Field of the invention
The present invention relates to a method and system for power
control of optical transmission span involving balancing power provideds
at the output of a network element handling multiple optical channels and
being part of a chain of network elements in a telecommunication line.
Background of the invention
Those skilled in the art are familiar with a problem that, in a point-
to-point optical communication line (which comprises a number of
network elements, such as optical fiber amplifiers OFA connected in
series, and possibly, one or more optical add-drop multiplexers OADMs),
one should pay attention to power equalization between multiple optical
channels outputted by one network element to be sent to another one. 15
The amplifiers may work according to one of the following principles:
1. the principle of fixed output power (when the fixed output power
is divided to a required number of the output channels - which leads to
underpowering if the number of channels is max, and to a very dangerous
overpowering where the number of output channels is low, say in a case
of fiber cut before an OADM that adds few channels into an OFA.
2. the principle of fixed power dependent on the number of
channels, wherein the amplifier output power is controlled with respect to
the number of channels, so as to provide a fixed output power per
channel. (for example, one channel - lmW, 15 channels 15 mW, 40 0125
more channels - 40mW max). In the known technologies, the number of
channels is usually stated (say, by an operator) from time to time.
3. the principle of fixed gain; according to that, the amplifier has a
fixed gain so that each channel, whatever input power exists in it, is
amplified with the same gain as other channels. To maintain the required
output power, it is known to arrange a feedback loop measuring the input
and the output power and, based on the result, capable of regulating the
output power of the amplifier. In the fixed gain scheme, in order to
obtain equal and fixed output powers of the channels at the output of the
amplifier, it is only possible to ensure that the input powers be equal and
fixed. To maintain the condition, it is kaown to insert a VOA (Variable
Optical Attenuator) into the input optic fiber, to adjust the input power of
the optical channels which arrive to the amplifier.
The prior art does not comprise any idea of a dynamic span power
equalization or control in a chain of network elements, comprising OF As
and OADMs. The problem stems from the facts that a) no dynamic
monitoring of the chain parameters has been proposed, and b) elements of
the chain may be different and thus behave differently, but no common
concept of optical power control has been proposed for such a chain.
Object of the invention
The object of the present invention is to propose a technique for
power control in a multi channel optical transmission line, which at least
partially overcomes the above disadvantages.
Summary of the invention
The main idea is a method of real time control of power per optical
channel in a multi-channel optical communication line formed by a group
of optical elements connected in a chain by fiber spans, wherein the
group of elements comprises one or more optical fiber amplifiers (OFA),
and wherein each of the spans is characterized by its span loss, while each
of the OF A is characterized by its gain and its designed output power per
channel,
the method comprises the following steps performed either
periodically or continuously:
calculating, for a particular optical amplifier (OFA) in the line, an
expected total input power value ED?,
measuring a real total input power (measured input power MI?) at
said particular optical amplifier, and
if a difference between the expected total input power and the real
total input power at said particular OFA exceeds a predetermined value,
adjusting the gain of said OFA to maintain its output power per channel
constant.
Preferably, the step of calculating the expected total input power is
performed using up-to date values of NOC (number of active incoming
channels) and NOA (number of preceding optical amplifiers in the line).
NOC can be determined, for example, by providing spectral
analysis in real time at each particular optical element. Industrial
spectrum analyzers are known in the art.
Alternatively, NOC can be determined with the aid of messages
transmitted in the line via a supervisory channel, using the method
described in the Applicant's earlier patent application IL 145262 and the
corresponding US 09/962,337, priority of September 4th, 2001 which is
incorporated herein by reference.
NOA of an optical amplifier and of a passive optical element such
as OADM can be determined as follows:
For OFA: NOA out = [(NOAin +1) and (not LOS)] or (LOS)and
l;
For OADM: NOA out = [ NOAin and (not LOS)] or (LOS) and 0;
where NOA out - NOA at the output of the optical element,
NOA in - NOA at the input of the optical element,
LOS -Loss Of Signal alarm, accompanies a fiber cut condition.
The step of maintaining of the output power per channel constant
means maintaining it substantially equal to the designed one, i.e., to the
output power per channel stated by a preliminary design of the line (in
other words, by the line configuration).
