As a final project, I want to build a machine similar to the DVD checkout machines but for larger items. The motivation behind this is to have a machine that will automate the checkout process of items at my dorm.

This is a first prototype and it is my first time building something with mechanical components in it. While most of the key parts are there, the machine is far from useable in practice because of security issues. The most important missing part is a locking mechanism to prevent the door from opening while the machine is working. Also, the parts should be made from metal to make them more secure against misuse (people kicking the machine to get what they want, see https://www.youtube.com/watch?v=VBXxmISAMA8).

I designed the outer shell to be eptagonal so that it aligns with the compartment dividers making the other compartments inaccessible when one compartment is in front. There is an opening in the front to remove the items and a side panel where the motor and the rfid scanner is placed.

The machine is designed to be laser cut from 1/8 inch acrylic. I used a black acrylic for the sides of the outer shell and a transparent acrylic for the rest of the parts. I wanted to make the top transparent so that enough light gets into each compartment without the need of an additional lighting mechanism.

For the electronics, I used a stepper motor I had already bought from ebay. Moreover, I used an rfid card reader that has the mifare chip rc522. This communicates via spi with the microcontroller. I milled a board for power supply and a board to control the rfid card reader and the motor.

The most expensive parts were the acrylic sheets for the laser cutter. I used a 24x36 inch transparent acrylic sheet and a black 12x24 inch. I bought both of them from amazon.

As a president of the Westgate dorm, one issue that we face many times is how to distribute items among the residents and control who has what. We have many items like books, movies, video games, board games and so on, that we want to make them freely available to the residents to enjoy. However, we always get stuck in logistics since there is not an easy way to know who has what.

If we leave it freely available we risk falling into the tragedy of the commons situation where people selfishly take items they like and never return them. The way we currently deal with the issue is we give items to the front desk and people can check them out but this whole process requires a lot of effort and human labor. The ideal solution would be to have an automated system where people can check out the items they want anytime.

Motivated by this problem, I would like to create a machine that allows people to check out any of the available items automatically and keeps track of all information regarding who has what item.

One easy way to implement this system is to have a circular tray with items that are divided into different compartments and a servo motor that brings the appropriate compartment in front for pick up. There can be many trays stacked on top of each other using the same servo motor but different pick up locations.

For this week's assignment, I wanted to create something that would have been really hard to construct by hand-cutting the carton so that the laser cutter can show its best. My first thought was a sphere since it is something that requires high accuracy and precision (it's really hard for me to draw a nice looking circle by hand so I would imagine a sphere would be almost impossible).

I thought about how to make a sphere using 2d surfaces and came up with quite a few solutions but were not really satisfying since I wanted my sphere to have an outer surface so that I can draw on it whatever I want. For that reason, I chose to create something that can be made from 2d parts and approximates well the sphere, a truncated icosahedron.

The only issue I had was that some pieces were not cut all the way through. I looked at the carton and noticed that this was probably due to some slight curvature which was causing the laser beam to slightly get off focus at some points. However, I thought that I must have done something incorrectly since it is didn't seem that it should be so sensitive to such slight height changes.

Looking at the suggested settings for corrugated carton again, I realized that I should use 2 passes when printing and that did it. Next time I tried, I added an extra pass when it was finished cutting and then all pieces were perfectly cut. When I lifted up the sheet of carton everything stayed down! This was amazing.

I downloaded a shapefile that has the longitude and latitude of all the boundaries of all continents and did my math to project each path to each polygon and rotate them appropriately. The process was a bit cumbersome and buggy but in the end it worked out well.

I created an SVG with the projected paths and input them into the vinyl cutter using the fab modules that supports svg files. It cut all the paths with very high precision and then I had stickers for all the pieces I needed. At that point, I was left with a huge puzzle trying to see where everything fits, as well as a crafts project trying to put together all the different pieces and the many islands. But in the end, it was all worth it plus I improved my geography knowledge :)

For this week's assignment, we were asked to create a programmer for avr microcontrollers. We were given some options for different board designs and we had to print our own PCB using the milling technique. Although the designs that didn't require soldering a miniUSB component appeared pretty cool at first sight, I decided to go with the standard board traces that were provided in the lecture. The main reason for that was that I had already seen the same design in the lab tutorial and I was well aware of the challenges for that particular board. Moreover, it seems that the flexibility of having a USB cable might come in handy when the time comes to program a new board.

The milling process was quite straightforward. I uploaded the design in the fab modules and imported the configuration for the traces and the interior. For the traces I used the 1/64'' end mill while for the interior the 1/32'' to cut completely through the board.

To test that the board was soldered correctly, I looked at it really carefully to locate any shorts and then I tested all the connections with the multimeter. I tried to program the board but the other programmer was not detecting my board properly. Reading through the instructions more carefully I realized that my board should be connected to the USB to be able to program since the programmer doesn't power the board directly.

After plugging in the board to the usb to power it up, I followed the "make hex, sudo make fuse, sudo make program" approach and the board was programed properly. Luckily, I didn't run into any other issues while programming. A resource I found very useful for all the programming steps and debugging the board is the tutorial on assembling and programming the fabISP which explains the whole process.

