Harnek Gulati

This is the second week's project: Cardboard

1

Started out with a simple CAD model I found online.

2

I imported it to Solidworks and used "Thicken" to push each continent out further. Earlier models I cut out couldn't distinguish continents due to thin thickness.

2

I then imported it into 123D make. I needed to cut it according to a different perspective because 123DMake wouldn't work with the default.

3

Once I finished, I exported into a DXF, imported into CorelDraw on the laser cutter.

4

You can make the laser cutter raster out the print of the layer that lies on top of the bottom layer as well as the number on the layer. This makes it a lot easier.

4

Because carboard will naturally bow and so you won't get an even surface, it's good to try to make it as flat as possible. Place the cardboard so the edges curve up, and then tape it down like so.

5

Once it's all printed out, you can just compile all of them. The smaller pieces are harder to find a place for (some might float), but keep them just in case.

6

The starting pieces! Don't glob the glue, or else you'll get too much of width between cardboard pieces.

7

After you're done adding up all the big pieces, go through and figure out what small pieces you can add too. The small pieces helped finish up North America.

7

Now you have to clean up all the laser engravings, or else it looks weird. Create a small puddle of glue, and strip off the top layer paper of cardboard.

8

Awaiting the birthday girl's approval

❮ ❯

This week, I started the cardboard laser cutter. I know I was supposed to work with pressure fits, but after asking Rob, I decided to go with the globe project. The idea is really simple and was intended as a gift for a friend. It was a lot of fun. I started out with this sketch that you can download here

A huge part was just working with the DXF model. In particular, it's important to realize that models come from different softwares and each export files in a different way. I asked online on what I could do, and someone got me the answer, you can find it here .

Here are the things that I learned through this project:

• Measure the cardboard beforehand. It almost didn't look spherical, but it wasn't terribly off.
• When you use outside sources, you WILL have to change dimensions and move everything.
• Laser engraving is good, but you have to do a lot of cosmetic cleaning up afterwards.

Fixing the problem of mixed gender institutions through Vinyl Cutters: The Toilet

1

Started out with a image I found online. Luckily it was only two colors, so it was very simple to save it into a png file.

2

It was cut out, but the small stuff was very difficult, so I left a lot of it out.

3

Its final position!

❮ ❯

Here's the stuff I learned:

• Stuff that are small islands are extremely difficult, especially with the weening part. Avoid them when possible.
• When weening, pull it back at as low of an angle as possible, and poke the thing that needs to stick.
• Placing vinyl on tiles are difficult because of the curves, but I was surprised by how good it ended up being.

Soldering and surface mounting!

1

I started by gathering all the parts. Simply take a double sided tape (parts can jump around a lot) and label everything needed. Drop each piece one by one beside the part needed.

2

Use two-sided tape to hold the piece in place. This will help stop it from moving as you're soldering each piece.

2

Completely soldered! I've had experience with this in the past, so I got most of it done in the first shot. Biggest takeway is to just practice. It takes a lot of time to master.

3

It worked! Woohoo! Unfortunately, I didn't have more images because I was too excited about the soldering process.

❮ ❯

Here's the stuff I learned:

• Soldering is soothing until it doesn't work.
• When removing solder, use the hot air gun. If you don't, take the braid and try to remove as much solder as possible, and then use tweezers to remove the part, but this can be horrible.
• Ensure all the parts are in the right direction before you add it.
• If it doesn't work:
  1. Look at all the IC legs first and check if things are connected.
  2. Whip out the beep detector and check everything, especially under parts that jump over copper paths.
  3. Check if your parts are in the right direction.
  4. Change your ICs out.
  5. Bang your head against a table, start from 1.

3-D Printing something that can't be made negatively

1

I started with the following CAD Model. I simply used the gears given in Solidworks Tools.

2

I simply added it to Cura and changed the scale so it would just take 3 hours to print.

3

I decided to use the Ultimaker to make this model, and the MakerBot for something else different.

4

The Ultimaker started working, but the material feed broke halfway.

