Methodical Madness Caps the Semester

Well, this sucks. It's 1:20AM, December 19, and I am leaving for home in... 9 hours, driving 10 hours. I will be awake, in a car, driving, longer than I will sleep tonight. It's going to be terrible. As such, I can't write nearly as much as I would like to write for this final post.

I will say that this semester, this class, has been the most valuable class that I have taken at Michigan. The principles I learned, the people I met, the discussions I had... it was life-changing. Max, John, and Karl were right, not surprisingly. They are truly brilliant, in their own ways.

I came into this class resigned to the idea that I was going to be stuck a metallurgist as an engineer. I knew this wasn't actually the case, but it was difficult not to think so.

I will miss this class. I will hopefully keep in contact with as many of the people as possible. I know it won't happen with most, but the people I met were all rather great in their own ways. Even those I did not get along with quite well, I had great respect for.

For those considering taking this course next year, and thinking it may be too much work - do it. It may be a lot of work, but it is worth it. Trust me.

Now, for a shameless plug!

I am continuing with another blog: Concrete Theory

Hopefully at some point it will be worth the read.

The Final Product: The iWall

Well, it's over, and it's been fun, so I guess it is time to do a quick rundown of what our SmartSurface was. The nice part is, I produced the blurb for our group, so I can kick off with that!

The iWall Modular Light-Filtration System

The wall you see here is a modular, scalable, self-adjusting multi-aperture window, designed to optimize light flow into a room by manipulating the position of rotating wood panels. You could call it an 'automatic venetian blind', as that is the gist of the purpose. The system consists of 30 individual modules, with one panel per module, which allow a variable amount of light to pass through. The structural elements are built modularly, but circuitry elements - such as wiring - are set up in a non-modular fashion. Individual module apertures open and close in reaction to the intensity of the light that passes through the wall as detected by light-dependent resistors (LDRs), seen on the projection side, holding a relatively constant light level on the ‘interior’ of the wall at all times - this gives 'heliotropic' and 'smart' characteristics to the wall - 'surface', fitting it into the premise of the SmartSurfaces course. The actual actuation is by a grid of servos powered externally and controlled through inputs from an Arduino Mega and the LDRs. The materials used in the assembly of the individual modules were 3/16" cast acrylic cut by laser and 3/16" 6061 aluminum sheet cut by water jet.

The design for the iWall was described in a previous post, so here I will mostly detail the construction. Fortunately, Pete was kind enough to provide us with a large number of photos taken while we were assembling the system in order to do just that!

I will not cover the electronics, as you can likely check out Neil's blog for the information on those. Suffice it to say, they worked, if not necessarily perfectly - no fault of Neil's, it was just 'quantum weirdness', as John put it.

First, our final product:

A Module Rendered:

Module rendering, with panel rotation shown.

Exploded to Show Pieces:

Fairly simple, each module only has 13 pieces:

8 Acrylic Rectangles

4 Aluminum triangles

1 Servo for operation

Step 1: Produce Module Pieces

All acrylic is laser cut.

All aluminum is cut with the water jet.

Step 2: Lineup

All of the pieces lined up pre-assembly.

Step 3: Assembly

Connect aluminum to acrylic cross-bars. Tight fit!

Put sides on triangles and glue sides together.

Took a little brute force, the fit wasn't exact.

Place in rack, attach third side to each half.

After this step, panel is placed in center and two halves are glued together, producing a full module. The glue had to set for several hours before it was structurally ready.

Step 4: Build a Wall

The modules simply bolt together, and wiring is laced down the sides. The setup process for the wall built from individual modules was a couple of hours, which was impressive but not altogether surprising. We did expect something to go wrong, but nothing really did until everything was built and wired - there was a circuitry issue (see Quantum Weirdness in introduction).

Step 5: Complete!

The reason the iWall worked so well was the attention to detail put into perfecting the design of each of the individual pieces, and into optimizing the production process of the modules. From final design to gallery piece was about three weeks, which shows how quickly our group could work once we had a concrete idea to unite around. Everyone put in a ton of hours over the last couple of weeks of the project in order to ensure that we got results, and our wall was a resounding success! It can be seen in the Architecture Gallery in the Taubman School of Art and Architecture for the next few days, and then will be moving somewhere in the Duderstadt Library after that point.

