Matt Pearson
mpearson at mcds dot org

Hello World!

San Rafael, California

This is my first entry and it took some time to get here! This project takes me into rich curricular design in an environment of K-8 students and teachers which has ongoing remixing of disciplines, materials and tools. Some call it transdisciplinary learning, others call it project based learning, still others call it inquiry based learning, surely it is all of those blended with the principles of Design Thinking in a classroom where failing rapidly and with low cost is the single most important objective.

I hope to evolve this web space through my learning and found it appropriate to start with "Hello World!" Mercurial, our primary tool for exchanging data with our primary server hosting this web space, is straightforward for the most part. My laptop admin account would not cooperate and when I created a fresh admin account on the laptop, followed the tutorials, corrected some tutorial typos I was pulling, updating, etc. in under five minutes. Fortunately, my journey took hours of tinkering and diving down many a rabbit hole to uncover the inner workings of Mercurial, the Bash shell, SSH to name a few. I learned many new tools and dusted off my aging UNIX command line skills. I have had a myriad of emails and calls with my mentor all to great learning success! The problem lied on my Mac OSX Lion installation and the permissions, which are still not allowing me to function in the Mercurial environment in my main account, having issues with private and public keys. I continue to try new things and come up with ideas, but that rabbit hole, which I find interesting, must wait. For, in the spirit of pressing onward in this project and in this class we must leave behind some rabbit holes to allow the learning to progress in other areas. Rainy days will come and the rabbit hole will be opened again in the future.

This concludes part 1 of entry 1 using HTML in Nano at the CLI of Bash. Part 2, I am sure, will reveal some inner thoughts and a bit about who I am as I pursue a final project idea that is tied both to education, the Fab Lab, and the Grand Challenges of Engineering. My search unfolds with a deeper understanding of the challenges and a search for a tool to enhance these entries, automate my updates, and ease the use of graphic and video media insertions into this journal of my Fab Academy journey.

[website later updated with Sandvox - see week 2]

mtp

"Design, make and document a 'press-fit construction kit" 

The press fit construction kit is taking shape.  Not having ready access to a laser cutter makes things difficult.  Iterative design is more my style and I will be enabled in that realm tomorrow when we have 2 two-hour sessions at the TechShop with the laser cutter.  This morning at a coffee shop some napkin design happened, you need to always be ready to capture inspiration OR little moments of time left to dream.

I realize now, that I left out the flex holes at the base of each slot on the triangle vertices.  As it turns out, the slot was quite a bit thicker to accommodate the corrugated cardboard used.  I used Inkscape to develop the design from the napkin and then imported into Adobe Illustrator (AI).

I wanted to make sure the design came up in AI before I started the Laser cutter time at the TechShop.  We only get two hours on the cutter and I didn't want to spend it on the initial design.  As it turns out, Inkscape is useable, but AI seems more robust.  I am new to both tools and feel I should have started in a CAD program and then brought it over to AI.  I was able to clone parts for the parametric build, but my design is built off of geometric shapes that should be much easier to specify and input.  I started the download and install of the Fab Mod tools and hope to explore those in future weeks. 

The design has evolved to include pressure points for insertion into the center circle as well as squared tips to allow for stronger “prongs”.  I call it a Tring short for Triangle.  It came out "Twing" in the wee hours of the night and others overheard me…it stuck.

The Twings may or may not be the only piece.  What I would like to include is an articulated system that can have additional pieces constructed under a standardized form ala LEGO style.  While LEGO is complex in its scale and ratio system, and difficult to deduce from the user perspective, the LEGO go together intuitively.  Intuitive plus a basic system that is easily understood, while both visual and tactile in nature, is the goal.

The Twings will allow the build out of a model in four directions, but lacks a system that is structurally strong while also easily fit into our world of rectangles. 

Any educational modeling system needs to have both.  I hope to try a press board and an acrylic in the weeks ahead when I have time to return to the project.

