Ice core v0.1

The Ice Core is a physical model of the type of ice cores that scientists extract and study. When a user touches the ice core, sensors tell a computer to display relevant information about the touched area. For example, if the user touches the surface of the core near the bubbles embedded in the ice, the computer will display information about how scientists use the gases trapped in these bubbles to learn about the atmospheric composition during that era.

To make the experience more interactive, the ice core will be cooled by a "Peltier Junction" and will feel cool to the touch. A Peltier Junction is a device that uses electricity to heat one side of it, while cooling the other. This thermoelectric effect is known as the "Peltier effect."

It might seem obvious; we want to know when someone has touched our ice core. Wouldn't we want sensors that react to touch? Maybe, but there are other options that may work better. For practical reasons, we will be using the products manufactured by "Phidgets" as they are easy to interface with computers and to write code for. Phidgets create a wide range of sensors, each with specific purposes. Four have potential for our needs:

Capacitive Touch sensors

Top view These might seem the obvious choice. These sensors trigger when contact from your skin completes the electrical circuit. The advantage of these sensors is that very little can go wrong because of the way they work. As long as you touch the sensor pad, it should tell the computer to display the correct information.

Unfortunately, that is the extent of the benefits. Touch sensors in this project have many downsides. The most inhibiting is that most touch sensors are rigid and sense only a small area. Our ice core is a translucent cylinder; this makes it quite difficult to hide a black plastic sensor that won't bend to fit the shape of our core.

One solution is to put the sensor inside the core. This might work, but has its limitations. The sensor is only sensitive enough to work through 1/8" of plastic. That means we have to nestle it very close to the surface of the core. This not only requires a very precise and complex design for the shape of our ice core, but chances are the sensor would be visible through our translucent material. To hide it, we would either have to use opaque plastic or cover it with a sticker or paint. This detracts from our goals of creating a realistic looking core, and prevents us from using LEDs to light up the section of the core that the user has touched.

The second major issue is that the sensors only cover a small fraction of the ice core's surface area. Each is only about the size of a penny. That means the vast majority of the ice core would not respond to the touch of the user.

Currently, we are experimenting with ways to maximize the utility of these sensors.

Force Sensors

These work exactly like a bathroom scale. You put force on the sensor, it tells you how much. In our case, even a small amount of force would tell us someone had touched our ice core and we could trigger the computer program. Some companies sell flexible force sensors, but unfortunately, none are discrete enough to hide on a translucent ice core.

The force sensors made by Phidgets would likely work well enough to make the program function, however they look like small black buttons mounted on a black plastic board. Even if we were able to make the complex design required to nestle the sensor into the inside of the core, it would still have a black button poking out.

Just as with the Capacitive Touch Sensors, the sensor itself would be visible through our translucent core, and the percentage of the surface area covered by the sensors would actually be lower than it would have been with touch sensors.

Vibration Sensors

Top View Vibration sensors operate just as your doctor's stethoscope. A flat pad registers vibrations that come into contact with it and tell the computer of what it finds. The advantage of these over the previous sensors is that they could be hidden from view. Since the vibration would resonate through the core, the sensors could be placed deep inside, or around back in a housing that disguised them. This allows us to create the aesthetic experience we aim for.

The other advantage is that the sensors would sense vibrations across the whole surface of the core. This is far more desirable than the other sensors which only responded to touch on a small fraction of the surface.

The sensors do have their downsides. A light touch, which would trigger the capacitive touch sensor, would likely not trigger the vibration sensor. This encourages users to be rough with the device and might reduce its lifetime.

Another problem is how do you determine where the origin of the vibration occurred? If sensors three sensors all register the same vibration, which information about the core do we tell the computer to display? Would the sensors be sensitive enough to consistently register to the closest sensor first? Or would we have clashes between multiple sensors?

Infrared Distance Sensors

Side ViewThese sensors send out infrared beams, much like the one found in your TV remote. If something gets in the path of the beam, it will reflect back and be registered by the sensor.

This may not seem like the best device for our needs, but it may actually prove the most effective. If the sensors are mounted at the end of the core, pointed along the surface towards a plate mounted at the other end, they will register a constant value equal to the length of the core. Any time someone's finger touches the core, it will break the plane of the infrared beam and the sensor will trigger. Because the sensors can calculate distance, we could use the output to determine at which point on the core the user touched.

In order to cover the entire surface of the core, this project would require multiple distance sensors mounted around the end circumference.

Side view showing housingThe sensors can be hidden within a housing and since they are only accurate at distances between 10cm and 1m, the housing could keep the sensors greater than 10cm from the closest point a person could touch the core, thus providing more accurate results. Because everything is external, we can make the core out of any material we want and no complex hollowing out of the core in order to insert components would be required.

These sensors are not very consistent, often give false readings and rely heavily on how people touch the core. Using anything but a straight finger coming in at a right angle can give false readings. Margins of error are in the range of 30cm.

Ribbon Controllers
Needs testing.

Frustrated Total Image Reflection

LEDs shining into the edge of a hollow plastic tube will not emit light. This is known as frustrated total internal reflection, also used in the Harvey Glowing Arm project. As far as I can tell, the curvature of the cylinder should not impact this phenomenon. In theory, it is possible to tell where someone has touched on the surface because the point where they touch causes refraction of the light backwards. A sensor inside the tube registers this light flare and triggers an event.

The index of refraction of the material determines how well total image refraction will work. The higher the index the more light will remain in the material. Air has a value of 1.00. Acrylic has a value of 1.490. Silicon is 4.

At this point, we know that this phenomenon works on a sheet of acrylic, though no testing has been done to register a finger touch by using a rear mounted sensor. If this proves successful, this model might be the best bet.

http://en.wikipedia.org/wiki/Total_internal_reflection
http://en.wikipedia.org/wiki/Refractive_index

Theramin

Needs testing.

The code for this project will be custom written in the Python computer language. Once it has been completed, it will be made available and explained here.

The touch sensors work from far away when connected to a wire. (Alligator grips insulated and not touching the sensor pad).

• Plastic tube using Frustrated total image refraction and a wii mote to receive the location data?
  

• Theramine
  

• LED sensor panels http://evilmadscience.com/majors/46-tables

• http://www.atmel.com/dyn/resources/prod_documents/qmatrix_white_paper_100.pdf