This past month, we have been working on a wooden mechanical model. We were hoping to make something that would be unique and iconic and we think we nailed it. What is it you ask? Well, what could be a more iconic movement than the bouncing DVD logo?
Brainstorming and Choosing a Mechanism
To start brainstorming, we first reviewed all of the 507 mechanical movements to generate ideas. After some deliberation, we decided to create a mechanism that would emulate the bouncing DVD logo that TVs would display as a screensaver. We reviewed the mechanical movements, specifically looking for back-and-forth linear motion with instant direction changes (as opposed to easing in and out), and found mechanism 114.
After discovering this movement, we quickly drafted an idea of using two of these mechanisms perpendicular to each other, in which each mechanism would cause a slot rack to reciprocate back and forth, where our DVD logo would interact with these slots and move back and forth on two axes. We would then use a series of gear ratios to make the gears inside the slots move at different speeds (so that the slots would then move at similar speeds to each other). Since we would want a long side of the screen and a short side, there would be two different sizes of the mechanism.
Gate 1
This gate required us to produce drawings, sketches, and vector sketches of our proposed design. We decided to create a CAD assembly of our design, so that we could better visualize the layers of moving parts and their ranges of motion. The assembly (shown below) shows a crank (left), two driving bottom gears, and two mechanisms of the movement we chose. The pinion and the teeth of the rack were generated by an online gear generator, which were edited into a mutilated pinion and two-sided shaped rack on Illustrator.
While this is the design we submitted for our Gate 1 assignment, we noted several problems with it, even before creating it in CAD. Firstly, the mechanism and design were very space inefficient, and would have required our “TV” to be much, much larger than the actual screen of moving parts. We thought of different solutions to this, such as rotating the mechanisms to exchange height and width for depth (left), or stacking the mechanisms on top of each other and changing how the slot was attached (right), both of which somewhat reduced the size at the cost of complexity.
We were also skeptical about whether or not the mechanism would even work as intended, given how the GIF of the mechanism showed teeth perfectly avoiding other teeth, and it would also take much more added complexity, parts, and time to ensure that each part is constrained properly.
Gate 2
These problems became especially evident when we laser-cut a miniature version of the mechanism. The rack with the teeth would not line up with the gear in the middle as its spacing had to be too precise. We tried multiple iterations of tooth count and sizing on the gears and each had its own set of issues. Additionally, since we were using wood to create everything, the level of precision in spacing we would need would not be possible without creating large amounts of friction. We attempted to make a base to fit the system in and we determined that this simply would not work on a larger scale. The system getting stuck because it skipped a tooth on a gear or became misaligned happened pretty frequently and multiplying the error by combining multiple racks and gear combos would not turn out well. After deciding to use a different design, we looked for a mechanism that could provide a similar movement.
We found a design online that was very similar to mechanical model 123.
In order to achieve the random looking bounce, we needed to connect the two gear and rack systems together in a certain way. First, we decided that we wanted the dimensions of the “screen” to be 5×7, hence one set of gears is 5/7 the size of the other set. If we were to then power these with the same rotational speed, one would move 7/5 faster than the other, so we needed the pairs of gears connected with a gear ratio close to 5:7. We decided on a 25:36 gear ratio. This is both very close to 5:7 but since these numbers are large and relatively prime, it would require more “bounces” off of the wall for the DVD logo to hit the corner again.
We designed some different shapes for the racks so that they would fit on a base, and slotted them in some dowels to let them move side to side. Finally, we decided to cut out the base from cardboard and the gears from wood for the gate 2, prototype.
Gate 3
For gate 3, we basically realized that we needed a good way of holding all of the gears and racks together. We decided to create a whole bunch of layers that would basically sandwich everything in place instead of it being free-floating.
Here is the final assembly we created in Solidworks, a couple layers at a time.
For the final steps, we placed all of the layers together and added a lot of washers between gears and on the racks in order to allow smoother movement. We had a bit of trouble with gluing the gears to dowels. At one point, we had basically finished the whole project but one of the gears in the center of everything was slipping on the dowel (so we could not turn any of the mechanisms). Because of this, we had to take the whole thing apart and reassemble and glue everything. Most of our time was just spent making minor adjustments to the assembled piece in order to reduce friction, optimize spacing, etc. For instance, one thing we noticed was that we needed a spacer that was the thickness of about half a sheet of wood. Therefore, we cut a large disk spacer on the waterjet as the metal sheet was thinner than the wood sheets.
For our primary metal cut piece, we decided to waterjet a handle for the box. We also put a vinyl sticker on the handle that was an arrow indicating the direction to turn.
We are very happy with how the final model turned out. One of the main things we learned in this project was the idea of measuring twice and cutting once. Since our CAD model kept growing as we added layers to it, we underestimated how much space was needed apart from the gears. Because of this, the model turned out much larger than we originally had planned it to be (it was originally the size of the crate and ended up being about 2 inches larger on one side).
Here is a video of the final model:
Finally, we cleaned up our work areas (as we did throughout) and put away any unused material or garbage.
Cost analysis:
Plywood (Home depot ¼’’ x 4’ x 8’ sanded pine plywood) – $28.62 (This is a large piece of plywood that we could have cut up)
Wooden dowels (Home depot; individual hardwood dowels) – $1.24 x 5 = $6.20
Epoxy (Clear Gorilla Epoxy Home Depot) – $5.98
Wood Glue (Gorilla Wood Glue Home Depot) – $3.98
Set of bearings (Amazon 20ct from YAMASO) – $6.49
Sheet metal (for waterjet) (12×24 aluminum sheet Home Depot) – $14.63
Black vinyl sheet (Only used very small corner of large roll; rolls found for around $10 on Amazon) – $0.50
Tools (Lasercutter, vinyl cutter, waterjet, drills, dremels, sanders, bandsaw, etc.) – Since there is not a good way of estimating how much all of this would cost if you were trying to recreate it, we are going to base the cost on the relative time spent (3 weeks of work) to the OEDK access fee (for certain groups). This comes out to be about $300.
Labor – We spent a total (combined) time of 180 hours on this project (Yes, it actually took this long). At a rate of $15/hour this comes out to $2700. This is a bit of a higher rate than we normally consider but since most of the project ended up being done mostly with the laser cutter and CAD modelling (which we are both quite good at) we are giving more value to our labor.
Total cost: $3066.40 – Almost 98% of this cost is labor and tools so the actual cost of making this is quite arbitrary.