5

The Ultimaker's feed is in the back of the 3D printer, unlike the Makerbot.

6

The gear was good, but unfortunately, I couldn't get to the structural material on the bottom of the gears, so I had to change models or printers.

7

Say Hello to my little friend: Stratus. This baby uses structure that can be dissolved.

8

Here's where the structural material gets dissolved.

9

Here is the final product! Everything is free and working. Woohoo!

❮ ❯

Here's the stuff I learned:

• Be patient with CAD. There's always a guide out there that will help you.
• Making a "part" is often very restricting. Start with an assembly and add parts to it. It'll help you make the bigger piece.
• Use a better 3-D printer if you can't figure it out in the first place.
• There's almost always an easier way of doing what you want to do in Solidworks. Think about the process before you implement it.
• If you're stuck:
  1. Press Control + Z. This can sometimes lead to huge failures though.
  2. Rethink the process. Is there a better way to do it?
  3. YouTube it. It's probably been done before.
  4. Talk to a friend who knows better
  5. Bang your head against a table
  6. Start over

3-D Scanning and Printing

1

I started by using this Sense

2

This is the setup. It was important to have the Sense be fixed and the person to be on the revolving chair, otherwise the scan would be imperfect.

3

The final product on the program!

4

I decided to use the MakerBot for this one to learn it better, but also ran into the same problem of having to change the material.

5

The MakerBot material is changed from the inside. You have to say "remove material" before you do "add material" or otherwise it'll stall.

6

The final product! Woohoo!

7

To scale. It's a spitting image of me!

❮ ❯

Here's the stuff I learned:

• Start Sensing with the model showing the side of the face if you don't want to lose as much as possible.
• It's incredibly easy to 3-D print someone's face.

Learning how to use Eagle

1

I started out Reading the Sparkfun tutorial, which you can find here . It helped a lot actually. A few things are outdated with the newest version of Eagle in terms of icons, but it's mostly the same.

2

This was my first attempt at drawing the schematic, but it became obvious that this was the worst way to do it.

2

This was my second attempt. Ah. Much cleaner. Just name the wires that should be connected the same thing. In the future, I need a convention to name two wires that have different port names (for example PB1 going to MISO. What do you call the wire?)

3

Off to the routing races! I did this horribly, actually. I should've 1. used the grid function earlier and 2. Used an earlier schematic and built on top of that. I tried to do the entire thing by hand and it was a mess. Here you can see a cut that's too small.

3

Fixed the problem and here you can see the final product.

3

I put it into fabmodules and got this. It was very easy plug and play.

3

Got it good. The outline could've been a bit further from the actual circuit, but it didn't cut into the actual copper, so I got lucky.

3

Quickly soldered. Just did the good ol' double sided taping to the back of the circuit. Then I solder anchored everything.

❮ ❯

Here's the stuff I learned:

• Eagle is great. Use it a lot.
• I would love to know how to make custom parts.
• USE THE GRID FUNCTION IN EAGLE!
• USE OTHER PEOPLE'S EXAMPLES AS INSPIRATION.
• Draw it out before assigning pin numbers. Sometimes things are easier if organized all on one side.
• If it doesn't work:
  1. Use a different drawing tool (Square, line, etc)
  2. Hold ALT when making the line
  3. Change the routing.
  4. See how auto-routing does it
  5. Change the schematic.

Building a Wiggle Chair

1

I wanted to build this wiggle chair I saw online. I was advised against making it all cardboard (it would use a lot of cardboard). So I went with something else.

1

Time to Learn Rhino (the grabcad I used was a .3dm file). Major props to Johnathon for helping me out. Turns out it's super intuitive.

1

This was the new design I decided to work with. The parts would press-fit into one anothe through holes on the side-frame

1

I had these DogBones originally. The outside circles were actually 6mm in diameter to begin with, but I changed that later to 3mm. That was a problem because the part couldn't get into the hole.

1

These ones were a little better. I set the Diameter to .2".

1

This is the simple joining piece.

1

Final Sketch! Woohoo!!!

