I decided to revisit my Copper Tape project since I had a bit of trouble originally completing the structure and getting the circuit to work, as the copper tape is very difficult to get working in series due to limited voltage.
This time, I wanted to add in a bit of locomotion, adding in the canards and ailerons to the aircraft. To accomplish this, I would need to make use of Arduino and actual (thank gods) wires, building a more robust circuit and working with a more stable power source. Ideally, I wanted to find an Arduino PS2 joystick so I could add in some interactivity (i.e. all surfaces pitch up when the controls stick is pulled up, etc.), but I had to settle for having the control surfaces sweep back and forth. Since I wanted to add in space for servos and an Arduino board, I had to work with a bigger surface, prompting me to switch to the Laser Cutter and a 12″ x 24″ plywood board for the baseplate. As such, I had to upscale and modify my original paper template, sectioning off the control surfaces.
Build Process and Modification:
My first iteration was to get a basic feel for the circuit layout and size, so I used a 12″ x 24″ board as the base. It would later turn out that this board wasn’t large enough to accommodate the servos and Arduino Uno controller comfortably, so I would need to choose a bigger base for the later iteration. While laser cutting, the large board had warped slightly and this caused the laser to be out of focus at times due to an inconsistent z-distance. For example, note the minor offset cuts on the baseplate in the picture below:
The first issue was to attach the canards (forward winglets) to the baseplate via a servo. Since the canards were cranked at a certain angle, I had initially intended to tilt two servos, one for each canard, and rotate those independently. However, since space on the baseplate was limited, I decided to use one servo and link the two canards together via a bent paper clip. To prevent the paper clip from rotating in place from the servo’s control arm, I had to secure it in place with a bit of hot glue…
Okay, a lot of hot glue.
The first thing to get used to was working three servos in parallel. This usually would be a bit problematic with only two ground pins in the Uno board, so I had to hook them all up to the breadboard to get them working in parallel
For the prototype, I decided to keep it simple and simply attach the servos to the control surfaces (instead of using control horns and push rods as is the standard case with RC aircraft). After working with the Arduino IDE I was able to get the servos moving independently simultaneously:
Instead of doing something that mimicked walking, as seemed the case for most others, I decided to work with something more mechanical instead. Now, obviously wheels and propellers were out of the question, since the servo could only rotate 180 degrees at most. Instead, I happened upon the idea of a rowboat, and realized I could obtain a loop-like motion by chaining two servos together (https://www.youtube.com/watch?v=bpvC5AVfEVY for inspiration)
I wanted to use the base of a Grecian/Viking warship, with a large number of oars as propulsion. The ‘oars’ would be made using basic wooden dowels, and for lightness’s consideration I wanted to use cardboard/paper for the body. However, the main issue was the ‘slots’ for the oars to be sturdy enough to withstand the oar’s rotating motion, so I decided to stick with cardboard for the first prototype and hopefully laser-cut a housing for the second one.
The basic sketch/conception of the robot.
Build Process and Modification
There were a few things that came up on the prototype. The first was the fact that the servos had tiny extensions on the side to allow them to be screwed or bolted on external surfaces. This becomes problematic when attempting to lay the servo on its side, so I had to use some extra cardboard to elevate the servo so it stayed flat:
Padding at the bottom of the servo to allow it to lay flat
I also used thin-gauge wire to substitute in as a push rod/extension for the servo arms. The thin gauge allowed it to loop and slot into the servo arms nicely, but when placed under pressure and movement it also began to bend easily.
A good view of the arm extensions and oar mechanism. Note the bending in the extensions (green wire)
I attempted to rectify this by reinforcing the center areas with thicker-gauge steel wire, but its effect was limited. This are is probably the weakest part of the design, with the best alternative to use actual push rods found with servos.
The overall result was a bit lackluster (demo video found here: IMG_5476). The primary issues were that the servos were not properly secured to the cardboard, the extensions bent too much and the overall robot was a tad too heavy for the wooden dowels to move without sufficient leverage. I will likely transition to using paper and glue for the next iteration to help secure the servos more, as well as reduce weight. I may also have to acquire more servos and push rods in order to complete this, so until I find push rods the second iteration may have to be held back a bit.
