload

at a fraction of the commercial solenoid’s mass and volume The commercial part weighs 70 grams compared to the EPAM’s 8.

This small piece of hardware uses EPAM

technology to auto focus optical lenses

in mobile/cell phone cameras,

according to Polyakov A 9.5 mm EPAM

actuator positions the lens based on

data from a CMOS sensor The

hardware is encircled by a flex circuit

power supply at 1.3 kV The unit is

specifieded to fit the standard SMIA

9.5 mm phone camera form factor.

causing movement.”

These actuators are made up of

only the EPAM membranes, a spring,

and end caps, which are cheaper,

simpler, and lighter than the several

parts and the electric motor otherwise

required to do the same job

Further Applications

Like motors convert rotary

movement to linear today, EPAM

technology will be used to convert

repeated linear motion into rotary

motion, such as is seen with a crank

and pistons in an internal combustion

engine — but much more cheaply

and efficiently in a more compact

size and space

“We can also perform accumulated

or step-by-step linear motion One of

the simplest examples of this is in theinch worm, where the actuatorrepeats a motion many times to produce linear travel much greaterthan the singular stroke of the actuator,” says Polyakov

In a sort of opposite application,EPAM is used to convert linear torotary motion This technology is arotary, one-way clutch which enables

a shaft to rotate in only one directionwhile the other side of the clutch isconnected to an arm which, in turn,

is connected to an EPAM actuatorthrough a linkage apparatus

“As the actuator oscillates backand forth, one direction of movementlocks the rotary clutch to the shaft,resulting in the shaft rotating a few degrees, while the other lineardirection unlocks the clutch andallows the actuator and linkageassembly to move free of the shaftreturning to the beginning of thecycle,” summerizes Polyakov Everystroke of this EPAM actuator

implementation turns the shaft a littlebit A continuous, high speed shaftrotation is therefore achieved throughmany thousands of cycles

Conclusion

While there is a view to the future

of EPAM with robot muscles in mind,there are some near-term obstacles.One is that the range of motion doesn’t scale for EPAMs as easily asthe force emitted by them “Our actuator stroke range is on the order

of 10 percent,” says Polyakov Robotsclearly require a much broader range

of motion SV

GEERHEAD

This is a demonstration model of a

buckling frame actuator which applies

extreme tension to the EPAM film The

frame is pliable enough that the film

buckles (the actuator’s starting form is

flat as compared with the apparent

curved wing-like form shown here) “This

forces the bent frame to act as the bias

spring in equilibrium with the film,”

explains Polyakov As the actuator

activates, the frame flattens out the film.

“The wings have been added to amplify

the effect/result of the motion They also

make for a very cool demonstration!”

This large stack of several, round actuators has been formed to activate

an automobile accessory clutch such

as an air-conditioning compressor pulley connected to a serpentine belt

on certain car models (my Buick Skylark, for one) The advantage is the reduction in weight, size, and required power over a legacy electromagnetic clutch The unit — shown next to an

AA battery — is quite miniscule.

This is a planar actuator configuration for driving a rotary shaft The circular black plate consists of eight pie-slice electrodes When any one of these electrodes actuates, the center of the disk shifts in the plane, engaging a crank

on a shaft By actuating the electrodes in circular succession clockwise or counter- clockwise, the configuration moves a shaft in a circular motion, creating an EPAM version of a stepper motor

EPAMs are a competitive technology

when compared with electric motors and

other actuators Their benefits include:

Virtual silence The muscle uses

no moving parts, so the actuator is

virtually silent; applicable where noise

is undesirable.

applicable under much tighter size constraints than electric motors, so therefore useful for applications in small cameras and consumer electronics.

More jump per joule “If you compare energy vs work output, we get more work per joule than piezo or

Advantage: EPAM

These are two miniature pumps based

on EPAM actuator technology One is a

35 mm pump, shown with a US quarter for reference, the other one is a 12 mm pump, shown with a dime The pumps use the same size pump head “That’s the cylindrical part with the two tubes exiting it,” explains Polyakov.

Photos courtesy of Artificial Muscle, Inc.

Know of any robot competitions I’ve missed? Is your

local school or robot group planning a contest? Send an

email to steve@ncc.com and tell me about it Be sure to

include the date and location of your contest If you have a

website with contest info, send along the URL as well, so we

can tell everyone else about it

For last-minute updates and changes, you can always

find the most recent version of the Robot Competition FAQ

at Robots.net: http://robots.net/rcfaq.html

— R Steven Rainwater

De ce mber

4-31 ROBOEXOTICA

Museumsquartier, Vienna, Austria

In this event, robots are tested on servingcocktails, mixing cocktails, bartendingconversation, lighting cigarettes/cigars, andother achievements in electronic cocktailculture

www.roboexotica.org/en/acra.htm

5 Robotex

Tallinn University of Technology, Tallinn, ESTONIA

Autonomous robot must clean a room by gathering up socks, empty cans, and other strayobjects The objects must be placed into a specialgathering area

www.robotex.ee

6 Penn State Abington Robo-Hoops

Penn State Abington, Abington, PA

Robots pick up foam balls and shoot or dunkthem into a basket to score points

www.ecsel.psu.edu/~avanzato/robots/

contests/robo-hoops/

17-21 IROC International Robot Olympiad

Kuala Lumpur, MALAYSIA

Includes a wide range of events including line-following, maze solving, legged robot contest,robot prison break, FIRA soccer, robot dancing,and an open category

TBA PAReX Autonomous Robotics Competition

Shiloh Community Center, Phoenix, AZ

This event includes mini-Sumo, maze runners, and line-following

www.parex.org/autoevent1.shtml

Ja nua r y

20-22 Singapore Robotic Games

Singapore Science Center, Republic of Singapore

Another big multi-event competition Some of theevents planned are autonomous Sumo, RC Sumo,robot soccer, wall climbing, pole balancing, underwater robots, legged robot marathon,legged robot obstacle race, and others

http://guppy.mpe.nus.edu.sg/srg/

24-26 Techfest

Indian Institute of Technology, Bombay, India

Autonomous bots compete in Micromouse andthree other events, called Pixel, Full Throttle:Grand Prix, and Blitzkrieg

www.techfest.org/

29 Robotix

IIT Khargpur, West Bengal, India

Now at nine years old, this is supposed to be the largest robot competition in India Expect tosee over 1,000 teams competing and over 2,500spectators this year

www.robotix.in/

TBA Kurukshetra

Guindy, Chennai, INDIA

This event includes Designer’s Quest, Solenopsis,Freightocog, and Catch Me If You Can

Q.For a year now, I have been researching a solution

to the problem of reliably determining a robot’s

position to within one inch in a 25 square mile

area Cost is a major design consideration (ideally less than

$1,000) Lasers cannot deal with obstacles (and I’m also

pretty sure my dogs won’t tolerate the goggles) GPS

augmentation is prohibitively expensive I’ve found a few

RF designs, but none that can be built on a budget that

would be approved by my CFO (wife) Hopefully, you can

provide me a solution or point me in the right direction

P.S I really enjoy your column!

— Thanks, Roger

A.First off, thanks for the kudos! I don’t know why you

need such exacting accuracy; it will definitely be a

challenge since I don’t think there is anything that

can provide that level of accuracy that is built into a

roving vehicle You’re going to need to “instrument” your

environment some how

Even differential GPS can only get an accuracy of — at

best — one meter One paper that I read — www.gdgps.

net/system-desc/papers/ion_paper_2000.pdf — suggests

a sophisticated Internet-based correction method for differential

GPS that can get an estimated accuracy of about 20 cm So

clearly, GPS by itself cannot get you what you want Regardless,

that will set you back way more than your $1000 limit RF

systems are very uncertain for determining ranges well and

clearly IR rangers are of limited value outdoors Okay, that

takes care of what WON’T work for you Now, let’s step off

of the deep end and speculate on what might do what you

want in your limited area of a quarter mile

My first thought would be to position two video cameras

at right angles to each other on tall towers aimed down on

the area within which you are tracking your robot I realized,

however, that I was thinking a quarter acre, not quarter mile

That wouldn’t work because you couldn’t get the whole space

in the frame at a useful resolution So, strike that idea

dig a hole, some of them will dig down to the correctdepth and place a spinning laser to show the depth allaround the hole But as you’ve noted, your dog and probablyyour neighbors wouldn’t be too happy about that Too bad,since that laser would be fairly easy to detect with a camera

on a stepper motor, whose angles could then be comparedwith each other to get an absolute location of the source

of light to a pretty good resolution With more sensors,even more resolution could be gained So, what else can

we use? Sonar is out since the sound waves will spread outtoo far as they move away from the robot

