Wow! I see I’ve narrowly avoided going an entire year without posting — by 3 days. There’s been things to blog about, too. It’s just my nature to start projects and not finish them. Which leads us, in an ironic way, to this post. It’s time to throw myself the gauntlet again in an attempt to self-motivate by posting this article so that all my friends can mock me when I don’t finish.
The Raspberry Pi is a fascinating thing. It’s a tiny computer, the size of a credit card, for around $25 bucks. It’s British, and was started by some fellows from Cambridge in the hopes that it would encourage kids to pick up programming, and even electronics, because it could be bought cheaply enough to give to them with no thought of how they might destroy it while playing. A bicycle, rather than the family car. If a kid wrecks his bicycle, they’ve only inconvenienced themselves; when they wreck the family car (read: computer) then everyone is hobbled while it’s out of commission — with the result that no kid is encouraged to muck about with said car/computer. With the RasPi, all you get is the computer iteself — it has no case, no peripherals, it’s just a tiny little motherboard with chips and ports. And it’s got a General I/O port, for wiring it up to all sorts of things. It’s ideal for mucking about with robotics.
I was reading up on it last weekend, and musing about robotics. Because, let’s face it, robots are infinitely cool. Who hasn’t dreamed about making robots? Or at least owning them? Movies and the SF books of the 50s and 60s all prepared us for the notion that the odd fellow living at the end of the block could very well have a house full of robots he’d built himself, in total defiance of the accepted wisdom that robotics is hard and expensive. Really, this has been going on since Ms. Wollstonecraft-Shelley gave us the good Doctor Frankenstein. One solitary genius leaps ahead of the world. That sort of thing can and does happens in mathematics, since all that really requires is your brain and it many ways math theorems stand alone. But in the other sciences? Sadly for our megalomaniacal daydreams, it really is true that most science comes from standing on the shoulders of giants. Robots, it turns out, are hard. And expensive. Japan spent millions of dollars and the time of industry and universities to make a humanoid robot that could walk up stairs. That’s all it does, by the way — walk up stairs.
Here’s where my wandering brain started last week: what if I could make a little toy robot that could climb up my bookshelves and pull out a book to barcode scan it, then put it back? That would be something to show off! And not too ambitious (said my lying and foolish brain at this point). The Raspberry Pi certainly has the processing power to do it, and webcams are cheap. We could even tether it to a power supply at first, meaning it wouldn’t have to carry the weight of batteries. Well, I want it to access my bookshelves, so wheels are out. It has to be able to climb. Which looks cooler anyway. More legs is better, since it can brace itself more easily. In fact, with the cheapness of the Pi board, one could design a robot leg that had its own controller, so they were modular. Let’s scale it back — I just want to make a robot leg, something that can move as ordered, has its own computer brain, and a cell phone battery (self-powered, and distributed weight for the overall robot should it ever happen). Trouble is, robotic legs are slow-moving. Things that can topple over, or fall, or need to navigate uneven terrain (like my couch) need fast reflexes.
Well, what about solenoids? A solenoid is basically a magnetic bolt surrounded by an electromagnet. Put on the current and it shoots out about an inch. Turn off the current and a spring pulls it back. The magnetic force makes them strong, and fast. What if you used a solenoid as a single muscle fibre, in say some sort of robot tentacle like you’d see on Kevin O’Neill’s conception of the Nautilus? Multiple solenoids strung together would allow more than an inch of travel; also a tentacle doesn’t need a lot of motion per-segment because it adds up over the length of the arm. Would that work? Frankly I have no idea. It may be impossible, or unworkable, or impractical, but it would have been fun to find out. But what it certainly would be is expensive. Solenoids tend to start around the $10 mark, and this project would need literally hundreds of them to be even vaguely practical. So it will have to be Bill Gates’s kids who do that little experiment. Even if I was willing to try it, no guarantee that it would work. Also, it would be heavy, since those (hundreds of!) magnetic solenoid bolts are solid metal and by no means lightweight as a consequence.
Now let’s be clear: at this point I’m just daydreaming while reading up on the Raspberry Pi on their website. None of these needed to be particularly practical or achievable. But I was reaching the point where these were obviously dead ends even for the purposes of daydreaming. Now for the coincidence: this discovery of the Raspberry Pi coincided with a friend coming over to pick up my Lego Robotics kit, which I bought a decade ago and never used. So I was happy to send it off to a good home, where kids could get some use out of it (plus, it was exchanged for beer, which is always a good investment). But I’d looked at it opened up again and was reminded why I wanted it in the first place. Robotics are cool! But I knew from that decade-old experience that just building a robot that can roll around the floor and avoid obstacles for the sake of having a robot that can roll around was uninteresting to me. Had it, never used it. A tension, then, between what is interesting to me and what actually happens in my life.
So now that I’d finished imagining movie-style crawling robots in my home (no doubt to eventually turn on their creator in a bloodbath of Brobdingnagian proportions), I was still intrigued by these tiny, absurdly cheap (yet quite powerful) computers. What if I could think of a practical project that would have some sort of use to me when it was completed? That might be the carrot that could carry me through to the end of a project. Well, the thing in my life that I’ve taken up and actually carried through on with some success is cooking. Conveniently enough, there are a couple of possibilities there.
My friend Doug, a masterful amateur chef, has taken up the Modernist sous-vide cooking method with dedication. This is the method of cooking vacuum-sealed food in a bath of hot water of constant temperature over a long period of time. There needs to be electronics to control the temperature and flow of the water. Doug’s got a nice rig made out of a cooler and some third-party hardware, but one could build one’s own pretty simply. Also, I’ve been looking at a temperature sensor for my Big Green Egg barbecue to have a digital thermometer that you could read via a wireless network, that could potentially also control a forced-air rig that could adjust the temperature, for unattended cooking over long periods of time. Truthfully, the Raspberry Pi is pretty overpowered for such an application. Not only could it control the hardware, but it could also serve up a pretty website for displaying and controlling it, while allowing someone to play Quake III in HD at the same time. (Quite literally; Quake III has been ported to the Pi and it has HDMI out.) But again, the cost is not extreme.
Well, this is something that I’ve never tried before. I’ve never been a hardware geek; not once in my life have I touched a soldering iron with intent to use. But the deal is done! I have ordered a Raspberry Pi, a controller expansion board, some digital thermometers (including a high-temperature glass one for the Egg), and a power supply. Now I shall poll my electronically-savvy friends to see if anyone has a soldering iron I can borrow. Further posts shall explore my foray into physical computing and electronics. Oh, and cooking — I have 2012′s Vernal and Autumnal Geekquinox parties to document as well. Onward into the unknown!