Making Simple Robots (2014)
Chapter 2. Robots That Get Around
How often I found where I should be going only by setting out for somewhere else.
— R. Buckminster Fuller
One of the most interesting challenges in robot design is figuring out how to get from place to place. Traditionally, robot locomotion has fallen into two categories: bipeds that walk in something approximating a human gait, and rolling robots that drive around like vehicles. But in recent years the options for robot locomotion have taken off. Today it’s not unusual to see robots that can fly, swim, climb, creep, or crawl—and some that can move in ways that are hard to categorize. The advantage of these alternative modes of transportation is that they allow robots to access places humans can’t easily go, from collapsed buildings to other planets.
From a design point of view, robots that move in unexpected ways are also fun to play with and fascinating to watch. Just look at the popularity of tiny drone copters, which can be remote controlled or preprogrammed with a flight pattern, or which carry onboard sensors that let them avoid obstacles on their own. The ability to shrink down electronic flight components to the size of a deck of cards has created a boom in the DIY arena as well. Hobbyists can assemble their own onboard stabilizer, compass, gyroscope, accelerometer (a sensor that measures changes in speed and direction), and GPS, for just a few hundred dollars and attach them to the body of their choice. Brooklyn Aerodrome sells a kit on the Maker Shed website that fits the flight deck on a corrugated plastic board that can quickly be switched to a new construction foam body after heavy use. One standout exhibit at the Emma Willard Mini Maker Faire in Troy, New York in 2013 was the flying pizza box put together by 10-year-old Emma Edgar, with help from her dad,Marc. The flattened large-size pizza box can easily maintain heights of 100 feet or more for more than 30 minutes and needs minimal human input to stay aloft.
Even rolling robots can move in unconventional ways. The Segway personal transporter is actually a robotic platform that can balance itself on a single axle. Designed by inventor (and FIRST Robotics Competition founder) Dean Kamen, the Segway was designed to carry a human being, but it is also used as a base for other kinds of mobile robots. Or consider the Sphero Robotic Ball (Figure 2-1). Controlled by a smartphone, the Sphero can self-propel across a room by rolling, and play games with both humans and pets. Its waterproof casing contains a gyroscope, accelerometer, and LED lights.
Figure 2-1. Sphero is a robotic ball that can be programmed to jump or roll with a smartphone. Credit: Orbotix.
Researchers looking for new ways to make robots move often turn to the animal kingdom. MIT’s RoboTuna and other robotic fish that swim like their real-life counterparts have been around for a long time. Harvard’s RoboBee is smaller than a paperclip and can flap its wings, hover, and turn on command. Still in the development phase—they’ve yet to come up with a battery small enough for the RoboBee to carry—the microdrone might someday help with environmental monitoring, or crop pollination. Boston Dynamics, which became a part of Google in 2013, is famous for Big Dog, its aptly named experimental rough terrain quadruped. Resembling a headless Doberman on steroids, Big Dog can trot along through rubble, mud, snow, and shallow water carrying more than 300 pounds on its back. The company’s fastest robot to date is the Cheetah, which can reach 30 miles an hour on a treadmill. But Boston Dynamics is also working on RiSE, a compact, six-legged robot that can scale brick walls and rough treetrunks at a much more leisurely pace. Inspired by the gecko lizard, RiSE uses microspines on its toepads to latch onto surfaces, balancing with the help of its flat triangular tail.
A lot of the cutting-edge research into robot locomotion has been supported by the U.S. military through DARPA, the Defense Advanced Research Projects Agency. Driverless cars, robotic drone aircraft, and remote-controlled miniature tanks used in battle zones and rescue situations are just part of DARPA’s legacy. A series of design competitions known as the DARPA Robotics Challenge has also produced some of the most mind-blowingly strange modes of transport in the robot world—among them undulating robo-worms and amoeba-like machines able to ooze and creep along thanks to flexible materials.
The projects in this chapter will give you the chance to build two different systems of robot locomotion: the tensegrity robot, which uses tension, compression, and vibration as modes of transportation; and the wheel-leg, a hybrid that combines some of the best aspects of both types of traditional robot motion.
Segway Personal Transporter
DARPA Robotics Challenge
Project: Make a Compressible Tensegrity Robot
Figure 2-2. Finished tensegrity robot
A tensegrity robot is made up of rigid struts suspended by a web of cords under tension, which let it move and respond to changes in its surroundings.
What It Does
A tensegrity structure can flex, stretch, compress when dropped or pressed, and then spring back into shape. It also has a high degree of compliance, a robotics term that means it won’t harm people or equipment around it. That, together with its resilience, makes the tensegrity a useful framework for robots that need to withstand jolts or squeeze and twist themselves through irregular spaces. But it also offers a range of unique ways to mobilize a robot, including compression and vibration of its flexible cords.
Where It Came From
The word “tensegrity”—a portmanteau combining “tension” and “integrity”—was coined by renegade architect Buckminster Fuller. Many of Fuller’s inventions, including the geodesic dome, play around with tension and compression in their structural design. Although Fuller is the person most closely associated with the concept, most scholars agree that he was inspired by artist Kenneth Snelson, whose sculptures incorporate what he prefers to call “floating compression.” Snelson’s massive sculptures do seem to defy gravity as they spread and soar skyward. His most famous work is “Needle Tower,” a 60-foot-high tapering sculpture of aluminum tubes and steel cables on display at the Hirshhorn Museum and Sculpture Garden of the Smithsonian Institution in Washington, DC.
Roboticists are looking at ways to employ tensegrity in robot design. At the NASA Ames Research Center in Mountain View, California, scientists Adrian Agogino and Vytas SunSpiral were playing with a squishable stick-and-elastic baby toy when they got the idea to use the tensegrity’s forgiving structure (Figure 2-3) in the design for a mission to Saturn’s moon Titan. Unlike the 2005 Cassini spacecraft’s Huygens lander, which scientists believe sank below Titan’s crust upon landing, their Super Ball Bot lander could bounce safely and tumble along Titan’s surface. And the Super Ball Bot could be dropped from the spacecraft without the need for an elaborate skyhook mechanism like that used by the latest Mars Curiosity rover. They are now studying how to protect the instruments a tensegrity lander would need to carry.
