Optical: systems and elements – Optical amplifier – Optical fiber
Reexamination Certificate
1999-11-30
2002-01-15
Moskowitz, Nelson (Department: 3662)
Optical: systems and elements
Optical amplifier
Optical fiber
C359S199200, C359S337100, C359S341320, C372S031000
Reexamination Certificate
active
06339495
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to fiber optical WDM transmission systems and optical amplifiers used therein, and more particularly to an optical feedback resonant laser cavity (OFRC), including a power dependent loss element (PDLE) for optical gain control (OGC) or optical power control (OPC), and to a method for implementing such control, which is particularly useful, although not so limited, in amplified wavelength add/drop multiplexed (WADM) transmission nodes.
2. Technical Background
Wavelength division multiplexing (WDM) is a demonstrated technology for increasing the capacity of existing fiber optic networks. A typical WDM system employs multiple optical signal channels, each channel being assigned a particular wavelength or wavelength band. In a WDM system, optical signal channels are generated, multiplexed to form an optical signal comprised of the individual optical signal channels, transmitted over a single waveguide, and demultiplexed such that each channel is individually routed to a designated receiver. Multiple optical channels can be amplified simultaneously in optical amplifiers such as erbium doped fiber amplifiers (EDFAs), facilitating the use of WDM systems for long distance transmission.
Add-drop multiplexers are used, for instance, at nodes in a WDM communication network to extract one or more channels from the multiplexed stream, letting the remaining channels pass through unaltered to the next node, and to add to the multiplexed stream a new channel for transmission. Another application of such devices is for routing nodes of reconfigurable optical networks, namely rerouting certain information streams as a result of changed traffic conditions or to remedy a failure downstream from the node.
As schematically illustrated in
FIG. 1. a
conventional WADM node
120
consists respectively of gain-controlled input and output amplifiers
121
,
123
, a pair of 1×N and N×1 multiplexers/demultiplexers
125
,
127
, and an array of add/drop switches
129
. This type of WADM node is herein referred to as a 1×N×1 node since there is a single input amplifier and a single output amplifier, both of which are likely gain flattened and gain controlled amplifiers. Wavelength add/drop multiplexing permits signals to be routed from different networks or propagate through different spans. As a result, however, the per channel power after each add/drop switch can vary significantly, say by YdB. An approach to equalize channel power at the output of each node is to monitor per channel power and to use a variable optical attenuator (VOA)
131
in each channel path to maintain constant channel power. Due to the characteristically slow VOA response, however, the settling time for channel adding varies from milliseconds to seconds depending upon the VOA technology employed. Although VOA response times are not currently problematic (as some traffic interruption is expected for switched channels), there is a significant pump power penalty in the output amplifier imposed by the required protection of the surviving channels. For example, a VOA with a feedback control has to level the channel power which may vary between channels by a value of YdB (that is, by 10(
Y/10
) times as much on a linear scale), as mentioned above. To protect the signal in a worst case where N×1 channels are added simultaneously. all with a YdB excess power, the pump power of the output amplifier needs to support (10
(Y/10)
×(N−1))+1 channels of power before the VOA can respond. In other words, the pump power penalty to protect the surviving channel is almost YdB if N is large.
One suggested approach to address this problem consists of replacing the VOAs and the output amplifier of the 1×N×1 architecture with multiple parallel power equalization amplifiers (PEAs) as schematically shown in
FIG. 2
, to form what is referred to herein as a 1×N×N architecture because there are N optically controlled outputs. Each PEA can be designed to operate in its saturation regime so that the output signal power is determined by the pump power and is substantially independent of the input powers. Simulation results have shown that it is possible for the output power of these PEAs to differ by only 0.5 dB for an input power difference of 6 dB, and a ldB difference has been experimentally demonstrated. Although this approach is cost effective in that the VOAs and a complex output amplifier are eliminated from the system, and separate pump diodes for each PEA are replaced by a shared pump source, there is a recognized need for transient (as channels are added/dropped ) power control of each of the parallel amplifiers. Without such control the inversion of the amplifier is higher when a channel is dropped, and there is a large transient power spike when another channel is added back into the amplifier. Repeatedly amplifying this transient spike along the amplifier chain may result in component damage or an even greater pump power penalty to protect the surviving channels. However, because these parallel amplifiers share the pump source, individual transient control is not easily achieved by conventional electrical control such as by regulating the pump power.
One way of addressing this issue is to incorporate an optical feedback resonant laser cavity (OFRC) in each PEA so that the transient control is individually applied to these parallel amplifiers, even though they commonly share the pump power. For controlling the power transient of the PEAs, the OFRC is configured so that an optical power control (OPC) laser turns on (lases) when a signal channel is dropped from the PEA. Ideally, the OPC laser turns off (stops lasing) when a signal channel appears in the PEA, so that the PEA is only saturated by the signal channel and the signal channel extracts all of the available energy provided by the pump power. However, because there may be a power variation of YdB among the signal channels, the OPC laser has to be turned off by a signal channel having the lowest possible channel power. In order to achieve this operating condition, the cavity loss of the OFRC must be high. However, a high loss produces a low OPC laser power when the laser turns on, and thus a high amplifier inversion. As a result, if a signal channel with a high power is added into the highly inverted PEA which is saturated by an OPC laser with low power, there will be a transient power spike due to that high inversion. To eliminate the transient spike, the OFRC requires a lower cavity loss and higher OPC laser power. This presents the paradox of having a high loss and a low loss in the OFRC.
In addition to the transient power control of the PEAs, discussed above, the WDM input amplifiers shown in FIG.
1
and
FIG. 2
are gain-controlled to reduce steady-state (DC) gain error. A common technique to implement such control in a WDM optical amplifier involves configuring each amplifier with an OFRC such as an optical gain control (OGC) laser cavity. It is well known for such a configuration that the optical gain must equal the passive loss at the lasing wavelength. As a result, for a homogeneous medium the optical gain at all wavelengths in a given spectrum is locked once the gain is fixed at any particular wavelength. Thus the gain spectrum of the amplifier is determined once the OGC laser wavelength and the passive loss at that wavelength are determined.
It is now appreciated by those skilled in the art that an erbium doped fiber (EDF) is not a purely homogeneous medium for light amplification; rather, it exhibits a certain degree of inhomogeneity. This circumstance gives rise to the phenomenon of spectral hole burning. When the power of the OGC laser is increased, for example, by dropping channels or by increasing the pump power, the spectral hole at the lasing wavelength gets deeper and results in a steady-state (DC) gain error in the signal band. This is illustrated schematically in FIG.
3
(
a
). Accordingly, there is a need to so
Cowle Gregory J.
Hall Douglas W.
McNamara Thomas W.
Wang Chiachi
Corning Incorporated
Moskowitz Nelson
Short Svetlana
LandOfFree
Optical amplifier with power dependent feedback does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Optical amplifier with power dependent feedback, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optical amplifier with power dependent feedback will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2854765