The method is hybrid since it combines the approach of fixed gain
and the approach of fixed power per channel i.e., each OF A works with
the fixed gain up to the moment when its output is to be corrected.
Actually, the Inventors propose continuous monitoring of both the
expected total input power of an OF A, and the real input total power in
order to control the OFA, to maintain its output power like it received the
expected input power and not the real one. It should be noted that when
designing the line, the capability to introduce possible corrections should
be taken into account say, when selecting gains and/or types of OF As,
ranges of power loss of the fiber spans, and control circuits of some
passive optical elements such as OADMs.
Preferably, the method is performed at all optical amplifiers OFA
in the line, to maintain constant output power at the fiber spans associated
with said OFAs, thereby enabling to avoid overpowering and under-
powering at the outputs of OFAs by taking into account possible changes
in the NOC and NOA,
The method is most advantageous when the group of elements in
the line, in addition to the optical amplifiers, comprises one or more
optical add drop multiplcxers (OADMs). The reasons for that are as
follows.
Firstly, the method is intrinsically adapted to take into account
changes in NOC which are most usually caused if a fiber cut occurs
before OADM, and/or if any rearrangement is performed in OADMs.
Secondly, the method will be suitable for equalizing output powers of
channels outputted from OADM, which is essential for the proper
operation in the line.
It can be recalled that any OADM is adapted to receive one or
more incoming optical channels, to drop one or more of the received
optical channels, to pass through the remaining received optical channels,
to add one or more new optical channels and to output the added optical
channels and the though (passed) optical channels. Any OADM is
therefore characterized by three values of insertion loss: the first one for
the through channels (EL through, from line input to line output), the second
one for the dropped channels (EL drop, from line input to drop output) and
the third one - for the added channels (EL add, from add input to line
output).
For example, if an OADM receives a weakened input signal while
it should not have been weakened (say, the number of channels has not
changed and there is just a contact or a fiber degradation due to which the
real total input power is lower than the expected total input power), and
no measures are taken to take care of these effects, the outgoing through
channels will be essentially weaker than the outgoing added channels.
Such a result is highly undesired for further transmission. To prevent it,
the input power of the added channels (i.e., the power of the added
channels) can be controlled accordingly.
Based on the above, the preferred version of the method
additionally comprises the following steps performed periodically or
continuously:
calculating, for a particular OADM in the line, an expected total
input power value, preferably using up-to date values of NOC (number of
active incoming channels) and NOA (number of preceding optical
amplifiers in the line),
measuring a real total input power at said particular OADM,
in case of a difference between the expected total input power and
the real total input power at said particular OADM exceeds a
predetermined value, controlling power of each of added channels of said
OADM to follow said difference, thereby to equalize output power of all
optical channels outgoing from said OADM.
In other words, adjusting the power of the added channels is
provided to make the output power of each of said added channels equal
to the output power of a through channel of said OADM, or to the
average of the through channels output power.
Preferably, all OADMs in the line are monitored as proposed
above, to ensure the equalized output power per channel in the fiber spans
associated with said OADMs; provided that the spans associated with the
OAs are also controlled, all fiber spans in the line will thereby be
controlled to have constant power per channel for the OF As and
equalized power per channel for the OADMs.
As has been mentioned, the method formally comprises a
prelirninary step of designing (pre-configuring) the optical
communication line and stating parameters from the following no
exhaustive list including: gains of the optical amplifiers, span losses,
stating initial values of NOC and NOA for any point in the line, and also
an expected output power per channel for each optical amplifier OF A, In
our method, this expected output power per channel is further maintained
to be constant
Generally, calculating the expected total input power (EIP)
comprises determining it for the input of any particular optical element,
based on information about gains of the preceding optical amplifiers,
span loss from the previous amplifier up to the particular optical element,
and updated values of NOC andNOA for said particular element.
Preferably, the calculation takes into account also noise figures of the
OF As, or the average noise figure thereof.
The proposed formula for calculating EIP will be presented in the
detailed description of the invention.
So, if the main idea of the invention is to have reliable means to
calculate the actual (up-to-date) expected total optical power at any point
in the network and to control the elements in the line, to compensate for
any difference from the expected value, the formula of calculating EIP
contributes to implementing the idea.