Other possible uses of tensegrities include TetraSpine, a multisegmented tensegrity robot being developed at Case Western Reserve University in Cleveland, which can crawl over uneven surfaces. At Union College in Schenectady, computer scientist John Rieffel and his students are working on using vibration frequencies to steer ball-shaped tensegrities (Figure 2-4). And Erkin Bayirli, an architecture student in Vienna, has been building elegant tensegrity robots out of bamboo that can walk or crawl by pivoting on their corners, without legs.
Figure 2-3. A prototype of NASA’s Super Ball Bot built by Ghent University’s Ken Caluwaerts.
Figure 2-4. A tensegrity robot developed at Union College which can be steered using vibration. Credit: Steve Stangle.
How It Works
A tensegrity structure maintains its shape by balancing the push and pull of its various components. In a standard tensegrity, cords do not touch cords, and struts do not touch struts. Each cord is stretched between two rigid struts. Each strut has cords at either end pressing inward upon it. The tensegrity transfers these forces throughout the entire structure, giving it exceptional strength and resilience. For example, a bicycle wheel is a kind of tensegrity. The thin metal rim is kept perfectly round, or “true,” because it is being pulled evenly by all the spokes that connect the rim to the hub in the center. That gives it as much strength as a wagon wheel, without the need for heavy wooden spokes. But the tension and compression in a tensegrity must be in balance or it will pop and twist like a bicycle wheel with a broken spoke.
The complex interplay of forces in tensegrities make them intriguing to mathematicians and physicists. And they may also be useful in explaining how bones and connective tissue work together in living things. “Biotensegrity” is a term used by orthopedic surgeonStephen Levin to describe a model of the skeleton that he believes is more accurate than the standard way of looking at it. In the biotensegrity model of the human skeleton, the struts are the bones, and the cords are the web of tendons and ligament that hold them together. As Levin points out, the bones in a body are not connected to each other directly by hinges, like a door swinging on a doorframe. Instead, the bones float in a network of connective tissue called the fascia. They slip past each other as they move, like the parts of a tensegrity sculpture. The fascia help to distribute the force applied on the bones by the muscles throughout the entire system, resulting in a body that is much more resilient to stress than it would be without it. With this in mind, roboticists are now trying to use the biotensegrity model to make robot bodies more springy and lifelike.
One interesting quality of a tensegrity is the way it can be pushed flat and then snap back into shape. So a robotic tensegrity can move by bouncing, like the Super Ball Bot. Or it can be compressed and then expand in waves. Robots like TetraSpine are made up of a column of self-contained segments connected to each other by cords. Each segment can compress in relation to its neighbors, like discs in a human spine. To move, TetraSpine uses DC motors and an Arduino microcontroller board to tighten and loosen the elastic connectors, causing waves that propel the robot along. The waves are created by central pattern generators, similar to the rhythmic signals that help living vertebrates crawl, walk, swim, and fly automatically, without input from the brain. A virtual simulation of TetraSpine has been able to use its rhythmic motion to traverse several kinds of irregular terrain.
Because the cords in a tensegrity are stretched tight, they can also vibrate like the string on a guitar. The Union College tensegrity robots move by setting up vibrations along the elastic cords. This is done by attaching one or more vibrating motors to one of the struts, making the strings resonate. Different speeds create different wave patterns. These patterns can be used to make the tensegrity jiggle and shake in one direction or another. Yet another way of moving is that used by Erkin Bayirli’s bamboo tensegrity robot,Shi. On the top of the robot is a servo motor that oscillates back and forth. As it moves, it twists a strut at the rear from side to side. Like the stick controlling a marionette’s legs, the moving strut pulls first one corner and then the other slowly forward.
Perhaps the most exciting aspect of research into tensegrity robots is Rieffel’s idea that mechanisms can act as minds. He is exploring whether control normally associated with a robot’s “brain” can instead be outsourced directly into body dynamics. In other words, the structure of the robot would have its own intelligence. When stress from the outside is applied, the robot structure reacts in a way preprogrammed by its physical design. As Rieffel explains, this mechanical programming can free up a robot’s computational resources for higher-level tasks, like tracking objects or detecting survivors trapped in rubble. The idea of programming through physical structure takes the importance of robot body design to a new level.
Making the Project
To understand how tensegrities fit together, you have to build one yourself. But as Rieffel notes, assembling a tensegrity can be challenging because of what is called “pre-stress stability”—when you move one part, everything else moves, too. So to start off, you’ll make a version that’s flimsier but easier to handle. The directions for assembling this six-strut tensegrity out of drinking straws and rubber bands are based on a DIY tensegrity holiday ornament project from Bre Pettis that appeared on the Make: website in 2007. It cleverly eliminates the need for an extra pair of hands by using small rubber bands to hold the straws in place until you have assembled the entire tensegrity. Then the support rubber bands are cut away and the structure pops open into its final form.
You will be building this project in two steps: the tensegrity body, and the electronic circuit. With such a lightweight structure as your base, you’ll be able to take advantage of another timesaver—the littleBits line of electronic building parts. Once you’ve built your tensegrity, you can quickly put together a circuit to make your robot move. The circuit will consist of a tiny vibrating motor, a dimmer switch to make it run faster or slower, and a bar graph indicator that shows roughly how much power you’re supplying to the motor. Varying the speed and placement of the motor will produce different kinds of motion, giving it a kind of physical intelligence.
Don’t forget to document your work!