How about this idea: put a bright rotating red (blue,green, whatever) light on the top of your robot This will bereasonably directional with minimal dispersion of the light —and you can perhaps make the point source even moredirectional with a tube limiting the direction of the lightfrom the reflector Next, put a video camera with a limitedfield of view and a filter on the lens that will admit only thecolor that you are looking for at the four corners of yourarea Place it high enough so that there is nothing blockingthe view of the area This assumes, of course, that you don’thave much in the way of shrubs and trees Some barriersare okay as long as at least two adjacent cameras can seethe target Now, scan the cameras using a motor that can

be precision turned like a stepper motor or a servo I woulduse an encoder to further enhance the accuracy of the posi-tioning By using a Cartesian coordinate system and somebasic trigonometry, you can calculate the location of therobot within the property by comparing the readings fromtwo orthogonal (fancy word for “at right angles”) cameras.The biggest resolution limiter is the cross section of thepoint source (the light) You can’t be more accurate than thewidth of that point source as seen by the camera I think thiswill work but you’ll need a computer to coordinate the readings,calculate the angles, and come up with the actual range.The light will need to be strong enough to reach the cameras

at a brightness that will stand out against the ambient light

Tap into the sum of all human knowledge and get your questions answered here! From software algorithms to material selection, Mr Roboto strives to meet you where you are — and what more would you expect from a complex service droid?

lots of things that will prevent this idea from working all of

the time, like light interference, the sensitivity of the camera,

or the expense of the light filters Other tuning (like the angle

of the cameras to the ground) could also cause issues

The problems in this type of system may be

insurmountable, so let’s think even MORE inside the box

Suppose we came up with small RF transmitters that were

carefully designed to have a limited range of about a meter

when used with a carefully tuned receiver Then, suppose

we “planted” these transmitters in a grid at one meter

intervals You could calculate your position by knowing your

range to each of these beacons by checking their field

strength Since it would take thousands to do this at one

meter intervals, you could put them at longer intervals, but

then their resolution will suffer and your accuracy will not

be very good Hmm, RF is just too unreliable

Note that I don’t even give odometry a nod here Since

you are outside, there is just about no way you will get any

kind of positional accuracy when you consider ditches,

bumps, wheel slips, and the like If you are going to use

some very sophisticated programming, you could use a full

IMU unit and calculate acceleration, deceleration,

inclina-tion, pitch, and yaw to create an inertial navigational unit

But I still doubt that you could get a one inch in a quarter

mile accuracy It would be fun to try it though! Google for

“Inertial Navigation” on the Internet I have seen some

interesting papers on this topic; some by other well known

hobby robotics folks in the well known clubs

Unless you can come up with something from my

rambling thoughts here, I’m afraid that all I’ve done is

convince myself that this just can’t be accomplished Atleast not within the $1000 budget restriction! If anyone out there reading this has some ideas, let’s hear them!With my roving robots, I’m happy if I can just keep theminside a set perimeter A GPS will get me within 10 meters

or so of the robot where I should be able to see it

Q.In last month’s column, I described a simple

PID algorithm and showed you code that used avoltage read from the motor CEMF to determinethe wheel’s rotational velocity Some who read that likedthe idea of a “sensorless” feedback system I received arequest to show how such a feedback mechanism could

be built and used, so, here it is

A.Last month, you saw the code and how I averaged

the readings and used the values read by turning offthe PWM and looking at the CEMF (Counter Electro-Motive Force) at those intervals This month, I’ll post theelectronic circuit that I used to get those values Once again,you’ll see that it is actually very simple to do this, and the nicething about it is that you won’t need finicky wheel encodersfor feedback These inputs come from the motor itself.Figure 1 shows the schematic of the circuit that matches the code that I gave in the November issue

How the Circuit Works

Some explanation of how this works is in order Whenyou drive a motor, you feed voltage to the windings on thearmature (the part that rotates) through contacts on the

Figure 1 Sensorless motor speed

commutator This voltage induces a current in the windings

proportional to the voltage divided by the resistance of the

windings The electromagnetic field then generated in the

windings will attract part of the armature to a magnet in

the motor can, which causes the armature to rotate

There are several sets of windings on a motor; how many

depends upon the motor Each set of windings will have its

own pair of contacts on the commutator Brushes are

connected to our PWM output and transfer the current

from our H-bridge to the motor windings I don’t want to

get into how a DC brushed motor works any more than

that; it is a fascinating topic all by itself We know that

rotating a coil of wire in a magnetic field (in this case, the

one generated by the magnets in the motor can) will

induce a current in the coil This induced current will be in

the opposite direction as the current that generates the

field in the motor windings The end effect of this induced

current is that as the motor spins faster, the current

generated in the windings by rotating in the magnet’s

magnetic field will get larger While we have the motor

turned on, the effect of this CEMF is that the current

required by the motor will be reduced as the motor spins

faster When we turn the PWM off to the motor, this

induced current will still be there and proportional to the

velocity of the motor rotation We will read that current

with the analog to digital converter (ADC) inputs on the PIC

to determine the motor speed

The resistor networks R1/R3 and R2/R4 will divide downthe CEMF voltage proportionally so that we can get as much

of the full range of change as possible The zener diodes D1and D2 will limit any voltage spikes we could get to a levelthat won’t damage the ADC circuitry The diodes D3 and D4will shunt negative voltages that may become present when

we shut the PWM off so that we don’t really damage our ADCinputs The output of all this limiting and filtering will then

be fed to AN0 and AN1 on the PIC to provide the readingsthat we need to determine the relative motor speed

In last month’s code, you saw that we looked at bothsides of the motor input to determine which way the motorwas turning On one side of the motor, there will be a negative voltage, which we shunt to ground so only -0.6V

is present; the resistor dividers will reduce this even further.The PIC will see that voltage as 0V; the other side of themotor input will be a positive voltage If the motor spinsthe opposite direction, then these voltage polarities willreverse and thus we can tell which way the motor is spinning Listing 1 shows how this is done in the code

Choosing Components for the Circuit

Your choice of resistor values depends upon your motorvoltage Here, I am using a 6V to 7.2V battery to power mymotors My choice of resistors (2.2K and 4.7K) gives me a

stepdown of about 0.7 which means the mum voltage that I’m likely to see is about 4.9V(which is pretty close to the ADC full swing of 0V

maxi-to 5V) You calculate this resismaxi-tor divider as lows: R3/(R1 + R3) * Battery Voltage = MaximumVoltage seen at ADC input So, my example is4.7K/(2.2K + 4.7K) = 0.68; 7.2V * 0.68 = 4.90V.You might see some healthy

fol-current spikes to the negative direction on yourmotor contacts so diodes D3 and D4 should be

at least one amp; I like 1N5817 diodes They willwork well for smaller motors; you’ll want to scale

up for really big ones The zener diodes D1 andD2 won’t be handling more than a few milliamps

of current, so just about any 5.1V zener diode thatyou have in your junk box will work well here.There are many enhancements you could do

to make this circuit work even better than it does

If you used an op-amp between the sensor resistorsand the ADC inputs, then you could adjust youramplitude and even offer some active filteringthat will give you more consistent readings Thequality of your motors will also have some effect

on how good of a job the circuit does, but I’venoticed that even cheap motors work well.Well, that wraps up this month’s reading sessionwith Mr Roboto I hope that you’ve learnedsomething or at least have enjoyed what you haveread As usual, if you have any questions aboutthings robotic, please drop me a line at roboto@

signed long getError()

/* Find the difference between where we want to be

and where we are */

{

signed long error;

unsigned int16 ma,mb;

setup_ccp1(CCP_OFF);

//Turn off PWM delay_us(500);

//wait for steady state set_adc_channel(0); //get forward side voltage delay_us(20);

ma = read_adc();

ma <<= 1; //This “amplifies” the

//signal, making it more //relevent

set_adc_channel(1); //get reverse side voltage delay_us(20);

mb = read_adc();

mb <<= 1; //This “amplifies” the

//signal, making it more //relevent

currPos = (signed long)((signed long)ma - (signed long)mb);

// Uncomment to debug //printf(“ma= %lu mb= %lu\n\r”,ma,mb);

setup_ccp1(CCP_PWM_PLUS_3);

error = calcPos - currPos;

Listing 1 ADC motor speed readings.

power supplies or the

CS220001 for European power supplies These kits

provide everything necessary to build a complete

Mechatronics Laboratory

They include the 48-pin C Stamp, a μC 101 BOL, PC

programming cables, power supply, instruments, and the

Mechatronics Parts Kit Learn programmable mechanical

control, infrared speed and direction detection, and a

host of other Mechatronics concepts

The Mechatronics Parts Kit includes sensors, display,

motor, etc The electronic components required for the

Mechatronics activities are also included in the Parts Kit

The activities introduce motor control and interfacing,

electronics, and transistor theory, just to name a few

concepts All necessary software, documentation,

curriculum, and highly-illustrated instructions written

by college professors are included in the kit CD

These Mechatronics kits are unique in the sense that

hobbyists, students, and autodidactic users alike work

with and develop embedded systems using professional

software development tools and a professional

programming language in a very user friendly

environment

The power and versatility of the C Stamp guarantees

that learners can expand upon their Mechatronics

knowledge with their own ideas, projects, and

imagination Well thought out and pedagogically sound

examples developed by college professors make sure that

learning Mechatronics with these kits is a powerful and

enjoyable experience

All items are new, fully tested, factory direct, and all

software is included at no extra charge

For further information, please contact:

RA-02 Robotic Arm Kit

The RA-02 RoboticArm from Images

Co teaches basicrobotic principles TheRA-02 kit includes allcomponents necessary

to build the roboticarm shown here.The robotic armkit also includes the SMC-05 five-position servomotorcontroller kit The servomotor kit includes the PCB and allcomponents necessary to build the servomotor controller(soldering is required) Using the five-position servomotorcontroller, one can control the robotic arm using theonboard three position switches to manually grab, lift,lower, wrist rotate, and pivot In addition, the Windows(98/XP) control software is available as a free download.This allows PC control through one of the PC’s serialport The Windows program has a script writing featurethat allows one to program the robotic arm with thousands of movements with full, precise, and repeatable control of each movement Scripts may hold

a maximum of 10,000 movements (including pauses)with a maximum repetition of 10,000

New AX-12 Smart Arm Web Control from

CrustCrawler Interactive

The CrustCrawlerInteractive grouphas released a simpleyet robust Telerobotics

EDUCATIONAL KITS

N E W P R O D U C T S

Tel: 718•966•3694 Website: www.imagesco.com

Images ScientificInstruments

656 Ironwood Dr.

Williamstown, NJ 08094 Tel: 800•985•AWIT or 800•985•2948 Fax: 800•985•2948 Email: info@a-wit.com Website: www.c-stamp.com

A-WIT

Technologies, Inc

TELEROBOTICS SOFTWARE

for remote, web control of any robotic platform.