Returning to the proposed method of maintaining the expected
power per channel, whenever the up-to-date EIP is calculated and
compared with the measured total input power MIP, the difference will
indicate the required correction to bring the optical element to a regime
for controlling the output power per channel (for example, for ensuring
the expected output power per channel the OF As, as preliminarily
designed).
The second important idea of the invention is to provide a universal
way of determining the required correction for various types of optical
elements in the line.
The difference between the updated expected total power and the
real total power at a particular point in the line constitutes the required
correction RC (R C = MIP - EIP [dB]) to be applied to the optical
element to ensure at its output either a constant power per channel (for
OFA) , or the equalized power per channel (OADM).
It should further be noted that a response time (time passing
between the moment of detecting a difference between the EIP and MIP
and the moment of introducing a compensating correction - Hold Off
Time, HOT) is to be selected so that every element in the line can be
updated on the NOC and NOA and be able to correct its output power not
simultaneously, but after the previous elements have done the corrections
if such were required.
As has been mentioned, elements in the network communication line
may be of two types - optical amplifiers OFA and optical add drop
multiplexers OADMs. According to the preferred version of the method,
the proposed concept of power control is common to all elements in the
line, though has its specific features for any one of the element types.
The OFA output power is corrected in inverse proportion with the
measured input total power. In other words, in case the measured total
input power is, for any reason, higher than the expected input power
(MIP>EIP), the correction should be calculated to reduce the OFA gain,
thereby to reduce the output total power and to avoid the dangerous
overpowering. In the opposite case, if the measured (real) input total
power is lower than the EIP (for example, there is an unexpected extra
span loss before the optical element), the OFA gain is to be increased so
that to compensate the extra span loss.
Gain new = Gain old- RC [dB]; where RC=MIP-EIP [dB]
It should be noted that OADM can be of two lands - a
conventional OADM and a so-called VMUX being an OADM formed by
a combination of a demultiplexer (DMUX), a multiplexer (MUX), drop
fibers, through fibers and fibers for add channels with variable optical
amplifiers (VOAs).
In OADM and VMUX, the power of an added channel is corrected
in a direct proportion with the measured input power.
Therefore:
APPC new = APPC old [dBm] + RC [dB],
where:
RC=MIP-EIP [dB]; APPC - added power per channel.
Though the above-described option to equalize power of output
channels of OADM is preferred since it is accurate and universal (i.e.,
uses most of the calculations required for controlling output power per
channel of OF A), there are other simpler options which also exist and can
be used for controlling output power of the OADM added channels.
For example, for optical communication lines transmitting a great
number of channels the noise produced by amplifiers is relatively low
uniformly and thus can be neglected in the calculations. Therefore, the
power of an OADM added channel can be calculated by measuring MIP
and updating NOC, but without taking into account the NOA and noise
figure parameters:
APPC = [MIP - 10log(NOC)] - IL oadm ± CO [dBm],
where APPC - is the added power per channel,
[MIP-10log (NOC)] - gives an average input power per one
incoming channel, which serves an indication of an output power of a
through channel of the OADM;
IL oadm - insertion loss introduced by OADM; this parameter
takes into account the attenuation created by OADM to the add channel
when the output power thereof is equal to the output power of the through
channel; it can be estimated as IL oadm = IL through - IL add.
± CO is a manually introduced channel offset, which is usually
selected for a particular channel in order to give it a pre-emphasis.
There is further provided a system capable of implementing any
version of the above-described method,
According to yet another aspect of the invention, there is proposed
a module suitable for controlling output power per channel of an optical
element in a telecommunication line comprising at least OF As, and
optionally OADMs interconnected by fiber spans, the module being
capable of performing operations of the proposed method to serve the
corresponding optical element.
For example, it can be an optical module comprising an optical
element (OFA or OADM) and a control unit for controlling output power
per channel of the optical element in real time,
said optical element being designed to be coupled to an optical
communication line via optical fiber spans and capable of receiving an
incoming multi-channel optical signal to form an outgoing multi-channel
optical signal,
said control unit, in real time, being capable of:
calculating a value of expected total input power (EIP) of the
optical element based on a number of parameters stated by a preliminary
design of the line, and a number of parameters changeable during
operation and including at least a number of active optical channels
(NOC),
obtaining a value of measured total input power (MIP) of the
optical element,
comparing the EIP with the MIP and, if the difference there
between exceeds a predetermined value,
producing a signal of a required correction to be applied to the
optical element for controlling the output power per channel of the optical
element.