§ Time Needed: 1 hour
§ Cost: $75–$100 (for reusable littleBits modules, purchased separately or in a kit)
§ Difficulty: Easy to moderate
§ Safety Issues: None
What You Need to Know
§ Skills You Already Have: Cutting, following directions
§ Skills You Will Learn: Assembling a circuit
Gather Your Materials
§ Six or more drinking straws (it’s good to have extras)
§ Six rubber bands, roughly 5 inches long
§ Six additional rubber bands (preferably shorter)
§ Masking tape, glue dots, double-sided mounting tape or other removable adhesive
§ littleBits modules
§ 9V power supply
§ Dimmer switch
§ Bar graph (LEDs)
§ Vibrating motor
Step 1: List Your Requirements
The goal of the tensegrity robot project is to build a self-propelled object based on the push-and-pull design of a tensegrity structure. Its movement should be controllable by altering the arrangement of its parts and the speed of its motor.
Step 2: Plan Your Project
Following the drinking straws-and-rubber band design outlined here is the most reliable way to get started successfully building tensegrities. The area most open to improvisation at this stage is the design and placement of the littleBits circuit. In particular, think about how to attach the motor so that it interferes with the motion of the tensegrity structure as little as possible. You can list or sketch out different possibilities to try once you have everything assembled.
Step 3: Stop, Review, and Get Feedback
Some helpful tips for building your first tensegrity are to take your time and refer back to the pictures often. If you’re not sure what you’re doing, just go with the flow. It may not be clear how it works until you reach the last step.
Keep some spare straws on hand while you’re working. If a straw bends, you’re better off replacing it than trying to fix it.
Step 4: Build Your Prototype
1. Cut six pieces of straw to no more than about 5 inches.
2. On each straw, cut a slit on either end, making sure that the slits are aligned (i.e., both vertical). The slits should be around a quarter of an inch deep—enough to hold the rubber band in place, but not so much that the straw begins to weaken and bend.
3. Line up two straws and wrap a small rubber band loosely around each end of the pair. Do the same to a second pair of straws and slide them perpendicularly between the first two straws to form an “X.”
4. Take the last two straws and wrap a small rubber band around one end. Slide them through the intersection of the other straws so that they are perpendicular to the first two pairs. When they are in place, wrap a small rubber band around the other end.
5. Add the long rubber bands. Line up the slits in one pair of straws so that they are horizontal and hold them so that the ends are facing you, one above the other. Fit a long rubber band into the slit of the upper straw end facing you. Hold the remainder of the rubber band up and stretch it over the ends of the pair sticking up, fitting it into the slits at the top of the straws. Then stretch the rest of the rubber down and latch it onto the horizontal slit in the other end of the first straw. The result should look like a suspension bridge.
6. Do the same with the all the remaining straws. Adjust the rubber bands so they are even.
7. Cut away the small rubber bands so that the tensegrity springs open. Make sure none of the straws are touching. If needed, adjust pairs of straws again so they are parallel.
Now it’s time to assemble the littleBits electronic circuit that will make your tensegrity go.
1. Plug the power piece into the 9V battery.
2. Next, attach the dimmer switch. Technically, this component is called a potentiometer. It lets you turn the voltage from the battery up or down.
3. Connect the bar graph to the dimmer switch. This is a piece with five rows of miniature LED light bulbs. The more power going through the bar graph, the more LEDs will light up. It doesn’t tell you the number of volts going through, but you can record different voltage readings in terms of the number of LEDs that are lit up.
4. Attach one or more wires. The wire modules are short, so use two or three if you have them. These give your tensegrity robot some space to move without having to drag around the rest of the components.
5. Finally, add the vibrating motor. This is a small disc, about the size of a vitamin pill, with two thin wires attaching it to a base that snaps onto the wire. Be careful not to break the wires or tear them out of the component.
BUILDING ROBOTS WITH LITTLEBITS
The littleBits electronic building modules are one of the hottest Maker tools around. Invented by MIT Media Lab alum Ayah Bdeir, littleBits are an ever-growing library of postage-stamp-sized components that snap together magnetically. All the circuitry needed to make the components work together is already built in, and they are all open source. (Open source means a design is made public for anyone to use, improve, and distribute to others.) The modules come in sets, but you can also buy them individually. Originally started as an education initiative to help children and other beginners design simple circuits and electronic projects, littleBits are being used by artists, designers, and inventors to aid in rapid prototyping.
What makes littleBits different from other electronic toys for kids is their stripped-down appearance. That has good points and bad points. On the plus side, they’re small and light, so they fit into tight places. All the components and connections are exposed like a circuit board, so you can see how they’re put together. Tiny labels help you figure out what each component does, and some modules can be adjusted with slider switches or a tiny screwdriver (included). But unlike Lego or other construction toys, littleBits are not themselves building blocks. In order to make something with them, you have to attach them to something made out of materials you provide. And because there’s no casing to protect the circuit boards, they’re a little less rugged than traditional electronic building sets.
That said, getting started with littleBits is easy. Each piece attaches magnetically to the next—just click them together and you’re done. If the magnets try to repel each other instead of attracting, you know you’ve got a piece the wrong way round. Creating a circuit is also a simple process, because of the color coding. Blue pieces are for power—9V or coin battery or USB plug. Pink is for input, such as switches, sliders, buttons, light sensors, touch sensors, etc. Orange pieces are connectors—wires and logic gates. And the green pieces are the output devices, including LEDs, motors, and speakers of various kinds.