(This Post is currently a WIP, documenting my progress on the project so far)
Drawstring Bag (in-class)
Embroidered Design (in-class)
Motivation and Initial Design
I went with a plushie design for this assignment, primarily as a present to my younger sister. Although I initially decided to go with a Pikachu plushie design, after some consultation with my sister I decided instead to opt for the fox plushie found on Sew Desu Ne. This saved me a bit of trouble of having to modify the Eevee plushie to make it suit Pikachu’s design.
Build Process and Modification
Hoo boy, where to start? Right from the get-go, there were a few difficulties. The first was that plushies are typically made with minky or fleece fabrics rather than traditional felt fabrics (this allows them to be a little more fluffy and ‘squished’). To do this I had to make a stop by Joann’s Fabrics, which severely limited the time I got to work on the project. After picking a few fleece fabrics, I began work by first embroidering the face design onto the fabric before cutting it out. It’s here where most of the issues began:
The first part came into play while the eyes were being embroidered. I noticed that the position was slightly off, and then I realized that the fleece fabric was rather thick (especially compared to felt) and was being held down by the foot of the machine, causing positioning issues.
Note the eye position inconsistency between left and right
This also left us with a rather incomplete embroidery. I tried to compensate for this by slightly lifting the foot up during sewing, enough to flatten the fabric without pinning it to the board. With this, I tried again.
The second try worked a bit better, with the embroidery beginning to take shape. However, A few more issues became known at this point:
The most major one occurred when I began to embroider the border of the muzzle fabric to the face fabric. Although it started out very well (especially compared to my previous attempt), I noticed that the machine was having a lot of threading issues, suddenly unthreading from the needle or having the thread catch on an unknown surface. After pausing the embroider I noticed that the back of the embroider had gotten progressively messier, and worse still, some parts of the thread had failed to go through the fabric. Instead, it began knotting in on itself and gnarling its way up the machine. Concerned about jamming, I canceled the embroidery and spent the next few minutes untangling the thread from the machine. I gave it a quick test afterward to determine if it still worked, and once done I realized I had used up far too much time already. With a test the next day, I decided to retire for the night and go back to review. I think when I resume this I might have to do so with a different fabric since fleece has proven too thick and not stiff enough.
I like to work on plastic models as one of my hobbies. I started around junior year of high school, working on plastic models of Gundam Mechs (often known as ‘Gunpla’ – a portmanteau of ‘Gundam’ and ‘Plastic Model’). While their size is often small (the most common ones stand at a height of ~4-5 inches), they are rather complex:
An example of a typical Gundam inner skeleton (~30 parts)
This small size and complexity can make them a challenge to work with at times, and often require a steady hand as well as precision cutting and finishing to give the parts a decent fit. To do this I often touch up parts with an X-Acto knife to eliminate excess plastic after cutting them out of the sprue trees.
For one of my recent models, I ended up pruning a little bit too much plastic and eliminating a connection point by accident:
The model in question – note the hole in the upper arm section
Since I would not be able to order a replacement part (at least not easily or cheaply) I decided to use the 3D resin printer to patch up my mistake.
I started by taking the damaged part into Fusion 360 and attempting to reconstruct it, but it soon became apparent that the part in question was far more complex internally than expected:
The damaged part (seen in white). Note the grooves and small details
Without a reasonable way to measure the spaces between the lines, coupled with the part’s large amount of curvature, I decided against reconstructing the entire piece, as I could only make relatively inaccurate estimates using the calipers provided in the Fab Lab. Instead, I sought to create a small part to simply patch the hole for me. This led me to this relatively simple piece:
Kind of underwhelming, huh?
Although the design was simple, it was also small (dimensions are 1.5mm x 1mm x 2.5mm). Nonetheless, I was about to find that even this small part would give me quite some trouble.