1

Quick Cardboard printout check. It worked =D.

1

The PressFit Worked!

1

Darn OSB is terrible for structure. Everything broke really easily. This quickly turned from a chair into a museum art piece.

1

Started the full cut, but this time drilled down the piece so the wood wouldn't damage itself.

1

I messed up the tabs, so I had to remove the pieces immediately, otherwise they would make large noises due to the vibrations.

1

It crashed halfway through. Lucky, Rob was there to help me out, otherwise I would've had a half-finished piece.

1

Assembly was difficult, but not super hard. This was just fixing one side.

1

OSB gives you mad splinters.

1

The half-finished piece.

1

Rob got it to work again and we kept cutting!! Unfortunately, the press fit wasn't perfect, so I had to use wood glue.

1

Front view of the chair. It was still gluing, so I didn't dare sit on it.

1

Side view to see the entire thing in action.

❮ ❯

Here's the stuff I learned:

• Rhino is great. Use Rhino.
• Always measure beforehand and ask a TF if you have questions. You rarely get that experience in real life.
• Give yourself ample time to figure out what you want to build.
• Don't utilize other people's resources. Change your design if you can to accomodate fewer materials.
• Figure out what kind of structural material you'll be using. OSB shears extremely easily, so a chair wouldn't necessarily work.

Using CC2650 for BLE and Microcontroller work

1

I used the LaunchPad CC2650 because I need to use BLE peripherals for my Final project and wanted to be sure I understood everything. It ended up being a lot of code and mostly just understanding and fiddling with the code.

1

I started by trying to follow the guide online, but it's very unwieldly. An online video tutorial ended up helping me set up by environment

1

TI has a great cloud interface that actually is more intuitive and has a cooler UI than their desktop application. However, it doesn't really have the features. For some reason, flashing with this was much easier.

1

First thing is that this didn't just have a main.c function, but rather a Stack and an App (The App has a main.c function). Together, it took up most of the RAM and all Project Zero did turn a couple LEDs off, so it was a little annoying.

1

One of the best things I learned was just how the OR function worked. Basically, it just sets a function to 1 if that bit is 0. All this is doing is setting Board_LED's type.

1

Although I didn't work with a microcontroller as much, I did learn a lot on my own just by looking over each function and then cross-refrencing things with the reference manual.

1

The Reference Manual is immense. Like seriously too long.

1

After a lot of work, I got the LED to blink! Yay! Cross-referencing manuals was a lot of work.

1

You can also get into talking with the device my using Putty and erading in COM4.

1

Getting the BLE to work took a long time too. Here you can see the device ID.

1

And here is how you can read and write things onto a BLE device. It's pretty cool, but I definitely need to understand it on a systems level more.

1

This is how Project Zero allows you to change the LED.

❮ ❯

Here's the stuff I learned:

• Don't bite off more than you can chew, or if you do, chew slowly.
• Use the youtube tutorials to really understand, but in the end, the reference manuals will tell you what you need to do.
• Chips that are Power-Saving are Power-saving because they force the coder to do extra pre-processing work like clock gating.
• BLE is really hard, but definitely worth it.

Molding and Casting

1

I'm starting to get the hang of Rhino! Took an STL file, found the center point, then added two cylinders on both sides. Then, I copied the cylinders, rotated each half outwards, and extruded/extrudecut the cylinders from the base.

1

Then, I placed it into the Shopbot 3DCarve software to get the following rough cut. It ended up taking a longer time than usual.

1

The simulator then said this is what my final product would look like. Looked good to me!

1

I had the orientation wrong the first time, but thanks to the aircut, I was fine. However, I had glued my piece to the wooden surface, which was inconvenient for all the other shopbot users. I ended up gluing it to a piece of wood first.

1

Begin the Milling! The offset was actually very close the edge, so during the finished cut, it created a small crack that I later used wood to fix.

1

I was afraid the Drillbit was not long enough, but it actually ended up being just right. You can see here it was close.

1

Oumoo was applied to the mold after it was cut with the Shopbot. Simple 50/50 mixture.