The changeable parameters may include also NOA (number of
preceding optical amplifiers), and the control unit is preferably operative
to obtain up-to-date values of NOC and NOA for said optical element.
The control unit is preferably capable of calculating EIP according
to the formula mentioned in the description of the method.
The optical module can be adapted for serving either OFA or OADM.
However, OADMs may be served by optical modules of a second
type, such a module comprising an OADM and a control unit for
controlling power of added channels of the OADM, the power of an
OADM added channel can be calculated by measuring MIP and updating
NOC:

Brief description of the accompanying drawings :
The invention will further be described and illustrated with the aid of the
following non-lirmiting drawings, in which:
Fig. la illustrates a telecommunication line comprising a number of
OFAs connected by fiber spans.
Fig. lb illustrates a time diagram explaining the proposed principle of
optical power control in the line shown in Fig. la.
Fig. 2 illustrates a line similar to that in Fig. 1 a, but comprising a passive
optical element being OADM.
Fig. 3 illustrates a flow chart of one version of the method according to
the invention, for controlling optical power in the line.
Fig. 4a schematically illustrates a block diagram of OADM with insertion
loss characteristics thereof with respect to added, dropped and through
channels.
Figs 4b- 4d illustrate power spectrums of various channels of OADM at
different conditions in the line.
Fig. 5 - is a block-diagram illustrating one way of controlling the power
of added channels of OADM .
Detailed description of the preferred embodiments
The method described in the present application starts from a preliminary
step of designing (pre-configuring) an optical communication line,
comprising optical amplifiers and OADMs, by stating gains of the optical
amplifiers, span losses, stating initial values of NOC and NOA for any
point in the line, and also an expected output power per channel for each
optical amplifier OF A. This expected output power per channel is further
maintained to be constant.
The line is preliminarily designed (configured) and calculated to
ensure, at each point thereof, the balanced (i.e., constant or equalized)
power per optical channel in each span of the line at normal (predictable)
conditions of the line's operation. The normal or predictable conditions
are characterized by the following groups of parameters:
a) configuration parameters which must not change during the
routine operation (nominal gains of the amplifiers; noise figures
introduced by the amplifiers; spaa losses i.e., values of
attenuation of the fiber spans);
b) parameters of the line which are pre-set for a particular
ccnfiguration, but may change during the normal operation and
should therefore be undatable, for example: a number of active
optical channels (NOC) at a particular point of the line may
change due to new settings at any preceding OADM or due to a
fiber cut in any preceding span; a number of optical amplifiers
before a particular point (NOA) may change due to a fiber cut
in any preceding portion of the line.
According to the invention, changes of the updatable parameters
(b), if occur, are used for updating the expected total input power.
Updating the NOC and NOA can be done automatically via a supervisory
channel in the optical line, either by element to element communication,
or via a central control block calculating the NOC and NOA for each
element in the line. In other words, each of the elements (or the central
control block) should be capable of calculating the NOC and NOA based
on the incoming NOC and NOA and LOS condition (the fiber cut
condition) as well as information on the add and drop channels at each
particular element.
Changes (if any) of the configuration parameters (a) related to the
amplifiers' and OADMs' (elements') hardware and to the attenuation
values of the fiber spans are always random, and usually form a group of
factors causing the mentioned difference between the expected up-to-date
total power and the real measured total power. Such factors can be, for
example, a fiber bending, a contact degradation, an amplifier's internal
fault, etc.
Fig. la illustrates a simplest communication line 10 comprising four
OFAs marked A,B,C and D connected in a chain by optical fiber spans
AB, BC and CD. The spans are characterized by respective span loss
values x, y and z, known in advance and taken into account when
designing the line. Bach of the OFAs is characterized by its gain. For the
further explanation, each of the OFAs will be associated with the
following power values which, in the drawing, have indexes
corresponding to the symbol of the suitable OF A:
EIP (expected input total power), which is continuously calculated,
MIP (measured or real input total power), which is continuously
measured,
MOP( measured output total power), which will also be continuously
measured to indicate how the OFA reacts to changes in the line.