The littleBits modules and kits are on the expensive side compared to many kids’ building sets and to raw components you need to wire up yourself. (The basic starter kit costs about $100, with individual modules ranging from around $8 to $40.) And once you start figuring out what you can do with the pieces, you’ll quickly find yourself in need of extra modules, since the sets don’t contain multiples of the same part. It pays to design your project carefully to make the most of the littleBits you’ve got on hand. When attaching modules to a project, think about how to make it easier to reuse the parts later. Here are some ways to attach littleBits parts to your project so they can be removed later with minimal harm to the piece or your model:
§ Masking tape—take a piece about 3-4 inches long and roll into a loop, sticky side out, fastening it to itself
§ Adhesive dots (such as those sold under the brand name Glue Dots)
§ Modeling dough (homemade or Play-Doh)
§ Velcro straps
§ Clear mounting tape, such as 3M VHB or Scotch Permanent Clear Mounting Tape
Step 5: Test Your Design
To try out your tensegrity robot, attach your electronic circuit to your straw model. You’ll want to situate the vibrating motor so that none of the littleBits modules get in the way of the tensegrity structure’s motion around the table.
1. Decide where you’d like to attach the disc end of your motor. Use tape or another adhesive (see Building Robots with littleBits) to hold it onto one of the struts. Play out the motor wire along the strut. Attach the motor base and the wire base (which you have connected magnetically) to the same strut.
2. Thread the orange wire modules through the tensegrity to keep the remaining components out of the way. You might want to try holding them above the structure so they don’t touch any of the other struts or the table.
3. Turn on the vibrating motor. Slowly increase the power to it using the dimmer switch. You should start to see the rubber bands vibrate in sympathy. Notice how different speeds produce different movements. At a certain point your tensegrity should start to shimmy along the table. See if you can steer it right and left by turning the power up and down.
4. Experiment with placing the motor in different locations on the tensegrity—in the center, off on one corner—to see which position produces the most reliable and interesting movements.
Step 6: Troubleshoot and Refine
If your tensegrity doesn’t move, make sure that it isn’t caught up on any of the components. You can also try putting the weight of the components higher up or lower down on the tensegrity. You may need to move its center of gravity a little off-center to overcome its inertia.
Step 7: Adaptations and Extensions
You can build a more permanent device using materials that are sturdier but take more effort to prepare, such as wooden dowels. Other suggestions:
§ Elastic cord
§ Fishing line
§ Clear plastic beading thread
§ Wooden dowels
§ Bamboo garden stakes
§ PVC pipes
§ Eye screws
§ Plastic end caps
To propel it, you may have to use a more robust motor, such as the littleBits DC motor. Convert it into a vibrating motor by taping a small weight (such as a bead or metal nut) onto the motor’s shaft. You can also add additional littleBits parts—including the remote control modules, the number module with a digital readout, or even the programmable Arduino module—to make either the prototype or more permanent tensegrities more interactive. For a real challenge, do some research on tensegrities and tensegrity robots to find more advanced models to try to animate.
Vytas SunSpiral of the NASA Ames Research Center
Dr. Stephen Levin’s biotensegrity website
John Rieffel at Union College
Erkin Bayirli’s Bamboo Tensegrity Robots
littleBits electronic building modules
Project: Design a Wheel-Leg Hybrid
A wheel-leg hybrid is just what it sounds like—a blending of the basic wheel/axle design with one or more legs.
Figure 2-5. The finished wheel-leg robot
What It Does
The wheel-leg makes it possible for robots to benefit from the efficiency of wheels on smooth surfaces while using its legs to keep going over rougher terrain.
Where It Came From
There are several variations on the wheel-leg design in the world of robotics. One design, the Quattroped, is a four-wheeled/legged robot developed at the Bio-Inspired Robotic Laboratory at National Taiwan University. In a move worthy of Optimus Prime, it can transform its wheels into legs and shift from rolling to walking in under a minute. Each wheel folds itself in half vertically into a “C” shape to become a leg. The central hub then slides up to become a “hip joint.” The legs can rotate 360 degrees around and grab the terrain with their rubber treads. Although not particularly graceful, the Quattroped can pull itself up stairs and over tree roots, and boogie on the laboratory’s smooth floors.
An even simpler design is the Wheg from Case Western Reserve University in Cleveland. A Wheg—another portmanteau word that blends “wheel” and “leg”—works a little like a wagon wheel with the rim removed. Each spoke functions as a leg, but the whole thing revolves around a central axle, just like a traditional wheel. More elaborate versions of the Wheg have been developed, such as the Loper from the University of Minnesota, which has a rounded lobe at the end of each of its three legs, allowing it to roll a little with each step. The MSRox wheel-leg designed from the University of Iran actually has a separate wheel on the end of each spoke. But the basic Wheg design by itself is simple and effective.
BREAK THE CODE: TRISKELION
The concept of a wheel made of legs goes back much farther than the electronic age. It appears in the ancient Celtic symbol known as a triskelion, which is Greek for “three legged.” In most depictions, the legs of the triskelion appear to be joined literally at the hip as they gallop around the circle. A bare-legged version with a leaf-framed face in the center still appears on the flag of Sicily. Another interpretation, legs clad in medieval armor, graces the flag of the Isle of Man.
How It Works
The invention of the wheel is considered one of humanity’s greatest achievements for a reason. The mechanics and control involved in rolling a cart downhill are so much simpler than those required to run after it. Given a relatively even surface, a vehicle on wheels can usually move faster and more smoothly than even the fastest sprinter.
That said, if you’re a land animal, legs are the way to go. Legs have an advantage when it comes to climbing over obstacles. But getting robots to walk like people or animals isn’t easy. They need programming and sensory feedback to tell them which leg to move when, respond to differences in terrain, and stay upright while doing it. That’s why a design that combines the two can add up to more than the sum of its parts. On smooth surfaces, the wheel-leg can take advantage of the efficiency of a standard wheel. Off-road, it can handle changing surfaces more easily and climb over obstacles.
For robots, the Wheg-type wheel-leg has another advantage over the traditional hip-socket type limb: it can spin around an axle like an ordinary wheel, so it doesn’t need programming to help it with balance and alternating stride.
National Taiwan University Quattroped
Case Western Reserve University Whegs
Loper from the University of Minnesota
How Does a 3D Printer Work?
