Advanced Circuits is absurdly fast. I placed the order halfway through Monday and the boards showed up on Wednesday morning. I haven’t tested them yet, though I see at least one obvious goof on my part, so we’ll see how they work.

Updates to follow.

This is all a huge step up, technically and aesthetically, from the logger Jenkins and I made for Computers for the Rest of You last semester (though that was a great start!).

As I blabbered about in a previous post, I’m working on taking this platform:

And putting it out where people can use it. Part of that has been going on for the past two weeks or so, with my little monitor hanging out in the ITP lounge:

and

(This may be the first chance I’ve had to show how damn small this thing is)

It’s been great to watch, actually — I’ve explained it to a handful of people, and they’ve explained it to a handful of others, but it’s still largely a blinky light to most. However, it seems to work — once they know how it works, people will of course run over and set it off on purpose, but then they’ll often look up after loud noises to see the effect it’s had on the logger. Better yet, half the time I’ll catch people looking at it in surprise when it finds something not-as-loud that they hadn’t noticed.

So, for now, the physical side holds and is doing well. The data side is coming along, though slower. The logger, after some wrangling, outputs straight ASCII values of the ports it’s reading, which after a quick find-and-replace to get rid of spaces, comes up nicely in graphing programs.

to

I’m still trying to find a better way to deal with visualization, but for now Apple Numbers is doing an okay job with the smaller sets.

Though I should note that trying to read the whole initial 40k-entry set I fed it caused it to eat up 5 gigs of hard drive space and crash. Ooops.

I’ve also done up a quick Processing app that can read the data coming out of the Arduino side when patched into a computer:

and

Part of the question is to see if it helps or hurts the cause — will detailed metrics help people quantify their experiences or just add to the hassles and distractions? Time will tell!

Thanks in no small part to the excellent write-up from my friend and Socialbomb compatriot Adam Simon, I’ve shipped my first round of boards out to Advanced Circuits for printing. The thesis marches on!

The boards for the personal unit are both done — microphone circuit on top, Arduino on bottom — and should be here quick (Advanced Circuits RULES). The mic circuit board will make the next steps easy, and the Arduino board will make it possible to make cheap, tiny arduinos as a part of nearly everything. There are pin-outs on each board so that they can interface with others fairly easily. The sensor boards (in this case the mic board) has a six-pin out that could go either to a microcontroller or straight to an Xbee or Xport, and the Arduino boards have a six-pin out that could either go straight to a computer (through a TTL-232 cable like the ones Limor Fried sells) or to a logger unit, and there’s access to the analog ins, RX/TX and digital pins 2/3 (for SoftSerial) for any GPS/location-aware tech.

(The aforementioned TTL-232 cable)

I’ve also revised the previously-kinda-huge logger circuit to include the new uSD logger from Sparkfun — it crams the same DOSonCHIP functionality into a teensy package that fits on a board the same size as the ones I’m using, so my original goal of having a ton of add-on boards for extra functionality is even more feasible now.  Either print one up, use one of mine, or just go to yer local RadioShack and solder up your own!

(Bottom-to-top: that’s the old logger, the new logger, and a 232 board for size reference)

I’m only going with the BareBonesPCB option they have, as it’s dirt cheap and fast, but if this all works I’ll print up prettier versions. I’ll also need to add

I’ve also based the size on the small square boards Radioshack sells (which I used for the prototype), as they’re juuuust big enough to accommodate an Atmega8/168 and readily available so additions to the platform are quick and easy to prototype and implement.

Of course, as things happen when one is facing a deadline, that it’s taken me the better part of a week to mock up the boards, test them out on paper, plan them around kit-available-and-cheap parts and fix the errors, but they should be here on Friday — yay!

Disclaimer: EAGLE is fun, functional, and free — but it may drive you a little crazy. If you’re thinking about doing up boards of your own, you might want to search Instructables for a tutorial (there are a bunch of good ones!) or talk to someone who’s done boards before.

Cambridge Mobile Urban Sensing (CamMobSens)

and

Of course, they’re doing things on phones, and half of my goal is to avoid that, but still. They showed several “google map visualisation showing some of our Carbon Monoxide data collected in Cambridge area (using Nokia mobile phones,each phone connects wirelessly to CO sensor and GPS receiver)” as well as the same tracking audio.

Of course, Intel Research’s Urban Atmospheres Group — Ergo
On-the-Go Air Quality Readings delivered to your mobile device

Part of their larger Participatory Urbanism movement.

And our own ITP favorites, Botanicalls Twitter!

And I just found this one, but I can’t believe how amazing it is:

SFEarthquakes on Twitter!

What you can’t see there is that I’m running around the room, getting further away and closer up, but I did, and it works! There’s a longer demo if you want to watch that too:



Full demo from Mike Dory on Vimeo.


After messing with it a bit on my own, I gave it to a few other people to test out and see what they thought. I didn’t give much more explanation than “it’s listening, and it gets brighter when things around you are louder,” because I was hoping people would fill in their own ideas about how to use it beyond that. The results were both reassuring and highly amusing.