Currently, the Tele-Robotics application contains control

and monitoring services for CrustCrawler’s AX-12 Smart

robotic arm If you do not own an AX-12 Smart Robotic

Arm, you can test drive the AX-12 Smart Arm through

CrustCrawler’s interactive Telerobotics web page Move

the arm through several preset movements while

monitoring in real time the position, speed, load, voltage,

and temperature of each of the AX-12+ Smart servos

from anywhere in the world

CrustCrawler’s Tele-Robotics software is ideally suited

for robotics applications that use servo motors which

can provide real time feedback for position, speed, load,

voltage, and temperature Such feedback features are an

integral part of the AX-12+, RX-28, RX-64, and EX-106

series of servos from Robotis

CrustCrawler offers the software as a free open

source download which contains fully commented client,

server, and AJAX code modules Help and additional

code can be found on the CrustCrawler robotics forum

Visit the website for more information and code updates

For further information, please contact:

HS-5485HB Economy Digital Servo

Hitec's new HS-5485HB digital servo features longlife components and zero wearKarbonite gears The HS-5485HB

is an inexpensive, fully programmable standard sizedigital servo

Features include:

• 89 in/oz of torque at six volts

• Backed by a two year warranty and Hitec’s service and

support

• Suggested retail price $25

For further information, please contact:

Servo Commander

The Servo Commander is a portable, hand-held devicethat generates the command signal required to driveservos and Electronic Speed Controls (ESC) without theuse of an R/C transmitter and R/C receiver You canmanually adjust the servo position or the ESC speed,

or continuously cycle between two slewing endpointpositions of your choice

Features are interference-free servo checking whenamong groups using live R/C equipment; test servo travel, positioning, and throttle adjustment; check forservo alignments, binding, damaged gear trains, or weak motors; enable mechanical adjustment of precisezero position; check step response; permit first order frequency response; check for vibration problems;

facilitate mechanical break-in; perform endurance testing;input is protected against reversed voltage; conveniencepower to driven equipment fused at 1A; accommodateseither 1.0 or 1.5 ms zero signal; always powers-up producing a zero output signal; output signal is short circuit protected; allows normal or reversed positionalcontrol; has teach function to enter arbitrary slew endpoints; slew rate is variable from 0 (motionless) to100%; has quartz crystal controlled accuracy; and typespecific power cables are available

CrustCrawler, Inc

Is your product innovative, less expensive, more functional, or

just plain cool? If you have a new product that you would

like us to run in our New Products section, please email

a short description (300-500 words) and a photo of your

product to:newproducts@servomagazine.com

Show Us What You’ve Got!

12115 Paine Street Poway, CA 92064 Tel: 858•748•6948 Website: www.hitecrcd.com

Hitec

Multiplex

3662 Brandi Drive Sterling Heights, Mi 48310 Tel: 586•274•9792 Website: www.sviind.com/servo

SVI

TOOLS & TEST EQUIPMENT

MOTORS

Featured This Month:

Laying Out Your Work

Area by Blake Hooper

ROBOT PROFILE – Top

Ranked Robot This Month:

26 Micro Drive by Kevin Berry

The wedge is one of the oldestand most successful types ofcombat robots The effectiveness

of a simple ramp with good drivepower has yet to be matched byany other robot design However,throughout the years the wedgehas managed to become themost hated design as many claimthey lack any sort of originality

or excitement There has been, however, a small community offast, well driven wedges that provide a good amount of excitement as they bouncearound the arena It was theserobots that inspired me to onceagain jump back onto thewedge bandwagon, thistime in the three poundbeetleweight class

Like I have been doingfor all of my recent robots,

I did all of the robot’sdesigning in SolidWorks3D CAD Before I couldbegin drawing up the

chassis however, I needed toknow how large it would be, asopposed to simply tinkering with

it until everything fit To do this,

I took all of the parts I had, (substituting paper cutouts withthe correct dimensions for theparts I did not have) and laidthem out in the most compactconfiguration possible, while leaving some space for wiring

It is important to have yourrobot’s dimensions as small

as possible Not only does itdecrease how large of a targetyour robot is, but since your chassis components will be

● by Thomas Kenney

Cloud of Suspicion

BUILD REP RT

FIGURE 1 The robot is drawn

up in Solidworks CAD before construction begins.

smaller, they will weigh less, and be

stronger since the mounting areas

will be closer together After all of

the basic dimensions were finalized,

I drew up the rest of the robot in

SolidWorks The completed design is

shown in Figure 1

The first part of the robot to be

designed was the drive system My

first choice was two Banebots 36

mm gearboxes with the RS motors

I was on a very tight schedule, and

hadn’t done belt drive before, but I

didn’t want to go to two-wheel

drive either My solution was to

switch to the 28 mm gearboxes

opposed to the 36 mm ones I

would lose a bit of durability, but

there was also a weight difference

of four ounces per gearbox With

this weight difference, I was able to

go to four-wheel drive, even though

the drive system took up half of the

robot’s weight With 1.875” wheels

and on 7.4 volts, I would get a top

speed of 6 mph and near

instanta-neous acceleration; certainly zippy

enough for a 8 x 8 arena Next, I

began on the chassis Because of

the huge amount of weight the

drive motors, gearboxes, and wheels

took up, there was only 1.5 pounds

out of three left for everything else

— including the chassis, electronics,

and the wedge itself To make

weight while still maintaining a

strong frame, I used a material I was

familiar with, and more importantly,

already had copious amounts of —

UHMW polyethylene

UHMW is known in combat

robots for its great structural

properties It is relatively soft, but

just enough that it won’t crack

under most circumstances UHMW

is also very light, weighing in at just

75% of the weight of polycarbonate

and acrylic It also comes in several

different colors, including white,

black, and blue The material I had

was a white, five foot long, 1.5” x

3/8” strip, which was the perfect

height for my robot’s chassis For

the top and bottom, I needed as

stiff and light a material as I could

get My first and final choice was

carbon fiber, a compositematerial made up ofstrands of carbon clothbonded through an epoxylaminate It is very stiff,which is a good counterfor the flexing tendency

of UHMW One of the only drawbacks is its price,which can total more than

$50 for a 12” x 12” x 1/6”

thick sheet

Since my UHMW wassimply a long strip with thecorrect width and thickness, all Ihad to do was to cut the correctlysized pieces To do this, the UHMWstrip was marked at the right intervals, and then moved down

a table saw Once this was done,

I had all of the basic pieces used toform the chassis Cutting the carbonfiber wasn’t as easy

When carbon fiber is cut, theblade generates a hazardous finedust, so extreme caution needs to

be taken as inhaling this dust can

be a catalyst for lung cancer

Getting carbon fiber in your skin isnot as hazardous, but once it’s in,it’ll itch for weeks To negate theaforementioned spray of dust, shaving cream was placed down the cutting line When the materialwas cut, the cream absorbed all ofthe motes of dust

After cutting the carbon fiberpieces down to their basic dimensions, they were placed onthe table saw again to cut out thewheel wells After this was done,the only work left on the chassiswas to drill and tap 12 holes to hold the frame together, and thensome more to attach the top andbottom pieces; overall about 45minutes of work With this done,

I had a finished chassis

Next, I began work on thewedge My first wedge robot had its wedge fixed to the chassis Though it was very sturdy, wedges with their leadingedges dragging on the groundgot under it easily In order to nothave this problem, I chose to go

with a hinged wedge this timearound As opposed to using a hardware store hinge, I cut a 3/16”titanium shaft to length and pressed

it into the UHMW side rails

Once the wedge was cut out

of the 3/32” 7075 aluminum, I cutseveral small UHMW blocks,attached them to the wedge, anddrilled holes in them a bit largerthan the shaft diameter I then simply slid the wedge assembly ontothe shaft This created a durable andeasy to fix hinge system that is stillrelatively light The completed wedgeassembly is shown in Figure 3

With all of the major construction finished, it was time

to mount the internal components.The Banebots 28 mm gear motorsall have a corresponding mountingpattern, which makes it easy tomount them to base plates and also makes it a simple task to switch them out between fights,

if needed To mount them, I created

a PDF file of the mounting patternand printed it out It was easy todrill through the marked areasstraight into the base plate where

FIGURE 2 The chassis is assembled with all internal components mounted.

FIGURE 3 The hinged wedge assembly is completed.

the motors would be mounted.

