Wednesday, December 10, 2008

We waited...

My Advanced Dynamics and Controls class was supposed to have a final this morning, but nobody showed up to administer it. The Prof was out of town, his secretary was not in, and the TA was MIA. After a while we decided to leave...but we did leave a note:


Kalman filter that.

Wednesday, November 26, 2008

Camera Purge

One of the benefits of having a video camera with a 30GB hard drive is that you can store 7 hours of video, but that also means that people leave a lot of raw video on it. (I say "people" because it's been a while since I've actually shot video with my own camera.) So to test my video converter and new computer, I took some of the left over clips from the Cap Kart project and clipped them together.

If you were wondering what NASA-like engineering procedures are employed in my summer projects, these should clear things up. Happy Thanksgiving!





Thursday, November 20, 2008

New Machine



You know that part in Apollo 13 where they finally shut down the command module and transfer all control to the LEM? That's what it felt like to move all my stuff to a new laptop. In any case, it's done.

Wednesday, November 12, 2008

The End of an Era

Yesterday, I decided that the era of the Dell Latitude D800 that I've run into the ground since June of 2004 was over. I am a mechanical engineer, and relatively critical of the computer-savvy in general, so for me to define an era by my laptop says something, I think. Maybe it's a symbolic gesture and a sign that I'm eager to come to terms with other changes, with a new role as a graduate student, with the fact that I honestly forgot my brother's age this week. In any case, I thought I should say a few words...

Not that the D800 is dead; in fact, I'm writing this post with it. My new laptop (a Latitude E6400, since this one outlasted the entire D generation) won't be here for some weeks and I'm sure this one will survive way past then. Once, I was riding a train with this laptop out and tried to pour myself a drink, only to have a bumpy northeast corridor rail cause me to spill it into the keyboard. Once, I left my window wide open to cool down my room while I went to take a shower and a freak thunderstorm soaked everything on my desk, including this laptop. Once, I tripped on the power cord and pulled this laptop off a table, causing it to fall on its corner on the concrete floor of a machine shop. The plastic is cracked, the hinges are shot, and the latch has been replaced with a piece of duct tape; but it still starts up quickly, all 1920x1200 pixels still work, and it still eats SolidWorks files for breakfast. If I had a dollar for every time this laptop has gotten either me or somebody else out of a jam by virtue of its uncompromising functionality, I'd be able to afford a much fancier successor.

More importantly, though, it is a digital archive of the last 54 months of "me," my entire undergraduate career and more. It's spanned so many of my endeavors, including the birth of some new traditions: the "crazy summer project," the "long-weekend-coding-project," and the "media machine." The experiences and people it's seen have changed my life and the media it's captured and stored are priceless to me. Even if I am (poorly) disguising my acceptance of change on many fronts with this new "flagship tool," I'm going to use the resilience of my old machine as a symbol for the many things I want to carry over in the process.

Tuesday, November 11, 2008

30,000 Words per Second

If a picture is worth 1,000 words, it follows that video is worth roughly 30,000 words per second, right? (Which means that my S.B. thesis could have been summarized with approximately 7/10ths of a second of well-thought-out video.) I have been capturing my projects on video since I got my first DV camera in 11th grade. Short of a live demonstration, it's really the best way to show somebody what I've been working on. In fact, contrary to the rumors that I have turned into an Apple fanatic, the main reason I bought an iPod Touch is because I can now show people my pictures and video without taking out my laptop. That, and episodes of House look better on it than on TV.

But there are so many video formats: MPEG-2 from my camera, MOV and MPEG4 for iTunes, H.264, WMV for editing and showing my Windows friends. And occassionally I like to edit frame-by-frame with my own software. I have fought with codecs for years and previously used at least three different third-party tools to do video conversion. No more! The long weekend inspired me to take on one of my rare pure-software projects. (Last time this happened, I wrote a program to create photomosaics.) I think there is still a CS person hiding in the back of my head. Anyway, I present SCV, which, if you want, stands for "Shane Converts Video."


It's an MPlayer/MEncoder front-end. MEncoder is a great piece of software that was suggested to me for my video editing needs since it can convert between virtually any format. The one downside, for me anyway, was that it is made for/by Linux losers with nothing better to do than run things from the command line. Hear me out: In the 21st century, we have GUIs. So this simple VB (yes, VB) program takes care of all the command line switching and lets me get on with my life.

Ooooooh, code. It can handle the formats I work with most often (WMV, MPEG) in different flavors (AVI, MP4). It can actually make video that works on an iPod. It can both extract frames to JPEGs and recombine JPEG folders into video. It can trim clips. And it can strip/merge audio. I'm sure there's tons more I could do with it, but that's all I need and my software projects are pretty much driven by necessity. If you want to play around with the source or somehow think the executable will actually run on your machine (needs .NET Framwork 2.0 and a working version of MPlayer, at least), here they are:

http://web.mit.edu/scolton/www/scv.zip
http://web.mit.edu/scolton/www/scv.exe

And if you want to see why I spent part of a four day weekend coding, check out my TechTV collection:

http://scolton.techtv.mit.edu

Sunday, November 9, 2008

Take a Lesson from Jesse James

Wow, this is starting to turn into a Discovery Channel show blog. I promise the next few posts will have nothing to do with Discovery. But I had to get this one out: Last week's episode of Prototype This was pretty cool in that the team finally admitted defeat after they tried fairly unsuccessfully to make a 6-legged walking vehicle. It's much better than pretending that everything always works out okay, even when things seem impossible at 12AM on the night before the deadline. But, I have a suggestion for the prototypers on how to deal with prototype failures, which comes from a show the precedes theirs by quite a bit, and here it is:



I wish they'd go into a bit more technical detail on why they failed instead of just blaming the overweight chassis. But I caught enough hints in the video and dialog that I think I get it: First, these suck. They finally switched to something like a Sevcon MillipaK for motor control. Of course it's hard to blame the components when the designers have an infinite supply of money. Try doing 300A DC motor control on a real budget. Also, they bought some fancy batteries, but in the last few minutes you can see they gave up and are using lead-acids. And they very briefly mention "optically isolated" controls being an important fix. I'm pretty sure that all this points to massively large loads - far larger than they expected - which demolished the electronics system. Sounds familiar.