Let us assume the line provides continuous communication of signaling
information between the OFAs which ensures recalculating the NOC and
NOA parameters for each particular amplifier, thereby enabling each of
the OFAs to calculate its suitable EIP. The functions of calculating NOC,
NOA, EIP, comparing it with MP and issuing an instruction of the
required correction are schematically shown to be performed by
individual control unite 12,14 and 16 respectively serving the amplifiers
B, C and D.
Fig, lb illustrates a time diagram explaining the main idea of the
invention using one example of operation of Line 10 shown in Fig. la.
Amplifier "A": Let OFA "A" generates a total output power MOPa
which is constant (graph 1).
Let the number of channels NOC does not change between the OFAs A
and B, and there is no fiber cut in the span AB; however, the span loss
"x" of the fiber span AB has unexpectedly grew and now constitutes "x-
3" [dB]. la other words, NOC and NOA do not change, as well as the
expected parameter ELpa stays the same, while MIPb which was equal to
(MOPa-x) will change now.
"What happens at amplifier "B": In view of the above, the expected input
total power (EIPb) at the OF A "B" stays the same, while the measured
total input power (MIPb) will be reduced due to the fiber span
degradation (MIPb = MOPA-x-3dB; graph 2). Let a local control unit 12
of the span power control reacts to the difference between the EIPb and
MIPb and issues an instruction of required correction (RC) to adjust the
gain of the amplifier "B". Due to that, the output power MOPb of the
amplifier "B" (see graph 3), which firstly reacted to the reduced input
power MIPb, will restore its value at the end of a hold off time period
HOT1. Let the restoration is incomplete due to some undetected internal
fault of the amplifier "b".
Behavior of amplifier "C": The expected input power EIPc is constant,
since no changes took place in NOC and NOA. However, the MIPc
follows the shape of the MOPb, since MIPc= MOPB-y. (graph 4).
Owing to the fact that the amplifier "C" has a longer response time than
the amplifier "B", ( HOT2 > HOT1), a correction block 14 will react to
the difference between the MIPc and EIPc which resulted from the
incomplete correction at the amplifier "B", and at the end of the HOT2
period, the gain of the amplifier "C" will be adjusted so as to provide
output total power MOPc proportional to the expected total input power
EIPc and not to the real total input power MIPc (graph 5).
The amplifier "D" will just repeat the waveforms received from the
amplifier "C" since its response time HOT3>HOT2>HOTl, i.e., at the
end of HOT3 there will be no difference between the EIPd and MIPd.
The response times HOT are selected so as to ensure fast correction in the
network, while allowing to take into account updated NOC and NO A.
The above example demonstrates how the proposed method allows
controlling the constant output power (and consequently, the constant
output power per channel) at any optical element of the line in a case
when unexpected faults occur (degradation of the fiber span AB, internal
fault in the amplifier B). The example is built for NOC, NOA = const.
Though known methods of maintaining fixed power per channel provide
similar results for NOC, NOA = const, they are already useless when
NOC dynamically changes. For example, if MIPb in graph 2 has the
shown shape not due to the increased attenuation of the fiber span AB but
due to a reduction in the number of optical channels (NOC), and (as in all
known methods) no EIP is calculated, the amplifier B will make
unnecessary corrections to compensate the input signal degradation.
Contrary to that, in the proposed method, the calculated EIP will change
with the change of NOC (if any), and no unnecessary corrections will be
made at the output of the amplifier B. A similar process will take place at
other amplifiers of the line.
Fig. 2 illustrates a modified optical communication line 20 comprising
three optical elements: two optical amplifiers (OFA) "A" and "B" which
are connected via one OADM "F" by optical fiber spans AF and FB. The
information about NOC and NOA is available at any element of the line,
for example it is recalculated at the OADM like it is recalculated at any
other optical element. It should be mentioned, that when calculating
EIPf, the ELpa forF (ELpaF) is equal to the span loss of the span AF.
However, for calculating EIPb, the ELpa for B (ELpaB) is equal to the
span loss of the span AF + the span loss of the span FB + the insertion
loss of the OADM for the through channels (i.e. the total span loss from
the output of amplifier A up to the input of amplifier B). Functions of
calculating NOC, NOA, EIP and RC are performed by individual control
units 22, 24,26 respectively associated with the optical elements A, F
and B.
Methods for detennining NOC and NOA will not be described in
detail in the frame of the present application. NOC can be determined by
a spectrum analyzer installed at the card of an optical element and
providing the channels count in real time. Industrial spectrum analyzers
are known in the art. Alternatively, NOC can be determined with the aid
of messages transmitted in the line via a supervisory channel, as it is
proposed in the US 09/962337. NOA of an optical amplifier and of a
passive optical element such as OADM can be determined as follows:
For OFA: NOA out = [(NOAin +1) and (not LOS)] or (LOS)and
l;
For OADM: NOA out - [ NOAin and (not LOS)] or (LOS) and 0;
where; NOA out - NOA at the output of the optical element,
NOA in - NOA at the input of the optical element, and
LOS -Loss Of Signal alarm, accompanies a fiber cut condition.
A formula for calculating EIP will be presented with reference to
Fig. 3. Ways of controlling output power of OADM to equalize its output
power per channel will be described with reference to Figs. 4 and 5.
Fig. 3 is a flow chart 30 schematically explaining how different optical
elements of the line 20 can be regulated to enable control of optical
power per channel in the line spans. The main function of the control
block implementing the flow chart is continuous calculation of EIP for
each particular optical element, comparing it with a continuously
measured real power (MSP) and obtaining an instruction of Required
Correction (RC) for controlling power at output channels of the element,
whether it is OFA or OADM.
In order to obtain the up-to-date input expected power EIP of a
particular optical element, there is proposed to obtain the value of
expected power issued from the previous optical amplifier by active
optical channels, and subtract therefrom an expected (configured) value
of the span loss (ELpa) from the preceding amplifier up to the particular
element (amplifier or OADM) of the line. This ELpa may include, in
addition to the fiber loss, the insertion loss of a passive element (such as
OADM) located between two amplifiers.
Therefore, the up-to-date Expected Input Power EIP of a particular
optical element can be calculated as a sum of powers of the optical
channels which should come from the preceding optical amplifier via the
fiber span, plus the noise power created in the line up to the output of the
preceding amplifier, and minus the span loss ELpa of the mentioned fiber
span:
EIP[dBm] = 10 x log{SignalsPa[mW] + Noisepa [mW]} - ELpa [dB]
where:
Signalspa [mW] - is power, in mW, of all active optical channels at the
preceding amplifier's output;
Noisepa [mW] - is the noise, in mW, at the output of the preceding
amplifier. Assurning that average noise figure of OFAs is 6dB and that
the noise is estimated over the C-band of the optical spectrum of the
OFA, the coefficient 27[dBm] is obtained; if the efficient bandwidth of
OFA is different, and/or the average noise figure is different, the
coefficient can be adjusted;
ELpa - expected span loss [dB] from the preceding amplifier;
and where:
Signalspa
NOC ia- is a number of incoming optic channels of the particular
element;
EPPCpa - expected power per Channel of the previous amplifier (stated by
the preliminary design)
AVGpa- average gain of optical amplifiers in the line, up to the particular
element at which EIP is calculated;
Noisepa
NO A - the number of preceding amplifiers in the line.
It should be noted that additional parameters and criteria, for
example gain tilt, can be introduced in the formula for more accurate
calculation of EIP. Likewise, in some practical cases the formula can be
simplified, for example from the point of considering noise.
Returning to the proposed method of maintaining the expected
power per channel, whenever the up-to-date EIP is calculated and
compared with the measured total input power MIP, the difference will
indicate the required correction to bring the optical element to a regime
for ensuring the expected output power per channel, as preliminarily
designed.
In Fig. 3, Block 31 is responsible for performing the EIP calculation for a
particular (i-th) optical element. Block 31 indicates obtaining results of the
real time measurement of the incoming total input power at the i-th
element, If the calculated and the measured values differ more than by a
predetermined threshold value (block 33), the hold off time HOT of the
element should expire before any corrections will be started (block 34).
If, however, the MIPi differs from EIPi during the whole period of HOTi,
the required correction RC is calculated as indicated in block 35. It goes
without saying that RC should be more than a predetermined value to
start any corrections. However, if the RC exceeds some maximally
accepted value (Track Tolerance - TT), it means that such a difference
between MIP and EIP cannot or must not be corrected by the optical
element. For example, the system may state that correction in a case of an
input signal weakened by 10dB is useless. In such circumstances, no
correction will be made (blocks 35a, 35b).