After the motors were mounted, the last thing I needed

to do was secure the electrical

components I mounted

my two cell 1,320 mAhThunderpower batteryunder my Spektrum BR6000receiver and simply zip-tiedthem to the base plate Ithen drilled through themounting holes of myScorpion XL speed controllerand used screws to secure

it After this was done, therobot was finished Soonafter the robot was finished, I gave

it the name “Cloud of Suspicion,”

which I found in a list of “SuggestedNames for Racehorses Expected to

Have Undistinguished Careers.”

In the end, the concept of alightning fast, overpowered wedge

in the beetleweight class ended upbeing very successful A few flawswere found, but easily corrected

It went on to win the rumble at its first tournament (destroying the arena hazard in the process)and then took home second

at its next event, where the aforementioned flaw prevented itfrom placing first Now that theseproblems have been fixed, I havehigh hopes for this robot in thefuture SV

When it comes to entering

combat or otherwise knownrobotics the easiest thing a builder

can do to help themselves get

prepared is to work out their layout

By layout of course I mean the site

in which they will be working

How you set this up is based on the

builder, but there are certainly some

tips which can make this process

much easier

The first and most importantpart is to figure out the space that

you will be working in Many people

have set up their own shops in places

like their garage only to discoverthey hate working there and end upwith hand tools on a chair in theirkitchen Pick a place with plenty oftable space that you enjoy working

in, and make sure it can handleheavy vibrations and flying scraps

Another important thing is tomake sure you have good airflow Ifyou can swing it, air conditioningwill be a godsend, otherwise makesure to have plenty of fans or open windows When working onprojects, heat quickly becomes burdensome Light is also an important consideration I have seenmany builders work in their garage

when the lights aren’t working andconsequently, blood starts flowing.Once you have your work spacelocation figured out, you’ll need

to choose what tools you want

No matter what, you’ll need a basicset of hand tools The actual sizes,measurement standard, and type isdependent on the builder The moreimportant thing to consider is whatpower tools you will need If you are on a budget, this can quicklybecome an issue The first factoryou’ll want to consider is whatweight class your robot will be in.This way you’ll know the type ofmaterial you will be using, the

thickness of it, and the size

of tools you will need.For a cutting tool, youhave several options If youwill be working with severalsections of long bar stock

or are cutting particularlythick material, a tablemounted circular sawmounted to the floor will

do fine Make sure you areusing the proper blade Ihave seen many builders

and ready for battle.

Standard shop layout Hand tools and

electrical work stations are clearly within

arm’s reach of each other The bot is Ze

Uber Wedge, of Team Raybotics.

Drill press and accessories are near by — messy, but easy to get to.

use aluminum blades on steel

only to discover a few weeks

into it that their blade can now

only cut soft cheese

If you are cutting thin

materials such as aluminum

and need specific shapes or

corners, a jigsaw can be quite

useful The same goes here for

the blade, but also make sure

to have the material you are

cutting firmly mounted Finally,

if you are cutting thin material

in small amounts and cannot

afford anything powered, a

hacksaw can work, just be

prepared to have tired arms

Drilling holes in your robot

for various reasons can be

done with several different

tools, depending on your

budget If you can spare the

cash, a drill press is a good

option; it will always stay

steady and can adjust to most

needs However, this can end up

being a bit too expensive, so a

simple power drill with the right

bits should do just fine If you

want good quality at a lower price,

“corded” will probably be your

choice Cordless drills — while nice —

can be pricey due to the battery

and the mobility it gives you

For any drills, however, you

will want to invest in a good set

of bits This can overwhelm many

beginners, but if you narrow down

what bit sizes you will use the most,

it can reduce your costs greatly

Also, make sure to invest in a punch

or two in order to mark your holes

properly A drill will easily wander

on you if you haven’t punched the

hole marking first

A tap set would be a good

investment if you want to be a bit

more advanced in your assembly

of pieces without too much of a

pocket drain This will allow you to

thread directly into the metal itself

For those of you with a larger

budget, a good addition to your

work space would be a mill This

tool will allow you to make many

modifications to both the shape and

depth of your robot For example,

if your bot weighs too much youcan mill out some slits from non-important pieces of metal or if youhave specific plates or patterns tomake, this tool will do the job Inthis range of tools, there are severallevels and scales to choose from, soyou’ll want to know what type ofmaterial and work you will be doingahead of time This will simplify your choices A simpler alternative

is hand-operated mill or you could

go the route of a computer drivenCNC mill which can do precise andsmooth work based on the designyou feed into it To find out about areasonably priced CNC machine forhobbyists, check out the Decemberissue of Nuts & Volts (www.nuts

volts.com), in the Personal Robotics

column

On the same end of the CNCspectrum is a lathe You can turndown anything from a wheel to ashaft with this machine A skilledoperator can add threads to a shaft

as well It is better to have a skilledoperator show you how to use alathe or else an accident can occur

Once you have the space andtools selected, it is important to

determine what the organization

of your set up will be If you are anorganized person by nature whoalways puts things back, then acomplex array of labeled drawerswill probably work fine for you

However, most builders I know lackthis quality However, I have found

a happy medium

First and foremost, keep tools

of the same skill set next to eachother, along with any support equipment such as bits or keys (Thiswill greatly simplify things when youhave lost something and are madlysearching through a messy pile.)After this, a simple set of bins orcubbies can serve to contain thingsseparately That way, you only need

to throw it back into the right pile

Tools are loosely placed on wall racks making them easier

to organize. Vises and jigs are all placed in one secure area.

Tools arranged on wall board keep everything nice and centralized.

Spare and useless bits kept in boxes off to the side.

Mounting boards arranged wall to wall, opening up the amount of readily available storage space.

Finally, do the same kind of looseorganization with your hardware.

Nuts, bolts, and metal scrap doesnot need to be organized by exact

size, but if you can at least keep thespecific type separate then you will

be ahead of the game

The last thing to do is to position the different work surfaces where you can get tothem easily Once you have done this, it is easier to work oncomponents where multiple toolsets are required After this, make sure to have an appropriateamount of mounted grips and jigs

in order to secure your variety ofcomponents as you work on them.You should be all set to begin SV

It was the fifth Robothon event for

Western Allied Robotics at the

heavily trafficked Seattle Center’s

Center House Large crowds

circulating throughout the day were

entertained by three and 12 pound

fighting robots The three lb

Beetleweight division was mostly

populated by returning builders but

there were a few new designs

Greg Schwartz of Team LNWbrought Idiot Savant, an undercutter

robot with a 16 inch titanium blade

clocked at over 3,300 RPM Doug

Brown of Team Dinobots came with

a heavily updated undercutter T-Rex

2.0 with a more powerful weaponand a more durable design DanOdell of Team Death by Monkeysbrought a redesigned GutterMonkey — a four wheel pushy-botwith a chromoyl steel wedge

T-Rex 2.0 quickly worked itsway up to the top of the winner’sbracket by knocking out the titanium brick Monkeying Around,the previous Robothon championHurty Gurty, and the full body spinner Melty B(eetle) On theloser’s side of the bracket, it wasGutter Monkey’s steel wedge vs

the titanium blade of Idiot Savant

In a series of big hits, both botswent flying with Idiot Savant being

Event Results

I3 lb BEETLEWEIGHT

● 1st — Gutter Monkey, Team Death

by Monkeys, driven by Dan O’Dell

● 2nd — T-Rex 2.0, Team DinoBots,driven by Doug Brown

● 3rd — Melty B(eetle),SpamButcher, driven by Rich Olson

12 lb HOBBYWEIGHT (12 lb)

● 1st — Organ Grinder 12, TeamDeath by Monkeys, driven by Rob Farrow

● 2nd — Shag, Team Gausswave,driven by Rob Purdy

● 3rd — Defiantly Daft, Team LNW,driven by Travis Schwartz

Circular saws placed in one spot,

keeping all extra pieces next to it for

loose organization.

Pictures From Dr.

Joseph Immel, home workstation Shows example of both tools, and a non-garage based shop.

A standard lathe chuck setup.

A standard mid-line table mill.

Door leading to garage from workspace.

● by Rob Farrow

EVENT REPORT

Robothon Rob t Combat

2008

unlucky and landing in

the arena pit Gutter

Monkey also managed to

defeat Melty B(eetle) by

keeping its steel wedge

facing the spinner and

pushing it around until

smoke started to come

out of it

In the Beetleweight

final, Gutter Monkey

faced off against T-Rex

2.0 Sparks flew as T-Rex’s powerful

blade struck the wedge With some

skilled driving, Dan O’Dell managed

to push T-Rex into the wall which

caused it to ricochet off and land

behind Gutter Monkey With its

wheels exposed, Gutter Monkey

spun around and got its wedge

pointed the right way again With

the help of some speed and luck,

Gutter Monkey managed to deposit

T-Rex into the arena pit and take

home 1st place

In the 12 pound Hobbyweight

class, the normally dominant

Fiasco’s weapon was as dangerous

as always ripping the titanium

wedge off of the four-wheel drive

Defiantly Daft As Kevin Barker

struggled to drive Fiasco which was

having drive motor problems, Travis

Schwartz did a masterful job driving

Defiantly Daft He positioned his bot

behind the wedge lying on the

arena floor and used it to push

Fiasco around even though the

wedge wasn’t actually attached

to the robot

Despite the damage to

Defiantly Daft, it won the judge’sdecision based on its control ofthe match Defiantly Daft’s luckdidn’t hold as it went up againstShag driven by Rob Purdy of TeamGausswave With the aid ofpowerful neodymium magnetsincreasing the robots down-forceand traction, Shag was able tocontrol the match and push thewedge around at will Shag thenstruggled against Organ Grinder 12

— a wide, wedge shaped bot with avertical tool steel disk spinning athigh RPM Rob Farrow of TeamDeath by Monkeys used the spinning weapon to flip Shag upsidedown, taking away Shag’s magneticassist on its drive train With the botupside down and its wedge hanging

in the air, Organ Grinder 12 took

the win resulting in a three-way tiefor first with Shag, Defiantly Daft,and Organ Grinder each having a3-1 record

In a first for a Western AlliedRobotics event, a three-way tiebreaker match determined thewinner The three bots met in thecenter of the arena, each jockeyingfor a good position Defiantly Daftslammed Shag into the arena wall

A three lb undercutter named Idiot Savant created by team LNW.