Three-dimensional or 3D printing (see Figure 2-6) has transformed the way researchers and hobbyists make prototypes. It’s now possible to create, test, and tweak a new design in minutes or hours instead of days or weeks. Even more extraordinary is how quickly 3D printers have gone from an industrial tool costing thousands of dollars to something tinkerers, artists, and small startup businesses can buy or make themselves for just a few hundred dollars. Companies like MakerBot and open source projects like RepRap are bringing 3D printing to homes, offices, and makerspaces around the world. And for those who don’t have access to a machine themselves, online services like Shapeways, Ponoko, or Sculpteo make it possible to upload a file (or choose a design on the website) and have it printed and mailed to your home.
Figure 2-6. A 3D printer producing the wheel legs from this section
Robot designers use 3D printers to create custom parts as well as entire bodies. Bob the BiPed is a 3D-printed robot created by hobbyist Kevin Biagini. Aside from a boxy body and two short stubby legs, the only other parts needed to build Bob are a small microcontroller board for a brain, a couple of ultrasonic sonar sensors for the eyes, and two servo motors to move the legs. Directions for making Bob are available on Instructables. More ambitious is Jimmy, an open source cartoon-like humanoid robot designed byIntel Director of Future Casting Brian David Johnson. As part of the 21st Century Robot project, anyone will be able to access the files to print their own Jimmy body and modify his design as desired. Plans call for a kit users can buy to animate their own Jimmy.
But the most exciting use of 3D printing in robots so far has to be the Robohand, a prosthetic developed by South African carpenter Richard Van As. He started working on creating his own artificial hand after losing four fingers in a work accident. Using donated MakerBots, Van As and Ivan Owen, an American collaborator, came up with a design that was cheaper and more functional than experimental models available commercially. They released the plans for free on Thingiverse, a file-sharing site for users of MakerBot. Then, in 2014, Jon Schull of MAGIC, a collaborative innovation center at the Rochester Institute of Technology, founded a group called e-NABLE that matched volunteers with 3D printers and children with missing hands or fingers to create custom-fitted Robohands at a fraction of the cost of traditional prosthetics.
The standard 3D printer used at home or the neighborhood makerspace works by taking a spool of thermoplastic filament—a plastic thread that becomes soft enough to flow when heated—and feeding it through a movable printer head, or extruder. The computer-controlled extruder squirts out a thin stream of softened plastic, building it up layer by layer to create the shape you have programmed in. It’s akin to squeezing frosting out of a piping tube—in fact, one of the first home 3D printers was called the Cupcake CNC. (CNC stands for “computer numerical control,” a term also used to describe computer-controlled routers that can be programmed to cut out patterns from wood and other materials.)
The two most common types of plastic used in 3D printers are ABS (the same plastic used to make Lego bricks) and PLA (a biodegradable plastic made from cornstarch, tapioca, or other plant material). You can also get filament made from a wood-plastic mixture, and more varieties are being developed all the time. The Pancake Bot, a 3D printer made using a Lego Mindstorms Robotics kit, was a big hit at World Maker Faire New York in 2012 (see Figure 2-7). Its computer-controlled ketchup-bottle extruders squeezed batter onto a griddle platform, creating tasty sculpted hotcakes. Other materials like plastic, glass, ceramic, and metal can also be 3D-printed using a method called sintering. This technique involves aiming lasers at a bed of powdered material to fuse it together into a specific shape with incredible detail.
Figure 2-7. The Pancake Bot in action
With schools, offices, and copy places beginning to get their own 3D printers, it may seem like 3D printing should be as easy as making a copy on a Xerox machine. Not quite. There’s still a lot of preparation and cleanup needed to produce a successful print. And just like the early days of the Xerox machine, 3D printers still require a fair amount of troubleshooting. When something goes wrong, you may have to check the temperature of the printer platform, how fast the filament is being fed to the extruder, and on and on.
Before you can even print a new design, it’s important to be sure it works within the limitations of the machine you’re using. Things to think about include:
Does the object have a good base?
Since the printer builds objects from the bottom up, you need to make sure it adheres to the printer platform. The printer software can help you with this by generating a removable “raft,” a thin base of plastic underneath your object.
Is the orientation optimal?
Each new layer of your object has to be supported by the layers below. Holes or gaps can be a problem, but sometimes just flipping an object on its side will make it possible for a design to print successfully. When parts of the printed object stick out enough to sag before the plastic has hardened, the printer software can generate a lattice of support material to hold it up. On printers with two or more extruder heads, you can set up the machine to print the support in a different material, making cleanup easier.
How much time and material will it take?
You can adjust the resolution on the printer, just like you can on a 2D machine. Lower resolution prints have less detail, but take less time, too. Depending on the design, you can also save time and material by choosing to make the object hollow or honeycombed inside instead of solid. In addition, you can specify how many skins (aka shells) to print, which tells the machine how thick to make the outer walls.
Your print may be a little rough when it’s done, with ridges where the layers were laid down. These bits of excess material, along with the raft and any support material, can be snapped off by hand and smoothed with a nail file or sandpaper. A more advanced method is to give your printed object a nice shiny finish by exposing it to acetone (nail polish remover), a solvent that dissolves ABS plastic. This has to be done carefully and with good ventilation, because the fumes are both flammable and toxic.
Although they’re not quite foolproof enough for the average person, it’s clear that 3D printers are quickly becoming the go-to tool for creating instant models, one-off parts, and unique toys and artwork. It won’t be long before they take their place in most households, alongside the paper printer/copier/scanner and other devices that are now a part of our lives.
Making the Project
The Wheg type wheel-leg makes a perfect subject for practicing prototyping techniques. It can easily be built in one piece, yet it offers an infinite number of variations to try out. And it provides a great opportunity to learn how to use 3D modeling software and 3D printers. You don’t even need to construct an entire robot to test it out. Just make a rolling platform, like a child’s pull-toy, or hack an old RC car by replacing the standard wheels with wheel-legs, and see how your design performs in different environments.
