Here's a quick SX project I put together in an evening a few weeks ago. Shoot, it took longer to figure out how to draw the schematic in Eagle, than to build.
Note that this is *not* my idea -- I saw a photo at MakeZine.com -- but the implementation and code is 100% my own.
Uses an SX (no resonator) and 7 LEDs. Push the button, wave it back and forth rapidly -- text and graphics appear in mid-air.
A few comments:
- the RCTIME circuit was used to tune the speed of the "columns". You could eliminate these components and hard code some value between 1900us and 2500us and it will be probably be fine for the average person. I just didn't want to have to reprogram with my SX-Key to tune
- I use an alligator clip to jumper the pins of the pushbutton when I need to program (to keep the power on while an SX-Key is connected)
- there are 32 message slots -- 16 are straight hard coded bitmaps, 16 are text strings of any length. The text strings are mapped to ASCII character bitmaps with one extra trick -- the engine will look for a break in the text string (a space character) after a set number of characters, and show JUST THAT PORTION of the string for 4 seconds or so, then it moves on to the next portion of the string. Why? We found that most people will comfortably display 12-18 characters on an arm swing -- so strings longer than that were not legible (you'll never see the whole thing on one right-reading swing of your arm). We didn't want to be limited to such short text messages, so there you go.
- messages are shown randomly with no repeats till all messages have been shown
- I used my own web-page page utility to generate bitmaps and such. You can try out that utility here: 1uffakind.com/robots/povBitMapBuilder.php
- Code, schematic, photos attached.
28 Nov 2012 -- I've uploaded photos of a latest iteration, along with photos of the POV in action.The newer iterations are functionally equivalent to the first version, EXCEPT that the RC circuit for adjusting the speed with a pot has been eliminated, and I run the SX directly from 2 AAA batteries rather than building a power supply.
We've been working on a (non-accelerometer-based) sensor for detecting end points of the arm wave, but results have not yet been good enough to put into play.