Driver-builder Dan O’Dell holding the

3 lb Beetleweight 1st place winner Gutter Monkey.

Driver Rob Farrow holding the 12 lb Hobbyweight 1st place winner Organ Grinder 12.

The youngest competitor Alex Conus holding up his three lb robot Vision #1.

The 12 lb robot Fiasco built by Kevin Barker

as it is spinning on its side with some sparks.

knocking it on its side and turning it

into a two bot fight As Organ

Grinder and Defiantly Daft tangled,

a hit by Organ Grinder threw the

wedge into the air and knocked

Shag back on its wheels, returning

it to the match As the three botscontinued to fight, the two wedgessustained minor damage to theirdrive systems making it difficult to

avoid Organ Grinder’sweapon

With 30 seconds left in the match, OrganGrinder knockedDefiantly Daft onto the edge of the pitimmobilizing it and continued to hit away

at Shag In the end, itwas a unanimous judge’sdecision for Organ Grinder

With the conclusion of the

12 lb final, it was another excitingcompetition with everyone lookingforward to the next event SV

The 12 lb robot

T-Bone built by

Dylan Feral-McWhirter.

The 12 lb pusher Shag built by Rob Purdy The plow is

brick-nearly 5” titanium (WOW! -Ed.)

ROBOT PR FILE

● by Kevin Berry

TOP RANKED ROBOT THIS MONTH

Weight Class Bot Win/Loss Weight Class Bot Win/Loss

150 grams VD 26/7 150 grams Micro Drive 10/3

1 pound Dark Pounder 44/5 1 pound Dark Pounder 23/3

1 kg Roadbug 27/10 1 kg Roadbug 11/4

3 pounds 3pd 48/21 3 pounds Limblifter 12/1

6 pounds G.I.R 17/2 6 pounds G.I.R 11/2

12 pounds Solaris 42/12 12 pounds Surgical Strike 19/7

15 pounds Humdinger 2 29/2 15 pounds Humdinger 2 29/2

30 pounds Helios 31/6 30 pounds Billy Bob 12/4

30 (sport) Bounty Hunter 9/1 30 (sport) Bounty Hunter 9/1

60 pounds Wedge of Doom 43/5 60 pounds K2 14/2

120 pounds Devil's Plunger 53/15 120 pounds Touro 14/2

220 pounds Sewer Snake 46/13 220 pounds Original Sin 12/5

340 pounds SHOVELHEAD 39/15 340 pounds Ziggy 6/0

390 pounds MidEvil 28/9 390 pounds MidEvil 3/0

Top Ranked Combat Bots

History Score Ranking Micro Drive – Currently Ranked #1

Historical Ranking: #2 Class: 150 gm Fleaweight Team: Team Misfit Builder(s): Zachary Lytle Location: Santa Rosa, California

BotRank Data Total Fights Wins Losses

Lifetime History 17 13 4 Current Record 13 10 3

Micro Drive has competed

at Marin Ant Wars 5,RoboGames 2006, SRJC Day Under

The Oaks 2006, Marin Ant Wars VI,

Halloween Robot Terror 2006,Smackdown in Sactown III,RoboGames 2007, SRJC Day UnderThe Oaks 2007, RoboGames 2008,

and SRJC Day Under The Oaks

2008 Details are:

● Overall configuration: Micro Drive

Robotic combat is a dangerous

sport in its very nature, and if

participating, one must accept the

safety risks that are all too often

present in the sport In most cases

however, simply accepting the risk

isn’t enough, and something must

be done to negate this risk as

much as possible One of the most

dangerous variables in robots is the

safety features of their radios

Most FM receivers can easily

interpret noise from motors, speed

controllers, and even microwaves as

valid signals With the amount of

destructive spinners and otherdangerous designs in the currentcompetition circuit, it should gowithout saying that even in thesmaller weight classes, this is aserious safety hazard

That’s where failsafes come in

Featured prominently in FM-PCMradio receivers, failsafes willautomatically have all channels go

to zero throttle or any otherposition that the user specifieswhen the transmitter’s signal is lost

However, this does not solveanother all too prominent problem:

frequency conflicts In large roboticcompetitions or conventions,another robot is bound to have the same frequency as yours

Fortunately, with the recent 2.4GHz radios, this has become a thing

of the past

Recent spread spectrum radiossuch as the Spektrum DX serieshave taken the R/C world by storm.Unlike conventional FM transmittersand receivers that connect throughcorresponding frequency clips,2.4 GHz transmitters bind to theirspecific receivers, meaning there will

● by Thomas Kenney

PARTS IS PARTS:

2.4 GHz Radi Fail-safes

is a low to the ground wedge

robot, with sloped sides and a

titanium lifting arm The robot

can operate upside down or right

side up (however, the robot has

better armor when it's right

side up), and it can use its arm

to self right

● Frame: Carbon fiber sheets

and rods sewn and glued

together

● Drive: The 56-to-1 Sanyo gearbox

from Solar Robotics

● Wheels: Modified Light Flight

foam wheels from Dave Brown

● Drive ESC: Barello Ant 100 dual

5A speed controller

● Drive batteries: 7.4V 145 mAh

Kokam li-poly

● Weapon type: Lifter.

● Weapon power: 7.4V 145 mAh

Kokam li-poly

● Weapon motor: Mini metal

gear high tech servo

● Weapon controller: Servo.

● Armor: 1/16 inch polycarbonate

sheet, with 020 inch titaniumreinforcement in critical areas

● Radio system: Seven channel

programmable JR radio

● Future plans: Reinforce the

mounts that connect the armor

to the frame and raise thebottom plates to give the robot

a higher ground clearance

● Design philosophy: Let children

and rookies drive my robotagainst saw robots duringdemonstrations or exhibitions and find out what breaks, thenfix it “If it doesn’t fit, force it

If it breaks, you needed a newone anyway.”

● Builder’s bragging opportunity:

Micro Drive has been my mostsuccessful and reliable robotfor the last three years Hehas won two RoboGames

championships and got the silvermedal this year He has alwaysplaced either first or second inevery competition he has entered.The first Micro Drive was built

in two days and I’ve spentprobably a grand total of a weekrebuilding or repairing it in itsthree years of service When I dodemonstrations at high schools,Micro Drive is the primary robot Iuse and I let the audience drive

it Micro Drive was actuallybuilt using nothing more than aheavy-duty pair of scissors, a hairdryer, needle and thread, glue,and soldering iron At next year’sRoboGames, Team HammerBrothers better watch out,because Micro Drive will be back

in full fighting form ready toreclaim his title as RoboGameschampion! SV

Photos and information are courtesy of Zachary Lytle All fight statistics are courtesy

of BotRank (www.botrank.com) as of

October 10, 2008 Event attendance data

is courtesy of BotRank and The Builder’s

Database (www.buildersdb.com) as of

October 10, 2008.

presentedRobothon Robot

Combat 2008 in

Seattle, WA on

September 21st

Twenty-two bots were registered

NERC presented FranklinInstitute Robot Weekend inPhiladelphia, PA on October 11th

Forty-twobots wereregistered

RobotsLive presented a show inHuddersfield, West Yorkshire onSeptember 20th and 21st SV

There are 80 channels in the range

of 2.4 GHz radios The transmitters

will search for an open channel, and

once this is found, connect to the

receiver it has been bound to via

this channel

The only problem with mostspread spectrum receivers is that —

like many R/C receivers — they

will only failsafe on the throttle

There are no failsafes on any of

the other channels, and they will

simply continue in the positions

that they were in when the signalwas lost

This is ideal for most R/Cplanes, boats, and cars where thethrottle is the only channel thatcontrols any large movement; thatbeing the motor or engine thatrotates the propeller on modelairplanes, boats, or the wheels onR/C cars The rest of the channelsmerely control servos that moveflaps, rudders, and steering whichhave no potential for any sort

of danger

In combat robots — in addition

to weapons often controlled by thethrottle channel — all sorts of

motors and other weaponsmay be controlled by theaileron, elevator, rudder, orany other channel

