Methods For Controllable And Flexible Imaging
Summary
Introduction
My name is Brian
Heflin and I am a graduate student working part time, around 20 hours per week,
at the Vision and Security Technology (VAST) laboratory at the University of
Colorado at Colorado Springs (UCCS) campus.
I took over the Methods For Controllable And Flexible Imaging project in
June of 2006. My preliminary work included research on a LCD screen light
filter, a MEMS microshutter light filter, and using a Spartan3 evaluation board
to drive test patterns on a VGA monitor at resolutions from 640x480, 60Hz to
1024x768, 70Hz. Unfortunately the Elphel camera was stolen before any work
could be done using it’s embedded FPGA. However, research on both the LCD
screen light filter and the MEMS microshutter light filter continued. The
following is a summary of the preceding report.
LCD
Screen Light Filter
The first solution
for a variable light filter is to place a semi-transparent LCD screen in front
of the Elphel camera’s CCD sensor. My first choice
for a LCD screen is the CyberDisplay 1280M. The abbreviated technical specifications of CyberDisplay 1280
are show below:
|
CyberDisplay 1280 Mono |
|
|
Display Type |
Monochrome Active Matrix Liquid Crystal
(AMLCD) |
|
Display Mode |
Transmissive, normally white TN (twisted
nematic) |
|
Columns and Rows |
1,280 x 1,024 (SXGA: 1.3M pixels) |
|
Resolution |
1,700 lines per inch (15 um pixel pitch) |
|
Grayscale |
Continuous |
|
Interface |
Digital timing and control; analog video |
|
Display Logic |
3.3V to 5.0V |
|
Video Inputs |
8 |
|
Power Consumption |
90mW |
|
Active Display Area |
0.96" diagonal (19.2mm by 15.36mm) |
|
Pixel Optical Aperture |
50% |
|
Frame Dimensions |
1.2" by 0.9" by 0.13" |
Synthesized Verilog
code will be used to generate the digital signals necessary to drive the
CyberDisplay 1280M. Additionally, an
algorithm will need to be developed and synthesized into the FPGA to provide
the grayscale values to the CyberDisplay for each pixel, based data from the
Elphel camera’s CCD sensor. Hardware additions such as a connector between the
FPGA and the CyberDisplay, and an n-bit digital to analog converter, will also
be needed. The CyberDisplay 1280M has analog grayscale (256 levels for 8 bit
input). Kopin also has two other monochrome CyberDisplay modules, the 320M and
640M with screen sizes of 0.24” and 0.44” respectivley. Kopin also has a WVGA
color module with screen sizes of 0.58” and a resolution of 854x480. An Internet search for information and
timing diagrams for the CyberDisplays yields few results. The only information
that can be found is on a few web sites by a few people making home-brew head
mounted displays (HMD) with the 320M.
However, I did find the complete specifications for CyberDisplay 320M
and 320C on the Kopin web site. The
specifications for the 640M, 1280M, and WVGA were not available, but they
should be similar. Unfortunately, when I spoke with Kopin they told me that the
640M and 1280M are for military applications only and they do not sell them to
the public. Special permission will need to be obtained to purchase the
CyberDisplay 1280M screen from Kopin. The lower resolution 320M displays is
available from Kopin to the public for $95.00. Luckily, I did recently bid and
win 42 CyberDisplay 320M displays on Ebay for $32.00.
Conclusion
In conclusion, to
date I feel that the CyberDisplay 1280M is the best LCD screen choice for this
project. However if a 1280M cannot be obtained, the WVGA screen is a comparable
second choice. Even though the WVGA is
a color module, it can display 256 levels of gray. In order to get a gray scale the red, green, and blue channels
will just have to be set to the same intensity. The recently acquired 42
CyberDisplay 320Ms will allow for preliminary work to begin, until a CyberDisplay
1280M or WVGA screen can be purchased.
MEMS
MicroShutters
A
second solution for a variable light filter is to develop a
Micro-Electro-Mechanical Systems (MEMS) MicroShutter system that could be
placed in front of the Elphel camera’s CCD sensor. The MEMS microshutters
differ from the LCD screen light filter in that the microshutters will be
either “on” or “off” as compared to the LCD screen that has an n-bit
grayscaling capability. My initial MEMS
research consisted of searching the internet for introductory MEMS information.
