Optical Fiber Injection Alignment System
EE492 Spring 2010
Project Advisor: Dr. Wataru Nakagawa
Group Members:
David Weir
Matt Strathman
Project Description:
In laboratory testing of Silicon-on-insulator (SOI) optical waveguides, light must be injected into
the waveguide under study from a laser source, normally through an optical fiber. Due to
vibrations or other environmental factors coupled with the dimensions of the guide (from
hundreds of nanometers to a few microns), maintaining a stable coupling between a tapered
optical fiber tip and the waveguide can be a challenge. This project involves the development of
a computer-controlled alignment system using the signal detected at the output of the waveguide,
and a piezo-electric nanopositioning stage. The primary goals of the project are to build a
prototype injection set-up, and develop the control system for injection alignment on a platform
compatible with existing laboratory automation hardware and software.
Our solution to this problem includes the use of a 3-axis piezoelectric stage that is computer
controllable. With this stage the position of the fiber tip can be controlled by a computer
algorithm. We built a testing setup to test the system as well as developing an algorithm that
uses a PD control loop to maximize the alignment of the system. Our algorithm moves the fiber
tip in each direction in succession measuring the slope of the power vs. position curve. This
information is used to determine both the direction of movement and the magnitude of movement
for the fiber for the next cycle of the algorithm. The operation of the algorithm was first
tested on various sized pinholes to simplify the optical setup and possible sources of error,
and will soon be tested on a more complicated waveguide setup.
Design Challenges:
The major design challenges that we have encountered so far on this project came from the
algorithm design and interfacing between the piezo stage controller and the computer. For
the algorithm, we needed something that would both align the system quickly and when it gets
to the position of maximum alignment move only very slightly to minimize the noise in the
output power. This was accomplished with an control loop that moves the fiber tip by a small
amount and determines the slope of the power curve and calculates following movements based
on that data. This algorithm allowed the system to make large position corrections when it
is far away from the maximum alignment and very slight movements when it is aligned. The
second major design challenge comes from communication between the computer and the piezo
controller. Because the piezo controller is an analog system and the computer is a digital
system, the use of digital to analog and analog to digital conversion is necessary. At first
when the system was tested on very large 25 micron and 50 micron pinholes the fiber was moving
enough that the resolution of the 16-bit digital to analog converters was not a problem. But
with the 5 micron pinhole the minimum movement of the stage was large enough to greatly
increase the noise in the system. The minimum movement of the stage comes from the minimum
resolution of the d to a converter which is about 3mV. This meant that the minimum movement
of the stage was about 25nm. To solve this we will design a circuit that divides the voltage
out of the d to a converter by a constant value, this decreases the maximum range of the system
but it also increases the precision.