Spektrum RC addressedthis problem with therelease of their SpektrumBR6000 bot receiver, whichreplaced their AR6000 thatwould only failsafe on thethrottle channel At therequest of combat robotbuilders, Spektrum created

a 2.4 GHz radio receiverspecifically for combatrobots that would failsafe

on all channels

Today, the BR6000 isthe standard for spreadspectrum receivers Thoughrecently, there has been arelease of 2.4 GHz modulesmade by Spektrum forinstallation in oldertransmitters such as the Futaba9Cap With this, hobbyists are able

to take advantage of the newspread spectrum technology withouthaving to adapt to a new radiosystem

As technology grows, somerobotic competitions (such asRoboGames) are requiring theircompetitors to use spread spectrumradios to eliminate the complications

of standard radio frequencies Now that the robotic communityhas begun to embrace this newtechnology, it’s impossible toimagine what more advantages

it might bring in the future SV Spektrum DX6 and

BR6000 bot receiver.

email: sales@crustcrawler.com

This month, we’ll add yet another motor and motor

driver project to the pages of SERVO The motor in

the spotlight this time around is a direct descendent

of common universal brushed DC motors However, this

motor type contains no brushes as its brushes have been

replaced by Hall effect sensors and electronic circuitry A

brushed DC motor contains a set of brushes, a stator

assembly, and a rotor assembly Commutation — as it

relates to brushed DC motors — is the process of switching

current and thus, magnetic fields between the stator

assembly and rotor assembly of the motor The motor

brushes play a large part in the process of commutation as

the position of the brushes (like the stator) is constant with

respect to the rotor’s magnetic fields Thus, the brushes are

actually part of the motor’s stator assembly

The magnetic fields of the stator always react with the

commutating magnetic fields of the rotor in such a way

as to coerce rotary motion via the motor’s rotor shaft Our

new motor type has nostationary brushes thatprovide a reference pointfor commutation

Instead, the motor typewe’re about to discusscontains Hall effect sensors, which are positioned for use ascommutation referencepoints by the brushreplacement circuitry

This type of motor iscalled a brushless DCmotor or just BLDC

After we go to school onthis new motor, we’regoing to design a

that takes the place of the Hall effect sensors

I attempted to disassemble the BLDC motor you see

in Photo 1 with no joy I suspect that its components aresealed to preserve the integrity of the positioning of themotor’s integral Hall sensors I didn’t have a second motor

on hand So, I didn’t work too hard at pulling it apart Themodel BLY171S-24V-4000 BLDC motor shown in Photo 1 ismanufactured by Anaheim Automation This particularmotor has a permanent magnet rotor, a three-phase stator,and a built-in trio of Hall sensors Picking apart the modelnumber tells us that this motor is a NEMA size 17 typeBLDC motor The 1S denotes a single shaft motor with 11oz-in of continuous stall torque The BLY171S-24V-4000’smotor windings are rated for 24 volts and are able to spinthe rotor shaft at 4,000 RPM

I f you were to go back and survey past SERVO articles that I have written, you would probably conclude

that I have this thing about motors and motor drivers For instance, we recently tackled Universal Motors and constructed a Universal Motor controller I also presented more than one SERVO stepper motor controller project In these pages, we’ve driven linear actuators, rotated hobby servo rotors, and built circuitry to oversee the direction and speed of simple brushed DC motors.

Photo 1 This BLDC motor comes

ready to work with drive circuitry

that employs the use of its

built-in Hall effect sensors These

sensors are ignored when the

motor is spun using sensorless

LEARNING TO DRIVE:

BLY’s motor interface wires are terminations for the three

stator phases The remaining five wires service the BLDC

motor’s built-in Hall effect sensors Here’s how our BLY

Hall Ef fect Sensors:

- Hall Effect Supply: Red

- Hall Effect Sensor A: Blue

- Hall Effect Sensor B: Green

- Hall Effect Sensor C: White

- Hall Effect Ground: Black

If we want to deploy the BLDC motor is a

standard manner, the Hall effect sensors are used in

the commutation process In a sensorless BLDC motor

implementation, we’re only interested in the motor phase

wiring Sensorless motor control commutation is a product

of the BLDC motor’s back electromotive force (BEMF),

which is produced in the motor’s stator windings as a

result of the movement of the rotor’s permanent magnets

past the stator coils Before we formulate a method of

getting rid of the Hall effect sensors, it would be helpful to

understand how a BLDC motor works with them

BLDC Operation 101

Although we have identified only three major coils in

our BLDC motor, a typical three-phase BLDC motor contains

a multi-coiled stator and a permanent magnet rotor The

more stator coils one can cram into the stator assembly of

a BLDC motor, the smaller the rotational steps Smaller

rotational steps result in less torque ripple The same

principle applies to the rotor A BLDC motor’s rotor supports

an even number of permanent magnets The more magnetic

poles associated with the rotor, the smaller the rotational

steps You know the rest Regardless of how many stator

coils and rotor magnets a BLDC motor has, we can still

gain an understanding of how a BLDC motor works by

examining only three coils

Like a stepper motor, a BLDC motor commutates

according to a predetermined coil activation sequence

That’s about where the similarity ends Stepper motors have

higher step counts and require a higher operating voltage

BLDC motors are designed to operate with Hall effect

commutation and variable voltage drive Precise rotor

alignment is not something a BLDC is particularly good at

Conversely, a stepper wants a constant voltage drive and is

designed for precise angular positioning of its rotor

A three phase BLDC motor has six discrete states of

commutation When the correct coil activation sequence is

direction of the shaft’s rotation I have labeled the BLDCmotor coils A, B, and C in Figure 1 As you can see, eachcommutation state consists of one coil driven positive, onecoil grounded, and one coil open This type of commutation

is called block commutation

Let’s walk through a BLDC motor commutationsequence visually Figure 2 is a simplified view of a three-phase BLDC motor’s coils and its rotor The simulated rotor

is centered in the figure as an arrow The pointed end ofthe arrow is positive while the opposite end of the arrow

is negative This positive/negative arrow arrangement simulates the magnetic fields generated by the rotor’s permanent magnets The stator coils in Figure 2 are shown

as rectangles, which are spaced at 120° intervals Thearrow (rotor) will move from commutation state to adjacentcommutation state depending on the direction of the current applied to the energized stator coils

Reference the coil voltage drive levels in Figure 1 andapply them to Figure 2 Commutation state 1 is a product

of stator coil A being driven positively, stator coil B beingdriven negatively, and stator coil C floating Driving statorcoil B negatively actually means that stator coil B is grounded.Thus, the stator coil current is flowing between stator coils

A and B while stator coil C is electrically disconnected Thepositive magnetic pole of the rotor will be attracted to thenegatively-charged stator coil B while the negative magneticpole of the rotor will be drawn towards the positively-charged stator coil A If we specify the positive pole of therotor as our commutation state reference, the positive pole

of the BLDC motor rotor is now positioned in BLDC motorcommutation state 1

Let’s transpose the next set of coil drive voltages inFigure 1 to our imaginary BLDC motor in Figure 3 Moving

Motor Drivers

Figure 1 This reminds me of a stepper motor coil activationchart An application that uses the Hall-Effect sensors wouldtrigger a commutation at every sensor logic level change

floater and the stator coil C drive voltage level transitions

from floating to negative This stator coil drive voltage

pattern forces the positive field of the rotor to move

towards the negatively charged stator coil C The negative

field of the rotor is still under the influence of

positively-charged stator coil A

I think you get the idea However, to set the BLDC

commutation concept in stone, we’ll walk through one

more commutation using Figure 4 Again, moving 60° to

the right in the Figure 1 commutation chart, we assign

stator coil A as the floater Stator coil B is no longer the

floater and moves to a positive drive state The coil current

direction in stator coil C remains as it was in the last

commutation state illustrated in Figure 3 Stator coil A is

electrically null and is powerless to influence the motion of

the rotor The magnetic forces of stator coils B and C now

determine the position of theBLDC motor’s rotor,which is now positioned in what

we have defined as commutation state 5

I’ve pinned the Hall effect sensor states to the bottom

of Figure 1 Note that the sensors generate a unique binarycode for each commutation state The binary code is used

by a BLDC controller or microcontroller to determine thecommutation position of the rotor which, in turn, determines the present and next set of commutation drive voltage levels Now that we possess the knowledgenecessary to drive the stator coils of a BLDC motor, let’s look

at the hardware required to process those drive signals

BLDC Hardware 101

A typical BLDC motor’s coils are driven by athree-phase half-bridge circuit like the one you see

in Figure 5 Block commutation with a twist is used

to drive this multi-phase collection of MOSFETs Thetwist is PWM, which is applied only to the bottomMOSFETs Applying the PWM there leaves the possibility of only one of the top transistors to betotally ON at any time I’ve used the coil drive levels from the commutation chart in Figure 1 toformulate the hardware commutation chart you see in Figure 6 Just as in Figure 1, each hardwarecommutation state in Figure 6 is associated with

a binary Hall effect sensor value So, the BLDCmotor controller knows exactly where to pick upthe commutation sequence it needs to rotate theBLDC’s rotor

From what we’ve seen thus far, the BLDCmotor’s built-in Hall effect sensors are always available to tell the BLDC controller how to correctly

Motor Drivers

Figure 5 You’re used to seeing the Phase A and Phase B

circuitry as a pair of half H-bridges commonly used to drive

stepper and universal brushed DC motors Phase C is a

Figure 4 We’ve gone halfway aroundelectrically and I think you know wherethe next set of stator coil voltages will take us The commutation table

in Figure 1 represents electrical revolutions, not physical revolutions.The number of magnetic poles in therotor and stator determine the physicalnumber of revolutions versus the number of electrical revolutions

Figure 2 The Hall effect sensors

(which are not depicted here) are

mounted in such a way as to sense

the position of the rotor’s magnetic

fields in relation to the magnetic fields

emanated by the stator coils

Figure 3 We could reverse the BLDCmotor’s rotor to commutation position

1 by backtracking the drive voltagechart in Figure 1 by 60°

sensors and still successfully commutate a BLDC motor?

