Final Project Progress

Fall 2017 | By Casey Evans

“The best way to do things is to do them.”

Assignment

Group Assignment: None.

Individual Assignment: Complete your final project, tracking your progress: what tasks have been completed, and what tasks remain? what has worked? what hasn't? what questions need to be resolved? what will happen when? what have you learned?

My plan is to create a puzzle box. That is, one of those boxes that you have to solve the puzzle of the box in order to open. I think I will use the different aspects of the course to create a series of puzzles that contribute to opening the box and giving the box features that warrant taking the time to figure out how to open it. I selected this idea because it sounds manageable and I've always liked clever puzzles. And boxes. Boxes are great. I also think this box will be useful either for myself or as a gift. I really want everything I make in this class to be useful in some way. That's just who I am. Very practical and not wanting to waste things. I wish I could be a touch more creative but I'm pretty slow in that regard. Maybe this class will help me to find avenues where I am more expressive. We shall see.

This is 100% not what I did. But more on that below.

Progress Check 1

10/1/17 So, I've been pretty caught up getting adjusted to MIT and intimidated by how smart and driven everyone is. I feel like a slacker amongst these gods. That was the point in coming here to an extent but it's more "extra" than I expected. Anyways, my first steps towards my final project was the planning and CAD work I did week one, which you can see on the Week 1 page. I think my next step will be to start thinking more about how to combine the elements of the whole course into the one project, especially so that I may be able to start tying my weekly projects into my final project. I'm all about efficient time management. I know quite a bit about embedded systems but I'm not sure what I'd want my box to do. Lock? Tell time? Pandora's box a decision like an 8 ball or something? I think there's a networking aspect that's supposed to be included. Maybe it could be like a music box radio?

10/4/17 Preparing: So the "make something big" might be a good time for me to make the box itself. I'm not sure if it's really big enough to count--but I could make it so. I could make it a box to be a metronome and tuner and store like 3-4 "sample songs" that I can play for reference while I'm practicing. Then I could just carry the box with me when I go places. Perhaps it could even be a music stand of sorts. Except that I already have plenty of those. But it could just be that the lid can posture up and have a lip on the outside. That'd be pretty nifty. Let's see, inputs and outputs. Well outputs could be sound in an audio jack (no speakers!). Input could be buttons to select metronome/tuner/reference song. Sweet. For networking then I could put things in a certain folder online that the box can grab from? I'm not really sure how networking works yet.

Progress Check 2

10/10/17 I decided to add a buzzer to this week's PCB to get a feel for how they work in case I want to actually make a metronome. Then I could use this board. I decided to also include a switch to turn on and off sound and still have an LED blink in case I wanted to test in a populated area. It's a feature that's on commercial metronomes as well. I realize now though that I didn't create a very effective design. I should have had both the current limiting resistor and the LED before the switch but I only have the LED before the switch and then two separate resistance paths. It's an artifact from thinking that the switch would just create a short, not actually flip between two paths. I had thought I would need to keep the resistance in the branch high enough to be non-negligible so that the buzzer (an 8 ohm resistor) wouldn't go off when the switch was shorted. As a result I had paired the current limiting resistor with the buzzer rather than the LED. Oh well. I may be able to fix that by getting creative with the board...or, looking at it again I should probably just suck it up and redraw that corner of the board. I'd also intended to make the chip in the shape of a music note but let that fall through because it required more effort than I thought I had time for back when I hadn't yet gone past EDS' open hours. In retrospect, it's also a bad idea because it would use quite a bit of copper that would be unused (mostly in the stem of the note). So a square is great. Maybe a cool shape like Niki did, but squares are also great and modifying in Paint is easy and I don't need to prove anything there.

10/11/17 I decided to redo the schematic since it would be easy enough and well worth the work. So you can see more on that on my Week 5 page.

Progress Check 3

10/16/17 I'm trying to add a secret compartment to the bookcase to keep with the "secret agent" theme. I'm also going to try to get a draft of the box up. The only problem is that the box I'd imagined would use normal woodworking rather than computer aided cutting (at least as a 2D process). As a result I may modify the final project to just involve my bookcase. I'm also considering switching to some kind of design assistance thing like a backlit table or something.

10/17/17 Hmm. I couldn't think of anything in time to get it on the sheet this time. I think I'd want nicer wood anyways. I should stick with a box since that's what I've been planning. But the ShopBot won't contribute. Perhaps I will talk with Gavin or another TA about the actual specifications for the final project. My roommates have spoken: I will be making a lamp. It will change color according to our TV remote (assuming I can get my hands on a logic analyzer) and respond to touch using a vinyl cut circuit with sensors down the length of the lamp stand. One of their suggestions is to make a setting that cycles through colors, like "party mode." We'll see.