Work flow

1. I started with a sketch that ended up on a napkin in a restaurant.
2. I took this sketch and modeled it in Adobe Illustrator, a vector based drawing program where the drawing is represented by formulas rather than pixels as in a paint program.  My experience with Illustrator is limited, but I worked extensively with Corel Draw 6 years ago and it is similar.  Once the learning curve was conquered the initial Twing was born.  Six iterations later Twing was ready! Here is the AI file of a Twing: http://bit.ly/1tK3Nh3
3. I made the design a quad image so I could printout multiples.
4. I needed to take a two hour class at the TechShop to use the Epilog Laser Cutter.  I would need to come back to do the cutting of the Twing at a later date.  The class was thorough and moved through all the settings, setup and materials you could use.  We finished with creating a little dog tag engraving.
5. In a few days I reserved the machine to make my cuts.  I moved the file to the control machine and realized I should be making a whole lot more than four at a time.  The laser cutter is setup as a printer and you send your print directly from Corel Draw using a the settings of the printer driver to control what happens on teh laser cutter.  First I needed to setup the machine: 
   1. Use a USB drive to take your finished work to the computer connected to the laser cutter.
   2. Prepare the computer:
      1. The computer is normally on and should be left on. Make sure the monitor is also on.  If in doubt, reboot the computer to start fresh.
   3. Prepare the laser cutter:
      1. Make sure the fan is on.  A fire may start if the fan is not running.  Never leave the area of the cutter when in operation.
      2. Turn on the laser cutter. The On/Off switch for the laser cutters is on the left side of the machine.
      3. Check the lens. A dirty lens can affect your power settings significantly and can also be a source of fire. If dirty, alert the TechShop staff so they may clean it.
      4. Check the metal bay for debris.
      5. Set the material in the bed of the laser cutter and set the home position.
      6. Set the focus by using the metal guide to come down on the material to be cut.  This will set the distance and focus.
   4. Open the project (convert from AI if necessary in Corel Draw) and setup the print and print settings:
      1. There was not a setting for corrugated cardboard or fiberboard, but there was a set of settings in the documentation for mat board with a 45watt CO2 laser: http://bit.ly/1nTXEBu  I started there and used the following settings:
         1. Speed = 50
         2. Power = 50
         3. Frequency = 500
      2. The above settings turned out to work almost perfect and I landed at a power of 40 with speed and frequence unchanged.

The Design
The design was planned to be used with cardboard that I had collected from the packaging of laptops.  It is corrugated fiberboard with two walls.  The same stuff used in packing boxes typically sent in the mail.  These particular sheets were neatly cut for laptops and fit nicely in the laser cutters 18"X24" bed.  This is the type of cardboard used: http://en.wikipedia.org/wiki/Corrugated_fiberboard

What I learned and would do differently
It was my hope that the cardboard would flex.  It does flex, but will not flex more than a few times.  The cardboard will also pinch and not retake it's shape.  I would like to try using a wood board and/or some tougher acrylic.  This will require a thinning of the design so the legs flex.

This week's material was done during Week 4.  Access to resources were not available for Week 4 and you can check that post for more info.

For this week we are tasked with 3D printing a small object that could not be made subtractively as well as  3D scanning an object to make it into a 3D model.

3D Printing
3D printing is called an additive modeling technique unlike a Computer Numerical Control (CNC) machine that uses subtractive modeling.  In additive modeling the machine, or artist, adds material to the model to build it up.  In subtractive modeling the machine, or artist, subtracts e.g. carves material away from a mass of material.

In review of subtractive modelers, I needed to come up with something where a milling bit (on left) could not access a surface to mill it.  Making something hollow seemed without challenge.  Making something static also seemed without challenge.

Building off of Week 3 and the theme of toys and puzzles, I have designed a puzzle that can only be made through additive modeling.  To be clear, you could assemble the puzzle from parts that were created subtractively, but this puzzle is created in one 3D print and has moving parts.  I call it Hells Balls.  The puzzle is constructed out of a double helix that is then copied and inverted below itself.  This construct creates a cage for the puzzle balls to operate in.  The balls will be printed inside the cage.  It would probably be sufficent to create just the two double helices, but I wanted to have something fun to engage with, not just look at.

Hells Balls v 0.5 (only one ball)

The object of the puzzle is to get two balls into cups at different locations.  This version has just one ball and one cup.  I am soliciting user input today to see where success lies in a design. Hopefully there will be time to play test and iterate this design. I would like to have easy, medium and hard difficulty and may number the balls and cups to that end.  Ideally, colors would be added for the parts so numbers are not needed.  The 3D printers I have access to do not easily support multiple colors.

As I dive deeper into design that necessitates parametric capacity, I continue to look for digital tools that are both intuitive, main stream, well documented and supported, and above all else, can output designs that I am after.  To that end, I have landed on Autodesk's Inventor. Unfortunately, I am not sufficiently skilled in this tool to use it for this week, nor do I have a PC at home capable of running the software (Inventor requires a Microsoft Windows operating system).  While I remedy these shortfalls, I continue my use of Tinkercad to design 3D models.  Tinkercad, recently acquired by Autodesk, is far and away the easiest 3D modeler to use in my experience.