The answer is BEMF

BEMF 101

When the BLDC motor shaft is spinning, BEMF is

generated by the movement of the rotor permanent

magnets past the stator coils BEMF is just a fancy way of

saying opposing voltage The BEMF varies according to the

speed of the motor The faster the motor spins, the higher

the BEMF that is generated

Recall that one of the three BLDC motor phases is

always electrically disconnected during commutation

Consider Figure 7 Matching up the phase drive voltages

with the commutation chart in Figure 6 tells us that this

drive pattern is representing commutation sequence 4

The floating phase in commutation sequence 4 happens to

be phase C The floating phase provides a portal for the

measurement of the BEMF voltage for this commutation

period Another look at Figure 6 shows that all three of the

phases are electrically disconnected at one time or another

Thus, we are able to measure the BEMF during any of the

six commutation periods using the available floating phase

Since the BEMF is directly proportional to the motor

speed, we can use the sensed BEMF levels to control the

commutation of a BLDC motor The goal is to commutate the

BLDC at the required speed and torque while maintaining a

safe voltage and current level in the motor’s windings

How Are We Going To Do This?

Now that we have some BLDC drive theory under our

belts and we know what we need to accomplish in relation

to driving a BLDC, let’s try to determine what we need on

the hardware side It’s pretty obvious that we’ll need a way

to generate PWM signals The easiest way I know of to do

this is to push some values into a set of microcontroller

registers and assign a PWM output pin We’ll also need a

way to activate and deactivate the MOSFETs in the

three-phase bridge Not only do we need to do this, we’ll need to

energize the correct set of MOSFETs in accordance with the

commutation table laid out in Figure 6 Logical activation

and deactivation of MOSFET switches sounds like a job for

a microcontroller to me

To implement a system that monitors BEMF and applies

the captured BEMF to commutating,the BLDC motor will

require a microcontroller with A-to-D converter capability

We’ll need at least four A-to-D converter inputs to monitor

the three phases and the overall current drawn by the

BLDC motor

A means of controlling the starting, stopping and the

speed of the BLDC motor would also be a nice thing to

have So, let’s add an additional A-to-D converter input

channel and a couple of I/O pins to our

microcontroller-capability shopping list

components we’ll need to build up our trio of phase drivers The MOSFETs for each phase can be realized using an IRF7309 The IRF7309 is an eight-pin device thatconsists of a pair of MOSFETs that matches the half-bridgeconfiguration of each of the phases outlined in Figure 5.Microchip’s TC446X series of Logic-Input CMOS drivers willprovide sufficient gate drive for the MOSFET pairs

I came across the word “BLDC” many times while reading through the PIC18F2431 datasheet So, I’m leaningtowards using the Microchip PIC18F2431 as our BLDCmotor control microcontroller The PIC18F2431 can providesix PWM channels and five fast A-to-D converter inputs

Next Time

I’m really anxious to get started on the hardware andfirmware design for our BLDC motor controller So, I’ll make this short and sweet Next month, we’ll build up our own BLDC motor

controller hardwarefrom scratch andexplore what it takes

on the firmware side

to successfully commutate ourAnaheim AutomationBLDC motor SV

Fred Eady can be reached via email at fred@edtp.com.

Motor Drivers

Figure 6 I used the voltage levels in Figure 1 to fill in this chart The PWM signals were substituted for the

–V signals given in Figure 1

Figure 7 The coil configuration is called a “wye” because

of its Y shape BLDC motors can also have their coilsarranged in a delta configuration The BLDC drive theorywe’ve discussed thus far is valid for both the wye and

delta coil configurations

Motor Anaheim Automation www.anaheimautomation.com

SOURCES

This month, we wrap up our four part series by

covering practical use of PWM on the Propeller using

our multi-controller board In last month’s article,

Kevin McCullough covered stepper motor control and made

schematics available on the project web page These

schematics use the L293D Quad Half-H Driver IC which

provides bi-directional drive currents of up to 600 mA from

4.5V to 36V We will use the L293D in this article for both

motor control and light dimming This IC is included on the

Propeller™ Professional Development Board as well, so those

who have the PPDB can follow along even if they haven’t

built a multi-controller board

Control Methods

Controlling the speed of a motor or dimming a light

can be done by varying the voltage going to the device

More voltage and the motor goes faster or the light gets

brighter Less voltage and the motor slows down or the

BJT or MOSFET) operates in its linear (or active) regionwhen not fully on or off When this happens, the driver has

to supply necessary current while limiting the voltage onthe output The power dissipated is wasted in the form ofheat which could also be harmful to the driver or othernearby components, and often requires large heatsinks

By using the driver as a switch, it is either on or off at anygiven time This reduces the voltage drop across the device andtherefore the amount of power dissipated as heat In order

to control the speed of a motor or the brightness of a light,

we have to modulate the output to vary the power to theload PWM can accomplish this and reduce the total amount

of power to the load without the losses typical of resistive/analog means In the off state, the driver is not conducting anycurrent; in the on state, there is little voltage drop across it

PWM Control

Using PWM, we will essentially modulate our signal at a

by CHRIS SAVAGEGETTING CONTROL

WITH THE

Propeller PART 4:

Propeller Multi-Controller

speed of a motor or the brightness of a lamp much more

efficiently Our multi-controller board is Propeller-based so we

can take advantage of the counters to help generate our PWM

signal The Propeller has eight cogs (processors) and each cog

has two counters for a total of 16 The counters operate

completely independently of each other and are advanced

modules having 32 modes of operation For this article,

we’ll be using the PWM mode which will be explained next

PWM Generation Using the Counters

The CTR.spin object (included with the Propeller Tool) does

a good job of going into the details of the counters and all

their modes and parameters, so I will focus on a small piece

of example code that will provide a working example of using

a counter to generate a PWM signal for the L293D First, let’s

have a look at the code line by line Please refer to Figure 2

for the code listing This code can also be downloaded from

the project web page listed in the Resources sidebar

Line 2 sets the clock mode and sets the PLL multiplier

to 16x Line 3 sets the frequency of the crystal we are using,

which is a 5 MHz crystal With the PLL at 16x, the Propeller

is now running at 80 MHz CLKFREQ is a command which

returns the current system clock frequency in Hertz This

means your programs can determine what speed the

Propeller is running at, as well as make calculations based

on this value Line 6 declares a global variable called

parameter, which we will use to change the duty cycle of

the signal by setting it to some percentage of the period

Line 8 starts our program code and defines the name

of this method as Main It also declares a local variable

named index which will be used as a counter Line 9 launches

the core PWM code into a new cog and passes the code

address (entry), as well as the address of the global variable

where it will obtain the duty cycle The @ symbol specifies

that we’re passing the address of a

symbol Line 10 starts a repeat loop that

never ends The lines indented below it

will run indefinitely or until the Propeller

is reset Line 11 starts a repeat loop

which uses the local variable index to

count from 0 to the current period Each

pass through the loop index is copied

into parameter and there is a small delay

before the loop continues When index

equals period, the routine will fall down

into the next repeat loop, which is

essentially the same as the first, except

that we’re counting back down The

effect is that we’re ramping the duty

cycle from 0% to 100% and back to 0%

in a continuous loop To ramp faster, you

can reduce the waitcnt value from

80_000 down to a lower value Likewise,

a higher value will cause it to ramp slower

Line 20 is the entry point to the

assembly code that handles our PWM

pin we will be using to make it an output This is accomplished by moving the value of diraval into DIRA Ifyou look down at line 30, this value was set using thedecode operator This effectively sets bit 16 high in the 32bit variable (all other bits are low) Now our PWM pin is set

to output so we drop down to line 21 which moves thevalue of ctraval into CTRA This value was calculated down

on line 31 by taking the bit pattern we want for the PWMmode (%00100), shifting it left 26 bits to get it into the correct field of the counter A control register, and adding

16 to specify which I/O pin to use for output This worksbecause the APIN is specified by the lower six bits of thecontrol register Line 22 moves a 1 into the FRQA registerwhich tells the counter to increment every clock cycle Line

23 moves the current contents of cnt into a variable called

25%

Period Duty Cycle

50%

75%

FIGURE 1 PWM example waveform.

