Fundamentally, what we are building is a mechanical version of Build Project 2 - a robot that can follow a pre-determined path - but with a purely mechanical memory storage and retrieval system. This system is the focus of our prototype.

We considered two different approaches. In Da Vinci’s original design, he seems to use a system of fan-shaped pegs, inserted into a rotating mechanism. Those pegs would then hit levers as they are rotated, which cause the steering column to turn:

Open

Fig 1. Original Da Vinci Car

Open

Fig 2. Pardon the Handwriting

However, this largely resulted in a trinary system - right, left, and straight. We thought that using a cam-and-follower concept would be more interesting and allow finer control over the path. In our full version, a cam would also allow a 3rd dimension, enabling much longer ‘programs’. Ultimately, for our prototype, we arrived at what is essentially a simple inverted slider crank, with a spring-return to hold the follower to the (temporarily) 2-dimensional cam:

Open

Fig 3. Original Sketch

Open

Fig 4. Early Analysis

Open

Fig 5. Physical concept, made from scrap parts

We started on CAD and detailed design:

Open

Fig 6. MotionGen Model of Inverted Slider Crank

Open

Fig 7. CAD Model of Slider Crank with Follower

We realized during design that we would need to make this length adjustable - in order to “zero” our steering column on the cam, and avoid the robot biasing left or right during movement. This feature did not make it into our prototype but will be necessary in the future.

Open

Fig 8. Critical dimensions for tuning

Cam and Follower Design

As mentioned, for the final version, we plan to make the Cam a 3-dimensional feature - a multi-layer spiral to allow for longer programs. For the prototype, we went with a simple egg-shape that will show the full +/-45deg deflection of the steering column.

Our linkages translate a +/- 2.5cm deflection to a +/- 45 degree turn angle at the steering column. In designing the cam, we elected to have a “middle” diameter of 10cm - therefore, a full left turn would be at a diameter of 5cm (radius of 2.5cm), and a full right turn would be at a diameter of 15cm (radius of 7.5cm). Future iterations will include a program to translate a desired robot path into a cam profile - however, for the sake of this prototype, we plan on only demonstrating full deflection in both directions.

We also decided to parallel-path the hardware - testing both laser-cut cams as well as 3D printed. For the follower, we simply ordered a patio roller off of Amazon.

Open

Fig 9. Cam Sketch for Full Deflection

Open

Fig 10. Lasercut Cam, linkages and spacers

Open

Fig 11. 3D Printed Cam Design (with handle)

Open

Fig 12. Prototype follower

Fit Checks and Cam Iteration

We required two iterations for final assembly. Fortunately, both the 3D-printed and laser-cut cams functioned well with our follower, so we elected to use the 3D-printed model to future-proof the design. However, the housings for the roller, as well as the shaft-linkage adapter, required an additional iteration - fit-checking showed some incorrect assumptions had been made on shaft diameters and space required for the roller. Our first assembly also demonstrated that we had redundant bearings in our shaft-to-linkage adapter and the follower housing, so we redesigned and reprinted the housing and adapter to use fixed shafts, thereby simplifying assembly and reducing part count.

Open

Fig 13. Fit Check

Our fit check also revealed an incorrect measurement - a miscommunication of the diameter and radius of the cam. A second issue proved the requirement to make the slider mechanism adjustable, when full deflection was not achieved on the 2nd version of the cam. Altogether, we needed to print 3 different cams to reach the full-scale deflection that we were looking to demonstrate.

Open

Fig 14. Incorrectly sized Cam

Open

Fig 15. The three Cam profiles, for comparison

3D Cam and Follower Iteration:

For our initial design, we were inspired by a previous group from Spring 2024, who also used a 3D Cam and Follower, to have our 3D Cam rotate along the shaft. We then had our follower follow the 3D Cam as it rotates on the X-axis while moving along the Y-axis as the 3D Cam splines track up and down. We accomplished this by having two sets of shafts running along both X and Y axes that allow the follower’s 2D motion. The follower is tensioned to the 3D Cam with rubber bands attached to the shaft supports, so it is in constant contact with the Cam as it rotates. The shaft supports are fitted with linear bearings to ensure smooth movement along the shafts for the follower and follower base.

Open

Fig 16. Group 18, Spring 2024

Open

Fig 17. Sketch of the concept

Open

Fig 18. V1 of the Assembled Steering Mechanism with Stand in Cam

 

Open

Fig 19. Video of Steering Mechanism

Bill of Materials: