Due to his arrogance and disobedience in battle, the mighty Norse god Thor was cast out of Asgard by Odin and sent to Earth. He lost his powers of his mighty hammer, Mjölnir, until it returned to him. As the hammer plunged down from the sky, it landed in a mystical arena of bizarre gadgets reminiscent of Xandar. That arena, which has become my second home over the past week, is called the OEDK.

As you can tell, I am quite the Marvel fan (pre-Endgame era). I joined forces once again with my teammate from a previous project, Shifan Liu, to work on this project of unbridled creativity. Once we were unleashed to design the contraption of our dreams, Shifan asked me, “Are you a Marvel fan?” A grin appeared on my face as I admitted to being one  (pre-Endgame era), and so the journey began.

We took off to Pinterest to brainstorm ideas about what to build. From Iron Man’s mask to Loki’s scepter, we looked at dozens of props from Marvel movies until we came across our beacon of expression: Thor’s hammer. Inspired by the scene in Thor: Ragnarok in which Thor’s sister, Hela, shatters his hammer into fragments, we decided to build a mechanical model of the broken hammer that move the broken pieces around, assembling and disassembling them. Even though Thor: Ragnarok was our inspiration, I could not come up with an OEDK-related pun with Ragnarok, so I picked the lesser movie Thor: Love and Thunder to make a pun out of and be the title of my blog post.

Instead of chronologically describing the design and manufacturing process, I will first present our final product and explain how it works, as that is the star of the show.

 

In the final product, a model of Thor’s broken hammer made out of wood is placed on a base containing all the gears and mechanism pieces. The hammer is tilted, giving it the impression that it is sunk into the ground. As the crank is rotated, one of the pieces moves apart from the other, which remains stationary. Continuing the rotation of the crank brings the moving piece back, thus brining rise to a translational path of the broken piece, with its velocity being a sinusoidal function of time. Rotating the rod simultaneously causes five white lightning bolts supported on rods attached to the master gear to rotate around the hammer while the broken piece is moving. Two of the rods supporting the bolts are attached to bearings on the gear, which allow the bolts to rotate in place. The complex mechanistic system of gears withing the base has multiple functions: to allow for the rotational motion of the lighting bolts, to allow for the translational motion of the broken hammer piece, and to support the center of the top of the base underneath the weight of the hammer, all while ensuring that no component of the system interacts with another in an undesirable way (like two components blocking each other’s movement). Three black vinyl stickers in the shape of different bolts were also placed on the top of the base, giving it a more aesthetic look. Finally, a name plate is attached to the moving piece of the hammer.

I will also outline the components of the mechanistic gear system because I will be referring to specific components with specific names (bolded) throughout the blog:

 

The master gear is the biggest moving component in the entire model. It controls the movement of all other components (except for the crank and the crank gear). It has a bearing in the middle, in which a rod that supports all the other moving components is placed. It also has 5 holes for bearings that hold the rods supporting the lightning bolts. There is also a little hole next to the central bearing that holds a turning rod that plays a major role in moving the other components.

 

 

The crank gear, as the name suggests, is connected to the crank. Its rotation causes the master gear to rotate, which subsequently causes everything else to move.

 

The bottom of the base has two central holes: one for a central rod and one for the bearing. The central rod is what holds the master gear and all components controlled by the master gear, and the bearing holds the rod holding the crank and the crank gear. There are also peripheral holes for rods that will support the periphery of the top of the base.

 

The periphery of the top of the base serves no mechanical function. It is only there to partly conceal the mechanistic system and make the model more aesthetic. The hollow circle is where the center of the top of the base is located, and the solid disk is where the crank is positioned.

 

 

The center of the top of the base supports the weight of both pieces of the hammer. A central hole allows for a rod to support it, and two linear holes allow for the broken hammer piece to be connected to a component in the mechanistic system by two rods. A little pocket hole is also placed underneath, whose function is described below.

 

The support structure is glued to the bottom of the center of the top of the base. It provides more structural support to the top with its inner disk and protruding rods that extend to hold as much of the top as possible. It was designed because we were concerned that the central rod and a washer would not be enough to support the weight of the top of the base and the hammer combined. The rods were also designed this way so that they would not come in contact with any other component like the rods moving the hammer of the pocket hole connected to the top of the base.

 

The cross is the crème-de-la-crème of our design. It has two perpendicular linear holes: one allows for the relative motion of the central rod, and the other allows for the motion of a turning rod. The turning rod, which is connected to the master gear, revolves in a circle. This turns rotational motion in the gear to translational motion in the cross. For this to work, the rotational motion of the cross has to be restricted such that only translational motion is allowed. This is achieved by the thick, rectangular end of the cross which enters the pocket hole. The pocket hole is there to restrict the cross’ rotational motion, hence achieving the desired translational motion. Two small holes in the cross hold the two rods that are attached to the moving hammer piece. As such, the translational motion of the cross brings about the translational motion of the hammer piece.