I gave it first to my friend and frequent collaborator Adam, who’s known a good deal about the process and prototyping, but hadn’t seen the finished version yet. Adam’s first reaction was “I like this. I want to take care of it — like a pet!” After messing around with the mic a bit, we went back to chatting. Every time he spoke, the light brightened — something that caught his eye, and brought volumes to his attention. As he noted, “it finds sounds that your brain normally filters out” which is also much like a pet. He then proceeded to speak to it like a tiny pet and hold it like one might hold a small dog or hamster, which was absolutely awesome.

He had a pile of excellent suggestions, mostly that it needs to have some sort of reaction at its peak beyond the light just being being really bright. He suggested having the box make a sound of its own, turn on a second light, vibrate — something along those lines that would call out that it was its limits. As he noted, “it has emergent behavior, but it has no opinion on it.” He’d also suggested, if I wanted to really push the pet aspects, that users could have some sort of input which would “calm it down” quicker after it hit peak, which is something I’ll have to try out if I make another more pet-ish version.

I gave it next to Tom Jenkins, who tested it out as follows:



Tom tests the prototype from Mike Dory on Vimeo.


I hoped people would find the design appealing, but I didn’t count on everyone turning it into a pet. That, I think, is a sign of success — if people like it enough that they can anthromorphize it on first use, then I’m hoping they’ll like it enough to keep using it for further testing.

Now, it’s out in public on the ITP floor:

At Kate Hartman’s urging, I put it out in the lounge to see how people would interpret the lil’ device. I haven’t put a video camera on it yet, but after two days of monitoring, it’s been getting a fair amount of attention. It’s done a pretty good job of relaying the volume level of the lounge (which during a typical ITP day goes from library-quiet in the early mornings to Grand-Central-at-rush-our insanity in the afternoons), and it’s gotten a bunch of people asking questions about it. I think it’s still too small and too impersonal to be of much use to anyone who hasn’t been given a 10-minute pitch about what it does, and even then it’s too far to actually be much good, but it’s certainly an assuring start.

Of course, this being ITP, it’s also getting attention from people who want to play with it. I’ve caught people running up to it and stomping, clapping, whistling, etc. People across the room have seen it blink and yelled to see it spike again. People have gone up to speak baby-talk into it and make sure it’s okay. To me, these are signs of success as well — people have to care about the object before I can expect them to learn much, and if they use it as a toy, they’re still augmenting the sounds they might otherwise not be paying attention to.

The lm386 circuit I’d mentioned a bit ago here did the trick, so I just transferred the whole thing to a breadboard and slapped an Atmega168 (bootloaded with Arduino) under it.  The unit has the full mic circuit on top (with mic and warm white 8,000 LED) and the Atmega circuit on the bottom with a 5v regulator-conditioned power in on one side and a 4-pin power-ground-RX-TX male header set on another.  The two boards connect by a short ribbon-style cable and two header sets, as is better visible here:

I wholly underestimated how long it would take putting this version together — getting all the components into a space this small and hand-soldering all the connections took two full days.  I always forget how long wiring circuits on perfboards can take if it’s this dense.

I’m still planning on making a proper Arduino shield eventually, but this idea hit me while I was doing the documentation:

This wouldn’t work with the pin outs I’ve used, but it could be made to work in the next version quite easily, and would be outstanding to have.  The ability to swap-and-play with different sensors on the fly is something that just blows my mind.  Of course, whatever I think is cool will be made infinitely better by the general hacker populace, so this will be made freely available for further mucking about once I’ve got the circuit 100% done and put on PCB’s.

And as these pictures don’t explain it that well, it’s really tiny:

This is small, but the PCB version will be much smaller (and hopefully integrate the battery into its design a bit better).  I still have to figure out a way to package them attractively, which will take some shopping around and some prototyping.  I want the final version to be cased in a translucent skin (to let the light through but hide the science), and it’d be nice if it was pleasant to touch.  I’ve had this discussion a few times lately, but it’s especially relevant here — if you want people to really attach themselves to an item and make full use of it, it has to be something they can relate to.  Good-looking artifacts do much better in most cases than ugly, awkward devices.

I’ve taken to calling this little cube Soundbox, and it seems to fit. I’d been struggling with a name during the earlier stages of prototyping, but it became clear during the testing of this one that what it was really doing was amplifying my own sense of hearing — it was turning the sounds I was actively or passively ignoring around me into light in a way that caught my attention. As I thought more about the amplification idea, the name became clearer — in musical instrument terms, a sound box is the part of the instrument or speaker that augements the sound and gives it color. From Wikipedia’s entry:

A sound box or sounding box, (sometimes written soundbox), is an open chamber in the body of a musical instrument which alters the instrument’s tone quality by modifying the way the instrument resonates. Most instruments respond more strongly to vibrations at certain frequencies, known as resonances. The sound box typically adds resonances at lower frequencies, enhancing the lower-frequency response of the instrument.