To get you started with CAD (Computer Aided Design) software, directions for re-creating a basic three-wheeled wheel-leg are outlined below. It uses a free online program called Tinkercad, so you don’t even need to download any software to your computer.
Don’t forget to document your work!
§ Time Needed: Several hours or more
§ Cost: Roughly $10 per piece for plastic printing from Shapeways; cost of filament for your own printer (or a borrowed machine) is negligible
§ Difficulty: Easy to moderate
§ Safety Issues: 3D printers and laser cutters can be hazardous—if you have access to one but have never used it before, take a class or ask someone more experienced to help you get started. See 3D Printing Linkbox for other titles that explain how to use these machines in more detail.
What You Need to Know
§ Skills You Already Have: Drawing, cutting, assembling a model
§ Skills You Will Learn: 3D computer modeling
Gather Your Materials
§ Pen and paper
§ Computer and access to 3D CAD software or online tool
§ Access to 3D printer or online 3D-printing service
Step 1: List Your Requirements
When designing your own version of a wheel-leg, its appearance will depend on several factors:
Where it needs to operate
Think about what you want your wheel-leg to do. Will it be traveling over hard ground or soft? The feet of a wheel-leg meant for rocky hillsides or rubble-strewn city use should be sturdy but small enough to avoid getting caught in holes. In deep sand or mud, on the other hand, a larger footprint will act like a snowshoe and distribute the robot’s weight over a larger surface area to keep it from sinking into the soft surface.
How it will be incorporated into a robot body
If you’re going to test your wheel-leg on a pre-existing robot or rolling toy, make sure it matches the dimensions of its regular wheels and can attach to the axle or gears.
How it is built
A 3D-printed wheel-leg has to fit on the printing platform and work within the limitations of the softened plastic filament it will be made out of.
Step 2: Plan Your Project
To create an original 3D model, you may want to sketch out a rough idea of what you want first—by hand on the back of a napkin is fine at this stage. Written notes and labels will help clarify your idea. Look at photos of existing wheel-leg designs online for inspiration.
Next you need to choose the type of 3D-modeling software you want to try. There are several popular programs you can access for free, including Autodesk 123D, 3DTin, and SketchUp. Each works a little differently, but once you have used one, it’s easier to learn others. You will use the free online program Tinkercad for this project because it is designed for children and other beginners and works with 3D shapes, eliminating the need to turn a 2D drawing into a 3D model.
HOW TO USE CAD SOFTWARE
CAD software has been around for a long time. Powerful (and expensive) versions for professionals are used by architects, engineers, and designers of every sort. It is also used by illustrators and animators to make two- and three-dimensional drawings and videos. Computer games are built using 3D modeling software. If you ever created a world in Second Life or built a castle in Minecraft, you’ve used a type of CAD software.
With the advent of 3D printers, laser cutters, and CNC routers, CAD software has become more popular than ever. Today you have a choice of several programs, each of which work somewhat differently. Many of them offer a basic version you can use for free, as well as a paid version with additional features. Some can even be used right online, just by signing up.
3D-modeling software lets you see exactly what your design will look like while you’re drawing it. Depending on the program you use, you can start with a 2D representation and then “extrude” it to pull it into the virtual third dimension on your screen, or build directly in 3D. Typically the design sits on a grid that looks like the graph paper you may remember from high school geometry. Just like in geometry class, the program shows you where your design is located using coordinates on the horizontal x-axis and vertical y-axis. The z-axis runs parallel to both of these, poking out above and below the plane of the grid into the third dimension. The program also lets you change your point of view in relation to the grid, so you can view it from any angle.
Starting a new design from scratch is as simple as assembling basic shapes (e.g., square or circle in 2D, cube or sphere in 3D). You can drag the shapes from the sidebar and drop them on your grid background or “floor.” Keep adding shapes, connecting them by making them touch or overlap. Some programs also let you upload a scan of a drawing, 2D digital design or photograph, or 3D scan of a model. You can then sculpt your shapes by stretching, squashing, or otherwise deforming them. Corners and edges can be beveled using the chamfering function, or rounded off with the fillet tool. To create an opening or indent, switch a positive shape into a negative space. You can also use a negative space as a tool to drill a hole or slice away a section of a shape.
Once you’ve created your 3D object, the discrete shapes used to create an arm or a torso can be joined together into a group. A group of shapes can be cloned, flipped, and otherwise manipulated as if they were one big shape. One group of shapes can also be subsumed into a bigger group, creating multiple levels. Ungrouping reverses the process, allowing you to back down through the levels until you are working with the individual shapes again.
When you’re finished, your design can be downloaded in a file format such as STL, which is compatible with 3D printers and online 3D-printing services. You can print out your design at home, your local makerspace, or even some neighborhood copy shops. You can also send it to an online 3D-printing service like Shapeways or Ponoko. Or share your design with friends or the public on sites like Thingiverse.
For other types of 3D construction, some CAD programs will also slice your object according to your specifications, creating layers that can be reassembled into an approximation of the original 3D shape. The 3D software will create printable patterns, much like a dress pattern, to show you how to cut the slices out. These patterns can be laid out on paper, wood, Plexiglas, or other flat material to be cut out by hand, a CNC router, or a laser cutter. The cut-out pieces can then be stacked or fitted together to make a good physical approximation of your design.
Step 3: Stop, Review, and Get Feedback
Before you start building, review what you’ve done to make sure it fits your criteria. If you can, ask a knowledgeable friend to look over your design for any obvious problems you may have missed.
Be sure to include some way to attach your wheel-leg to a vehicle or robot body for testing. For comparison, you can look at the wheel-leg I created on Tinkercad. You can also see it on Thingiverse and Shapeways.
TINKERCAD CHEAT SHEET
Tinkercad is a free online CAD program that lets you build designs from three-dimensional shapes. Beginning tutorial videos help you get started, and more videos are available on Tinkercad’s YouTube channel—the Shortcuts video is particularly helpful. You can also find information about design features on the Tinkercad blog. Although Tinkercad is relatively easy to use, skimpy Help and Search functions are weak points.
