Remote Over-saturation Ampliﬁcation of Optical Fiber Monitoring

Pulses

Breno Perlingeiro

1,∗

, Pedro Tovar Braga

, Felipe Calliari

, Guilherme P. Temporão

Gustavo C. Amaral

1,2

and Jean Pierre von der Weid

Center for Telecommunication Studies, Pontiﬁcal Catholic University of Rio de Janeiro, 22451-900 Rio de Janeiro, Brazil

QuTech and Kavli Institute of Nanoscience, Delft University of Technology, 2600 GA Delft, The Netherlands

Keywords:

FPGA, Coherent Optical Ampliﬁcation, Semiconductor Optical Ampliﬁer, Optical Network Supervision.

Abstract:

Optical ﬁbers constitute a staggering portion of the physical layer underlying modern communication net-

works. To extend the reach of such networks around the globe, long-haul links are necessary. In this context,

establishing a connection between two remote locations is only possible due to signal booster stations inter-

spersed along the way. Supervision of such long distance links is of the utmost importance for their reliable

operation. For multiplexed networks, high-ratio optical splitters are necessary to distribute the optical signal to

multiple users, diminishing severely the transmitted power for each network. In this work, an automated sig-

nal boosting remote station for monitoring signals is presented. A Field Programmable Gate Array (FPGA) is

part of the remote station and grants its autonomous operation. Making use of a topology capable of reaching

over-saturation ampliﬁcation of semiconductor optical ampliﬁers (SOA), a higher portion of the optical loss

experienced in the splitter is precompensated in the remote node allowing for supervision reach-extension.

Approximately 0.5 dB of increased dynamic range is experimentally achieved when comparing the proposed

remote station with another one using the same optical ampliﬁer. Even though the obtained extra gain is

a minor improvement, the proposed topology paves the way for scalable ampliﬁcation, allowing for longer

reaches.

1 INTRODUCTION

The globalization process creates the urge for long-

haul physical layer links that connect distant locations

with high data rates. Due to the broad bandwidth and

low attenuation provided by ﬁber-optic links, these

are often the choice when designing long-haul com-

munication systems (Yu and Zhang, 2016). This at-

tenuation, however low, is not zero and, even in an

unperturbed link, the optical signal must be boosted

to reach its destination with the expected power level.

The uppermost advantage of using optical ampliﬁers

for signal boosting is that the ampliﬁcation occurs

entirely in the optical domain, without the need for

any electrical conversion, the price of which scales

abruptly (Simmons, 2005).

Even with the inclusion of such signal boosters,

link maintenance is limited by power loss events usu-

ally associated with the mechanical fragility of the

ﬁber, and robust operation can only be achieved with

physical layer supervision (Urban et al., 2013). Un-

fortunately, the amount of Rayleigh backscattered

power, onto which most supervision methods depend,

is extremely faint: the Rayleigh coefﬁcient of stan-

dard telecommunication single-mode optical ﬁbers is

-72 dB/m (Bergman et al., 2016). In the case of in-

tegrity tests of an optical ﬁber with a standard Opti-

cal Time Domain Reﬂectometry (OTDR) device, the

power of the monitoring pulse can eventually be less

than the required to reach a desired distance (Lai et al.,

1994). In such cases, an optical ampliﬁer can be used

in a remote node to amplify the OTDR pulse, thus

making it capable of reaching longer distances.

As far as optical ampliﬁers are concerned, the op-

tical ampliﬁcation method can be classiﬁed, with re-

spect to the pumping, in two (Govind, 2002): Semi-

conductor Optical Ampliﬁers (SOAs) make use of

electrical pumping, i.e., the injection of current into a

semiconductor heterostructure junction induces pop-

ulation inversion; and Fiber Ampliﬁers (such as Er-

bium Doped Fiber Ampliﬁers – EDFAs), on the other

hand, make use of optical pumping, i.e., population

inversion is achieved through the excitation of a dif-

ferent optical transition inside the material that decays

Perlingeiro, B., Braga, P., Calliari, F., Temporão, G., Amaral, G. and von der Weid, J.

Remote Over-saturation Ampliﬁcation of Optical ﬁber Monitoring Pulses.

