18.2 - Prototype and Design Iteration

Introduction

Upon comparing preliminary ideas in the proposal phase, we narrowed down to using some sort of cam follower system to draw a longhorn. The high-level idea is that we have a pen/drawing instrument attached to the cam follower, and at the same time we move the paper. In the initial design phase, we spent a lot of time brainstorming, white boarding, to try to come up with solutions for the following challenges:

1. How to design a system with a single motor that gives us two degrees of freedom
2. How to switch between subsystems/mechanisms

Here we document the ideas that we have gone through both on paper and with physical prototypes.

Design Variations

Very early on in brainstorming sessions, we found that any sort of cam switching mechanism is hard without additional motors. Therefore, we try to think of ways to not have to switch between subsystems (e.g. two cams). In addition to cam + follower which gives us one degree of freedom, i.e. translational motion along y direction, we also need motion along the x direction by moving the paper. However, this introduces some challenges. We were faced with 3 options: a) move the paper from left to right to draw the bottom part of long horn (i.e. the shape of the face) and then from right to left to draw the top part of the longhorn (i.e. the top of the horn), b) move the paper from left to right, reset to origin, and then left to right again, c) have two drawing instruments that draw the top and bottom at the same time. Unfortunately, all of the three options are challenging to do with one single motor. Then, a brilliant idea came to us: make the paper a cylinder! Now, instead of translating the paper, we rotate the paper. But more importantly, after we rotate the paper for one full circle, it automatically goes back to the starting point.

Now, we have our refined high-level design: one cam that encodes the entire drawing somehow, plus a paper cylinder. 

Cam Variants :

Figures 1 & 2 : Iteration #1 - Flat cam with an inner track and predicted path

The first variant of the cam proposed was a flat cam, which would have an inner track cut into it. Following along with the process in figure 2, as the follower moved along the outer track of the cam (1, figure 2), it would draw the upper curve, as motion in the Y-axis correspond to positive motion based on increasing distances to the center of the cam. As negative motion corresponds to decreasing distances to the center of the cam, we could draw the bottom curve by introducing a "smaller cam" by cutting an inner track into the cam (2, figure 2). We could then finish out the image and reset the machine by putting the cam into reverse (R, figure 2).

Figure 3: Helical cam CAD model

The second variant of the cam proposed was a helical cam. Using this design, we would be able to draw the entirety of figure by following the entire path of the cam. This variant eliminates a number of difficulties when using the flat cam with a track. Many of these issues would be linked to difficulties with manufacturing, such as building a follower which could reliably travel inside the track. As well as design issues, such as when we reversed the input from the motor, how would we ensure that the cylinder would continue to travel in the same direction?  It is for reasons such as these that we elected to use a helical cam in our prototype.

Follower Variants:

Figure 4: Single rod cam follower, with follower and track groove considerations

The first variant of the cam follower is a rod fitted into a grounded pipe with a wheel that follows the track of the cam and a spring that maintains a consistent force between the follower and the cam to ensure consistency in the image.  A follower using this design would require careful tolerancing between the pipe and the rod, and would require that the cam moves at a consistent speed relative to the follower to stay on the track. 

Figure 5: Slider cam follower variant

The slider variant of the follower is similar to the single rod style, using the wheel and the springs, but has the important distinction of placing the rod with the follower on a rod with a sliding bearing. This allows the follower to slide freely as the force of a rotating cam moves it along the track.

We then considered different meshing between the helical cam edge and the follower wheel (see Figure 4). We considered V-grooved, U-grooved and flat-grooved wheels; for the cam edge we considered straight edge and angular edge that compliments a V-groove. With the help of physical prototypes and reasoning, we decided to go with a V-grooved wheel + an angular edge for the cam.

Cam/Follower Assembly Variants

One major challenge with the helical cam+follower system is that we only want ONE degree of freedom from the follower along y direction, provided that we are using a paper cylinder which provides the other degree of freedom along x axis. However, with the use of the helical cam, it naturally introduces translational motion along x direction too. Therefore, we need to design extra mechanisms to constrain or cancel out the x direction translation brought by the helical cam. We came up with two different approaches.

Figure 6: Spline gear+Helical Cam+Stationary Follower

As the grounded rod follower requires that the cam translates along the X-axis for it to move along the track, this assembly makes use of a spline gear to provide the input from the motor and turn the cam. As the cam assembly turns, it will move through a grounded threaded component, which will cause it to translate.

Figure 7: Helical Cam + Sliding Follower

The ability of the sliding follower to move along the surface of the cam as it turns means that it can be paired with a stationary cam. This design essentially switches the requirement for horizontal translation from the cam to the follower.

At this stage, it becomes necessary to make practical considerations, such as addressing concerns over the maufacturability of the different configurations. The stationary follower configuration faces issues such as the availability, cost, or manufacturability of a relatively small spline gear and the threading of the grounded component that is used to produce translational motion. The sliding follower configuration can be produced almost entirely from parts that are easily available or can be made using a 3d printer or laser cutter. As such, we selected the sliding follower configuration to move into the prototyping phase.

Prototype Fabrication

The fabrication of the prototype can be broken down in to three sub-categories: the base, the follower, and the cam.