 

The rotational disk is connected to the central rod through the bearing and is located between the master gear and the cross. The turning rod goes through the hole next to the bearing. The rotational disk’s function is dual: to provide additional support to the turning rod and to allow the cross to rest on a surface without floating in midair (since the cross is not connected to the central rod).

 

Putting all these components together, along with the hammer, the crank, and all the stationary support rods, the end result is spectacular. I still cannot believe that we were able to accomplish this. It is not perfect (as I describe a flaw below), but it still looks amazing. I especially love this contraption because of the sheer number of man hours we put into this: 87 HOURS!!! I realize that I am psychologically biased to love this model despite its flaws, but I still believe that it is an impressive feat.

The story behind the design and manufacturing process is a rollercoaster of ups and downs. Our initial idea was to have the hammer break in pieces that would fly in 4 or 5 different directions, and lighting bolts surrounding the hammer would revolve around it in a circle, and this entire motion would be controlled by the rotation of a single crank connected to a gear that lies in a vertical plane.

Obviously, this was unfeasible if not impossible for us to make. For one, having a vertical crank gear connected to a horizontal master gear required bevel gears which were unavailable to us and unfeasible to manufacture. As such, we decided to keep both the master and the crank gears in the same horizontal plane. This made the base bulkier, but it was necessary. We also decided to simplify the breaking of the hammer into only two pieces, and only one would move in a line. Accomplishing this by itself proved to be extraordinarily difficult, so I am glad that we chose the simplification route.

The dimensions of the hammer were obtained by scaling an a real-life prop down to half its size. Proportional measurements of the prop were made by using a ruler on my laptop’s screen to measure the length of each side of the hammer, as well as the length of the handle. However, the hammer was tilted into the ground; as such, the lengths of some of the sides (specifically, those in contact with the ground) had to be calculated. After a session of intense trigonometry, the lengths were finally determined, and an Adobe Illustrator file was ready to be prepared.

The master and crank gears were very easy to design and manufacture, as all we had to do was pick our desired settings in Gear Generator, download the Illustrator file, and use the laser cutter on cardboard (for low and medium fidelity prototypes) and wood (for medium and high fidelity prototypes). Wood glue was used to glue two layers of gears together. I did not realized how important it was to clamp the gears from all directions while the glue was drying until it was too late, so I had to redo it. It worked out in the end.

The final mechanistic system was based on mechanical movement number 93 in 507 Mechanical Movements. However, that was not always the case. Our original mechanical movement was number 114. This also converted rotational to translational motion, but the linear speed of the hammer piece was constant in both directions, and the change in direction was abrupt. This posed many different problems for us. For one, the only way to make an Illustrator file out of it was to trace a screenshot of it, which resulted in jagged edges. Second, we were warned by one of the TAs that it was very hard to achieve, and we might need to look into other possibilities. He was right. As we constructed our low and medium fidelity prototypes, they were made out of cardboard, so there was no way to reliably test the mechanism. However, once we got to wood, we realized that it was not going to work. We had wasted all of our time on a fruitless mechanism, and we had to start from scratch (this is a big reason why we ended up spending 87 hours on this project). 

 

 

 

 

We went back to 507 Mechanical Movements to look at other ways to convert rotational to translational motion. We ultimately decided on number 93, known as the “Scottish Yoke”. We swiftly went to the laser cutter to make a piece and test out a new medium fidelity prototype. That was when we realized that we were going to have to come up with a much more complex mechanistic system than a disk spinning under a yoke.

A number of problems was glaring at us. For one, the central rod interferes with the path of the yoke. The central rod also had to support the rotational disk, the yoke, the center of the top of the base, and the hammer by itself because the lightning bolts spinning around the hammer rendered the center detached from the periphery. As such, the periphery could not play a role in supporting any of the other components. Having all of our moving components lie in horizontal planes posed a serious challenge in making the model durable and stable. We also needed to find a way to restrict the rotational motion of the yoke so that it only moves in a line.

To solve the first dilemma, we redesigned the yoke into what is now the cross. By adding another linear hole perpendicular to the first, we created a space for the central rod to move relative to the yoke without interfering with its path. To restrict its rotational motion, we also came up with the idea of adding a pocket hole to the top of the base. We would then extend one of the sides of the cross to fit in the pocket hole, which only allows for translational motion. Regarding the issues of durability and stability, we took a gamble and continued on with prototyping, as they can only be tested with the high fidelity prototype.