Build Process and Modification:
During the first build of the patch part (referred to as a ‘nub’ from this point onwards), I wanted to use white resin since that was also the color of the part I was patching. However, I ran into an issue with the curing process for the resin, which according to Emile had been acting unusually recently:
Nothing solid came out – just globs of sticky resin…
Under Emile’s insistence, I decided to switch to the clear resin, which was much more consistent in quality and curing times. The result was an extremely small nub:
A nub being sanded using tweezers. The clear color makes it hard to see.
Yeah. It’s really small. After the initial prototype I realized a few things:
There are small inaccuracies in the printing process, even with resin, which made most nubs come out slightly oversized. Through testing, I estimated the error to be about 0.1mm off, which is fine for most prints but in my case made fitting difficult.
The nub’s small size and clear color made it really easy to lose track of.
I worked around the first problem by sanding down the nubs until they made a decent fit, using tweezers and 200-grit sandpaper. The second problem was more difficult to solve, given both the size of the piece and its material were not easily adjustable, so I decided to make several copies of the nubs as backups (I created 9 backup nubs, 6 of which were lost). Luckily given the piece’s small size and sole reliance on z-dimension for print times, (I flipped the nubs on their sides to make the z-dimension even smaller) each print batch (the single-nub prototype batch and 9-nub production batch) each took only 25 minutes.
Testing the nub’s fit. Note the darker piece poking out of the white one.
After sanding and testing the nub’s fit, I took the mostly-completed nub and coated it and the damaged piece using contact cement. It took a bit of time and some prodding with a paper clip to get them to fit in snugly (given the scale we were working at), but ultimately the patch was complete and I set it aside for a few hours for curing.
A pretty good fit, if I do say so myself.
Even though this project was fairly fun, (and helped me out with my hobby, after all) I decided that my next projects should probably be on a larger scale. I personally would like to reconstruct the entire model with CAD, but this time using my own designs and dimensions instead of trying to estimate another designer’s work. Other than that, I would definitely add improvements in considering the printing errors and manufacturing tolerances, as well as developing better ways to keep track of each component.
Let me first preface this by stating that the build is not 100% complete. And yes, my overambition was absolutely to blame. I wanted to make a paper model of one of my favorite aircraft – the ADF-01 Falken from the Ace Combat game series (shown below)
I first went about this by going into Inkscape and performing a bitmap image trace, allowing me to use the Silhouette Cutter to make precise cuts. I then traced out several cross-sections that I could use to add stucture to the paper model.
By this point I should have realized that 1 – the effort alone was already taking me far longer than expected, and 2 – the complexity meant that I would have to prototype at least 2 iterations for a comprehensive aircraft – a process I simply did not have time for. Nonetheless, I made the silhouette cuts and was ready to start wiring:
Build Process and Modification:
First things first – the Silhouette cuts were not perfect. It appeared the blade had been dulled from overuse so some parts were still attached to the excess. After some moderate difficulty removing the parts via X-Acto knife, I then started making the circuit on gray baseplate for the aircraft.
The first iteration of the circuit ran into some issues – Due to space constraints I decided to run the engine lights through a series circuit while the cockpit light was parallel to it. I ran into an issue where the engines had significantly higher resistance than the cockpit light, thus not activating when the switch was activated. After some calculations I realized I would need a second battery to run all lights effectively, something I simply did not have the space for. It was then back to the drawing board.
The second iteration went better. To save on space I integrated the battery into the switch itself, freeing up space for me to run a full 3-way parallel circuit. However, the setup was still relatively unstable, as the switch became heavier and bloated
Nonetheless, I decided it was sufficient enough for the time constraints and then moved on to defining the structure:
As of this moment, I’d say I’m about 80% complete. I’m not completely satisfied with this outcome and wish I had more time. Improvements I would add would be modifications to the template design to allow for smoother interlocking as well as a more robust circuit design backed by solid calculations.
