Decay! Decay! Decay!

I had a friend who introduced me to Helium, and I thought she needed a Helium of her own, so I set about making one.

My objective going in was simply to have some sort of spherical helium which said "Decay! Decay! Decay!" when it was manipulated in some fashion. All the more specific design decisions were dictated by expedience.

I started out with a very minimalist design for the speaker, involving a microcontroller and four resistors as a driver. The microcontroller used was an ATMega8, because I had some in stock, and there was just enough memory for the sound clip. The resistors were 250, 500, 1k, and 2k, to form an 8-bit DAC. The values were chosen because 250 ohms was the smallest value which could be used at 5V without exceeding the 20 mA drive specification of the AVR microcontroller. As it turned out, I ended up running from 3V instead of 5, and therefore could have gone a bit lower, but I had already ordered 1% resistors in octave values starting at 250.

In the picture, you can see the resistors stuffed into the accessory connector of the STK500 programming board. With just the resistors, it did work, but it was very quiet. I wanted more volume, so I added an NPN transistor as an emitter follower, which worked quite well. I believe I used a pn2222, but it might have been a 2n4401. I don't think it really matters. There's a DC bias on the speaker, but at this level of quality, that doesn't really matter either. Here was the circuit used at this point:

	    -o|  |o-
	    -o|  |o-
	    -o|A |o-----------------------.
       VCC  -o|T |o-----------------.     |
	+   -o|M |o-----------.     |     |
	|   -o|e |o-----.     |     |     |
	-----o|g |o-    |     |     |     |
	-----o|a |o-    |     |     |     |
	|   -o|8 |o-    |     |     |     |
       ===  -o|  |o-    \     /     \     /     VCC
       GND  -o|  |o-    /2k   \1k   /500  \250   +
	    -o|  |o-    \     /     \     /      |
	    -o|  |o-    /     \     /     \      |
	    -o|__|o-    |     |     |     |    |/
						 |  __ /|
						 ^-|  | |
						 .-|__| |
						 |     \|

I had originally intended to use some sort of tilt switch, probably and old scrounged mercury one, for activation. Unfortunately, I couldn't find any old ones, and by the time I realized that, I didn't have time to order one anymore. Thus, I modified the design and ended up using a magnet in a tube with a reed switch. This part didn't go particularly well. I destroyed two or three reed switches bending the leads, and the positioning of the magnet enclosure such that it was close enough to activate the reed switch, yet not so close that it would get stuck, turned out to be extremely difficult. Since I don't recommend the sensing method I used, I won't go into detail about the construction of it.

The figure to the right shows the assembled circuit. It's powered from 2 AA alkaline batteries. Note the other kwality features, like the socket which is the wrong size because I didn't have any correct ones in stock.

To get the actual data for the music, I captured some audio from the flash animations on the web site, and then converted it into 8-bit raw format at 11kHz with Sox. The final conversion to four-bit audio in was performed with a Perl script. The following link contains the actual code I used. The audio data I used is not included, since I could conceivably be sued for copyright infringement if I distributed that, but the Perl scripts used to generate it are included, so you can put in your own from whatever source you want. Note that the pin assignments changed between the preliminary circuit diagram above (which was set up to use a convenient part of the expansion connector on the programming board) and the actual board, so don't wire it up as shown above and then use the code unmodified.

AVR Code, including one-liner converstion scripts
I left the "artistic" part for last. I started with a styrofoam sphere from a local craft supply store, and then cut out two arms from smaller blocks of the same styrofoam and attached them with wire. The entire thing was then given a coat of pink acrylic, which took nearly a day to dry. The red, white, and black details were added once the pink was dry.

There's a cutout in the back where I inserted the circuit. I had originally intended to cover up the hole with a piece of paper or felt or something prior to painting it, but then I realized that the intended audience was probably going to rip it out anyway to see how it worked, so I ended up leaving it exposed.