I've been fortunate to have a little time this summer to work on a project or two that are less than mission critical, while still being potentially useful...case in point a custom chess set design that I'd like to make available to support our engineering program, and would like student assistants to be able to manufacture for me. I'd finished the lathe programs for all the pieces and only had the slot in the bishop's mitre and the knight's nose to finish and was wondering how to make it simple. It occurred to me that amongst all the designs that are parametric and flexible for many situations, one of the strengths of 3D printing is to make that single, one-off, ultra custom, just-for-one-purpose piece.

So, here are the custom machining jig parts that simplify finishing the bishop and knight pieces for my custom designed chess pieces.

This is part of a study being done by Dr. Salim Azzouz of the McCoy School of Engineering at Midwestern State University. Senior Engineering students were challenged to find and/or design a gearbox that would transform a variable rotational input to an output at a constant angular velocity. This is the resulting study prototype, which uses a printed planetary gear train to accomplish the desired results.

If the single shaft is used as the input, it will drive the output (sun gear) at the same speed, however, if a different output speed is desired, the lower shaft on the other side (hub gear) can be driven faster, slower or even counter-rotated to produce the different desired output velocity.

This was a close collaboration between the students and myself in design and manufacture. Extensive work was done in SolidWorks, resulting in a functional digital assembly. The gears and carrier plate were printed on our Makerbot, the axles turned from aluminum stock on a CNC lathe and the support structure cut from Makrolon, an impact resistant, clear plastic. Delrin bearings were purchased from mcmaster.com.

Further study is planned, connecting the input and differential to motors, monitoring the velocities of each of the shafts with tachometers, recording the results at varying speeds and comparing the observed results with the theoretical calculations.

Marvin is a wall (and foot) avoiding robot made from printed parts, an Arduino board, an Adafruit.com motor controller board, 2 motors (from Pololu.com), a distance sensor (from Pololu.com) and a battery clip (Radioshack).

Video of the bot in action: youtube.com/watch?v=PfaEIkYA8fI

The body, wheels and neck are parts designed specifically to fit what we already had on hand (with the addition of the Pololu parts since we lacked matched motors or a distance sensor).

This bot was created as a mini-project of the MSU robotics group. (Midwestern State University, Wichita Falls, TX)

I've taken to printing the top of Wizard's jewelry box as a good way of testing surface finish, since it is a demanding shape that shows up many of the difficulties, without taking too long to print.

I created this part to help with the calibration of the Skeinforge Carve setting for my Cupcake CNC. It is a set of two single-walled, open, stair steps 10 and 20 mm tall and 20x20 mm length and width. By serendipity, the inside wall pulls away from the outside wall as the carve setting gets too high, allowing me to precisely set my Carve to 0.39

WARNING: Your mileage may vary!

The Klann Lego Spider is a LEGO compatible prototype of the Klann linkage (see mechanicalspider.com).

Video clips of the spider are available at:
youtube.com/watch?v=Bl_5wCbvxMU and
youtube.com/watch?v=hz4mflE2foM

Update to v1.1 at:
tinyurl.com/y8lhcr3

The legs for the vehicle were milled from 3/8" plastic sheet stock. All the rest of the components are from LEGO Mindstorms (see tinyurl.com/6nj6w), which we use to introduce first year students to robotic concepts.

The project evolved somewhat during the build, as can be seen in the videos and pictures. The turning video shows a simple timed turn where the vehicle moves forward for five seconds, reverses the direction of the left legs for five seconds, then moves both sides forward again for five seconds. The forward motion video shows a skinny version of the vehicle that had to be widened when turning was added, since the skinny version simply fell over on its side instead of making a turn.