Tuesday, November 30, 2010

Less than 2 days left...

I really should be spending this time soldering bits of wire together, but I'm taking this time instead to update whoever reads this on what in the world we've been up to this month. I will give a brief summary of each week leading up to now, give a detailed description of our final project and where we currently are, and conclude with a summary of what I know still has to be done.

Oct. 28: We showed our plan of what wanted to do - a wall of moving panels - as well as the inspiration behind that. At that point, the motion was created by solenoids moving each panel up and down. We had a small Merkur-set gadget with two stacked solenoids that moved one panel in a set of five up and down in response to different sensors. We also had LEDs on the side of the gadget that lit up in response to some of the sensors. We decided that solenoids seemed too weak and we would need far too many of them, so we started researching different mechanisms of movement.

Nov. 4: This time, our gadget consisted of several rods of shiny spinning panels, powered by gears and servos. Pat showed a rendering of what we want the wall to look like eventually. Dan and Betsy showed the different natural-looking tessellation patterns they had found and cut for possible panel shapes.

Nov. 11: I can't tell much from personal experience, since most of the work was done Wednesday night and I got horribly sick then and had to skip class the next day to recover. But, the rest of the group created a large (~7ft tall by 4ft wide) module with several strips of chipboard panels, spray painted silver and hot-glued to a metal rod. The framework was constructed from stiff strips of metal, but was a little wobbly due to lack of accurate measurements for the corner braces. The strips of panels were spun at the top by stepper motors controlled by Arduinos. I haven't watched the video for that class yet, but apparently their presentation went quite well, and Max thought we were going in a great direction when I talked to him the next day.

Nov. 18: We experimented with different methods of stringing the piles together to see which material gave us the best effect. We wanted a cascading effect. After trying fishing wire, cable, and different methods of attaching ribbon, we decided that ribbon strung through the tiles gave us the best effect. It didn't seem like much progress from last week to me, as I knew we had to really get on this stuff, but I felt like I couldn't say much after not being involved at all the previous week. We also figured out that we wanted to vacuum form the plastic tiles. After we figured that out, and realized how much time that took up, there was a sudden realization of how screwed we potentially were. During class, Simon went out to the site and saw the installation place...which was completely different from what we requested, so I'm not even sure if our opinion even mattered.

During the following weeks, much of the group spent time routing the woodboard needed for the vacuum forming, vacuum forming the plastic tiles, cutting out the tiles, poking holes for the ribbon, and building the gigantic framework. I soldered a row of circuits for the motors somewhere in that madness, but most of the work was done over Thanksgiving weekend, and unfortunately I had a family trip planned since the beginning of the semester from Wednesday afternoon until Sunday afternoon, so I couldn't do much. As soon as I got back, though, I rushed to the Architecture building and started washing tiles off (that was what I was told to do). We've been working around the clock ever since I got back, it seems.

Our final project has a giant 12ft wide by 9ft tall metal framework, with 33 strings of 14 tiles spaced equally along the strings and along the framework. Each string is connected to a stepper motor via a metal strip, which spins the string of tiles as directed by Arduinos. The Arduinos spin the tiles according to sensory information from 17 PIR motion sensors along the bottom of the framework.

I still have to solder several wires, and we still have to test everything, and we have to assemble our thing onsite, which will have to happen tomorrow. We still have the rest of tonight and much of tomorrow for circuit-building and assembly of parts that can be assembled offsite. I'm not sure how many circuits we have left to build, but most of the stuff has been for motors so far and I don't know where Chris is going to put the motion sensors in all of this. I suppose I'll just start soldering stuff and trust that he knows what he's doing.

Thursday, October 28, 2010

Thoughts on the current presentation format

At the start of every class, we present our projects. An order is chosen, and the first group presents their stuff, which takes maybe 5-20 minutes. Then the floor is opened up for questions.

The questions, I think, are supposed to be from the students as well as the professors. However, I've noticed a trend. Once the professors start critiquing the group, they continue on for quite some time, bouncing off each other. If students have questions or points while the profs are talking, the hands sometimes go unnoticed, and the points are forgotten or given up if the student is not extremely patient and/or assertive.

Also, it is difficult to stay focused on any one thing for quite some time. We have, as of right now, used 3 hours, half, of the class time today for presentations and critiques. I'm not saying that's a waste of time. I'm saying that I can usually keep alert through one presentation, but I start zoning out during the long critiques if there isn't a quick bouncing around of questions and ideas from several people. But several presentations, all in a row? Goodness, we're all zombies by the end of it. We need a break, maybe multiple breaks.

Ah, well. Food time.

Progress and Exciting Advancements in Documentation

Well it is nearly 3am and all that is left to do is the coding of the Arduino gadget monster thing. Since Chris is basically nocturnal, and is also the Arduino god of our group, we're just going to leave that to him. He's currently chatting with someone in another group who just got some radio Arduino communication thing to work.