10/18/17 Just to help out for the IR stuff - here's a link to my repo from my microcontroller class in undergrad. It's certainly not the best but it's in a very similar style to how I did this class so--here's the link. It's a private repo at the moment so I'd have to ask my old professor if I could make it public. They keep the projects the same throughout the years so it gives away possible solutions.

Progress Check 4

10/31/17 Happy Halloween! So on the lamp--I looked up some LEDs that are battery operated online--not really sure how to integrate them with a microprocessor, I've only ever done little lights before with clearly defined unpackaged leads. But I definitely want a quality light to fill the room. Also thinking laser etching on the circle boards that make up the stand (I have sketches). I'm thinking of three size sequence of circles (3" 6" 9" 6" 3" and so on). The top would have an Atlas holding up the lightbulb. Or maybe a cameraman with a light slung on his/her shoulder. It's hard to think of things that can be created with a 2.5D process. Maybe if Atlas is wearing a toga it'd be easier, but then I'd have to learn how to do mesh editing in a semester, which I certainly don't think I am capable of doing without failing in some other aspect of my life. I could also just make a lampstand from Zelda with a lantern on it. The Zelda lanterns are nice and symmetrical. I can still laser etch (raster) the sides. Except it would be kind of short. I did find a rupee STL and a half rupee STL! I could do that instead for now in molding and casting and try to find something else to incorporate into the lamp. Maybe I could make a Sisyphus rolling a stone up a mountain. I found some other pictures of art deco lamps I could try to replicate. We'll see.

Art deco lamps:

11/8/17 Um so clearly the Atlas men didn't work out. I'm thinking plate casting may be a good idea after all. One plate per roommate and there's three of us and they can decorate the top, though I was interested in demonstrating the reusability of the molds. Things to consider thinking about. Also I did some research on lights to use both battery powered and corded. I'm thinking battery may be best but brightness is huge so I'll keep looking into that. Color changing I found in a toilet seat of all things. LED may be the most promising. I could do a trail of LEDs up the stem of the lamp that are adjustable (make the lamp fun to play with) either using switches or sensors. Maybe if you get close they "shoo away" to the other side of the lamp. Just throwing around ideas still.

Progress Check 5

11/14/17 Shirin told me a plain lamp is not complicated enough. She suggested an app to control the lamp via bluetooth. I think that sounds awesome! Also ideas for the design: beaver balancing a ball/lamp on its nose, wood tree with a drystone beaver at the bottom, temple of time with an Ocarina. I should also do the week breakout but I need to finish this week's assignment first (machine building).

11/20/17 Breakdown:

• Introduction and Computer-Aided Design: This will be inherent in the design.
• Project Management and Computer-controlled Cutting: Cardboard flexure around the bulb. Vinyl cut built in circuitry for microcontroller battery.
• Electronics Production: I will be printing my microcontroller board.
• 3D Scanning and Printing: 3D printed battery holster.
• Electronics Design: On off switch. I will be designing my board.
• Computer-Controlled Machining: The podium for the bulb will be machined.
• Embedded Programming: The microcontroller will be programmed.
• Molding and Casting:
• Output Devices: Lighting (enough for a room).
• Machine Design: I don't think this is needed, but I should ask.
• Input Devices: Bluetooth reciever. Slide switch for master control.
• Interface and Application Programming: Application interface for bluetooth phone control.
• Networking and Communications: Bluetooth with phone.
• Composites:

11/29/17 So tired. But like...MIT App Inventor. Cardboard flexures. Peter Pan. Neverland. Sleep.

Real Actual Progress

12/13/17 Here is where the actual working begins. I'm going to finish my write-ups for the weekly projects up to this date (some will be finalized by the final project) and then start putting all my final project documentation here. The other final project page will be for answering the questions and giving a good clean presentation of the project. This is for the process.