For the scanning part of this week assignment, I decided to 3d-scan a small rubber duck. It is a nice curvy universally recognizable object and thought it would be cool if I could manage to make a 3d model of it.

I used the Sense 3d scanner which gave me a quite accurate representation after several attempts. The Sense scanner is quite unstable and many times confuses the background and the foreground. It also depends a lot on the lightning of the object which made it quite tricky to get it to work correctly.

After all the photos are taken they are processed so that the background is separated from the foreground. The goal is to reconstruct the surface using only this information.

My idea is to use Solidworks where I can start from a cylinder and subtract solid parts so that the projections in each plane is the same as computed. This approach seemed a bit complicated since I would have to parse each projection as a curve and import it at Solidworks. Another approach would be to write my own code and work on each z-plane separately to reconstruct the polygons that match the projections and then obtain the 3d solid. Unfortunately due to time constraints I was not able to complete my 3d scanner this week but I am really interested to continue this idea and see what results I can achieve.

After having all the important points of the knot, I imported them to Solidworks using the create curve from XYZ option. To create a solid from a curve and give it a rope feeling I used 3 circles and swept them along the curve.

I printed my designs using the Makerbot printer. Printing was quite fast since the knots are not very tall in the z-direction and don't require a lot of material to print. My only issue was with the support material. It was quite hard to break off and the real parts of the knot were quite thin and fragile which made them break at some points as I was cleaning them up and had to glue them together. Here is the final result including the famous "Trucker's Hitch"!

For this week's assignment, we had to make something big. I thought about making a custom furniture that fits my needs and I decided to make a rotating bookcase. It is a very useful piece of furniture for places like the corner of my desk that is hard to reach. It will make everything on my desk much more accessible!

For this task, I wanted to use a bearing that will make my bookcase rotate. I searched online and I found one called "lazy suzan bearing" that did exactly what I wanted. It was available at my local home depot so I drove there and bought it. I also bought some pieces of plywood to make my bookcase look nicer and be less rough.

With my plywood board destroyed, I decided to use my OSB board instead. For that however I had to update the design with the thickness of the new board. Luckily, with the help of David, I was able to do that very quickly and send my final design for cutting.

This time everything run smoothly and I got all my pieces cut properly. The whole process used 5 passes over all. The first two passes were used for cutting the interior (one rough and one for finer details). The next two passes were used for cutting the exterior (one rough and one for finer details). However the rough pass the second time didn't go all the way through. This was to keep all the pieces together and so that the pump pressure stays high during the whole time. The final pass is rough cut of the outline that is used to completely separate pieces from the rest of the board. With all the pieces cut, it was time to assemble everything.

For this week's assignment, we had to redraw the hello world board to add a button and an led. I didn't want to use those surface mount buttons in the pcb as they are quite small and hard to press! So instead I decided to use the buttons from an SNES game controller I had lying around since those buttons are designed to be very convenient.

The next natural thought that came to my mind was why not make an actual game controller since I am already using those buttons anyway. All I need is a TFT display and I am all set. I had already purchased a pretty cheap 2.2'' color display from eBay which would be ideal for this project.

Once I was done with the design in eagle, I milled a board in the Rolland Modela and stuffed it using all the components. The process was very similar to week 3. The parts that were mostly different was desoldering the shift register and finding a way to connect the screen to the board.

To desolder the shift register I used the heat gun where I pointed to the SNES board and slowly pushed the part to come off. I made sure that the orientation matched my design and I soldered the part on my board.

For connecting the screen, I had to remove the old connector that it had and solder a new header that is parallel to the board. I also mounted the matching part in my board. The part for the board was surface mounted while the part for the screen was dipped in.

For this week's assignment, we had to make a mold out of silicon rubber (oomoo) and then cast an object that is either plastic or cement. To make the mold we had to pour the oomoo in a machinable wax that had to be machined to create the two positive parts of the item we wanted.

For this task, I decided to create a replica of the Phaistos Disc which is disc containing hieroglyph scripts on both sides that are possibly written in Linear A. The disc is from the Minoan palace of Phaistos on the Greek island of Crete dating to the middle or late Minoan Bronze Age (2nd millennium BC). It is still one of the greatest challenges to decipher its meaning and while many decipherment claims have been made, we don't know for sure its true meaning.

Once the modela finished milling the wax it was time to pour the oomoo inside. I did some preprocessing steps to get the shape exactly as I wanted it to be. I had to clean some of the cuts since they were very thin and wax was trapped in there although it was cut. Moreover, I wanted to roughen the edges of my disc so that it looks closer to the original that is worn out, so I removed some of the wax manually to give this feeling.

To create the oomoo mix, I mixed the part A and B by volume. I used 2 small cups since I didn't need a lot of material and then I matched their heights. I stirred carefully not to create any bubbles and then I poured the oomoo in the wax. I added some extra oomoo even after it reached the top just to be sure and that ended up making the two parts of my mold connected. This turned out to be quite useful since I could just fold the two pieces to cast the materials, so I decided not to cut it.