I have exciting news from the land of documentation - although I have yet to start a continuous Google Doc or any sort of archive of emails (outside my inbox) or uploading stuff periodically to the CTools page, I did create a couple of pages in order to promote group uploading and publishing our work to the outside world. If you want to follow our group specifically, the links are below.

Flickr: http://www.flickr.com/photos/55236033@N08/ (no uploads yet since I have to grab them from facebook since those have a much smaller file size)

YouTube: http://www.youtube.com/user/ducktrap2 (3 videos so far)

Also I have to say thanks to Prof. Max Shtein who is letting me use his camera for the semester. Most of the pictures were/will be taken from my own camera, but some pictures were/will be taken on the Shtein Cam, such as when my batteries die or when it's just more convenient. Videos were/will be taken exclusively on the Shtein Cam since that camera is far better at videos than my older model. Both are slightly beat-up Canon point-and-shoot digital cameras. Mine is an older model with more settings and zoom, but the file sizes of anything are humongous. The Shtein Cam appears to be a newer model will fewer functions but much better video and possibly smaller file sizes.

The only thing I have to finish for tomorrow is the budget and the presentation. We probably should've decided to meet early tomorrow so we could go over the presentation. oh well...I'll just upload everything to CTools and send out a notification.

Tuesday, October 26, 2010

Tuesday Oct. 26 - 37 days left

We currently have the entire team gathered in DL1. It seems like no one has actually gotten their things to work in the time being - well, Chris has the temperature and humidity sensors working, but Pat and Betsy are having trouble finding the right code for their devices, and Dan's still trying to get the Arduino software to work on his Mac.

Chris had an idea for the tile wall where we'd have different "resolutions" by using different shapes. If we used triangles and hexagons like the first project, we could do "low resolution" actions with the hexagons and "high resolution" actions with the triangles. Theoretically we could have two types of motion going on, one level with the hexagons (sets of 6 triangles moving together) and then another level on top of that with the triangles moving individually.

He also had another idea inspired by our 489 group's chat with Max Shtein earlier today. Shtein brought up some windows that have groves in the glass/acrylic that allow light to travel along the straight grooves when there's an LED at the start of the groove. Chris had the idea of having curvy grooves so the light would travel in curves.

For Thursday, we need an operational prototype and a $500 budget. We're thinking that we want to make a simple interactive surface for this week - maybe a couple of chipboard panels and some sensors, and have the panels (attached with actuators) move different ways according to different stimuli. Maybe we could stick some LEDs on there too, and make them go off when the other sensors do something.

Now we are looking for a tessellation/tile pattern found in nature that is interesting and would appeal to someone's sense of wonder.

Looks like we're breaking up for the evening soon. The goal is to get everything working by the time we meet tomorrow. Chris, Dan, and Simon will get together to work on code at 6pm tomorrow. Betsy and Pat will work on finding a good pattern for our final project and getting some renderings of it.

First Thoughts on the Final Project

Goodness, I haven't updated this thing in forever - sometimes I wonder why Shtein wanted me in particular to be the Documentation Queen. I mean, I love organizing, I love taking pictures, but I'm not entirely sure I'm responsible enough and disciplined enough to handle this.

Two weeks ago we got the assignment to design our teams and figure out how we would work for the rest of the semester. That went pretty well, I thought. I'm not sure how it's helped us so far, but we also haven't worked much on the project yet. We also got a tour of the old Pfizer building, now the North Campus Research Complex (NCRC), where we'll set up our projects at the end of the 6 weeks. I kept having a weird sense of deja vu from when my AP Chem class toured it when it was still Pfizer back in high school, but it was cool seeing all the places where we might put our thing. The problem was, we really had no idea what we were doing then, so we didn't really know what to look for. At the end of class we got together in our groups and figured out a $200 budget, which at this point just consisted of different Arduino gadgets we could play with. Most of the gadgets we ordered were sensors, but we also got a data logger and some actuators.

Last Thursday we got the assignment of what we were actually supposed to build, which was about as vague as any of the assignments we've received for this class - build a biomimetic smart surface. I guess we knew that was what we were doing from the start, but it seemed like people were looking for an actual assignment before they could start thinking about it. After lunch we took a tour of the Botanical Gardens, where we got a close-up look of a lot of very cool plants and learned about different ways plants work - different forms of biomimicry we might be able to use in our project.

On Sunday we met up for a short time to distribute the Arduino gadgets and brainstorm ideas. We decided that we would work on getting the gadgets working and see what we could do with them, then meet again Tuesday night. We came up with the idea of a double layer of moving tiles that would shift to interact with people and let light in in different patterns (the wall of tiles would be up against one of the glass walls, so natural light will be coming in). That's where we left off...now I'll start a new entry for what happens tonight.