Tinkercad is an online tool and no installation is necessary since it runs entirely from a web browser.  All models are saved in the cloud, are private or public, easily downloadable in many formats, has tutorials, and is easy to work with.  You can also code custom design tools and/or build off of other custom design tools. I started with a helix tool and then modifed the parameters to stretch.  My goal was to create a sphere and as you can see, there is work to be done.  Click on image to enlarge.

March 1, 2014 - Update from the 3D printing trenches at TechShop San Francisco. 

Since the Richmond Fab Lab is still coming online we are doing much of our work at the TechShop in SF.  Our 3D Printer choices are the MakerBot and a machine called the Type A.  The Type A came highly recommended and there are two at the TechShop.  I went with the Type A.  The machine on their website is the newer one.  The version I used is on the left.  The printer is very similar to the Makerbot, but manipulates the Z axis by moving the printing deck.  It also has a very good printhead and filament feed system.  The print deck is not heated, but since I am using PLA that was not a problem.  After spending time with the Printrbot and the original Makerbot, this machine is golden.

I decided to design an alternate creation than the one above as Hells Balls would have taken all night to print and would have been tough to iterate given my limited knowledge of the printer.  I decided to create a box with a sphere inside.  This design also included cut outs on each side.  I used TinkerCad to design this in about ten minutes.

Create in TinkerCad, then download the STL file and put it on a thumb drive.  Now, it is off to the very small sub-notebook that is hooked up to the Type A.On this small computer we have the two pieces of software necessary to create the piece: Pronterface and Keep It Simple Slicer aka KISSlicer.  Pronterface controls ther printer and KISSlicer takes the STL file from TinkerCad and creates the GCODE file you feed into Pronterface.

Picture of Pronterface to the left. This tool allows you to control the printhead, zero to a home position, set temperatures and more.  This is the same tool that the Printrbot used when I built that although the version I had caused quite a bit of cutting edge bleeding if you know what I mean.  

KISSlicer makes a GCODE file of the small piece I will refer to as BoxBall.

Ok, time to print!  Unfortunately, the first print did not have adequate infill or a raft to build pillar supports for the ball that was suspended inside the box.  What does all that mean?  Infill is how solid the piece is.  KISSlicer can put air pockets inside to save both time and material.  It works great when you have the right amount.  A raft is a little later of plastic, like a raft in the water, to build upon.  This allows layers of plastic to adhere more easily.  In my first try, KISSlicer made a pillar to support the floating ball.  It was working great till the printhead tipped it over on its side.  In the second attempt, I added a raft in TinkerCad and increased the infill from 33.3% to 50%.  As you can see, the arches of the side cutouts were given a pillar to support them.  This was done because I had enabled this option in KISSlicer.

Box Ball final print with movies below.

3D Scanning

3D scanning was an adventure into Autodesk's 123D Catch. I loaded the application on to my iPhone 5 and searched for something to scan.  This application is free and relies on a server farm to process images you feed it from your phone.  The process can take from minutes to hours and is a quick way to get some interesting results.  It should be noted that this system can be inaccurate in the details, but gives some amazing and startling captures.  The application walks you through the process and uses the phones sensors to detect which orientations you need and are on as you circle the subject.  I used about 35 photos for this dragon.

Step 1: Start a scan.  The software is becoming more and more social as it incorporates other users scans and an interface for you to share your scans.

Step 2: Take a series of pictures as you circle the subject.  The subject should be uniformly lit and have a contrasting background.  In this picture you can see that I do not have a contrasting background.  I added white paper all over the place in a subsequent attempt.

Step 3: Review your photos and delete or retake any that did not work out. It is unclear to me what qualifies as a bad shot other than poor lighting or angle.

Step 4: Then submit your photos to the server farm and await your results!

Here is my "stage" for taking the dragon photos.  Quick and dirty and Yoda did not bother me.

Here are two scans, the dragon and my hand.  Pretty cool stuff!

Dragon

The Dragon bombed…not sure what went wrong.  It blended the entire stage.  I think placing the dragon on an outside table in the sun would be better…at noon!  Wait!  As it turns out…there is lots of scatter pickup from the paper "stage" and i was looking at it from the bottom.  Ta da!  You will have to maneuver the model to flip it over if you click on the 3D manipulate box.