The Base

The base consists of two 6mm thick wooden boards, which were originally designed to hold the follower assembly upright, above the cam, allowing gravity to help keep the follower wheel pressed against the track.  These boards were then laser cut to size and assembled.

Figures 8 & 9: The upper and lower parts of the base, respectively

The Follower

The follower consists of a two 8mm rails placed in linear bearings, which are held in place by two housings on each rail. The rails are kept from slipping through the bearings by a shaft collar at the top of each shaft. At the base of each shaft is a connector piece, one of which connects to the keyed shaft, and the other which holds a writing implement and connects to the keyed shaft.

Figures 10 & 11: Connector and connector with slot for writing implement

The two connectors are held together with a keyed shaft, which is used to prevent the follower from spinning freely as the force is applied by the cam. The follower rides along the the keyed shaft using a sliding connector, to which a rod is inserted, which itself is connected to the casing for the follower wheel using a 4 cm length of 6 mm shaft.

Figures 12, 13, 14: Slider for the keyed shaft, casing for the follower wheel, casing and wheel assembly

The Cam

The helical cam assembly consists of the cam secured to a 6mm shaft using an attached shaft collar. The shaft is then attached to the base using two brackets, which contain bearings that must be pressed on.  An additional 5 cm of shaft protrudes from one of the brackets, so that the cam may be manually turned.

Figures 14 & 15: An early iteration of the cam and a later iteration placed on the shaft

Issues to Address in Future Iterations

Figure 16: A collection of broken 3D printed parts

When designing and fabricating parts for future prototypes, two important things to consider will be the tolerancing of 3D printed parts and their print orientations. Designing parts for a press fit means that not only are the parts difficult to place on the assembly without significant force, which causes concern over breaking the parts, but it can also cause issues for fitting multiple parts together and disassembly. This should either be done more carefully to avoid these issues or avoided. Another thing to consider is the forces that a part will encounter in its final position. If the part is printed with the grain along the same line as the forces applied to it, it may very well break. It is also possible to have issues with the surface quality of the part based on there the supports are attached based on the print orientation.

Prototype Assembly and Testing

Figure 17: The prototype assembly

With the prototype assembled, we tested it to see how well the follower rode along the track and find any issues that needed to be addressed. Upon testing we found that there was binding between the keyed shaft and the connector resulting from the connector being too short to account for the twisting experienced when the follower rode along the cam, as well as the surface finish on the keyed shaft. To account for this in future iterations, we want to replace the keyed shaft with a connector that rides on two cylindrical shafts. By using a second shaft, we can ensure the proper ratio to avoid binding and eliminate the concerns over the surface finish, while resisting out of plane movement. We would also like to place the shaft connecting the slider and the wheel casing at an angle, as this would ensure greater contact between the wheel groove and the tracks on the cam. 

Figure 18: Prototype Testing

Final Design Decisions

We arrived at our final design by making a few key adjustments to our prototype. These were the implementation of a rail system to avoid out of plane motion in the follower, the addition of the motor and electronics, and the design of the gear train and drawing cylinder.

As can be seen in Figure 18, despite the use of the keyed shaft to avoid rotation, the moment placed on the follower by the turning of the cam would still result in out of plane motion. To compensate for this we initially considered and tested a dual shaft design, wherein the keyed shaft would be replaced with a regular 8mm shaft with a linear bearing on it to prevent binding and the shaft holding the follower passed between two parallel rails that would prevent the shaft from moving out of alignment with the slider. While this partially solved our issue, it meant that a taller overall design would be needed as the longer shaft would need to pass between the rails before reaching the cam. To solve for this issue the dual rail system and the linear bearing slider were consolidated into a single set of parts. This meant that we could avoid out of plane movement, prevent binding, and keep the shaft short to prevent large moments and material waste.

Figures 19 & 20: A prototype dual rail system with linear bearing rail holder (left), a consolidated rail holder design (right)

The second major issue was that, as of the design of our prototype, the device was still powered by the user. To automate the process of drawing the figure on the cam, we decided to use a stepper motor (and accompanying motor driver and arduino) that would be able to turn the helical cam a set number of rotations forward or backward with replicable results, so that the follower would draw the full figure and the device could be reset mechanically. By positioning the motor in line with the cam, we were able to take full advantage of this rotation and keep the design straight forward.

The final major design consideration that would lead us to out final design was the inclusion of the gear train and drawing cylinder. The major concept that we proved with our prototype was that we could get the follower to travel along the track of the cam, but this would represent a limited success without the ability to inscribe the data of the cam into an image. As such we needed a way to get our motor to turn the cylinder we had designed alongside the cam. Thus,  by iteration we conceived of system by which we would turn the cylinder using a gear train consisting of multiple sets of spur gears to turn the cylinder using the power of our motor and refined it into a design which used a belt and pulley system alongside a set of bevel gears.

Figures 21 & 22: An early spur gear and cylinder design (left), a final belt and pulley+bevel gear cylinder design

With these design decisions made, we were ready to fabricate and assemble our final device!

Bill of Materials

For the final prototype, our team has discussed both a gear train and a belt pulley to determine the optimal system for connecting the helical cam and the cylinder to the motor, taking into account factors such as the friction and size of our entire system. Once all dimensions of our final prototype are finalized, we will move on to purchasing either several gears or pulleys with belt. Materials on the last three rows are the tentative materials we haven't finalized yet.