Thursday, October 30, 2008

Prototype That

In addition to Time Warp, the Discovery Channel has another new show called Prototype This. Although you definitely get the impression that Discovery is trying to ride the Mythbusters wave a bit with these new shows, this one features an interesting balance of mechanical and electrical/computer geeky hosts.

This week's prototype, a truck that could navigate above a traffic jam and slide sideways into an occupied parking space, featured a bit of technology that I've seen in action before: Mecanum wheels. Rather than try to explain how they work, a short video:


(Check out the new TechTV player!)

These pretty cool sideways-action wheels have been gaining popularity in the FIRST competition for some time. So, if you don't have the budget of a Discovery Channel prime-time show, you can obtain a set of these wheels at reasonable cost from FIRST supplier AndyMark. They only have small robot versions for now, but are working on a 21" version which I think would suffice for making your car slide into a parking space.

Most of my projects fall squarely in the "prototype" category - the first take at something or other. After that is the "beta" phase, which is an interesting in-between ground for working out the finishing touches. Then, the final mass-manufactured product release. I've seen a bit of the whole sequence, but most of the time I enjoy working in the early "whiteboard" stages.

Monday, October 20, 2008

On Simplicity

I love simple systems. I think somewhere in the process of building a 300-amp motor controller from scratch, we forgot how remarkably simple the electric go-kart really is. It took a while to get there, but in what I think is a good example of engineering system design, the individually complex pieces (most of which I would have preferred to buy off-the-shelf if they actually existed) together make something so easy to understand that it must work.

When the kart engages its regenerative braking circuit, there are only really two places for energy to be stored: in the moving mass, or in the capacitor. The batteries are out of the equation, since the main controller, in an act of almost absurd simplicity, cuts them off and shorts the motor terminals across the capacitor. Of course there is friction, but let's say we idealize by using a 54lb steel flywheel instead. ("That is one of the scariest things you've ever built." -Matt R.)

The only way to move energy from the mass to the capacitor is by energizing the motor's magnetic field, forcing current to flow as it acts like a generator. The flywheel slows down, the capacitor fills up, and along the way a sizable portion of the energy is dissipated in the resistance of the motor. Since we can measure current the whole time, the energy dissipated can be calculated. And guess what? It almost exactly matches the energy difference between the final value of the capacitor and the initial value of the flywheel, every time. The data is up on the site.


I'm not sure why I was surprised by this. There's a huge spinning disk, a huge capacitor, and a huge copper motor winding. There's no where else for 10kJ to go in a matter of seconds without destroy something, so that must be all there is to it. I'm pretty sure I thought this all through and came to a similar conclusion before we even started building. But after months of fiddling with power electronics nuances, it was wonderful to see the conceptual simplicity come to life and to see the data confirm that physics does indeed work.

Take-away: If I draw a resistor and a capacitor and ask people if that's a simple thing, they say yes. If I draw a go-kart with regenerative braking, they say no. Hopefully I've taught a few people to say they're the same.

Monday, October 13, 2008

Time Warp

Time Warp, the new Discovery Channel show that was partially filmed in the Edgerton Center (where I work often) at MIT, premiers this week. Their cameras are very nice, but a while back I wondered what kind of high speed work you could do on a budget. Turns out there are now a few consumer-level cameras that can shoot at relatively high speeds. One is the Casio EX-F1 (and its newer, cheaper cousin the EX-FH20). Here is a compilation of random things shot with a borrowed EX-F1 and finger in one day:


Thursday, October 9, 2008

Large Spinning Disks of...Aluminum?!

I was wondering how UPS had managed to alter the material properties of steel to make it so light. A minor setback, steel disks are now on their way. I hope somebody out there is designing something really cool that can be cut out of the inside material of left-over 14" aluminum disks.

Monday, October 6, 2008

Large Spinning Disks of Doom (or Steel)

An annoyingly blog-like title to start with. Aside from just testing out how this blog thing works, I will attempt to catch my readers (hah!) up on one of my current fun projects, an electric go-kart with a 110F ultracapacitor boost. (Think railgun-you-can-ride.) Working with a crazy - but very competent - group of high school students from the MIT Edgerton Center Summer Engineering Workshop, we built it over the summer and have had some fun driving it around campus:



By the way, I do this kind of stuff because I sometimes get bored with business-as-usual in my life as an engineering student. But this particular project is interesting because I think, as fun as it is, it also has significant serious research potential. Since we've already proven it is a fun ride, I'm thinking it might be time to satisfy academia with some more controlled experimentation. Hence, the spinning disks of doom:

These 11lb, 14" steel disks will provide rotational inertia directly to the electric motor on the kart, enough to simulate the effect 500lbs of kart + driver. And all without leaving the table in our shop. Not as much fun, but with the wireless data acquisition system it should rake in results fast enough to meet some paper deadlines. And even if the papers get rejected, guess what, now we have a sweet electric go-kart and five massive steel plates.

If you are wondering what magic MIT technology we use to produce custom-cut steel disks, it is called an abrasive waterjet. But in fact you can use one yourself. I sent the order out at 2:45PM and they were cut and shipped by 5:00PM. Can't beat that.