So my first big push for the final project was the composite pole, shown here. I also started looking into what it would take to get the LED bulb and set up. I was going to buy a bulb until I saw this picture of the inside of an LED bulb:

The site I got that from also said they are connected in series. I can totally do that. The site I found that picture at was actually a good "someone did this before me" page (link). I counted 36 LEDs on that picture but I also saw other commercial bulbs on Google with ~18. So I looked up recommended lumens which wasn't actually very helpful since the data sheets don't give lumens. Basically I'll want to run them near full power though, I got that much. It's not helpful that red isn't much like blue and green. Oh yeah, so I was trying to decide between white LEDs and fancy "full-color" LEDs, and after looking at the data sheets I figured that the multi-color ones aren't very hard to understand (they have four pins and three LEDs - red, blue and green - that you can mess around with) so they would give me more flexibility if I have time to make improvements after getting on-off to work. Here's the clipping from the datasheet if you don't believe me:

Also they came with diffusers, which I think will help but I'm not totally sure. It could also just limit the total light that I can use. Before I committed though I set up a super patched together test to see what the LED would look like. I used a 1k resistor and several questionable connections that required two hands to operate to see the light I would expect. I was happy with what I saw. Here's the setup - you can hardly tell what is going on but I basically just messed around running current through each of the diodes. The wires are connected to the 5V power sections of an FTDI cable from a USB port.

Next step will be deciding how many to use (I'll probably start with 36) and how to imcorporate them into the BLE circuit. Which brings me to my next thing. I made a plan to finish my project. I have finals Tuesday, Wednesday, Thursday. I want to make sure I have time to study so I'm glad this class is first. At my old school final projects were due the last day of classes to avoid this sort of thing but I kind of need the time unless they got rid of a weekly assignment. Anyways, my projected schedule is:

• W-Finish documentation from prior weeks up to current state of affairs.
• T-Networking week and interface week. Bluetooth board printed. Bluetooth working to pull a pin high/low. Finish designs for tomorrow. Document.
• F-Electronics housing fabricated. LED bulb fabricated (reflow soldering). Lamp shade fabricated. Base fabricated. Find the cords. Integrating the entire system. Document.
• S-Tightening up the programming, adding new and fun stuff. Color change. Brightness. Keyboard. Recording a message. Adding the magnetic sensor. Battery and housing. Aesthetic improvements. Document.
• S-Tightening up the programming, adding new and fun stuff. Color change. Brightness. Keyboard. Recording a message. Adding the magnetic sensor. Battery and housing. Aesthetic improvements. Document.
• M-Finalizing documentation. Final video. Studying for finals!

So I'm currently working on getting documentation up to date (obviously) but I'll be back in business tomorrow. If I have time tonight I will start looking into the BLE module.

12/14/17 I spent all morning redesigning the BLE board to work with a 5V cable and then I was having scaling issues (see week 13) and then Niki gave me a 3.3V FTDI. Meep. I tried printing the BLE board just plain and simple and still had scaling but the lab closed. So I came home and have drawn up some sketches for the base, shade and bulb. But now it's late and I'm going to get some sleep. I'll get pictures of the CAD once I make them. Or at least the finished pieces. Hopefully I can get BLE working tomorrow. I may need to skip work, or the party, or both. Meep.

12/15/17 Good morning! I told my advisor I wasn't going to make it to the lab today because I really needed to knock this project out, so let's hope I can be effective with my time! Plan from here to the end is: stand design, support design, bulb design, shade design (when lab not open), BLE circuit machined, soldered and tested (hello, commands, and have it send something on its Tx), App Inventor with BLE, App to turn on existing LED with BLE, build stand, supports, bulb and shade, develop new BLE and LED control board, build new boards, update code for RGB control, tidy up (assemble, etc, anything else ambitious). So I'm in Fusion 360 working on the stand now. It's going to be pretty basic but show the 3D modeling I've learned. I'm using real twigs to upset the surface of it like roots. The base will be non-resined fabric and the rest will be a machined foam covered in composite (3 ply). The resin will stop before it hits the center to give flexibility in adding the pole. I will be able to vacuum this piece! Check! Here's a picture:

Now on to the support for the shade. I'm going to be making it out of cardboard. From cutting week I got a kerf of 0.02in so I'm going to start with that (though I'm designing in mm so I'll have to do math at some point). I'll use a parameterized design because I may need to change some things. Time to go to the lab. I got there and cut out my board...actually I set up the cutout for my board and then I tried to start the stand job on the shark but apparently you need to babysit the thing so Dave let me pause the shark job to go cut out the circuit but that meant I couldn't solder. So...totally useless. Plus the scaling was way off for whatever reason but I figure it will be faster to hack into the foam with a saw later to get the hole right than to try to redo the CAD (I did spend a few minutes on it but when I tried to change the scaling it decided to invalidate my lofts, which I couldn't then get to reappear. So. Yeah. I'm a useless no good who can't do anything right. But I already knew that so it's just for the benefit of the masses to have real genuine proof of the fact. Not that they care but, you know, I should keep tabs so that the day I decide to do nothing productive with my life I have plenty of examples of why I'm no good at anything to justify that existance. Also, to make my day better (in addition to vaccinations, missing meetings, traffic, things not working, it being cold and me continuing to be a useless lump of flesh) it turns out LEDs don't just go in series. They don't work like that. Awesome. Wonderful. I knew it was too good to be true. Grace is a savior though. She did a lamp and has a lot of spare parts. And with a relay I could still do Bluetooth. Bluetooth on and off would be "more than enough" according to Grace. If only I could actually build the thing. And I did. Wow. Elysa watched the Shark while I soldered and I got it to do the thing. I documented this progress in week 13 and week 14.