If the optical element (i) is OFA, the RC is applied to regulate its
gain (block 38). Since the OFA output power naturally changes in a direct
proportion with the measured input total power, the correction should be
applied in the inverse proportion. In other words, in case the measured
total input power is, for any reason, higher man the expected input power
(MIP>EIP), the correction should be calculated to reduce the OFA gain,
thereby to reduce the output total power and to avoid the dangerous
overpowering. In the opposite case, if the measured (real) input total
power is lower that the EIP (for example, there is an unexpected extra
span loss before the optical element), the OFA gain is to be increased so
that to compensate the extra span loss.
Gain new[dB] = Gain old [dB] - RC [dBJ, where RC =MIP-BIP;
Example: EIP = 10mW; MIP = 5mW, i.e. is twice reduced (MIP = ½
EIP), or
10log (1/2) =101og 1 - 10log 2 = 0 -3 = -3 [dB];
The required correction (RC) for the OFA gain is (x ½), or -3 dB:
Gain new = Gain old - (-3dB) = Gain old +3dB,
thus the amplifier will be capable to output the unchanged output power
and therefore to maintain the constant power per channel.
In some embodiments of OF A, the gain regulation can be performed by
using a so-called variable optical attenuators VOA, if provided at the
input of the OFA.
If the (i-th) optical element is OADM, the required correction is
applied to control the output power of added channels thereof (block 39).
In OADM (or VMUX), the power of an added channel is corrected in a
direct proportion with the measured input power.
It should be noted that outgoing optical channels of OADM
comprise those channels which have passed through the OADM and
those, which have been added at the OADM. The problem is to equalize
the output power of the through channels and the output power of the
added channels, since their ways via the OADM are totally different and
their powers never match to one another if not specifically adjusted.
Therefore, in case the total input power of OADM ( being a certain
indication of the input power of the through optical channels) is lower
than the expected value, the output power of the through channels will
also be lower due to a through insertion loss; it means that the power of
any one of the added channels should be also reduced to provide that the
output power of every added channel be lowered accordingly and be
equal the output power of the through channels, Therefore:
APPC new = APPC old + RC [dB],
where APPC - added power per channel
Example: EBP = 10 mW; MIP = 5mW, i.e. the MIP = ½ EIP or the
required correction RC = -3dB. To ensure the equal power to the output
channels of OADM, power of any added channel should also be reduced
(multiplied by ½ ) to be equal to the through channels at the output of the
OADM. In dB:
APPC new = APPC old + (-3dB) = APPC old - 3dB.
Mg. 4 explains the purpose of equalizing output powers of OADM
optical channels. It schematically illustrates a block-scheme of OADM 40
which receives incoming channels number 1,2 3,4,5,6 and 7 having
different wavelength and almost equal input power. Let the input power
per channel is designed to be 0 dBm, which equals to 1 mW per channel.
However, due to some reason, the optical channels arrive with the input
power 0.5 mW or (-3 dBm). The two levels of the input power are
shown on the spectrum diagram 4b as Pin ~?. The total incoming power
MIP is measured at the input of OADM with the aid of a photo diode 42.
According to this particular example in OADM 40, channels 1,2, 5 and 7
are dropped and will be attenuated due to the OADM insertion loss for
dropped channels (ILdrop). Let us assume it equals to 3dB. It is
understood that the dropped channels will be obtained weaker than
expected, but their amplification can be provided by respective
customers.