Here are some basics:
§ Create a free account so you can save your work and share it with others.
§ You can create “projects” in which to store several “designs.” When you start a new design, Tinkercad will assign it a random nonsense name, which you can change by going into Properties from the Design menu. You can also save or duplicate your work from that same pull-down menu. That’s useful if you want to create a backup that makes it easier to go back to a certain point and start again.
§ When you create a new Tinkercad design, you will see an empty Workplane. This is the virtual surface on which you will “build” your object. Surrounding it are menus and icons you will be using as you build. In addition, various pop-up windows will appear as needed.
§ To view the Workplane from a better angle, use the gray navigation icons on the left side of the screen to turn it or tilt it up or down. To zoom in on a particular piece, select it and then click on the little drawing of a box above the + and − signs. To go back to the center of the grid, click the little house icon in the center of the circle with the navigation arrows. You can also change your view by right-clicking your mouse to grab and rotate the grid. Scrolling your mouse wheel will zoom in and out on the grid.
§ The sidebar to the right of the Workplane has tabs that open when you click on them. “Geometric” has a selection of premade shapes that you can click and drag over to the Workplane. “Shape Generators” lets you create your own shapes (see the directions for this project). “Helpers” includes a ruler, which you will be using in this project. If you want to delete an object once it’s on the Workplane, just click on it and hit the Delete or Backspace key on your keyboard.
§ “Snapping” an object to the grid lines it up to the horizontal and vertical markings on the Workplane. By default, those markings are 1.0 millimeter apart. To make more exact adjustments, you can change the Snap Grid setting (in the bottom righthand corner of the grid) all the way down to 0.1 millimeters. You can also simply turn it off.
§ If you will be 3D printing your object yourself, you may be able to use one of the presets to adjust the grid on the screen to the size of your printer’s print bed, using the Edit Grid button on the bottom right.
Step 4: Design and Build Your Prototype
For this project you will make a wheel-leg from three rectangular solid legs connected in the shape of a “Y.” Each leg ends in a rounded foot, and they are connected in the center by a ring-shaped hub. The final size of the entire wheel-leg is only around 60 millimeters (a little less than three inches) across, so you can print more than one at a time on most print beds. Just a note—the default unit of measurement in most 3D CAD programs is millimeters.
1. To build your wheel-leg, start with a box shape. In the menu column on the right, find the Geometric tab. Click on it to reveal a palette of shapes. Find the shape labeled “Box” and drag it over to the grid. Notice that each corner of the box is marked with a tiny white rectangle, and each side of the box has a black dot in the center of its bottom edge. When you mouse over these spots, the length of that side is shown. By default, the box shape is a cube, and each side is 20 mm.
2. The basis of the leg is a rectangle 30 mm long by 5 mm wide by 8 mm high. To modify the size of the box, click on a black dot on the front to grab it. Drag it forward until it is 30 mm long.
3. To adjust the width, grab the black dot in the middle of the long side you just created. Move it inwards to squeeze the box to a width of 5 mm. To adjust the height, grab the white dot at the top of the box and drag it down until the rectangle is 8 mm high. Once you’ve got the leg/rectangle to the right size, go to the Inspector box in the upper right of the grid and click to lock the transformation. That will keep you from accidentally changing the dimensions as you move things around.
4. Now for the foot. When finished, the foot will be rounded on the bottom with a pointy toe and flat top. Its dimensions will be 20 mm long, 5 mm wide, and 8 mm high. To create this irregular shape, you’ll use a shape generator, a program that lets you set the parameters of your object. From the Shape Generators menu in the sidebar (which is above the Geometric menu in the sidebar), drag an Extrusion (which looks like a gray cylinder) over to the bottom of the leg, on the left side. Viewed from above, the round gray shape should overlap the rectangle so that together they look like a lowercase “d.” Before you start to modify the cylinder’s round shape, first shorten the height to 8 mm, using the white dot on the top.
5. Next, to change the shape from a cylinder into a foot, you’ll work with the smaller two-dimensional Profile grid that popped up when you clicked on the Extrusion. On it you should see an outline of a circle, which represents the cylinder as seen from above. The four dotted lines tangent to it on the top, bottom, left, and right are the Bezier handles. You use them to stretch, squash, or twist the outline. Once you’ve got the shape as close to finished as you can get in Profile, you can tweak it still more on the Workplane (the main grid). But before you make any changes, mouse over the little gear in the top right corner of the small Profile grid. A box labeled “Snap” will appear; check that box to make the shape line up with the grid.
6. Grab the top handle by the little square dot at the top of the circle and pull it straight down so that it lines up with the square boxes on the side of the circle. If you wait a second, you’ll see the 3D Extrusion on the big grid take on the same kidney shape as the outline on the little Profile grid.
7. Grab one of the round dots on the ends of the left handle. Shorten the handle by bringing the dot towards the middle. The other dot will mirror your movements until both round dots are overlapping the middle square dot. You have now created the pointy toe.
8. Grab the bottom handle by the square dot in its center. Move the handle straight up until it is level with the lower round dot on the right handle.
9. Center the right handle between the top and bottom handles by moving it down (grabbing it by the square dot). Shorten the right handle to fit between the top and bottom handles. Next, tilt the right handle until the top round dot covers the rightmost round dot on the top handle.
10. Lengthen the bottom handle by dragging the right round dot until it is directly under the square dot on the right handle. Then grab the bottom handle by the square dot and move it to the right until it connects with the right handle. You may want to uncheck the Snap function to get it straight.