The books “An introduction to
Microelectromechanical Systems Engineering” by Nadim Maluf and “Fundamentals of
Microfabrication The Science of Miniaturization” Marc J. Madou also proved to
be excellent books for an introduction to MEMS systems. After gaining a basic understanding of MEMS
systems and processes, I searched for any MEMS micro-shutter systems that are
currently or already have been developed. After I researched current micro-shutter
projects, I began looking for individuals and possible courses offered at UCCS
that may have be able to help with this project. Apparently, UCCS used to have
a MEMS class and laboratory. However,
the class has not been offered in over 2 years and the MEMS lab has been torn
down. I did find two professors at UCCS that have experience with MEMS
systems. They are Dr. Carlos Arajuo and
Dr. T.S. Kalkur. Dr. Arajuo e-mailed me
that he has a bible for MEMS that has been used in the past at his company
Symetrix, and a course on MEMS by the National MEMS Facility that he “would be
happy to speak with me about.” Additionally, I spoke with Dr. Kalkur in
person. Dr. Kalkur also said that a
MEMS micro-shutter system would be extremely difficult to make. In addition to, it would take around 2+
years to set up a lab at UCCS to even begin a MEMS process. Dr. Kalkur
suggested that I could do the design, analysis, and simulation of the
microshutters and a fabrication laboratory such as Sandia National Labratories
would have to do the actual fabrication of the microshutters. I have also tried
to contact Dr. Victor M. Bright, Ph.D.
Dr. Bright is the Associate Dean for Research and Professor of
Mechanical Engineering, College of Engineering & Applied Science, University
of Colorado at Boulder. Prof. Bright's research activities include micro- and
nano-electro-mechanical systems, silicon micromachining, microsensors/
microactuators, opto-electronics, optical, magnetic and RF microsystems,
atomic-layer deposited materials, ceramic MEMS, MEMS reliability, and MEMS
packaging. Unfortunately, to date I have been unable to contact him. Currently,
I am still trying to find someone to help to determine the feasibility of some
of the various ideas for a MEMS shutter system. Figures 1, 2, and 3 below show
some of the ideas for the MEMS micro-shutter structure.
Once
the feasibility and selection of a microshutter structure is determined the
project will proceed with a general design Methodology for a MEMS device which
is listed below:
1.
List of Specifications for the MEMS device and system
2.
Identification of the general operating principles and overall structural
elements
3.
Analysis and Simulation of the MEMS device (Simulation of mechanical, thermal,
and electrostatic structures)
4.
Layout of the lithographic masks
5.
Outlining the individual steps in the fabrication process.
6.
Fabrication of the Modules
7.
Post-Fabrication tasks.
Analysis and Simulation of the MEMS system
Once
the list of specifications for the MEMS device and system and identification of
the general operating principles and overall structural elements is complete,
analysis and simulation of the shutter and actuator system will need to be
performed. I have currently found two analysis and simulation tools for MEMS
devices. The first program is called
SNL MEMS Design Tool and available from Sandia National Laboratories. The
second MEMS analysis and simulation tool called MEMulator is available from a
company called CoventorWare. However, Sandia’s MEMS fabrication program is not
compatible with CoventorWares MEMulator software program. Once the analysis and
simulation using the software is completed fabrication of the MEMS microshutter
system can begin.
Fabrication of the Micro-Shutters
Without
any available MEMS fabrication tools at UCCS, a company such as Sandia National
Laboratories would have to perform the actual fabrication of the MEMS
microshutter modules. Sandia National Laboratories has a MEMS prototype program
that allows customers to have a share batch fabrication process of MEMS
modules. Customers receive at least 100 unreleased die, and fabrication costs
are shared among customers, $10,000 per module. Below is a picture of Sandia’s MEMs process:
Post-fabrication
tasks
After
fabrication, the MEMS shutter system will have “packaged” or put in a
protective housing. Additionally,
interconnects for the electrical signals and pins for external connections need
to be added. Sanida can include custom
packaging as a part of their SAMPLES program. Additionally, reliability,
characterization, and failure analysis must be performed on the shutters.
Reconfigurable software running on the FPGA test board can be used for some of
the analysis testing and experimentation.
Unfortunately, the need for multiple microshutter prototypes may arise.
Conclusion
In
conclusion, with the help of a company, such as Sanida National Laboratories,
design, simulation, and fabrication of a MEMS microshutter device is possible.
The price list for all of Sandia’s services is listed in Appendix A. Conceivably, engineering of a MEMS
microshutter system may prove to be exceptionally difficult due to the lack of
local help and resources. Moreover, this approach will take longer to complete
because interfacing and experimentation and with the Elphel camera can only
begin after design, analysis, simulation, fabrication, packaging, and testing
of the MEMS microshutter system are complete. Finally, MEMS microshutters are
an alternate approach to allow controlled modulation of light from ambient or
other sources. Numerous MEMS microshutter systems have been successfully
developed and have also been the top design choice for numerous light
modulation projects, including a project being developed by NASA for the James
Webb Space Telescope, which is scheduled to replace the Hubble in 2013.
.