Design and print a single-piece robot chassis that holds all electronics

[darozak]

Ideally the robots will have a single-piece chassis that tightly holds all of the electrical components without needing screws.

[darozak]

My first attempt at printing a chassis went fairly well. For the most part, the holes accommodated the necessary parts. However, I don need to make the following adjustments before proceeding with the next print:

1. I need to make a secondary cutout on the front and back of the circuit board to make room for the pins and solder points. There is only room to hold the circuit board in the groove along its edges.
2. For the servo insert, I cannot have the groove for the wire at the top of the print because this creates a situation that doesn't allow for bridging the gap at the top of the cutout. As a result, the top of the hole sags and doesn't leave enough room for the servo to slide into the opening.

[darozak]

I updated the OpenSCAD code to address the preceding points and to add a slot for the battery holder (https://github.com/Darwin-River/Robotics/commit/4e05967f879c342b0ace44a155587af23293d3ec). Printing the new chassis tonight.

[darozak]

The print still leaves plenty of room for improvement. Here are the things I need to adjust before running another print tonight:

1. Remove the groove from the bottom of the well that holds the circuit board. Because I printed the chassis upside down, this groove created problems with bridging and it really isn't needed.
2. The slot is too thin and probably not wide enough to hold the circuit board. I need to expanded some more.
3. The batter slot is a little too tight for the battery holder when it contains batteries. I need to make it bigger.
4. Increase the length of the servo slot by a couple mm. Also, allow more room for the wire in the groove.

[darozak]

Now printing an updated model that adjusts the dimensions, adds pilot holes for the photosensors, and a resting point at the back of the robot. (https://github.com/Darwin-River/Robotics/commit/9fb4534c911f6ad008cdd590915c7c42b115f0e2)

[darozak]

The lates version of the chassis worked fairly well. All parts fit snugly. I think the pilot holes will work well for mounting the sensors. I ordered some M2 screws to mount the sensor boards on the front of the robot. Here are some thoughts on what I could do next to further improve the design:

1. Create a trapezoidal offset for the sensor mounts. The region that holds the pilot hole will just out from the body so that the sensor board can be mounted without crunching down on the solder points on the back side. The offset will include slanted sides so that it can be printed without awkward overhangs.
2. Have the wire holes for the servos go straight through the body of the robot. This will make it easer to pry the servos out of the chassis if necessary.
3. Add a mounting point for the microphone on the top of the robot. The mic will be used by the robot to detect the start signal.
4. Add rounded corners to the chassis. This is entirely a stylistic improvement.

[darozak]

I modified the print so that it includes elevated (conical) hardpoints for each of the three sensors and has the through holes for the wire run straight through the robot (https://github.com/Darwin-River/Robotics/commit/c9b89913221bebe0a68bf4efda0a9ae69f9fe5dd). Still not sure I want to add rounded edges because this will complicate the OpenSCAD script.

[darozak]

I ended up removing the servo pass through holes because I realized that having the circular cutout at the top of the print would create bridging problems with the servo cutout (https://github.com/Darwin-River/Robotics/commit/f054b18c18f38dad702d02dbf3d4bb37a3ce150d).

[darozak]

The new design works great! Only issue is that the photo sensors are mounted a little too slow and push the robot off the ground. I need to raise the photo sensor hard points to where they're level with the mic hardpoint.

I also noticed that the pins on the sensor are arranged differently than those on the IO shield. Therefore I need to use individual wires to connect the sensors to the board. I ordered some individual female-female breadboard connectors to solve this problem.

[darozak]

Updated the OpenSCAD file so that the photodiode sensor hard points are higher up on the chassis and parallel with the mic sensor hardpoint. I also decreased the diameter of the hemisphere on which the robot rests from 9 to 8 mm so that the robot tilts back a little further (https://github.com/Darwin-River/Robotics/commit/7ace217c20d76ccbec2976bb2703c1f41db1a8b5). Hopefully, this will help stabilize the robot.

[darozak]

The revised design works perfectly. I was able to assemble a complete robot with two wheels, a battery pack, and three sensors!