1

Once it was out, I just cut it in half and aligned the keys. The pokeball fit perfectly!

1

Watching plaster dry.

1

And it's done! What a quality plaster of paris ball.

❮ ❯

Here's the stuff I learned:

• Start simple sometimes. I sometimes try to do overzealous projects, but this one was fun.
• Understand fractions. I messed up.
• When removing bubbles, it's both the suction and the replenishing of air that pops them, not just the suction
• Understand the entirety of what you're trying to accomplish before CADing it.
• Sometimes Rhino can be really weird and not work normally. In that case, remove the parts and start again.
• STL files work better in Rhino than in most other CAD programs I use (which is basically just Solidworks)

Making a circuit, the struggle

1

I tried to begin this week by using an IC (AFE4404) I would use later for my final project. This ended up being a very long expenditure that took up my entire week. I was basically looking at a 84 page spec sheet trying to understand everything, but to no avail. Getting this LED to light up took an entire day.

1

After much more work, I ended up getting the green LED to work, but the data I read wasn't readable. It was so noisy, I was wondering what I was even doing.

1

I found an old evalBoard setup to check if the processing would work, but because of the software not being compatible, I was lost again.

1

Not wanting to fail this week's project, I decided to move onto something else. I decided to make a thermistor that didn't use an H-Bridge but sent out the information to another pin using the NRF24L01+.

1

This looks good enough, but the problem is that the pins for the NRF24L01+ didn't give enough room for the ISP ribbon header. I didn't realize this until later.

1

Here you can see there wasn't enough space for the Ribbon header.

1

Here we moved it further to the right to accomodate.

1

Unfortunately, I didn't have the time to make it work, given the failure over the weekend, so I had to call it a day for this project. I'll try to get it done over the weekend.

❮ ❯

Here's the stuff I learned:

• Failure is inevitable sometimes. Try to have a backup working.
• Use a line width of 10 for modeling on CAD. Otherwise it can get really difficult.
• When exporting, you can lie about the diameter of the drill bit to get better cuts.
• Check out the spec sheet of the IC (AFE4404) in the case that you don't really want to use it.
• Understand that ICs are not magical. You still have to do a lot of post-processing with them to make it all work.
• Using an analog signal out of an IC is probably a better idea that having it stored as a digital signal.
• Measure twice, cut once.
• Don't put your ICs in a compartment that has a bunch of dust and residue. Things will short and your programmer won't work anymore

Making a Bottle Opener

1

I wanted to try something a little more complex than bowl and I wanted to test the limits of a composite, so I decided to try doing something small instead of something large. This ended up being the biggest mistake, but I learned a lot in the process.

1

I cut out the piece as you can see, but the problem became cutting it out. Foam isn't horrible to detach, but I slowly realized how bad of a piece I was making.

1

In order to try to work with the edges, I made small cuts in the linen and placed them over, but the problem soon became not being able to align the cuts with the actual piece.

1

I then took the piece and covered it with linen doused in a 2:1 ratio of epoxy and hardener. Linen ended up being a solid choice and in the future, I'll definitely be using that.

1

After adding the breathing layer and top layer, I moved it into the vacuum. I realized now there were a lot of problems with what I was doing.

1

All sucked up. Placing the breathing hole right next to the breathing hole sounded like a good idea.

1

Now it's all done! It looks really bad... None of the linen that I used to cut open the bottle really even stuck.

1

Time to cut it out.

1

Let's play a game called "What's that Composite?!". I promise you no one will guess it's a bottle opener.

❮ ❯

Here's the stuff I learned:

• Don't suck.
• Think (just a little bit) about the entire process before taking the next step.
• Don't make a bottle opener using composites.
• Linen is great as a composite! Although it started off seeming too thin, it ended up being as strong as the burlap with 1-2 more layers.
• Sharp edges for composites are a hassle. Even the cuts in the linen make it too difficult. Use it mostly for things that don't really have that.
• Large objects are probably best for the vacuuming portion, as well as the folds.
• Use the bandsaw to complete the composite!! It cuts pretty easily

Starting my final project: An attempt to Charlieplex

1

I started out with basically just eagle. Very simple design, but because it took so long to mill, I added a new thing.