A real problem appears with the channels which are to be further
transmitted in the line. Channels 3, 4 and 6 pass through the OADM and
are attenuated by the OADM insertion loss for through channels
(ILthrough). Let ILthrough is equal to 5dB. The expected input power of
these channels is lmW or 0 dBm, so their expected output power will be
0.33 mW or -5dBm (see the spectrum diagram 4c). Channel 5' is added
to catch the place of the dropped channel 5; the added signal passes the
add circuit 44 and the insertion loss of the OADM to the added channels
which is marked as (ILadd). Let in this example it is equal to 3dB. The
circuit of the added channel 5' is preliminarily adjusted to such a power
of the added channel (measured by a photo diode 46), to ensure, at the
output of OADM 40, the output power of channel 5' equal to the
expected output power of the through channels 3,4 and 6. For example, a
transmitter 48 of the channel emits the signal of 0 dBm, which is
attenuated by the add circuit and ILadd, to appear at the output of the
OADM with the same output power, as the through channels 3,4 and 6 if
obtained at the expected input power (-5dBm or 0.33 mW). It means that
a variable optical attenuator (VOA) 50 can be adjusted to 2dB, so that
together with ILadd =3dB it brings the signals to the power level of -
5dBm or 0.33 mW.
Since the input channels are obtained twice weaker than expected
(0.5 mW or -3 dBm), the output power of the through channels will
further be weakened due to the ILthrough = 5dB, and become equal to
0.166 mW or (-8dBm). However, the add channel 5' remains the same as
it was, i.e., twice stronger that the real through channels 3,4 and 6 (see
the spectrum diagram 4d). To compensate for such an effect, the add
circuit of the add channel 5' is to be adjusted, say by changing the value
of VOA 50 to 5 dB, so that the output power of the signal of the add
channel 5' becomes -8 dBm or 0.166 mW. However, the adjustment is
preferably provided upon preliminarily checking the add power by the
photodiode 46, and only then by affecting parameters of the circuit 44.
The adjustment described above can be arrived to, say, by using an
algorithm which utilizes obtaining an expected power per output channel
of OADM and correcting the OADM add power per channel using the so-
called Required Correction (RQ. This will further be explained using
Fig. 5.
However, the power of an OADM added channel can be
calculated less accurately, just by measuring the real total input power
MIP and updating NOC, without calculating title expected power and
taking into account the NOA and noise figure parameters:
APPC = (MP - 101og(NOC)3 - IL oadm ± CO [dBm],
where APPC - is the added power per channel,
[MIP-10log (NOC)] - gives an average input power per one
incoming channel, which serves an indication of an output power of a
through channel of the OADM;
IL oadm - insertion loss introduced by OADM; this parameter
takes into account the attenuation created by OADM to the through and
add channels thus enabling estimation of the power of the added channel;
if the output channels' power is to be equalized, so ([MIP - 10log(NOC)]
IL through) =( APPC -IL add), and APPC - [MIP-10log (NOC)] - (IL
through - IL add); thus IL oadm = (IL through - IL add);
? CO is a manually introduced channel offset, which is usually
selected for a particular channel in order to give it a pre-emphasis. This
practice is caused by the fact that any OFA has noticeably different
specific gains for different wavelengths. The channel offset thus gives a
means to slightly regulate fluctuations in the line caused by the above
fact
Fig. 5 illustrates a block diagram of a system 60 for adjusting add power
of different add channels in OADM, using the preferred version of the
method. The diagram of system 60 partially shows operations performed
for controlling the output power of OADM which are schematically
shown as block 24 in Fig. 3.
In the block diagram of Fig. 5, OADM 62 is shown as its add block
inserting four optical channels 1,2,3 and 4 into the line. At the input of
OADM, measurement of the real input total power (MIP) is performed by
a photo diode 64. At the output of OADM, the real (measured) total
output power (MOP) is obtained by a photodiode 66, A processing block
68 obtains data about the expected power loss from the previous amplifier
(ELpa) and about the expected power per channel from the previous
amplifier (EPPC pa), as well as'data on insertion losses of the OADM.
These values are known from the preliminary design and from the
OADM characteristics. Based on them, the expected power per channel at
the output (EPPCo) is calculated. Actually,, instead of the calculated
EPPCo an operator may choose to use a constant pre-configured value of
the power per channel at the output PPCo. The selected value is farther
sent to summing blocks 70 of the add channels, which modify the power
per channel at the output by a channel offset CO (blocks 72) which is
individual for each specific optical channel and depends on the degree of
attenuation which the particular wavelength undergoes in the line. Upon
modifying the reference values of the selected output power per channel,
they are respectively corrected in summing blocks 74 by adding the
Required Correction to each modified value received from 70. The RC
for this particular OADM is calculated based on the difference between
the MIP (total) and EIP(total), as explained above with reference to Fig. 3
(/RC/