Pololu Blog (Page 18)
Welcome to the Pololu Blog, where we provide updates about what we and our customers are doing and thinking about. This blog used to be Pololu president Jan Malášek’s Engage Your Brain blog; you can view just those posts here.
Continuous rotation servos like FEETECH’s FS90R are popular actuators for beginner robots because of their low cost and ease of use—since the motor controller is built right into the actuator, it can be controlled directly from a microcontroller or RC receiver. However, to complete the drive system, you need wheels, and that is something that we have not been able to offer for the FS90R until today. I am excited to introduce the new 60×8mm wheels for FS90R micro servos, which should make it much easier to get your FS90R-based miniature robot rolling.
All that said, we still generally recommend creating custom drive systems out of individual motor drivers, DC motors, and wheels over continuous rotations servos, since that gives you much more control over performance. Continuous rotation servos are more appropriate for projects where cost and simplicity are more important than performance, and with these wheels and the FS90R, this approach is simpler and more affordable than ever.
LVBots held a mini-sumo competition at Pololu on August 20. The goal of mini-sumo is to make an autonomous robot that pushes the other robot out of a 30″ ring, but this is not BattleBots: the robots cannot be controlled by a human, and they are not supposed to damage one another. Eighteen robots faced off in our head-to-head double elimination tournament. The video above shows some of the more entertaining matches and the full results of the contest.
The robots have become more sophisticated since our previous mini-sumo competition. Our new Zumo 32U4 Robot, which came out in the meantime, improves on the Zumo Robot for Arduino by adding IR sensors and encoders. This allowed some entries to do well just by programming a Zumo 32U4 robot (for example David’s Zumo Red). Also, people generally have gotten better at fabricating and programming their robots. Some people used 3D CAD programs to design 3D-printed and laser-cut chassis.
Kevin’s Roku won the competition, with the consensus being that Kevin won because he did not have enough time to make a gimmicky robot (like his line following hovercraft). His compact design used our new A-Star 32U4 robot controller and Sharp GP2Y0A60SZ 10 to 150 cm analog distance sensors, which kept the wiring minimal and the sight range long. Ben’s robot, The Big Ben, was unchanged since competing in the previous contest, yet it managed to do much better this time around, taking second place (though Brian was operating the robot in Ben’s absence, so he might want to claim some of the credit). Paul’s reigning champion, Paul Sumo 2, took third place despite also remaining unchanged since the last competition.
Update: Here are posts about some of the robots in the contest:
- Grant’s mini sumo robot: Rattata
- Patrick’s mini sumo robot: Covert Ops
- Kevin’s mini-sumo robot: Roku
- Brandon’s mini sumo robot: Black Mamba
Are you in the Las Vegas area? Check out the LVBots Meetup page to get involved.
We are having a big Labor Day sale throughout the weekend, with 15% discounts on over 700 products when you use the coupon code LABORDAY15. Note that we will be closed on Monday, so orders placed after 2 PM Pacific Time on Friday, September 4 will be shipped on Tuesday, September 8.
For more information, including all of the sale items, see the sale page.
Jay Doscher posted on his blog at Polyideas.com about his 2-axis solar tracker designed to provide the optimal amount of power output with a portable setup. In the build, Jay uses a Raspberry Pi A+ topped with our Dual MC33926 Motor Driver for Raspberry Pi to control the motion of the system, which is accomplished using a Concentric 4″ linear actuator with feedback. In lieu of a GPS unit, the tracker uses hard-coded longitude and latitude coordinates with Pysolar, an open-source Python library, to calculate the sun’s predicted position. The system keeps the solar panel pointed at the calculated position with the help of a Razor IMU from SparkFun. The video above is time lapse footage of a mechanical test of the system that shows the unit tracking the sun (although it is indoors).
In the picture above, you can see the Raspberry Pi and dual MC33926 driver board on the left and the IMU on the right. The Dual MC33926 Driver for Raspberry Pi fits on top of the Raspberry Pi mainboard, eliminating a lot of wiring and making it easy to use while also leaving the setup looking clean and organized. Additionally, the Dual MC33926 Driver for Raspberry Pi provides a set of three through-holes where an appropriate voltage regulator can be conveniently connected, allowing the motor supply to also power the Raspberry Pi. You can see one of our D24V10F5 switching step-down regulators mounted on top of the dual MC33926 driver board to serve this purpose in the picture above as well.
This project was also a 2015 Hackaday Prize entry and made it to the quarterfinals!
For more information about this project, see Jay’s blog post, which has additional photos and details including a parts list and links to his code.