Thursday, October 7, 2010

A Better sort of Sun Visor

Thursday, Sept. 30, about two weeks from now, we were given a new assignment: Build a 2-week surface with 2 axes of motion that can track light, align with light, and indicate when it's aligned. For this Thursday, today, we have to have an operational surface that meets those goals, a presentation featuring screenshots from Digital Project (a not-very-intuitive computer modeling program), and documented biomimetic strategies as potential solutions to the challenge.

Last Monday, we had a brainstorming session where we tried to think of a good application for a solar tracking device. After several strange ideas such as a window with gas inside that changes color to block out sun, a giant ant-burning laser thing, and an alarm clock that attacks you to wake you up, I raised the issue of the sun visors in cars not being sufficient to block the blinding sun at certain times of the day. The other people that were there (Pat, Chris, and Betsy) all agreed that that was actually a pretty good idea, or at least more viable than attacking alarm clocks and windows full of gas.

Last Tuesday, we met at The Desk - a collection of four work tables near Pat's desk in the architecture studio to brainstorm mechanisms and structure of this device. After trying to draw out and visualize the inside of a car, Simon brought his car around to the back of the Architecture building and we all spent a good amount of time chilling in it, looking at how the current car visor design worked, discussing what our device would have to accomplish, and sharing stories. After we seemed to know what we were doing, we headed back to the desk and started to sketch stuff out.

Last Wednesday afternoon, Simon sent out a text saying he and Pat were rethinking the design, but later, when we met that evening, nothing had changed. We built a kind of ugly frame that our visor ran along, with a motor on either end to pull it. Once again, we ran into the problem of motors being too weak and so they couldn't really pull the visor-car along the track.

When the professors saw everyone's projects last Thursday, Karl was surprised that we didn't all have working solar detector things, since he thought that was all they asked for. Maybe they should've made that more clear after interrogating every group about context and application each week for the previous assignment, because the result was that everyone had an application but their projects only kind of worked at best. While about half our group was in the sewing workshop, including me, the profs gave my group feedback. Apparently the general gist of it was to make things more "elegant". It's just so confusing and frustrating when for 3 weeks they say "WHAT'S THE CONTEXT? YOU NEED CONTEXT!" and then we give them context and application and everything and then they're like "what? context? we just wanted a pretty solar tracker!"

So we met in DL1 for a while and figured out how to make it more elegant (enclose the track, hide the wires, round the edges on the visor, make the track prettier) and how to eliminate the strings in the mechanism, since we just don't seem to have great luck with strings and fishing line and it's an ugly solution anyway (we now have a toothed gear pushing and pulling a pole which pushes and pulls the visor to adjust the angle).

Now, once again, we're meeting on a Wednesday night with a ton left to do still. We have a better-looking track which we are apparently going to hang from the ceiling - not sure how we'll manage that. Chris bought stepper motors and smaller servos since John's gigantic servos won't fit on our little track. Chris has the Arduino program ready to go, we just have to assemble the car (the piece that runs along the track) with the Arduino and everything on top of it and solder whatever needs to be soldered. I mean, yes, we're further along tonight than we have been previous nights, but Pat isn't going to get here until 7:30 and I think he wanted to recut everything again. I'm not sure why, when our "prototype" track seems perfectly fine, but as long as we eventually get everything assembled and working I think it'll be fine.

Now more people are here. Back to work!

Thursday, September 30, 2010


I'm currently too tired to update you start to finish on what we've done. Once again, most of our group is collected in the Architecture studio stringing strings through holes. I worry because it is very similar to the method we used before, and that method didn't work too well. The others seem pretty confident though.

The surface looks a lot better, more professional. It is all done in acrylic now, and spray painted, with hinges...so many little hinges with so many little nuts and bolts...of course if this was ever mass-produced robot arms would likely screw everything in precisely but it's very tedious to do it by hand. We've also solved the problem of the flaps not closing by using strips of elastic to hold the flaps flat. We've been using hot glue and super glue to attach a lot of stuff to save on time (vs. sewing or other methods).

Okay we just ran into a problem - the elastic holds the flaps too tightly and so we need a lot of force to pull the flaps closed.

Also - we haven't even connected the servos yet. All of the individual pieces work by themselves - the surface looks cool and does what we want it to do, the Arduino program works when it's not actually tied into the system, but we can't seem to figure out the right mechanics to get the Arduino program to move the flaps the way we want them to move.

A note on context - we are thinking of a roof of an outside pavilion now, like the farmers market shelter, that moves depending on activity and depending on weather. With no people, the flaps would form pyramids. With people, if it was raining or snowing, the flaps would form a flat roof. With people, if it was sunny, the tilt of the flaps would depend on the time of day - mostly shielding at noon but mostly open at earlier or later times. Chris has rigged up an awesome light and temperature and motion sensing program. We tried to get a humidity sensor in there as well, but Chris couldn't get it to work with the program. The temperature and humidity readings seemed really off - extreme negative values.

Okay - sorry I keep coming back to this after long periods of working on the project - we now have the strings all threaded through and we have to figure out how to do the "drive string" and how to make the servos work.

arrgh...the strings broke.