Handy Tools

Make Something Big!

This week we were charged with making something big.  Big being a relative term, I decided to interpret that as the size of a piece of furniture.  I will make some bar stools on the ShopBot.  The ShopBot (at right) is as big as a small car and likes to work with sheets of plywood or similar dimensioned material. I first had to schedule a 3 hour class at the TechShop to use the tool.  This was tough, but finally found a class in San Jose, CA (2 hours away).  I decided to book the ShopBot for right after the class in San Jose, since the SF TechShop was tied up until early April.  I guess Silicon valley likes working with small stuff and not big stuff, the ShopBot there is wide open for scheduling!

 The work flow for my design is as follows:

1. Come up with a low resolution prototype in a tool I am familiar with: I sketched it on paper. Low resolution means, very little of my time from concept to version 0.01 of the idea.  
2. Now that I have the idea, formalize it a bit more and do some research on stools.  What are the standard heights etc.  These stools need to fit under a bar height counter.  What is bar height, what is the height of a bar stool etc. 
3. Search for examples.  ShopBot's website has some great designs to serve as examples.  My design was created by taking an existing design and tweaking it to my needs.  This design was tweaked using ShopBot's software PartWorks directly.  Partworks is a very straightforward tool which is used to take a design and create the path files necessary for the ShopBot to make it's cuts.  The design could have been created in many tools including Inkscape and Adobe Illustrator, AutoDesk Inventor, etc.

ShopBot's Partworks Software in Action

Partworks is a fairly intuitive tool and with the mouse over help I was well on my way. The creation of the paths was a click away.  I resized the stools to take into account the thickness of the plywood and also increased their height and changed their edge look a bit.

The ShopBot Tool Interface Software

Used to calibrate the X-Y-Z.  This will also set the speed of the mill etc. There is quite a bit more to this tool that I would like to learn.

Loading the Plywood

I set the plywood with brass screws onto the cutting surface.  There were two sacrificial boards in place so I felt comfortable moving forward and leaving them there to protect the machines surface.  This turned out to be a mistake.  The boards were old and warped slightly which gave me an uneven surface.  We use brass since it is a soft metal the will be more gentle on the end mill if I accidentally cut into them.  I used a screw in each corner and on the long sides in the middle.  Next time, I will use more screws and put in my own sacrificial board.

The Cut Begins...

Mesmerizing.  Looks very good on top.  On the bottom, I should have used more screws and had it penetrate further.   The tool interface sends out a warning about depth of cut, but now I know it is ok to go a few extra hundreths of an inch, or even more, to get a good finish on both sides.

The Cut Continues...

Assembly of the Stools

Since the ShopBot did not fully penetrate the material in all locations due to uneven surfaces, the assembly required some sanding and filing to get the slots to be the right thickness.  Once this was done the pieces came together nicely! I am considering a new stool design and also seeing how the Twing from an earlier week might come out at a large scale.

"Design a 3D mold, machine it, and cast parts from it" 

Well ok then!  Steps 1 and 2 were done with Tinkercad and the 2Bot respectively.  I cover Tinkercad during week 5 and 3D printing here.  The 2Bot is a neat milling machine for Expanded Polystyrene (EPS), higher density foams, machinst wax and I am doing research with mushroom foam.

Tinkercad and the 3D mold

The 2Bot uses a tray that suspends the material in two tension clamps held right and left.  It defines the Z top by calculating the middle of the clamps less one half of the height of the material.  Not an ideal situation as no "bit drop" z calibration is available.  I am sure this is to keep things simple, but results in errors.   It does allow you to flip the material for two sided cutting as well as scaling the model to any size you want by calculating the material needed.  This is all done in software.  All path calcs are also done in software.

The model machined really well, but I left it a bit tight on the bottom.  You can also notice all the little "hairs" hanging off of everything.  This was done at a resolution of 0.01" with a 1/8" bit.  I am uncertain how the 2Bot will handle a different bit as it is a unique design.

As you can see in the below picture, after applying a little heat with a heat gun the "hairs" melted and the surface became more sealed.  You can also see where i manually expanded the bottom of the mold. I learned that if you do a two-sided cut (a small checkbox in the 2Bot software) it will ask you to flip the model to do the opposite side.  I did not need the other side cut and did not want it to cut out the mold (see model above in TinkerCad).  If you select two sided and never flip it will not cut it out.

Getting the work area ready...