12/16/17 Couldn't sleep last night because I kept dreaming about Bluetooth. And it was a touch too warm in the room. I'm procrastinating heading out to campus because it's cold outside but I have a review session for another class so I have to go eventually. I with I knew where I lost my roommate's phone. My next guess is the lab but it doesn't open until 12 so I guess we'll see. I'm going to need it soon. Anyways, here's a link to some kerfing ideas I got from someone else's page. When I got to the lab I pulled out all the pieces and did some work with the stand (documented in week 14) before having a heart to heart stare down with my lamp. See, the lamp shade thing might not work. I needed an extension cord with at least two ports. One for the relay and one for the BLE. But pulling them at the base would mean massacring my beautiful stand. I couldn't do that so I decided they both needed to go up top which didn't make much since because there's not that much space and the pole is too narrow for most of that stuff. So I figured I was going to have to rethink the shade. I spent time thinking about it by doing up the stand composite (week 14). Then Natalie and I went to Chipotle and I was muching on chips and staring at my lamp thinking about how to make a box look good (get a stuffed animal to be climbing the tree, somehow getting a "rock" to sit at the base) when I realized a box is like a house and a house is like a castle and a castle is interesting. This example is clearly overkill but the idea of an internally lit castle to house the electronics was appealing:

Think about it. Jack and the Beanstalk, am I right? I could even make a little house to put at the base and a little cardboard person climbing to be Jack if I had enough time. I can use tissue paper as a diffuser, laser-cut cardboard for the pieces, and blotting layer for clouds. I could even use a piece of cardboard to be a giant's silhouette. I will cut lines like bricks (obviously not fully connected or selse it will fall apart) to help let light through. There can also be windows and an open top. A castle in the air. Pretty neat. I guess we'll have to see how that works. Instead of cadding that right away though, I started working on the relay. I did test the BLE pin setting feature with a multimeter so I knew I was getting about 3.23V for logic high and 0V for logic low. But before I could work with a lamp I needed to find cords. I was going to use a cord I found in the apartment but then I found the lab had tons and tons of cables. Many of them precut. Admire them:

I then needed to pull the cord through the pole before I started adding things to it lest I make it impossible to thread. As is it was a bit thick so I tied it to another cord that went through fairly easily and pulled it through. My set up for that is shown here:

That was just the extension part though. For the relay segment I needed to cut up a wire, which proved to be rather challenging to do without removing the insulation of the internal wires. I burned through 2 cords by exposing wires I didn't want exposed. For the cables I worked with Black = hot, White = Neutral and green = ground (like earth ground). You use black and white to power the load. Earth ground is to protect the user by grounding the casing and such. I tore off the third pin with pliers though since the extension cord I found was only two prong. Plus I wasn't going to use ground anyways. One of Natalie's friend's helped me figure all this out. He also taught me about finding 'electrical consistency' with a multimeter, which was pretty neat. It beeps when things are electrically consistent and that helped me to check my assumptions about which cords went to which prongs. The wires were made of a bunch of smaller wires which were really annoying to work with when screwing them onto conductive plates. So I tinned the tips to help:

It was much better this way and I was able to get a good result. Somewhat just for kicks and somewhat to check polarity of the bulb socket Grace found for me I used another cable without a relay to test the bulb. It worked! The internet said that with CFLs reversing polaritywon't blow it up. It just won't light. But I guessed right: white to silver, black to gold. Here's the relay circuit. I am switching the hot cable as recommended by the course website.

And here's the lightbulb working all on it's lonesome with the socket:

And here is the relay circuit working when I apply 3.3V from a voltage supply to the control side of the relay. Ready to go!

I then hot glued it all into place once I made sure it worked. Unfortunately the lab was closing at this point so I got some measurements of the piece that would need to fir through a hole in the bottom of the castle (the extension head). But when I got home I just drank some seltzer water, closed my eyes a bit and hung out with my roommates. Until now. Now I'm updating documentation. But I'm really tired and will save cadding the castle till tomorrow morning before church. My roommates are making me do something for my birthday. They wanted me to not go to the lab at all but I made a compromise since the lab doesn't open till 3, we'll go to lunch together. At this point I just need to cut out the shade and assemble it, ensure the stand is functional, and get the Bluetooth to work with an app. Almost there.

