Optical: systems and elements – Optical amplifier – Correction of deleterious effects
Reexamination Certificate
2000-12-07
2003-06-24
Black, Thomas G. (Department: 3663)
Optical: systems and elements
Optical amplifier
Correction of deleterious effects
C359S337100, C359S337000
Reexamination Certificate
active
06583923
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention is directed to an optical fiber amplifier having a switchable or interchangeable fiber module for varying the effective length of an amplification fiber.
2. Description of the Related Art
Optical wavelength-division multiplex (WDM) transmission systems must be designed such that they can be operated on links having different amplifier spacings (physical). The problem that arises is illustrated as follows. Line attenuations between 33 dB and 14 dB are provided between the individual, multi-stage amplifiers. This results in the input powers per channel at the amplifier input lying between approximately −28 dBm and −13 dBm. The following presumes that the optical amplifier (inline amplifier) can be divided into a pre-amplifier followed by an attenuation element and a booster (power amplifier). Fixed levels (for example, 0 dBm/channel) at a photodiode are required for optimum detection, so that the power per channel at the output of the amplifier and, thus, at the input of the booster must be constant, i.e., independent of the gain in the individual amplifier sections.
In the above applied example, the gain of the amplifier, which is defined by the length of the doped fiber as well as by the average occupancy inversion in addition to being defined by the fiber parameters, must be capable of being varied between 13 dB and 28 dB. The variation of the amplifier gain is particularly easy via the pump power. When the pump power is reduced, the average occupancy inversion (and, thus, the gain as well) decrease. This, however, is unsuitable for WDM systems since the curve of the gain over the wavelength is highly dependent on the average occupancy inversion {overscore (N)}
2
. The average occupancy inversion {overscore (N)}
2
is defined as the normed average of the occupancy inversions N
2i
(for all ions with meta-stable levels) of all amplifier stages V
i
(for example, 0<i<3 for a pre-amplifier and a booster) over the entire length of the optical amplifier. The N
2i
can be functions of the location. Amplifiers for WDM systems must exhibit a flat gain curve that can be additionally realized with the assistance of specific filters in a defined an operating condition. When the average occupancy inversion is modified, then the individual wavelength channels experience a clearly different gain.
A flat gain curve can even be achieved without specific filters in the L-band (approximately 1570-1610 nm), in that an average occupancy inversion of approximately 35% (or approximately {overscore (N)}
2
=0.35) is set. This percentage is dependent on the erbium-doped fibers employed. Thus, for a gain of 30 dB, the gain differences amount to, e.g., only 1.8 dB. Noticeably greater gain differences occur, however, as soon as the average occupancy inversion changes. A pre-amplifier for WDM systems is therefore usually dimensioned such that it comprises a maximum required, constant gain. By inserting an additional attenuation, the gain can then be reduced to the value needed in a specific application. However, high noise coefficients arise given high attenuations for low gains.
The International Patent document PCT/WO98/36513, “Optical Fiber Amplifier Having Variable Gain”, discloses an optical fiber amplifier with a gain control for WDM signal transmission. The circuit is explained in
FIG. 2
of this document, in which a controlled attenuation element
5
is inserted between the two amplifier stages
3
and
11
. Three photoelectric elements
13
,
17
and
25
measure the light power along the amplifier, and regulate the attenuation via a controller. The goal of this regulation is to obtain a variable output gain of the amplifier given a constant curve of all output channel levels of a WDM input signal. Expediently, the “tilt” of the gain following the first pre-amplifier stage is compensated by a spectral gain smoothing filter
9
. A spectrally uniform gain can thus be achieved at the output of the amplifier for all WDM channels. An attempt is made in the general case to keep the gain and, thus, the noise of the first pre-amplifier stage low. The attenuation is boosted for high input levels. However, this still produces high noise coefficients.
The following paragraphs discuss a few versions optical amplification according to the Prior Art, illustrated in
FIGS. 1-4
. Their properties are explained and the disadvantages that are eliminated by the present invention are described.
As shown in
FIG. 1
, the amplifier is divided into two stages V
1
and V
2
between which a variable attenuation element DG is inserted.
FIG. 1
also shows the gain G and the signal power P (or level) along the fiber amplifiers for two different amplifications.
For low input levels of the WDM signal S, an amplification ensues with a high gain GG. The broken line refers to the high gain GG of a signal with low input power.
For high input levels, a low gain KG given a poor noise coefficient is achieved due to the high attenuation between the amplifier stages. The solid line refers to the low gain KG of a signal with high input power. The high levels are highly attenuated by the attenuation element DG between the two amplifier stages V
1
and V
2
.
The amplifier in
FIG. 2
comprises the same components V
1
, V
2
and DG as in
FIG. 1
, but the attenuation element DG is attached directly to the output of the second amplifier stage V
2
. This version enables a variable amplification with good noise coefficients for low and high gains G. The levels in the amplifier, however, will reach high values for signals with a high input power, resulting in non-linearities occurring particularly in the L-band for high levels in the amplifier stage V
2
. However, the average occupancy inversion {overscore (N)}
2
must be maintained given increased input power, requiring clearly higher pump powers. It can be clearly seen in
FIG. 2
that signals having high input power are amplified unnecessarily with high pump power and are in turn attenuated at the output. This version, however, solves the previously identified noise coefficient problems of the amplifier according to FIG.
1
.
FIG. 3
shows the typical structure of an L-band amplifier. The occupancy inversion is schematically shown as function of the fiber length L. It is critical that the occupancy inversion N
21
in the first section S
1
of the erbium-doped fibers EDF is very high in order to keep the noise coefficient low. The desired, average occupancy inversion {overscore (N)}
2
is then set with the following sections of lower occupancy inversion. The average occupancy inversion {overscore (N)}
2
is calculated as
N
_
2
=
1
L
G
∑
i
∫
L
i
N
2
i
(
l
)
ⅆ
l
,
where L
G
references the overall length of the erbium-doped fibers in the i various amplifier stages. N
2i
references the occupancy inversion of the i
th
amplifier stage with fiber length L
i
. Only two amplifier stages with corresponding occupancy inversions N
21
, and N
22
are shown in this example. The occupancy inversions N
22
should be as small as possible, so that the first fiber section becomes optimally long with a given, average occupancy inversion.
FIG. 4
shows a fiber amplifier that is adaptable to different input levels. A respective, doped fiber section EDFE and EDFA outside a module housing MG is located at the input and output. The gain is then varied in that these fiber sections are replaced by correspondingly shorter or longer ones. For a change of the lengths of the two fiber sections by &Dgr;L
1
and &Dgr;L
2
, where N
21
, and N
22
are kept constant and &Dgr;L
1
×L
2
=&Dgr;L
2
×L
1
is also valid, the average occupancy inversion {overscore (N)}
2
does not change and the relative gain differences continue to be slight. The outlay given this version is too great since an extensive set of fibers is needed for all required gain values.
Therefore, only a supply of a few fibers and, further, an attenuation element DG are provid
Bell Boyd & Lloyd LLC
Black Thomas G.
Siemens Aktiengesellschaft
Sommer Andrew R.
LandOfFree
Optical fiber amplifier 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 fiber amplifier, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optical fiber amplifier will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3102112