Zippy is an RC balancing robot created by Larry McGovern. It uses an Arduino Nano to read pulses from an RC receiver and accelerometer and gyroscope data from an MPU6050. After processing that information, the Nano commands two ST motor driver development boards, which each control a 30:1 37D mm gearmotor with encoder. The whole system is powered by a 3S LiPo (brand: Zippy, of course!). You can watch Zippy scoot around on pavement below:
In the video description, Larry mentions that he modeled Zippy after the Balanduino robot, but we would like to highlight one noticeable difference: he used his own pair of wheels, which are mated to the output shaft of his gearmotors with our 6mm scooter wheel adapters! I had a major role in designing these, so on a personal note, it is especially exciting to see someone get a good use out of them. (It also looks like our stamped aluminum L brackets are used to mount the motors.)
Customers have been requesting an assembled version of our Zumo 32U4 robot kit ever since we released it in March, so it makes me very happy to be able to announce that we now have three pre-assembled Zumo 32U4 robots to choose from:
The three options differ only in their motors, and while the speed and torque vary across the three gear ratios, the peak output power is the same for all of them. You could maximize speed (i.e. 50:1 motors) or torque (100:1 motors), or perhaps you are looking for something in the middle (75:1 motors). The following table compares the gear ratio in more detail, with the first four columns showing specifications of the gearmotors by themselves and the last showing the measured top speed of a Zumo chassis loaded to a weight of 500 g:
|Top Zumo Speed
@ 6V and 500g
|50:1 HP||625 RPM||15 oz·in||1600 mA||40 in/s||(100 cm/s)|
|75:1 HP||400 RPM||22 oz·in||1600 mA||25 in/s||(65 cm/s)|
|100:1 HP||320 RPM||30 oz·in||1600 mA||20 in/s||(50 cm/s)|
These three gearmotors are the ones we consider best suited for typical Zumo 32U4 applications (and many of our example programs are tuned to work with 75:1 HP motors), but we have many other gear ratios available that you can use when assembling the kit version of the Zumo 32U4 robot.
At this point, you might be wondering why it took so long for us to make an assembled Zumo 32U4 robot. Well, we have been working on several improvements to the Zumo 32U4 ever since releasing the kit, and we wanted to have them all in place before coming out with these more finished assembled products. The first improvement was to the sprockets, which changed from white with solid hubs to black with spokes. These new sprockets fit better on the motor shafts and make assembly and disassembly easier, and we think they just look cooler! They might also help you hide from your opponent’s IR sensors, but the color is of course no use against other sensing technologies like sonar.
The second improvement was to make a new component to hold and shield the IR LEDs used by the proximity sensor system. Without this, the LEDs are just supported by their leads and shielded by a piece of heat shrink (see the pictures above), and we wanted something better. Now the kit and assembled versions include a plastic LED holder that mounts directly to the front blade:
Finally, we have improved the blade. They are now stamped rather than laser-cut, and we have added cutouts around the general-purpose mounting holes so that they can be hand-bent to new angles as desired, independent of the blade angle. This new blade also has the chassis mounting tabs pre-bent to the appropriate angle, so that’s one less step required during assembly of the kit.
And speaking of the kit, we still strongly encourage people to get the Zumo 32U4 kit and build it themselves. We designed the Zumo 32U4 to be a starting point, and building it yourself will make you more comfortable with customizing and enhancing it. Making it yourself will also make it a little more meaningful when your robot triumphs over the competition!
Forum member spiked3, whom we previously posted about, has shared another robot with a custom laser cut chassis. The new robot uses his own custom Arduino shield, the S3-Pilot, which has sockets for an IMU and two of our MC33926 Motor Driver Carriers.
Custom Arduino shield created by forum member spiked3.
The MC33926 drivers control two 37D motors with encoders, and the encoder signals are processed by the Arduino. The robot also includes a lidar, PIXY Cam, and Raspberry Pi. The on-board IMU and encoders allow the robot to keep track of where it is and what direction it is facing, so spiked3 was able to implement a high-level interface for the robot that accepts movement commands like “go forward three meters” or “turn a certain number of degrees to the right”.
You can find out more about this robot and see some videos of it being tested on spiked3’s blog.
We are now finally carrying a cable for the Sharp GP2Y0A51SK0F Analog Distance Sensor 2-15cm. The GP2Y0A51SK0F, our shortest-range analog distance sensor, has a compact package with a unique JST ZH-style connector, so this cable will not work with any of our other distance sensors. The cable is 12 inches (30 cm) long, with wires that you can cut and terminate as necessary for your project.
For more information, see the product page.