Another problem arose: the cross was basically floating in midair. This is where the rotational disk comes into play, as it acts as a platform supporting the cross. This, however, brings about another problem: friction. The bane of every ENGI 210 student’s existence, friction was a grand challenge for many to overcome. In our model, we anticipated a lot of friction between the surfaces of the hammer and the top of the base, as well as between the surfaces of the cross and the rotational disk. Even though we were also crossing our fingers that it would work out in the end, we did have a plan to sand all touching surfaces with grit that goes up to 800, as well as applying layers of oil and wax.

We needed to test out the mechanism of the cross before we started manufacturing our final prototype, so we laser cut all the necessary pieces and used hot glue to stick them together. We were so nervous as this was our second attempt at testing a mechanism. To our wonderful delight, it worked, and we felt a huge sense of accomplishment. Given that we had heavily modified the Scottish Yoke for our model, Shifan and I jokingly dubbed our mechanism “Mechanical Movement 508”, which is not included in 507 Mechanical Movements.

Now that we had a glimmer of hope that the final model would work, we commenced the manufacturing process of the high fidelity prototype 3 days before the project was due. After strategically waiting for the laser cutter to be unoccupied (by strategically, I mean going in the middle of the night when almost no one is there), we cut all our pieces from the Illustrator files. One detail that I admire about the model is the Norse patterns engraved on the sides of the hammer. They are a huge eye-catcher for anyone walking by.

 

I distinctly remember that first day, or I should say night, of post-processing. All artefacts of the manufacturing process had to be removed before we could begin the assembly of the model. I went to the OEDK at 5 PM to begin sanding with successive levels of grit. I especially sanded all the touching parts. I also applied wood glue and some intense clamping to fix all the three layers of the bottom of the base together (I used 6 clamps simultaneously). I then realized a heartbreaker when I got to the hammer: some of its sides had to be glued at angles. As such, I spent hours working on sanding down some of the edges to achieve the desired angles. There was a myriad of failed attempts, some of which were due to mistakes in accounting for the width of the wood (as the trigonometric calculations assumed 0 thickness of wood). I ended up with a huge pile of failed parts.

I was eventually able to achieve the angles, and I then went to spray paint the hammer with a metallic coat. By the time I was at this stage, it was 3 AM (I did take a dinner break in the middle). I live off campus, so I decided to just spend the rest of the night at the OEDK. I continued post-processing other parts until it was 6 AM. At that time, I decided to go back to Lovett and take a “nap” in the OC lounge.

The next day, I stained and waxed all the wood while Shifan prepared all the rods, which had to be measured, sanded, and post-processed. We initially wanted to stain the wood with a dark mahogany hue, but the only colors available were black, white, and blue, all of which were undesirable. We thus opted to use linseed oil instead. When most parts were ready, we decided to test out the mechanism by gluing the components with hot glue, as opposed to super glue for the final model. To simulate the weight of the hammer, we simply placed the hammer pieces on the center of the top of the base, without having an actual hammer on it. We only got to test the mechanism 2 nights before the due date, so we were quaking with anxiety. Deep breath in, deep breath out, and we spun the gear. IT WORKED. We were jumping around the basement of the OEDK, and I was yelling like a monkey. I could not help it. Our efforts were bearing fruit.

On the eve of the project deadline, we were to bring everything to a close. Both Shifan and I freed up the entire day to work on the project. Starting at 3 PM, we continued post-processing and assembling. I glued the pieces of the hammer together while Shifan started assembling the model with super glue. Washers were placed under the rotational disk and the support structure to provide additional support, but super glue was not quite effective with washers. Instead, to our surprise, we found hot glue to be more effective, so we went with that. We also wanted to attach bearings that would allow all the lightning bolts to spin around in their places, but only two were available, so we laser cut fillers for the rest of the holes. As such, only 2 of the 5 lightning bolts spin in place. To increase the stability, we put wood shards in the holes with the supporting rods, and that turned out to be extremely efficacious. The sanding and waxing were also astonishingly potent at reducing the friction, as the mechanistic system could be set into motion, and it would take it a significant amount of time to come to a stop.

In order to support the lightning bolts, we used a jigsaw hammer to cut grooves into the rods that would appropriately fit the lightning bolts. In the Illustrator file, we added teeth to the bolts that would fit in these grooves. Super glue was used to keep them in place.