1

We started out with milling out the printed circuit board. I wanted to try milling the backside, but it became really tough.

1

We're on the Modela! Setting it up was hard. I had no idea how to actually cut it out on both sides. It wasn't until the next week that I realized I could just flip the board on the board.

1

This is how it looks flipped around. I realized this is where I failed. I had taken in the wrong trace. Charlieplexing was officially a failure.

1

The next part was studying up for my LEDs. I tried to test how much mA each LED ran at and what luminosity it stayed at. The majority of this week was trying to work on my final project. This was extremely difficult, as charlieplexing had failed and the big board I bought also failed.

❮ ❯

Here's the stuff I learned:

• There's a lot of research needed to be done before doing your final project.
• The output devices week of handling LEDs is very confusing.
• Charlieplexing requires connecting the LED array's ground to ground.
• Charlieplexing also doesn't allow you to change the brightness of the LED you use.
• Charlieplexing uses n ports get n^2 LEDs
• Please look at my final project for more details on output devices. I continued this week on that page.

Making a quick Processor Sketch

1

I found a website that had a processing sketch I figured I could imitate. I had used it earlier during research and wanted it to make it useful for everything else.

1

This is what the Processing Sketch looked like earlier. It wouldn't work with the format we sent in earlier because Processing has an interrupt based on a new line character. Hence, I had to change the code.

1

I made a quick board that gave me the temperature. I planned on making a graph that plotted it over time.

1

After MANY iterations, I settled on a starting "1234" frame as before, and then the data, and then the "\n" character. Processing still acted weird, as although I was only sending out 7 bytes, it read 8 bytes. I found out this is because Java takes in characters as 2 bytes, hence it had a filler.

1

It starts and asks what port to connect to. This was given to me by the PulseSensor, so I can't take any credit, but I was very impressed by the work.

1

Woohoo! This is me breathing on the temperature sensor. I decided not to do the post-processing to find the temperature because although I wanted it, I didn't actually

❮ ❯

Here's the stuff I learned:

• Don't reinvent the wheel. This was much easier to make. At least credit the site though.
• Read the differences between Java and others.
• Java doesn't take unsigned integers or bytes, so in order to make it unsigned, mask it with a FF short and then convert it. Otherwise, you'll be stuck with a signed byte.
• Read the documentation of serial ports!
• Use an H-Bridge for temperature. I was cocky and didn't use one, but it was obvious that the granulity, while fine, could've been better with an H-Bridge.

Networking

1

I found a faster way to make boards using the same board. Just trace everything first using the same board, and then switch bits, cutting everything else out.

1

I know it's not super hard to do this, but was a major breakthrough in how fast it takes me to get multiple boards out.

1

Quickest I've ever gotten 3 boards out at once.

1

Soldered up and ready to network.

1

This was the hardest part, I think. Each board had to be seperately programmed, but the programmer was being very finicky. I could barely get the cable straight to program it.

1

This was the final answer. I changed the code such that when it was personally addressed, the node would fade in and out. I was planning on changing it such that it woudl listen on the line for a while before sending, kinda like a CDMA type setup, but never got to that as programming took so long.

❮ ❯

Here's the stuff I learned:

• Networking is hard.
• Best way to avoid collisions in a wired network is to have two lines.
• Serial communication through different slaves on the same bus-line requires a change in input to output pins.
• When doing your final project and it doesn't relate to networking, work really hard on your final project.

Group Machine Work

1

I was part of the laser cutting team. We began by reading the instructions that were posted online. It stated we needed to cut four full pieces of posterboard, scoring some lines. After some initial calibration, we were able to get the cutouts and scoring just right.

1

Once we lasercut, we had to use a knife to cut out certain pieces. For the parts that were scored or needed double folding, we had to remove the top and middle layer of cardboard, leaving a very thin piece to hold it together.