11. Going back to the Workplane, the foot should be roughly shaped like a parallelogram. Move the foot if needed so it overlaps the rectangular leg at a right angle. You want a smooth “sole” on the bottom of your foot, so take a look from that viewpoint to make sure the leg isn’t poking through the “heel.” If you want, you can sharpen the corners of the foot by lengthening the handles on the Profile grid so they overlap slightly. Or you can point the toe or the foot a little up or down by rotating it on the Workplane using the curved double-headed arrow that appears when you click on the foot.
12. When you are satisfied with the design, open up the Helpers menu (right above the Geometric menu) in the sidebar, grab the Ruler and drag it over to the foot. The measurements of the foot should appear. You can change the measurements as desired by simply typing over the numbers. Make the foot 5 mm wide and 20 mm long. If needed, slide the foot around one last time to get it in the right position. Dismiss the ruler by clicking outside the foot, and then on the gray X circle.
13. Finally, select the entire leg by clicking on the grid and dragging the light blue outline around until the entire leg is inside. Click Group on the top right of the page to save the leg-foot shape as one piece and avoid changing it accidentally. The leg will switch to a single color. (You can also change the color yourself at any time with the Inspector pop-up box.) If you need to work with the pieces individually later, you can always ungroup them.
14. Now you can make the other legs. Click on the leg, then click on Edit at the top right of the page, then Duplicate. Click on the leg to slide the duplicate away from the original and up to the left. Then rotate the second leg 120 degrees, using the protractor that appears when you mouse over the curved double arrow. Move the rotated leg so its corner touches the top corner of the first leg like a hinge. If necessary, turn off the Snap grid function to give you more control. Duplicate and rotate the second leg the same way to make the third leg. Move it so that it touches the other two legs. The tops of the three legs together should form an equilateral triangle at the center. Zoom in as much as possible to check that the three edges are touching, without gaps or overlapping. Then group the three legs together.
15. The last step is to make the hub of the wheel-leg. Go to the Geometric menu and drag a cylinder onto the grid to work on it, a little way from the wheel-leg. Resize the cylinder to 8 mm high, 12 mm long, and 12 mm wide. Then go to the Holes tab in the right sidebar and drag a Cylinder Hole onto the grid. Set it to 10 mm high to be certain that it goes completely through the hub. Set the diameter of the hole to 3 mm. That should fit on an axle made from a bamboo skewer. (To adjust the size of the hole, see the section on troubleshooting.) When finished, drag the hole into the solid hub. Poke the hole down below the grid a little bit so the ends stick from the top and the bottom of the hub. To center it, select the cylinder and hole, go to Adjust on the top menu bar, and click on Align. On the grid, guidelines will appear. Click on the guidelines that cut through the middle of the cylinder horizontally and vertically. Then dismiss the Align tool, select the pieces, and group them together.
16. Finally, drag the hub into the middle of the wheel-leg. The Align tool will not work here, so center the hole as well as you can by eye. Group all the pieces together. Remember, you can always go backwards and ungroup them, layer by layer, if need be.
17. To download your wheel-leg for 3D printing, choose that option from the Design menu at the top right of the page. You will be asked what kind of file to save it as (this depends on the 3D printing machine you use). The Design menu also contains links to let you upload your file to Thingiverse or order a 3D print of your creation from online services such as Shapeways, i.materialize, Ponoko, or Sculpteo. You can also download the file for laser-cutting.
Step 5: Test Your Design
When your printed wheel-legs are ready, attach them to a bare-bones platform to test them out. You can make axles out of cocktail toothpicks, bamboo skewers, or wooden dowels. Poke them through one of the “tunnels” inside a piece of corrugated plastic or cardboard, attach the wheel-legs, and make them move by placing them on a tilted surface, pulling them with a string, or adding a motor. You can also try out your wheel-legs by substituting them in for wheels on a toy car or robot platform you have available.
Step 6: Troubleshoot and Refine
Most hobbyist-quality 3D printers are not that exact when it comes to tiny details like the hole in the hub, so getting the wheel-leg to fit on whatever you use for an axle can take a little trial and error. For a better grip, you can try making the circular hole more “D” shaped by overlapping a small box on one side. You can also turn any shape into a hole using the Inspector pop-up box. A star-shaped hole with points to grab a wooden shaft may allow for more variation in the thickness of the axle, for instance. Or, a less high-tech solution is to widen the hole once the wheel-leg is printed by carefully hammering in a nail of the right diameter.
If your wheel-legs can spin freely around the axle but don’t roll smoothly along the ground, you may need to go back and look at your design. Do the feet lie along the line of an imaginary rim? If not, your “wheel” may not be round enough. You can also try making the feet shorter or turning up the toes a bit more. Use the results of your tests to modify your drawing and make a revised prototype.
If the feet slip while trying to move on slick surfaces, you can give them some traction with a peel-and-stick craft foam “shoe,” or put treads on the bottom of your wheel-leg feet with lines of hot glue. You can also cut out the ridged fingertips from a rubber dishwashing glove and use it for the sole of the shoe.
Step 7: Adaptations and Extensions
Once you’ve got the hang of using the 3D CAD software, and have tested out the basic design, try creating a wheel-leg prototype of your own. Questions you should consider include:
§ How many legs per wheel work best?
§ Are legs with bent knees better than straight legs?
§ How much curvature should the feet have?
If you want to try making the wheel-leg by hand, using cardboard or other material rather than 3D printing, use a CAD program that lets you design in 2D as well as 3D, or which lets you convert your 3D model into slices. Programs to try include AutoDesk 123D and SketchUp.
To extend the project, you can go on and design the rest of the robot using 3D modeling, to find the best body and attachment options for your wheel-leg design.
3D PRINTING LINKBOX
Bob the BiPed
Jimmy, the 21st Century Robot
Wheel-leg on Tinkercad
Wheel-leg on Shapeways
Make: 3D Printing: The Essential Guide to 3D Printers by Anna Kaziunas France (Maker Media, 2013)
3D Modeling and Printing with Tinkercad: Create and Print Your Own 3D Models by James Floyd Kelly (Que, 2014)