We also utilized the plasma cutter and the vinyl cutter. The plasma cutter was used to make a part of the crank, and the vinyl cutter was used to make black lightning-shaped stickers that would be placed on the top of the base. For the metal part, we decided to incorporate it into the crank because incorporating it into the mechanistic system would add too much weight and jeopardize the stability of the model. The metal piece was post-processed with the surface grinder, 800 grit sandpaper, and spray paint.

 

 

 

 

 

 

We continued to work and assemble until we almost had the entire contraption assembled. One thing was left: attaching the moving piece of the hammer to the rods connected to the cross. When we did so, the results were… concerning… We had forgotten to account for the weight of the hammer, which was not a problem when it was made out of cardboard. As such, the weight pulled it down such that it was not flat with the top of the base. This caused it to interfere with the path of the lightning bolts, rendering the entire mechanism dysfunctional. It was 11:30 PM and we had to come up with a solution fast.

In hindsight, we could have added more pocket holes in sequence to support the cross more and prevent the cross and hammer from tilting. It was too late for that, though. We tried attaching supporting rods to a side of the hammer, and that rod would touch the base, effectively holding the hammer upwards. Not only did that ruin the aesthetic, but it also did not even work as the surfaces were too waxed for the super glue to be effective. After a long period of tense silence, I figured it out: we could simulate multiple pocket holes by simply gluing a stick of wood to the first pocket hole; this stick would then support the cross further and prevent it from tilting. It was not pretty, but it was an idea. After quick post-processing and holding the stick to the pocket hole for 10 min while waiting for the glue to dry, we were ELATED to see that it worked. The model was functional again. The moving side of the hammer does wobble a bit because the added stick does not provide the best support, but at least it moves.

As a finishing touch, we attached the rod of the hammer to its side. We covered it with multiple layers of heat resistant tape to give it a brown hue, and we drilled a hole in its top to insert an elastic band that was stained black to act like a leather strap. We also added our spray painted name plate to the side of the hammer. At 3:30 AM, we were finally free to leave the OEDK and exercise our various liberties. I spent the night at the OC lounge again, but I slept that night knowing that I had succeeded in such an ambitious endeavor. It was worth it in the end.

 

 

 

 

 

 

There was so much to unpack and so much to learn. This was my first time designing such an intricate contraption and not just building things according to a strict set of rules like a LEGO set. I was able to exercise my creativity to create something that, as Dr. Wettergreen puts it, “I am proud of”. That was my motivation for tirelessly working on it for 87 hours. It was also my first time applying trigonometry to a real-life model to find the dimensions of the sunken hammer. I also learned to adapt to a lot of problems that stemmed up along the way, from changing the mechanism to waiting for the laser cutter to using different stains to solving a last-minute dilemma. Engineering was truly an iterative process along the way. I also realized the importance of testing, as it acts as a traffic light that tells us whether we should progress forward or retrace our steps. Finally (and obviously), I enhanced my hard skills in laser cutting, wood working, and post-processing – skills that are undoubtedly useful in any prototyping adventure.

And now… the grand finale… MONEY TIMEEEEEEEE 🤑 (this one is going to be quite pricey)

• 1/4″ x 4″ x 4″ wood costs $5.75. 5 of these were used, amounting to $28.75.
• 1 roll of blue painters masking tape costs $2.17. An entire roll was used.
• 1 1’x2′ steel sheet costs $14. Approximately 1/10 was used, amounting to $1.40.
• 1 spray paint can costs $6.48. Approximately half was used, amounting to $3.24.
• A quart of Sunnyside linseed oil costs $10.29. Approximately a quarter was used, amounting to $2.57.
• A 4-pack of superglue costs $5.40. Two full packs (8 bottles) were used, amounting to $10.80.
• 1 vinyl sheet costs $1.49. 1 was used.
• 1 3/8″ x 48″ wooden dowel costs $1.26. Neglecting the dowels that were wasted in prototyping, 2 were used, amounting to $2.52.
• 1 Elmer’s wood glue costs $2.48. Approximately half a bottle was used, amounting to $1.24.
• 16 oz Feed-N-Wax wood polish costs $9.98. Approximately 1 oz was used, amounting to $0.62.
• 20 steel ball bearings cost $7.57. 6 were used, amounting to $2.27.
• The heat resistant tape and elastic band used are on the house.
• Labor costs $10/man hour (above minimum wage since I am not factoring in the cost of using any of the OEDK equipment). A whopping total of 87 man hours were spend, amounting to $870.

Andddddddd…………………. $933.61 is the final cost. It is Thor’s hammer, after all.

Here is a link to a very short slideshow outlining the various fidelity prototypes and manufacturing techniques used in the making of this model. Here is also a little blooper while we were filming:

 

brianbishara