1. Prep area on table:
   1. Oomoo 25
   2. Mixing cups
   3. Gloves
   4. Protect work surface
   5. Face mask
   6. Demolding release agent
   7. Mold (EPS microscope)
   8. Mixing sticks
2. Since I am working outside and there is a breeze, move mold inside on cardboard after pour (pour inside?)
   1. I may move the whole operation inside as the wind is annoying and may impact the pour.
3. Mix the Oomoo part A and part B by hand in containers with a "square" edged device per tech video
4. Measure out into cups in equal parts and mix a bit more
5. Combine cups into one or the other and mix some more
6. Pour mixed oomoo into another clean cup
7. Pour into mold in very fine stream very slow and let it level itself out (this is where the wind will mess stuff up).
8. Oomoo 25 sets in 75 minutes and Oomoo 30 sets in 6 hours.  I will use Oomoo 25 first to see how it goes.
9. Demold after 75 minutes
10. All of this is at ~73 degrees Farenheit

For Oomoo 25 I will not use the vacuum chamber.  Hopefully I can recast the positive if there are lots of bubbles.  I spoke to Smooth On Tech services and they said it should be fine except that the Oomoo 25 will creep into the little holes in the EPS.  I have applied heat to the surface which has melted and sealed the positive, but it is still rough.  I also did not make the mold big enough along the bottom of the microscope and carved it out by hand resulting in a rough top.  Hopefully this will not impact the actual negative.  I will also use mold release agent per technician's recommendation…here we go!

So, all setup and ready to go in my small workshop.  I decided to clear it out and get a windless area to pour.  Good idea as the wind picked up!  I did not have the recommended square steel stirrers as in the Smooth on videos and used wooden tongue depressors that the lab had.  While cleanup will be a pain with the metal stirrers I can see whay they recommend them.  The viscosity was very stiff at first and required quite a bit of careful stiring.  I was right at the edge of the temperature guidelines and maybe a bit under.  As I poured the Oomoo it became clear I was fighting a 15 minute window, again perhaps due to the temperature.  It came out in ribbons.  I was concerned I would not have enough, but it turned out to be the perfect amount.  Once I was done pouring I tapped the sides and bottom to release bubbles.  I didn't think any would come out in the thick liquid, but sure enough they started rising to the surface.  Now the 75 minute clock is ticking which I will probaby extend to 90 minutes. Since I am doing this at home, and do not have easy access to the lab and milling machines I hope this works out.  While I wait I am starting a composite bowl in week 11.

Releasing the Oomoo from the mold…

It was rather difficult to get the Oomoo out.  I ended up carefully cutting the sides of the mold.  The Oomoo was very sticky even though it cured for 90 minutes.  In the end, I was able to carefully release the mold and retain the positive, albeit without the walls.  There are many bubbles and next time I will use steel stirers as well as a shallow mixing container.  I used plastic cups the Lab had and it was challenging.

And on to Hydrostone…

I read up on this plaster/cement stuff and found a great video on technique. I decided to try out using soapy water in the mold to release and reduce air bubbles in the hydrostone.  I did also pre spray the Smooth On mold release.  I also decided to mix in the plastic bag as in the video.  I carefully weighed 100 parts Hydrostone to 32 parts water and ended up mixing 300 grams of the Hydrostone.  It seemed to come out a little thick but filled in the mold nicely.  I then scraped off the surface to get it flush with the top.  I think in a future casting I will build a border so it can be hung on the wall.  I am waiting for the Hydrostone to set to see the results.

Note that the Spray n' Wash is not that, I am using that bottle to hold soapy water made with a generous helping of dish soap.

Wow!  Demolding a success!

I had no idea things went so quick with the Hydrostone.  I remember Neil saying it is quick, you don't have to wait over night etc. etc.  Upon digging and some research I learned it sets enough to demold in just 30 minutes!  This entire process went slowly for me, on purpose, so that I didn't miss a step.  

I scraped the edges of the final casting with some aluminum scrap and am quite impressed.  No bubbles evident I guess as a result to the soapy water.

Some things I would do differently next time:

1. The EPS seemed to work fine and is very quick to mill.  I will give more clearance to the edges and possibly apply Styrospray 1000 to the EPS before making the silicone mold.  This will give it a very smooth finish, but adds a step.
2. I would like to mill something with the wax.
3. The Oomoo needs a better stirer than a tongue depressed and a shallow bowl.  This will help with the bubbles.
4. I would change nothing with the Hydrostone…the soapy water and the bag mixing is the way to go!