FIGURE 2 Program listing.

time The cnt value is a 32-bit counter incremented every

clock cycle In line 24, we add our period to time This

calculates the value of cnt one period from the time cnt

was read To summarize, lines 20 through 24 set up all the

parameters we will need in our assembly routine before

running the loop This includes setting up the direction

register for the output pin, setting up the configuration

register for counter A (which sets the mode and output

pin), and calculating the target cnt value for timing using

the system counter

At this point, we enter the main assembly loop at line

25 This line reads the current value from the address

passed in the par register which — as you may recall —

points to the global variable parameter where our main

code can alter the duty cycle Line 26 causes the cog to

wait for the target cnt value which was stored in time

Once reached, period is added to time again to recalculate

the new target and the code proceeds to the next line Line

27 takes the variable value and makes it negative before

loading it into PHSA

The reason for this is that in order for the output to be

high, bit 31 must be set By making value a negative number,

we set bit 31 and our output goes high Because value is set

to our duty cycle and the counter will increment every clock

cycle, then once the counter reaches 0 the output will go low

Line 28 loops back to line 25 where we fetch a new duty cycle

value and wait So, as you can see the output stays high for

the duty cycle, which is a percentage of the period and then

goes low The cycle restarts at exactly the number of cycles

in the period This happens indefinitely or until the cog, is

reset Since this routine runs in its own cog, the Main method

can change the duty cycle at any time and the PWM assembly

code will update at the end of the current period

Connecting the L293D

In this example, I will be driving a 7.2V DC gear motor

from a robot This motor draws approximately 190 mA

@ no load, which is well within the specifications of the

L293D If you look at line 32 of the code, you will see that

we have specified a period of 8,000 cycles This gives us a

PWM frequency of 10 kHz and a period of 125 μs which is

cannot use P0 through P15 on the multi-controller because

of the 3.9K series resistors Instead, we will use P16 whichhas no series resistor If you’re using the PPDB, you can simplyconnect P16 to Input 0 and tie the Enable (0, 1) connection

to the 5V just below it to keep it enabled Optionally, the enableline could be controlled by an I/O pin but we’re keeping itsimple for now If you are using the multi-controller board,you can breadboard this circuit and connect the ground andI/O lines using pluggable wires Please follow the schematicshown in Figure 3 PPDB users will still require an externalmotor supply and ground connection as shown in theschematic When you run the code, it should start rampingyour motor up and down between stopped and full speed

Dimming Lamps

Dimming lamps is done in the same manner as themotors, with the exception of not requiring such a highPWM frequency For light dimming, you can connect a bulb

in place of the motor and make the following changes tothe code Change the period in line 32 to 40_000 instead

of 8_000 This will reduce the frequency down to 2 kHzand a period of 500 μs You can also change the 80_000cycle delay in lines 13 and 16 to 4_000 to adjust the ramping rate A MOSFET is another option to drive a lampand there will be additional circuits and code available onthe project web page for this series

Other Options

So, now you have a good foundation to start generating PWM on the Propeller and add functionality tothe multi-controller board we’ve been working with Whilethis article covered the basics, more advanced code with bi-directional motor control will be available on the projectwebsite We may even go so far as to integrate all threetypes of control into one demo

For now, experiment, build, and learn Don’t be afraid

to try new things The chips are socketed and easily replaceable

if you should make a mistake that results in smoke Rememberyou can always visit the Parallax Object Exchange and theParallax Support/Discussion Forums if you need additionalcode or have any questions Feel free to email me at csavage@parallax.com if you have any questions, comments, or suggestions Take care and have fun! SV

1 2 3 4 5 7 8 6

What today’s robotics

engineers, rocket scientists,

and astrophysicists do for fun.

Exercise your skill, expand

your mind, and challenge your

creativity with everything from

games to gadgets from Edmund

Scientific Whether you’re a

hobbyist, scientist, or student,

Edmund Scientific has something

to make you smile and think.

For the latest trends in toys

and technology, call for your

free catalog or shop online at

scientificsonline.com

THE BASICS

Agait is defined as an ordered sequence of states that

typically result in locomotion To measure the gait of

a person, we measure the angles of the various parts

of the body with respect to a reference at each instance of time

during the motion The simplest reference that is available

everywhere and has the strongest effect on gait is gravity

So, we measure the gait by strapping accelerometers to

the hands, legs, hips, neck, and so on, and calculating the

angle made by the gravity vector with the axes of the

accelerometers Prior to data acquisition, we obtain a few

seconds of reference data in the rest position to measure

the angles made by the sensors with respect to the gravityvector Subsequent angles are all calculated with respect tothis reference Of course, when the limbs are moving fast,the accelerometers measure the centrifugal and radial accelerations in addition to the gravity vector So, it gets abit complicated Research systems typically use gyroscopes

to measure the angular velocity and compensate for suchrotations, but gyros are expensive To keep our costs low,

we built a system using only accelerometers It turns outthat even such a simplified system provides remarkably useful information (for example, even without the gyros,

we can measure the maximum forces on our joints whilerunning) and is routinely used in cutting-edge research

projects!

If you are looking for

an innovative science projectthat contributes to gaitresearch, this is your startingpoint The system presentedhere can also be used innumerous other applicationslike measuring the

accelerations of your bicycles, motion tracking

of slow and controlledmovements, or even in falldetection

HARDWARE DESIGN

The basic design ofthis gait detection systemconsists of one (or more)slave nodes daisy-chained

to a single master node

2

Ever wanted to monitor your rhythm and stride pattern while you run? How about the G-force on

your knees on the treadmill? Want to do simple motion tracking (inexpensively)?

This article describes a simple but powerful design for a gait detection system (of course, along with all the software you would need!) using the highly configurable Cypress CY29466 microcontroller for control and ADX330 accelerometers for the sensors.

ByBalakumar Balasubramaniam and Wendi Dreesen

Multi-Purpose, Daisy-Chained

Gait Detection System

FIGURE 1: CIRCUIT DIAGRAM FOR THE SLAVE.

slave nodes are strapped onto the anatomy, and the

accel-erations imposed upon the slave nodes as they piggyback

during body movements are measured to calculate the gait

The key sensor component residing in every slave node

of our current gait detection system is the humble triple

axis MEMS-based accelerometer ADX330 In lieu of the

stand-alone accelerometer chip, we chose to use the

DE-ACCM3D modules from Dimension Engineering

(www.dimensionengineering.com) as they carry

on-board regulators, power filters, reverse voltage

protection, and provide buffered outputs

The three analog outputs (one for every axis) from the

DE-ACCM3D module in each slave is then fed into

program-mable gain amplifier modules within a Cypress C8YC29466

PSoC mixed signal array microcontroller The amplified

signals are then digitized using an incremental ADC converter

(TRIADC8) capable of fast conversions Note that the

TRIADC8 was chosen for this project solely for superior

speed However, the PSoC provides a large variety of

incremental and delta-sigma converters that can be

FIGURE 2: CIRCUIT DIAGRAM FOR THE MASTER.

swapped for the TRIADC8 depending upon the application.

Once the analog data from the accelerometers is digitized,

it is sent to the master for accumulation

The operation of the slave PSoCs are controlled by a

single master PSoC on a PSoCEVAL1 evaluation board (see

Figure 4) We recommend the use of evaluation boards

instead of stand-alone breadboarded master units to ease

debugging and to facilitate further hacks on the present

design The data received by the master from the slaves is

finally transferred to a computer using the UART output

from the microcontroller (and an FTDI 232R chip) or by

connecting a serial-to-USB converter to the evaluation

board’s RS-232 port The data is recorded and saved to afile using Windows HyperTerminal

One of the difficulties we faced during our initial versions was the unreliability of data transfer while usingthe I2C connection between the slaves and the master Wetraced this problem to the long connection lengths betweenthe boards required to interconnect the slaves on a movinghuman to the stationary master In our case, the furthestdistance of the slave from the master was about five meters.With a rated capacitance of 460 pF over this length, the designexceeded the 400 pF maximum allowed line capacitance for

an I2C connection The trick to mitigate such length issues

in I2C connections is to use bus extenders We selected the Phillips P82B715 extender along with suitable pull-upresistors on the serial data (SDA) and serial clock (SCL) lines

It is estimated that the use of the extender chips allows us

to increase the maximum cable distances to about 30 m.While the design of each slave node is simple, the system obtains its enormous power by synchronously operating several slave nodes in parallel, controlled by amaster node equipped with the right software

MECHANICAL ASSEMBLY

After we got the hardware all set, we faced the formidable challenge of mounting the boards firmly on theappendages of the subject! After much trial and error, webuilt a reliable and low cost mount for the slave board byplacing it in a plastic container (soap container from Wal-Mart, $0.50) and screwing the board to the bottom of

it Initially, we experimented by sticking a piece of wood

in between the board and the container using epoxy to provide rigidity However, the epoxy was not enough tokeep the board firmly positioned under vibratory motions.Hence, four screws were used at the corners of the board,each with three bolts (two on either side of the board tohold it, and one to fasten the container’s surface) to make

a firm mount One could also use foam to fill the spacebetween the board and the container to damp out spuriousresonances between the board and the soap case Oncethe soap container was ready, we used an elastic neoprene

to fix the container in place

SOFTWARE DESIGN

The code associated with this project contains twocomponents: the hardware configuration component forthe selection and correct configuration of the various modules in the CY8C29466 chip; and the second component that takes care of the inter-chip communications, sends synchronization signals to controldata acquisition and transmissions

We configured the master PSoC with one each of

FIGURE 4: PHOTOGRAPH OF THE COMPLETED SLAVE MASTER.

FIGURE 5: SAMPLE SIGNALS FROM A HIP-MOUNTED

ACCELEROMETER SHOWING THE VOLTAGE LEVELS FOR

VARIOUS EVERY-DAY ACTIVITIES.