ECE445 Project Proposal
September 15, 2010
Today a computer display is thought of as a complex electrical instrument that is an
evolution of the CRT television. After seeing an example of how to create a mirror from
a decidedly non-reflective source of wood and light from Daniel Rozin, I became excited
for the prospects of using similar materials to display nearly anything from a computer
source, not just what was in front of a camera and mimicking a mirror.
The goal of the project is to reduce the complexity and labor intensive nature of
constructing a display from an array of motor controllers. Current motor controllers
control up to 8 motors at once. If a single motor is equivalent to a single pixel,
constructing anything resembling a full display would take a massive amount of
controllers. One objective is reducing the number of controllers needed to control a
similar number of motors.
After calculating which angles of light off a wood pixel corresponds to which level of
grayscale for a certain pixel, the corresponding angle is then transmitted to each of the
motors through the custom controller, which handles all motor functions.
Alternative to ordinary displays
Can be used as art
Functions as a display
Full knowledge of inner workings not required
Data Source Computer:
As the device will act as a display, the most
logical signal source will be from a computer
that can easily supply a signal from many ports.
Computer Also, it powers and supplies an operating
platform for any source and software.
Software The data source simply sends an image to the
software to be displayed.
The software portion will take in an image
and convert it to grayscale, then pipe the pixel
data to the controller for dissemination to the
The housing will provide physical support for
the controller and motors, as well as allow for a
power bus the controller and motors will need
Motors The controller takes in the pixel data from the
software and then sends the specific pixel
information to the specific motor controlling
the pixel. The controller does this for several
After receiving the information as to which angle to move to, the motors adjust
accordingly, altering the angle of light hitting the surface of the wood. This brightens or
darkens the surface proportionately to the level of grayscale in the pixel in the original
As a single pixel is a motor and thus parts costs increase with each additional pixel, the
system should be able to display for any number of pixels in the current system. When
additional pixels are affordable, they can be added to expand the display.
As the system as a whole isn’t as quantifiable as a simple voltage output, each subsystem
will need to be taken into account. For each subsystem, the testing procedure will be:
Computer – make sure it is sending data correctly to the controller using a test image of
simple black and white squares, this will allow a full bright and full dark test. Another
test would be a gradient from white to dark to test the software’s ability to determine
Data source – To make sure the source is sending data correctly to the software, simple
all black and white images can be sent through the system
Software – Sending the software a fully white image, then fully dark image and then a
color image with a varying intensity gradient will allow checking for accurate pixelation
and transformation to grayscale.
Housing – The housing is responsible for holding the controller and motors, so the only
way to accurately test the ability to hold the weight would be to hang a similar weight
before attaching any vital components to ensure the system won’t destroy itself.
Controller – Knowing which motor is getting signal and what angle it needs to be at (due
to a test image from the data source), the outputs can be viewed on a scope. RC Motors
use Pulse Width Modulation, so signals leaving the controller and inputting the motors
can be checked with an oscilloscope. Modifying the data source should be visible on the
Motors – Sending a signal of full bright, full dark, and scanning the middle brightness
range will check angles. A secondary consideration is testing to ensure the motors
themselves are strong enough to hold the wooden pixel and will not break apart from the
continual vibrations caused by several motors running at once.
Since the project is a display, the motors need to be refreshing relatively quickly. The
controller could become a major bottleneck if the system can’t update a massive amount
of motors without buffering the information. The system must be able to jump from one
extreme (all bright) to the other (all dark) quickly, as well as being able to differentiate
between small differences in grayscale. These responses can be viewed on a oscilloscope
with the Pulse Width Modulation of the motor inputs.
Cost Analysis: (Assuming 16 motors in initial build)
15 hours a week
180 Hours total
1*35*2.5*180=$15750 in Labor
16 Motors * $5 = $80
Metal Framework for motors = $50
PIC Microcontroller = $1.00
PCB = $20
Total = $15901
Since I am the only one on the project, all the following will need to be done by myself.
9/13 – Proposal and Research Software to Controller Connection
9/20 – Determine Specific Parts Needed
9/27 – Design Review and Order Parts
10/4 – Coding Software and Building Housing
10/11 - Coding Software/Controller and Designing PCB for Controller
10/18 – Coding Software/Controller and Designing PCB for Controller
10/25 – Individual Progress Reports and Order PCB
11/1 – Mock Up Demo and Testing PCB
11/8 – Mock Up Presentation and Continue Debugging
11/15 – Finish building and Continue Debugging
11/22 – Thanksgiving catch-up period. Finish debugging and prepare Demo/Presentation
11/29 – Demo and Presentation
12/6 – Final Paper and Checkout