While I initially decided to use the design I had set aside from last week (seen below), I quickly realized that converting it to a vinyl sticker would take 7 layers, all of which had to be different shade of the same color since I intended for it to be monochromatic. Since the lab did not have access to 7 layers of monochromatic vinyl, I decided to shelve the idea again.
Instead, I decided to use another patch I had previously designed for one of my engineering teams a year prior – a parody of the famous Lockheed Martin’s Skunkworks experimental team, modified with the UIUC mascot of a squirrel:
By simplifying the brown parts to dark orange, and dark brown/black parts to dark blue, I could reduce the required layers to 4, allowing for a simplified vinyl sticker.
Build Process and Modification
The build started off relatively normal, with me placing four vinyl patches in 2×2 pattern consisting of dark blue, dark orange, white, and salmon.
After the cutting began, however, I realized that a 4″x4″ was a rather small surface to cut on, especially given the details of my logo. However, instead of stopping the cut and moving to a bigger canvass, I decided to continue and use whatever practice I could. The result was a relatively small logo, although the details turned out better than I expected.
When peeling off the excess, I noticed that small details and patches – particularly around the letters, would sometimes stick to the excess as it was lifted up. I decided to proceed by peeling the excess off first, and then cut off the necessary bits stuck to it using an X-ACTO knife and a flathead screwdriver to nudge and modify the pieces back into place:
This process of reattaching and modifying the small details extended my work time by approximately two hours, yet once I was done the sticker looked relatively whole:
For this project, I decided to design a custom icon. I wanted to display my specialty, which was the ability to fuse two elements from different franchises and backgrounds into one design. I feel there is often synergy to be found in the most unusual pairs, and I’m always looking for common themes or motifs found in my favorite franchises.
The first design was made with Adobe Illustrator, as a reference to both Nier: Automata and Mobile Suit Gundam: Iron-Blooded Orphans, which were both franchises that affected me deeply. However, the different tones of grey made the icon difficult to implement as a laser-cut piece, so I shelved it for use as a vinyl sticker.
The second icon was created as a fusion of the video game franchises Dishonored and Xenoblade Chronicles. I took the iconic Outsider’s Mark and fused it with Shulk’s Monado icon (found in Super Smash Brothers) as a statement of how art and technology can blend together in interesting ways
Part II – The Nametag
For the name tag itself, I felt it would be appropriate to add more into the background of the design. Going along with the ‘Art and Technology’ theme, I decided to create the background using clean, rasterized shapes (based on the Monado’s circular designs). This creates a sense of modernity and cleanliness, while the more stylized icon and font would be overlaid on top, representing calligraphy and art built upon a technological base.
Build Process and Modification
Initially, I planned to create the tag using two plywood layers, using a background with rasterized designs etched into it, while the second layer would be the icon and font which would ‘pop out’ at the observer. I ran into an issue when cutting this second layer out, as the font size caused the laser to focus its heat on a relatively small area, causing the wood to burn from overexposure.
Since I did not want to redo the name tag to work with a larger font, I was instead advised by one of the TAs (Emily) to work with a clear acrylic rasterization and then fill the indents with paint, creating a finish on the piece. I decided to test this method, and when I saw the result, I decided to apply it to the background as well
The result was a very clean-looking name tag created from 2 1/8″ pieces of clear acrylic. Although I initially intended to fill the etchings with paint, I decided I was already quite pleased with the result and left it as is. To finish I had to seal the edges of both acrylic pieces with acrylic glue (a material that was a tad difficult to work with), and as of the time of this post, the tag still needs to wait a few hours to fully bond and seal. Additionally I intend to polish and sand down the edges to get rid of excess glue, but otherwise the tag is complete.
‘Testing, testing. Is this thing on? Can you see this? Sweet, good to know some things are still working.’
‘The name’s David, David He. Probably one of the most generic Asian-American names you’ll ever hear, and trust me, I’m don’t exactly subvert anyone’s expectations when it comes to that. Maybe that’s why the guys up top were so willing to throw me into the frontier up here.’