by Camille | Workshop Blog
In the video below we have documented our first “basic motion test” on the new Mars Rover. Using our radio-based handheld remote control, we are testing the forward, backward, and rotation of the robot to make sure that it moves smoothly and correctly. We then test the NASA-style rocker-bogie suspension system. The purpose of the rocker-bogie suspension system is to allow the robot to drive over obstacles (such as rocks) and remain relatively steady even if the obstacle is only on one side. The front wheel goes over first while the two back wheels remain firmly on the ground (rotating on the swivel-hub of the rocker-bogie hub, which is between the two back wheels). Once the front wheel is firmly on the ground again, the rocker-bogie on the back goes over the obstacle, one wheel and then the next.
You may also notice a green laser on the robot. The green laser indicates the scan region for the thermal sensor. The robot reports eight temperatures from left to right where ever the laser is pointing. This will be used by science center visitors to scan and identify targets in the exhibit. The results of this first test are described below the video.
The Mars Rover by Beatty Robotics from Camille & Genevieve Beatty on Vimeo.
Overall, the robot performed extremely well in its first motion test. The motors and wheels ran very quietly and smoothly with no noticeable issues, both forward and backward. The steering servos turned correctly and rotated the robot on command in both directions. The robot traveled smoothly with no noticeable noises, vibrations, or problems. All the tubes, screws, hubs, joints, and other components remained in place and performed their function. Mechanically, the robot met all requirements during the motion test.
We noted that the rocker-bogie suspension system worked very well, but as expected, it behaved a little bit differently compared to the first Mars Rover we built. On the first Mars Rover we installed a counter-rotating differential gear box in the central shaft of the robot. When one of the front wheels goes over an obstacle and that side of the robot lifts up, then the other side of the robot is pushed down in the opposite direction (keeping the main box more level). The exhibit at the New York Hall of Science is flat, so we won’t be doing any rock-hopping, so we didn’t install a counter-rotating gearbox (we wanted to keep the design simpler and easier to maintain over time). This means that when one wheel of the robot goes over a large obstacle, both sides lift up instead of just one side. However, overall, the rocker-bogie suspension system still works very well. In fact, because it is more rigid and less gangly, the robot is rock solid, very steady when traveling and when being picked up, and there appears to be no detrimental effects to the simpler design.
There are two known issues at this time that we need to work on: 1. Sometimes when the rover stops, the digital servos (which are known for being jittery) take quite a while to find an equilibrium and settle down. Sometimes they jitter/wiggle when the robot is at rest. This is not a critical problem, but it wastes battery power and we want to improve this situation. 2. The eight sonar sensors, which are intended to measure the distance to the nearest object in the direction of travel, are reporting erroneous distances. This is a critical problem that needs to be resolved, otherwise the robot may run into things when it is driving autonomously. We are working on that now.
by Camille | Workshop Blog
A few subscribers have asked about laser and the thermal array sensor on the new Mars Rover, so this post is dedicated to that:
We needed to add a targeting laser and a thermal array sensor to the Mars Rover robot, so we experimented with different wavelengths of lasers and types of lenses. After looking at the results of our tests, we decided on a 532 nanometer laser (which is bright green in color), with adjustable 5 to 25 milliwatts of power, and a 38-degree line-generating lens.
We plan to use a TPA81 Thermopile Array sensor to scan the temperature of the objects that are out in front of the robot (a few feet or a few yards away). This particular sensor provides an 8 x 1 array of eight temperatures corresponding to eight points from left to right in front of it (about 40 degrees across). But it’s hard to tell exactly where it’s pointing and which objects it’s measuring, so we plan to use the laser to identify the scanning zone. We cobbled together a prototype and our initial tests of all this worked well, so we moved into the final build stage and mounted it into the robot. The key was to make sure that the laser was lined up with the thermal array sensor, which necessitated a special mounting mechanism.
Here I am with a devilish grin as I experiment with different types of lasers. I decided on the green laser because it’s strongest, easiest to see, and “just plain the coolest.”
I used my favorite tool, the vertical mini mill, to do some machining to modify the laser mount. We then drilled and tapped various threaded holes into the mount.
Here is the finished Laser Mount. The laser fits in the large hole in the middle and is held with a set screw through the threaded hole on the side. The threaded hole on the bottom is used to attach the mount to the articulating arm. The threaded hole on the face (top in the picture) of the mount is used to attached the Thermal Array Sensor. This way the Laser and Thermal Array Sensor are screwed together and always pointing in the same direction/angle.
Here we have mounted the Target Laser (on top) and the Thermal Array Sensor (on the bottom) into the front of the robot (front plate removed). You can’t see it too clearly in this picture, but the laser mount sub-assembly is mounted on an articulating arm on a base so that we can adjust the pointing angle/direction. Please note that the line generating lens is tilted in this picture and needs to be adjusted so that the visible lines in the lens are vertical (which is weird because it generates a horizontal line).