12/17/17 So I started cadding up the castle and got a lot of inspiration from an online picture (below) on how to get light to come out of a castle. I asked my roommates without context if they thought giants lived in castles or houses and they decided castles. I probably would have done a castle anyways. Here's the reference photo:

First I made some sketches to get an idea of what I wanted and then I moved on to Fusion. Here are the sketches:

And then I made some CAD with Fusion/Inkscape to match (Fusion for the basic shape and Inkscape for the texturing).

12/17-19/17 (final push) Now to head to the lab to cut this out and check on the stand. All of that is documented in composites week. Or some of it, I guess. Because I eventually added a nice padded bottom to the lamp:

And a lasercut wooden label that I painted to go with the theme (so people other than me would get the joke). When I realized the labels made it seem like something that would go in a kid's room I kind of regretted it but then decided if I like the theme then it doesn't really matter if it's childish or not, because being childish in terms of what children are interested in is an admirable quality. Or I'll at least keep telling myself that. Wow, I got distracted. Here's that label:

Jack's Lantern. You know, like a Jack o Lantern but actually a replica of Jack's beanstalk so Jack's lantern. Haha. Actually, I'm probably the only person to find that funny but that's okay. It may help people understand the theme. A theme that originated from trying to get the ugly pole to have a purpose. Speaking of which, here's that castle that's going on top actually getting cut out of cardboard:

Once I assembled it it looked pretty nice atop the pole. I secured it through the hole in the center with tape from the house floor to the inside edges of the pole. Here's a picture:

Then I printed out the window frames on 1/8th in plywood:

But I did it a bit too agressively. There's a reason you stand next to the laser cutter:

Then I painted the castle black, added parchment paper to the window frames with Gorilla glue (Elmer's would have been better) and then hot glued them to my castle. I also Gorilla glued on some laser cut labelling ("Giants") to the doorframe. Gorilla glue was a bad idea. It dried slowly and made my hands miserable. I didn't want hot glue because I though it wouldn't lay flat and would end up being very visible but it turns out the gorilla glue is pretty noticable too:

Oh! And I added the batting as a cloud. Then I checked out the integrated system with my button press program and it worked beautifully:

The inside of the castle isn't very pretty but that's why I used parchment paper on the large holes to the outside. I tried to be organized but the castle was a bit small for careful positioning. I wanted to keep using FTDI cables in case I needed to use them for the BLE but I could have made a nice USB power supply circuit to connect with my wall to USB converter block (the white cube in the following picture). Here's the inside:

Biggest regret is that the lightbulb can't stand up. But the castle would have to be strangely tall for that to work well. Maybe I should have had one central tall tower instead of only having sorta tall walls all around. Also note that I added hot blue to the screws on the relay to keep them from coming undone and used the plastic safety case in the final design to protect against shorting the wally socket with your hands (or anything really). Here's what that looked like, you can't really tell but you can see the protective casing.

When it was all done, the finished exterior looked like this:

I know there will be wayy more impressive projects (esp from the Media Lab section) but I'm proud of my lamp. So I had to protect it with intense notes when I left it in someone's cubicle in my department:

I guess this is basically the end of the class for me. I'm going to program the board with a button press on/off switch for the demo. This was a great time. Can't wait to start making for not-class! Oh...I'm going to make one last push to try to get the button to be a doorbell. I may never document that since I'm in crunch time but it will require adding a pin header to the ground strip of my microcontroller and adjusting my code for the new pin. Wish me luck! The code for it is here. And here's a video of the button turning on and off an LED (same concept for the lamp but different pins):

Now it's just pin swap and assembly. I've got a few hours left, let's go!

Okay, so it doesn't rightly look like a doorbell so I put it on the side (actually, I just couldn't bring myself to poke holes in the castle for the wires), like so:

And here's what it's like in operation:

Concluding thoughts: I could have used the speaker on the board to make a doorbell sound but that would be kind of annoying. It can be programmed in many a way now. Maybe the doorbell being up high makes sense because they're giants. But it is their house so probably not. It's all kind of ironic because the giant's house is pretty small. It makes me think of the time I went to the Tokyo Sky Tree and they had an Attack on Titans exhibit that was poised as if a titan was head level with the top of the tower. When I looked down I was shocked by how massive something head level with the skyscraper would have to be. I kind of made the same thing. The giants are big for Jack but humans are big for the giants. And that's all for now folks!