The purpose of the sound box is to amplify the volume of the instrument,^[1] but it also gives the instrument its distinctive sound. A sound box is found in most string instruments.^[2] The most notable exceptions are some electrically amplified instruments like the solid body electric guitar or the electric violin, and the piano which uses only a sound board instead. Drumhead lutes such as the banjo or erhu have at least one open end of the sound box covered with animal skin (or a skin-like acrylic material). Open back banjos are normally used for clawhammer and frailing, while those used for bluegrass have the back covered with a resonator.

Loudspeakers also are mounted on a sound box to enhance their output, particularly bass speakers.

So, while that’s a bit all over the place, the point is that it works for the project: I’m giving people little boxes that augment the sounds around them.  The question is what they’ll find out from that, and proper user testing should give me a good deal more on that shortly.

(Thank you Soyoung Park and ITP Sensor Wiki!)

I’m using a 386 because they’re cheap, low-power, and easy (insert a “your mom” joke here).  They don’t require negative voltage, they can do a variety of functions, and they’re easy to find at pretty much any electronics retailer.  The circuit is very similar to what Tom Igoe has in the Physical Computing book, with a bridge rectifier added to turn negative-and-positive readings into all-positive (as mics measure input by swinging voltage both positively and negatively) and a low-pass filter to get rid of some of the high-frequency noise that was giving me ugly spikes in the readings.

Breadboarded, that looks like this:

Lots more pictures in that Flickr set too.

I keep finding a ton of better ways to do what I’m doing, so consider this to be one of a million work-in-progress listings, but here’s what I’ve been all up ons the past week or so:

From Rason.org — Designing Op Amp Audio Preamplifiers

At the moment I’m using an lm386 rather than 741, but for the second draft I think I’ll try that route.  It’s a much more powerful amp, so I can only imagine the improved clarity and response.  For now though, the 386 is so simple, cheap and low-power that it’s hard to argue with.

I’ve also been reading a pile of other op-amp things, including:

Wikipedia — Operational amplifier applications

Ron Mancini — Op Amps for Everyone (and free PDF here)

Don Lancaster — Active Filter Cookbook

I started with ye olde Radioshack PC Board condenser mic:

And eventually (mostly due to the pre-soldiered-ness and a slightly more attentive range) ended up with this guy, another Radioshack electet:

I was lucky enough to have some suggestions to check out the ITP Sensor Wiki, and found some fantastic schematics from SoYoung Park and code from Jeff Gray (thanks!!), which after some tweaking and testing brought me to this:

This is a Processing app that watches the incoming analog values from the Arduino (in this case, the audio signal), and outputs the points on the screen, moving from left to right. The red dots are the actual values, and the line is the average. For my purposes, this works splendidly, as mics output both positive and negative voltage — so in this case, with the signal coming out of the mic boosted 2.5v, watching for the average gets a really good picture of what’s going on.

This is the ambient room volume in the pComp lab:

And this is what happens when you clap a few inches away from the mic a few times:

The circuit is based off of the one Tom Igoe has in the Physical Computing book, with SoYoung’s rectifier and low-pass filter added in, the guts of which look like this:

And the whole thing like this:

So here, with better feedback examples, there’s this:

In this case, when you make a loud enough noise close enough to the mic (that blurry thing in the center), the big ol’ green LED in the upper-right corner of the breadboard gets tripped and goes fully bright.  It’s actually a pretty engaging interaction — I’ll have to take video and post that shortly.

Right now I’m working on getting some mic circuits together, testing them out in semi-public, and conning other people into testing them.

Updates to follow as things work.  Or don’t.

(Note to self: find job in R&D in near future.)

I started by ordering a pile of kits from the folks at Anykits — they have outstanding prices and a website that’s actually easily searchable (though, being located in India, their shipping is a bit pricey). They have a tiny board for nearly every application (amplifiers, switches, light solutions, sensors, etc.) and they aim to educate people while giving them a decent product.

Clearly aimed at hobbyists, the kits include all the parts, a printed circuit board, and a printed schematic with the layout and a a part list. It’s pretty easy to figure out what goes where and in what direction as long as you understand the mega-basics of components (and if you don’t Wikipedia will get you up and running quickly).

The board is spacious enough to not be too hard to work with (maybe still a little tough on absolute newbies, but hey), but still small enough for a lot of projects.

There are a lot of odd-ohm (to me) resistors, but they’re each also only used once, which is smart! I’m not sure if that’s due to the fact that they’re hoping to confuse people less or simply that the circuit works best that way — not all circuits ship like this, but it’d be smart of them to do so I’d think.

The holes on the board are actually a little too big for the wires and have small pads, so soldering isn’t as easy as it might be. Still it’s not difficult, and things are nicely spaced out so that people who aren’t used to cramped soldering conditions shouldn’t have a problem.

You’d have to know:

* Which side on caps is positive.

* Which resistor is which (i.e. know resistor colors)

* Basic schematic signs

* Basic soldering skills

But all in all, this kit was easy to do, and it left me with a nice-looking finished product. Granted, it turned out to be of little help to me for my audio work, but that’s not their problem.