Sergei Skorobogatov's student project suggestions
Here are some ideas towards Part Ib and Part II projects in
Computer Science. The descriptions provided are only vague ideas, not
ready-to-use proposals for a particular type of project.
I will take care after all of the hardware designs necessary for
the projects, so unless explicitly requested most of the projects will
involve only writing software (either for a PC or a microcontroller).
Joystick control of a motorised stage with displaying current position
Resources:
- Motorised stage controlled via GPIB or RS-232
- PC with GPIB card or USB-GPIB adapter and joystick port
OR
- Stand-alone microcontroller-based device with RS-232 or
USB-master interface and LCD (4x16 or graphic)
- RS232-GPIB or USB-GPIB adapter
Details:
For optical observation under a microscope it is easier to move a
sample using a motorised stage controlled by a joystick. The movement
should be sensitive to the angle of the joystick handle and the speed
should be increased if the handle is at the farest position. The LCD
should always display the absolute coordinate of the stage (can be
requested from the controller). For measurements, relative coordinates
can be set and displayed. It is difficult to keep the image always in
focus therefore compensation for the Z-coordinate should be provided
and automatically adjusted during the X-Y move. The software should
have convenient GUI for controlling the motorised stage parameters
such as acceleration, speed and travel limits and give access to most
of the controller commands.
Current problems:
- All the necessary hardware is ready for a PC-based project
(XYZ-stage with controller and PC with GPIB card), but the stand-alone
hardware necessary for microcontroller-based project is not designed
yet, so a stand-alone device could be a separate project once the
PC-based is finished
Laser scanning with a motorised stage
Resources:
- Optical microscope with laser pointer (less than 5mW, classII)
- Motorised stage controlled via RS-232 or GPIB
- Digital oscilloscope with deep memory and GPIB interface
- PC with GPIB card or USB-GPIB adapter
- Computer controlled data acquisition preamplifier
Details:
Laser scanning is used in semiconductor failure analysis for various
applications from detecting p-n junctions to reading the transistors'
state. The project is aimed on buiding a laser scanning system that
moves the sample using a motorised stage so that the laser hit all the
points on the surface. It is difficult to keep the image always in
focus therefore compensation for the Z-coordinate should be provided
and automatically adjusted during the X-Y move. The response is
acquired with an oscilloscope. Then the data is collected from the
oscilloscope and analysed by software. First, the linearity should be
restored because the stage does not move linearly in the beginning and
the end of the specified travel. Then the result should be presented
in suitable graphical form, for example, by using a MatLab.
Current problems:
- An oscilloscope with 'deep memory' option should be used to
acquire the whole scan image
- Stage movement is non-linear and should be compensated later in
software
Merging digital pictures
Resources:
- Picture files taken with a digital camera using a microscope
Details:
Once the surface of a chip is photographed it is necessary to combine
all the images together into a single database. A user-friendly shell
should be used to navigate over this large image with the possibility
to zoom in and out, measuring distances between any objects and
retracting relative coordinates for any point. The program should be
able to manage several planes (for multilayer chips) synchronising
them and displaying either one selcted layer or 4 layers (1/4 of a
screen) at a time.
Current problems:
- We assume that the images do not have any geometric and color
distortions
- If we combine the images taken with high magnification (>1000x) we
end up with thousands of images to be merged into a single file. Such
file will be very difficult to manage. It might be easier to cut the
combined image into smaller sections
Projects listed below cannot be implemented straght away because some
hardware issues are not sorted out yet.
Taking a mosaic of pictures of a semiconductor chip
Resources:
- Microscope with reflected light illumination
- Motorized stage controlled from a computer
- Digital camera with USB or FireWire interface
- PC with USB/FireWire and GPIB card or USB-GPIB adapter
Details:
A first step in reverse engineering of semiconductor chips is
preparing a high-resolution optical image of its surface. This
involves taking digital pictures under a microscope. Because the
pictures are usually taken with a high magnification, the amount of
pictures very often exceeds one hundred. This is almost impossible to
do manually and a computer controlled motorised stage and a digital
camera should be used.
Current problems:
- The microscope to be used is a Leitz Ergolux AMC. It has fully
automatic control over objectives, filters and motorised stage via a
GPIB interface. The command interface is not documented and it is
currently switched to the manual mode with a remote controller
- The digital camera we have at the moment is a Nikon CoolPix 4500. It
has a USB interface for downloading the pictures but it can be
controlled from a computer via an RS-232 interface only. The problem
is that both interfaces share a single connector at the camera and
therefore cannot be used simulteneously. Some sort of custom adapter
should be built to have both interfaces together or a new camera
bought
Laser scanning mirror interface and control
Resources:
- Laser pointer (less than 5mW, classII) mounted on an optical
breadboard
- X-Y fast steering mirror mounted on an optical breadboard
- Motion controller to move the mirror with RS-232 or GPIB interface
- PC with motion controller interface (RS-232, USB or GPIB)
Details:
For real-time laser scanning applications either rotating prizms or
voice-coil mirrors are used. This allows one to scan the surface with
2-5 frames/second speed which is enough for laser scanning
microscopy. The same idea is used to draw laser pictures onto a wall
where the laser is modulated with a video signal.
Current problems:
- The mirror interface is not documented but it seems to be bare coils
- It is not clear whether a standard motion controller (e.g. Newport
MM3000) can be used to drive the mirror
- We do not have the breadboard and the mounting brackets yet
Sergei Skorobogatov
<Sergei.Skorobogatov (at) cl.cam.ac.uk>
created 08-10-2004 -- last modified 09-10-2004 -- http://www.cl.cam.ac.uk/~sps32/