The finished result. We’re ready to scan some alien Martian life ! (Or at least some warm rocks!) 🙂
by Camille | Workshop Blog
We’ve been busy working on the new Mars Rover. Over the last few days we’ve been focused on the chassis, the thin-film solar panels for the top, the mast, and aligning the servos for steering. Pics below.
Here I am using my new power driver to work on the robot’s chassis. This is now my favorite tool.
Working on assembling the solar panels that will cover the top deck of the Mars Rover.
The completed “solar wings” and top deck. These are actual thin-film solar panels attached to custom-machined aluminum plates.
We ran into a particular challenge when it came time to build the mast. In the end, we decided to design and machine a custom servo plate using the CNC. The top of the plate will hold the pan servo. The bottom of the plate will hold the shaft tube (using a circular slot).
The custom servo plate mounted on the mast. The plate is held down with a threaded rod that goes down through the mast and threads into the bottom of the rover.
Here is the completed mast assembly. From the bottom up: The robot’s bottom plate, the mast flange, the mast tube, the custom servo plate, the pan servo, the top servo plate, the servo horn, and the mast head, which is made out of two custom machined plates of aluminum.
In order to align and calibrate the steering servos we created a special, full-scale drawing that indicates where the wheels should be at each of the various steering positions. This really helped center the servos properly.
by Camille | Workshop Blog
Here are some pictures as we continue work on our new Mars Rover.
I’m using the vertical mill to modify an aluminum part for the Mars Rover. Cutting, drilling, and machining metal is my favorite area of work now. I’m turning into a good machinist.
I’m using our mini drill press to drill holes onto the metal wheels in preparation for tapping the threads.
A new skill: I use our largest Milwaukee drill to tap 6-32 threads into the holes she drilled into metal wheels. The threaded holes will be used to secure the tread to the wheel (we don’t like glue). Tapping threads is very tricky (it’s very easy to snap the tapping bit), but she does a splendid job.
I’m mounting a side-firing sonar sensor on the the rocker-bogie suspension system. She’s done most of the mechanical construction on the robot.
I’m feeding wires through the chassis tube.
After fishing the wires through the chassis tubes, I solder the power, ground, and signal wires to the sonar sensors.
I use the electric heat gun to complete the heat shrinking on the wires she just soldered. (Red nail polish is optional when heat shrinking in our workshop, but always a plus!)
Slowly but surely, we’re making progress. We’ve completed all the wiring and soldering for the sonars, motors, and servos on the chassis, and we’ve assembled and installed the wheels. We sill need to complete the rover’s top, install the solar panels, build the mast head, and many other elements. Stay tuned.
by Camille | Workshop Blog
The Mars Rover needs to detect infrared in its camera so that visitors can identify infrared-emitting rocks that are hidden in various spots in the Mars Exhibit. We have completed a series of experiments to confirm that this part of the Rover is working properly. Pics and explanation below.
The Mars exhibit has many Mars rocks. A few of the rocks are hollow and have an infrared lamp inside so the rock emits infrared energy. The visitor’s objective is to drive the Rover around the Mars exhibit and find the hidden infrared-emitting rocks. (On the real Mars, infrared-emitting rocks indicate that the rocks may have been formed in liquid water, and therefore may be a clue to past life on Mars)
The Mars Rover is equipped with the inner core of a high-quality 640×480 WiFi camera (the Sharx Security SCNC2700). The video camera looks straight ahead from its front panel. By sending the camera a special web-post (thank you to the folks at Sharx for their special help on this), we can programmatically turn the camera’s IR cut-filter on and off, thereby switching the camera between “Normal Camera” mode and “Infrared Camera” mode.
We didn’t have a “Mars Rock” to work with in our workshop, so for this experiment, we placed an infrared lamp inside a black box.
We turned the infrared lamp on.
We then covered the box with six layers of black muslin cloth. To the naked eye and when viewing it through the video camera in normal viewing mode it looks simply black. This is serving the purpose of our “Mars Rock” for this experiment. This is a picture of what we see on screen in the camera’s video stream. It looks like normal black cloth.
We then turn on the “Infrared Camera” mode. Just as we hoped and theorized, the black box now appears as bright white in the camera. So, when visitors to the exhibit switch the Rover’s camera into “Infrared Camera” mode they will be able to easily differentiate the infrared-emitting rocks from the normal rocks.
Special thanks to the folks at Sharx Security for not freaking out when I told them I was tearing one of their cameras apart piece by piece, integrating it into our robot, and sending it to “Mars.” They were very helpful in providing the technical information we needed to get the camera to do what we needed it to do. All indications are that it’s going to work very well for us.
