How it works
The basics:
1) Person touches jello with finger or spoon
2)Indicator lever pulls sensor 2 + triggers servo
3) Servo jerks disc around on turntable
4) Sensor 1 is bounced up and down by ridges
5) BX chip turns values from sensor 1 into frequencies out
6) Speakers with built-in amplifier produce sound
Building it
Ironically, the mechanisms that make the Jello react to touch have
very little to do with the Jello itself. After exhaustively experimenting
with making Jello jiggle, I decided that Jello jiggles best when
it is shaken from side to side, which is why the disc is on a turntable.
However, this created some problems with the piezo sensors I was
using to detect vibration.
About piezo film sensors
It turns out that when Jello jiggles, the movement is concentrated
on the top of the Jello pile. Moreover, in order to maximize the
‘jiggle factor’ the plate on which the Jello rests moves
both laterally (around in a circle on the turntable) and vertically
(up and down on the scale). Unfortunately, piezo film sensors measure
differential voltages ACROSS the surface of the film by measuring
how much the sensor is bent.<
From www.piezo.com: ‘When a mechanical stress is applied to
a single sheet of piezoceramic in the longitudinal direction (parallel
to polarization), a voltage is generated which tries to return the
piece to its original thickness. Similarly, when a stress is applied
to a sheet in a transverse direction (perpendicular to polarization),
a voltage is generated which tries to return the piece to its original
length and width. A sheet bonded to a structural member which is
stretched or flexed will induce electrical generation.’ Basically,
then, the piezo measures the force applied to it as the difference
between one state and another.
In order to measure the ‘bend,’ the sensor must be fixed
in place at the base. It doesn’t take too much experimentation
to realize that the sensor that measures ‘up and down’
movement must be vertically fixed in place, while the sensor that
measures ‘side to side’ movement must be horizontally
fixed in place.
As well, piexo sensors work best (or work at all) when they are
bent in half on the horizontal axis (If you examine a piezo film
sensor closely, you will see a thin line across the bottom third
of the sensor. This is where the voltage is measured). I had originally
planned to use only one sensor to measure touch and vibration, but
it soon became clear that the functionality of the device had to
be split in two so that the disc could also move laterally
Because they measure minute fluctations in force, piezo film sensors
typically generate wide ranges of values over time, but differ little
from millisecond to millisecond. If the object is to detect the
amount vibration from one second to the next, code written for the
piezo should incorporate a ‘checking function’ that uses
the the difference between sensor values instead of the sensor values
themselves. That way, the results are relative, not absolute.
FreqOut!
FreqOut is a BX-24 behavior that outputs frequencies as faux-analog
signals that can be read by speakers. However, the signals are extremely
weak, and if they are to be loud or even vaguely audible need to
be run through an amplifier. The piezo values generate the freqOut
tones directly, but have to be arithmetically transformed to produce
frequencies high enough to be heard by the human ear. I also made
them extra high to simulate a panicky, screaming sound. Were I to
extend the project, I might condsider using MIDI – the tonal
range of the BX-24 is kind of default-sounding.
Making waves
The ridges along the bottom of the disc are there so that the piezo
sensor can vibrate semi-predictably, in a stable manner. Sensor
1 is a super-sensitive piezo with a weight at the sensing end. There
is also a wire attached to the weight. That wire acts just like
a record stylus: it is bounced up and down by the ridges as they
swing around on the turntable, producing reliably huge piezo values.
(This had the unforeseen effect of making the values from the piezo
very, very similar most of the time, meaning that the self-generated
character of the noise is largely unnoticeable.) Ironically, plain
old masking tape was the best ‘ridge material’: flexible
so as not to put undue strain on the sensor, but stiff enough to
push the wire up and down a little.
Go to www.piezo.com for general
piezo info.
}
|