Methodical Madness Makes an Update

Our group is humming along, and with Thanksgiving over we have been back in the saddle, moving towards completion of the final project at a brisk pace.

Currently, we are divided into three task groups, with each team having their respective tasks.

Build Team: Brieana, Josiah, Pete, Taylor

Coding & Circuitry: Neil

Documentation & Logistics: Me

I feel that this division plays well to the strengths of each of the teammates, and when I am not writing I help the build team as an extra pair of hands.

Module Design

As described previously, we are making a modular light-adjustment system. The modules are structural, and made of 8 pieces of acrylic with 4 triangular pieces of aluminum for bracing and aesthetics. I have included an image of the module in an assembled state.

Gallery Design

Our gallery piece will eventually look like this:

With light in front and LDRs behind the wall.

Project Status

Currently, we are producing the individual pieces of the module, and assembly will progress for the next few days. I will make a rather large informative update with step-by-step photography at some point in the near future.

Concrete Theory Has Found a New Home

My post line titled Concrete Theory has moved on to Here. I feel that it will evolve far beyond this course, and leaving it constricted to a course blog is just wrong. Also, I like the WordPress format. I haven't taken a whole lot of time to customize it, but perhaps Thanksgiving Break will allow me to start.

Methodical Madness is On Track

So, our final project is humming along at this point. We had a productive session on Friday, and plan to have quite a bit done by the time break hits Wednesday night. We constructed and tested three modules with LDR light detection behind them to determine variance of values. We also built some simple code that can get the job done, which will get more complex as time goes on.

Unfortunately, I'm not really able to directly contribute to much this week. I'm flying out tomorrow evening for an interview, and won't be back until Monday evening. This leaves Tuesday as my only available day. As such, I have moved back to coding from working with our build team, since I don't have to be as present and can get more done in less time. I just don't have the hands-on skills with the mediums we are working with in order to contribute significantly. It happens.

It appears everyone's blogging has fallen off at this point. I really don't feel so bad that I haven't done as much on my blog in recent weeks, after looking around at other peoples' blogs.

Methodical Madness Makes Progress

Our project still has issues, there is still a lack of excitement, but we have gotten to the point where we have a direction and we are just saying 'go'. It is more important at this point to have a finished product than it is to have a better idea.

Our plan is to produce a wall with multiple apertures that allow light to pass through in varying amounts in order to optimize the amount of light on one side of the wall. In theory, the wall would be external, probably south-facing. For the purposes of the gallery piece, we have to rely on a large lamp to produce the necessary light. The wall will adjust to light fluctuations externally over time, as well as to internal commands to increase and decrease light.

The apertures open and close by an array of servos commanded at their core by an Arduino Mega. The servos will require significantly more power than can be provided by one Mega, so there will be an external power source as well. It would be possible to power the entire thing with an array of solar panels in an outdoor environment, but in the gallery we will have to use grid-based power. Light detection will be handled by a color detector and an array of LDRs.

Overall, I feel that the project fits the requirements well for being a heliotropic smart surface.

Methodical Madness Returns

So, I just had a week-long hiatus of pure frustration with this class and school in general. The constant modification of the final project has taken its toll on me, as well as several extracurricular influences, and my motivation for the class that has been there all semester is basically gone. There is hope, however, in the latest iteration of our final project idea.

The current direction we are taking for the final project, which is likely the final direction, is having a variable shading system, which will allow a constant amount of light in from the sun. The design will operate by guessing positions based on intensity readings until it is correct, and then possibly (ideally) recording the proper positions for future use. The algorithm for this is extremely complex, and inspired by the concept of the SMS adaptive learning in the Hot Strip Mill at ArcelorMittal Riverdale, where I interned for the past two summers.

I feel that this plays into my interests well, and will make for a great design project. More detail will be given at some point as things actually start moving along.