[WP1] [WP2] [WP3] [WP4] [WP5] [WP6] [WP7]Objectives
The objective of this WP is to investigate new concepts for
POF transmitter and receiver components
Description of work
In POF transmission systems to date there are two major
bottlenecks regarding available components. Because of high
fiber attenuation and inefficient LEDs and fiber-chip-coupling
the transmission distance is limited to 50-100 meters; second,
the speed of POF systems is limited to max. 1 Gbit/s because of
limited performance of laser diodes and receiver photo diodes.
In this work package, the above mentioned bottle necks shall be
investigated in detail and measures for improving the
performance of those components shall be applied with the goal
to improve distance for low speed systems and bandwidth for high
speed systems.
In a first step, commercially available transmitter and receiver
components shall be reviewed and measured in detail in order to
extract the potential for performance improvements.
The following work will concentrate on the components LED and
LED-POF-coupling, laser diodes, laser-POF-coupling and receiver
photo diodes. DieMount already has a concept to improve
LED-POF-coupling with microreflector techniques. At present this
method works well with standard POF with a high NA. If future
edge network installations will use high speed fibers with lower
NA, the active components will have to be coupled to this fiber
even if the bandwidth of the fiber is not needed for all
applications. So it will be the main work for DieMount to
investigate the translation of their concept to coupling into
POF with lower NA (<0,5).
Limitations for high speed transmission with thick core,
PMMA-based POF arise primarily by the edge-emitting laser diode
operating at a wavelength of 650 nm. It is believed that an
essential part of the bandwidth limitations are due to the way
the naked laser die is mounted in a TO package. Therefore a goal
of this WP is to characterize the naked laser dies without
package and with deembedding techniques, respectively, so that
the performance of the diode itself can be estimated.
Furthermore, parameters for a laser diode model shall be
extracted in the frequency range up to 3 GHz by FHG. The method
also will be applied for blue laser diodes with a wavelength of
405 nm, since at this wavelength the losses in POF are even
lower. Currently, blue lasers are built in Blue Ray Disc drives
and will fall dramatically in price soon. The laser model will
be used in WP 2 for designing matching networks and driver
electronics.
In order to perform laser diode characterization, the laser has
to be mounted on PCB in such a way that parasitics can be kept
as low as possible to fully exploit the high frequency
performance of the laser crystal. Moreover, special attention
has to be paid to thermal matching of the laser crystal and the
board or submount material to avoid mechanical stress for the
laser. Finally, fiber-chip coupling has to be done including
optical simulations of coupling between active devices and POF
and design and optimization of an optical coupling system which
allows high coupling efficiency. All the work for mounting and
packaging of optical modules including fiber-chip-coupling will
be done by STMicroelectronics, that will study and develop a new
alignment method that allows high coupling efficiency between
POF and active components.
Another approach to laser packaging is to use microreflectors in
order to be able to couple the light from the back mirror into
the fiber, too. This approach will be taken by DieMount who will
have to modify their microreflector technique developed for LEDs
for the use with lasers. In order to maintain stable operation
of the laser (then without the monitor diode), a temperature
control circuitry will also be integrated as a part of the laser
module, where the laser diode temperature will be measured and
the laser current will be controlled accordingly. For cost
reasons, no active cooling shall be used here.
Another important part of this WP is the development of
high-speed, large area photodiodes and transimpedance amplifiers
(TIA). The idea is to use differential photodiodes (sometimes
called spatial light modulators in the literature) built of
pn-junctions in a standard CMOS technology. With this technique,
the bandwidth limitations due to carrier diffusion can be
overcome, performance in the Gbit/s-range has already been
shown. To solve the problem of diode capacitance, a large area
photodiode receiver will be constructed by a large number of
small photodiodes, where the diodes´ currents will be added via
cascode transistor stages and subsequent amplification by a
transimpedance amplifier. This technique can only be applied
with integrated photodiodes, where a large number (approx.
25-100) of photodiodes, the corresponding cascode transistors
and the TIA can be arranged closely together. Therefore, this
concept can only be realized with an ASIC solution. Moreover,
for high volume applications this concept offers the possibility
of low cost receiver front ends since a standard CMOS technology
is used. Design work will be done by FHG, ASIC fabrication will
be done via the EUROPRACTICE IC service.
Finally, the receiver chips with TIA will have to be implemented
in a package in an efficient way. Here again the microreflector
concept from DieMount and STmicroelectronic new alignment scheme
will be used and applied to receivers instead of LEDs. Extensive
measurements of improving efficiency with this packaging concept
will be performed with various POF types and compared to
mounting in TO package.
This WP will be led by STm. The components developed in this WP
will be essential for use in WP 1-2. For any of the subsequently
described WP3 tasks, intermediate version of the components will
be made available to WP1, WP2 and WP5 activities as soon as
possible.
|