Here is the final result after painting with nail polish to give it an antique feeling and making it a small necklace. Turns out that the leftover plastic from the hole I made to pour the plastic was quite useful as a hanger!

For this week's assignment, we had to program the board from week 8 using different methods. I used the libraries from high-low tech to add the attiny boards as targets in the arduino IDE.

Compilation and uploading was quite straightforward once I had everything installed. One thing that seemed a bit wasteful is that the arduino added its library which has a large memory footprint (~2KB) which is almost half of the available memory on the attiny and doesn't leave much space for complex programs.

For this reason, I also experimented using the avr-gcc and upload using the avrdude. The arduino IDE had already installed these software so I wanted to use them directly. I had a small issue with avrdude requiring a configuration file which was in a different folder, but I ended up looking at the commands that arduino uses and worked on adjesting them for my needs.

For this week's assignment, we had to make something using composite materials. The idea is that combining two materials, natural fibers (burlap) and epoxy resin in this case, we can achieve very good structural properties. I wanted to take advantage of this fact and make a flat surface that needs to be sturdy and robust. So my idea for this week was to make a small snowsled board that you can hop on and slide in the snow.

I created my design in solidworks. I drew some guiding curves and then asked it to do a surface loft to fit a smooth surface in my curves. Once I was satisfied with the result I went on to machine the surface onto the blue foam.

With my foamy mold in place I was ready to cast my composite. The steps I made were the following: I first waxed the surface and sprayed it with mold release. This was so that I don't need to use any layers of plastic sheets between my composite and the mold.

Then, I cut 7 layers of burlap in approximately the right size and spread epoxy on them one by one. I had to cut two small holes on each side to release the pressure and make the burlap take the double curvature. Finally, I placed the 7 layers of burlap on top of my mold, together with a breather and several layers of cotton to absorb the extra resin and put it in the vaccum bag. I was careful not to put the amount of resin necessary to cover all the parts that I am interested in.

In the interest of time, we placed 3 different projects all together in one bag at the same time with the help of David (the TA). We left it for 4 hours and then came back to pick up our composites!

For this week's assignment, I wanted to play around with step response and create a capacitative touch keypad. It seems exciting that you can create a keypad with so high sensitivity using only a resistor and copper.

To get used to the idea, I saw Neil's examples and worked through previous year's assignments. I tried it out by building a general ftdi board and connecting two pads of aluminium foil and measure its capacitance.

I programmed the board to send through serial every quarter of a second the state of the two pads. The state is just a number that shows if the button is pressed. This is done by comparing the value of the analog measurement to a set threshold. In the video, you can see it responding to my button presses.

The aluminium foil was taped to a plastic sheet to minimize interference. The cool thing is that it even detects presses on the other side of the plastic sheet.

For this week we had connect our microcontrollers with an output device. Since I am already familiar, with motors and lcds (see week 4 and week 8), I wanted to try something different. I thought about making a persistence of vision display using a led array that will be moved fast by hand to give the illusion of a 2d screen displaying a message.

To make this, I looked up the LED drivers from Texas Instruments. I really liked two of their parts. One is the TLC5916 which is basically a shift register with constant current output that is able to drive up to 8 normal leds. And the other part is a TLC5947, which is the same for RGB leds. It has an internal timer and does the PWM automatically to get any color you want. I ordered them to try them out. The plan is to use them together with the accelerometer to know the direction and speed of movement so that I can adjust what is printed to the leds.

Playing around with the components, I designed my final board with the leds. I included an accelerometer and a voltage regulator at 3.3V for it to work and the led driver together with the attiny44.

This was the first time trying the PCBdesigner so I was excited to see that all the parts were the right dimensions and everything fit perfectly.

For this week, I wanted to make my board communicate with the computer via USB. I wanted to do this ever since I attended the tutorial session on V-USB.

Since attiny44 doesn't have hardware usb support, the idea was to use the V-USB library that implements it in software and talk to the computer the same way the FabISP does.

For this week, I milled the same board as the FabISP since it has all the necessary parts for USB communication including the 3.3v zener diodes and the 20MHz crystal.

I programmed my board using one of the USB example codes that talks HID to the computer and simulates a mouse and a keyboard. The HID protocol seemed very interesting because it is already supported by all computers and doesn't need any special drivers.

One way to pass data to the computer is to simulate a keyboard and type the appropriate messages to the screen everytime. In general I spent a lot of the time this week learning how usb works and trying to implement that using my attiny. I learned a lot and was surprised to see that I got it working so nicely. Since this weeks purpose was to have at least two attiny's communicate with each other, I thought that a good way is to communicate via CAPS LOCK NUM LOCK and SCROLL LOCK presses since those are send to all the connected keyboards. I tested the communication by using my keyboard and sending messages to the board by pressing CAPS LOCK but I didn't use more than one boards communicating because of time. I'll try to mill a board in the next couple of days and get my experiment working!