1

We had difficulty in the beginning assembling. The double scoring was especially strange, but all it actually meant was that we'd be folding another layer of cardboard inbetween.

1

Here you can see our initial mistake with double folding. We should have folded that piece into the other one.

1

Once we figured out how to make one, we were able to make many more. Another difficulty was placing the end pieces, given that the bottom was different sizes.

1

Done with one! The threaded rod went through the center and the box in the middle required a screw-in.

1

Here we you can see we created more and more. Unfortunately, this was the end of our module for the project. Check out everyone else's for more information on how it all came together!

❮ ❯

Here's the stuff I learned:

• Ask questions before starting on powertools. I got into trouble because I couldn't find a razor, so I figured I'd use a bandsaw. Stupid idea.
• When something is scored, it usually means it wants to be folded. When something is scored with something scored next to it, it means you want to double-fold it.
• A threaded rod allows circular movement up and down the rod if the object is held in place by one or two rods.
• Stepper Motors allow for this kind of movement to happen, although we wonder if brushless motors can also do the same.
• A laser cut machine is a genius device. Serious props to Nadia for creating something so cool.

The Word Clock

1

Quick Intro. The backside contains the circuitry which control which lights turn on, and then you have the lEDs, and then baffles to ensure that no lights bleed into other segments, and finally the actual display itself. Then it's all framed with beautiful wood.

1

Where it all started. Margins: 1.5" form the top and 1.25" from each side. I had 12 letters per row.

2

Lined everything up and drilled 5mm holes through for the LEDs. Sketch drill press, but whatever.

3

Soldering the lights together. I didn't really know about mechanical fixes back then, but in general, soldering should have no free-floating parts.

4

Trying to learn about bit shifters. I bought the wrong bit shifter by accident and didn't know anything beyond how to plug and play circuitry.

4

A Few things learned here:

1. Always find the newest version of the spec sheet.
2. Googling a part doesn't mean the first link refers to it. Double check the spec sheet.

4

No more joking around. Time to study about bit shifters.

4

That LED took too long to light up. But at least I know bit shifters!

4

Put the bit shifters in line with darlington transistors. Back then, I thought they were just "power boosters". Didn't really know how they worked. God bless ES153.

4

Put the breadboard over the LED board I made earlier and started soldering. I should've been much more organized from the beginning in terms of wire placement, but whatever.

4

Do not buy 7 dollar soldering irons from amazon. I got burned just holding it.

4

Let's build the outside frame! This miter saw doesn't have a case, so I low-key could've died. Whatever, let's do this.

4

In order for the frame to fit, I also needed to cut the pieces at 45 degree angles. I used the table saw for this because the miter saw wouldn't move.

4

Look at that wood baby.

4

Not perfect, but I just needed two pairs that were close enough. Beggars with crappy tools can't be choosers.

4

Look at that wood!

4

Making the front cover! This took WAY too long, but Ponoko was really nice about it. With Ponoko, you HAVE to use illustrator, which kinda sucks, but still pretty cool.

4

Excuse the blurry picture. We also need the backlights to defuse the lights, so I got acrylic from a nearby plastics store.

4

Here you can see the lights and the baffles. The baffles are supposed to stop the lights from lighting up parts that should be dark. I used ASB plastic.

4

All the light baffles in place! This was horribly done, honestly. I should've used something precut, or used something more professional than scissors to cut it out. It ended up being a huge headache. Also, WAY too big of baffles. I could get by with quarter inch.

4

All put together, still looked like crap. Looked like something though! Time to fix all the small things.

4

All fixed!

4

In Ann Marie's Room!! That took way too long, but I learned so damn much =D.

❮ ❯

Here's the stuff I learned:

• Study a lot before starting something. Most of the time, you'll mess up if you think "I'll just wing it."
• Never give up
• If you really care about your friends, you don't need to build something for them. Spend time with them. Hang out with them. They'll love you way more.
• That being said, build something if you really love building stuff
• Circuits require patience and organization. I have one, I still need to learn organization.
• HTC Ones make terrible pictures.