DOI: 10.5220/0007365401390146

In Proceedings of the 7th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2019), pages 139-146

ISBN: 978-989-758-364-3

139

into the transition of interest. In both cases, for given

input and pump powers, all the population from the

excited state will be consumed in the ampliﬁcation

before the pump can once again repopulate this tran-

sition, what leads to a steady-state condition known

as saturation.

For multiplexed passive optical networks (PON),

such as wavelength division multiplexing (WDM) or

time division multiplexing (TDM), high-ratio splitters

are necessary to distribute the optical signal to multi-

ple users, which accompanies an intrinsic substantial

reduction of the signal power. Therefore, being able

to preamplify the monitoring signal to the absolute

limit of the optical ampliﬁer is interesting in order to

compensate a following a priori known high loss. The

network design in which the ampliﬁer is placed before

the splitter allows for the ampliﬁcation of the pulses

that monitor each of the users’ ﬁbers connected to it,

i.e., centralization of the ampliﬁcation, and extension

of the supervision reach.

Recently, an architecture that provides gain above

saturation of SOAs in the pulsed regime has been

presented (Amaral et al., 2016), with a subsequent

design simpliﬁcation and optimization in (Resende

et al., 2017). The underlying concept for both im-

plementations is multiplexing one optical pulse into

two or more pulses, amplifying them separately until

the SOA’s saturation level, and later demultiplexing

the ampliﬁed pulses into a single one, thus granting

an ampliﬁcation gain above the saturation gain of a

single SOA. There, however, the design was strictly

local, i.e., not suited for remote applications; fur-

thermore, its polarization alignment dependency pre-

sented a signiﬁcant challenge for remote deployment.

In this work, a remote over-saturation ampliﬁca-

tion node is assembled and its operation is veriﬁed

making use of off-the-shelf components and the de-

sign of (Resende et al., 2017). Using the simplest

implementation of the over-saturation system, with a

single multiplexing stage, a ∼ 0.5 dB dynamic range

gain could be achieved when compared to a standard

ampliﬁcation system. Even though this is still a neg-

ligible factor, the successful remote operation of the

system represents an important step towards its ap-

plication in a scalable system, with more than one

multiplexing stage. The paper is organized as fol-

lows. Section 2 details the system architecture and

each of its constituents: the over-saturation ampliﬁca-

tion system; the synchronization of the remote node

with respect to an incoming probe pulse; and the po-

larization stabilization system. In section 3, the re-

sults of photon-counting OTDR monitoring (Herrera

et al., 2016) are presented, demonstrating the trans-

parent adaptation and independent operation of the

presented system; comparative results are also dis-

cussed. Finally, Section 4 concludes the paper, with

further comments and conclusions.

2 EXPERIMENTAL SETUP

Throughout this document, and in previous works

(Resende et al., 2017), SOAs are used in the over-

saturation ampliﬁcation topology while the aforemen-

tioned ﬁber ampliﬁers are not. The reason behind the

employment of the former and not the latter is the fact

that, differently from ﬁber ampliﬁers, SOAs can act

as high-speed high-extinction-ratio optical switches

(Amaral, 2014), a feature that allows the SOA to be

triggered in a chosen instant for a short period of

time. This results in a time-localized ampliﬁcation

peak while also minimizing the emission of ampliﬁed

spontaneous emission (ASE), a serious concern espe-

cially for time-domain ﬁber monitoring solutions.

RemoteNode

Central

Office

AWG

Users

AmplificationSystem

Figure 1: Conceptual high-level block diagram of the im-

plemented system architecture. High peak power polarized

and narrow-bandwidth supervision pulses are generated in

the central ofﬁce (CO) and launched into the feeder ﬁber.

In reaching the remote node, the pulse is ampliﬁed before

going through a passive optical splitter (in this particular

case, an arrayed waveguide grating (AWG) compatible with

dense-WDM networks. A bypass structure composed of

two circulators allows for the outgoing pulse to be ampli-

ﬁed and the scattered signal to be routed back to the CO

unchanged.

The experimental setup, complete with a central

ofﬁce and a remote ampliﬁcation node, is depicted in

detail in Fig. 1. The OTDR probe pulse is generated

in the central ofﬁce using a polarized and coherent

light source in order to comply both with the ampliﬁ-

cation system requirements and with the requirements

of wavelength-division multiplexing (WDM) optical

networks monitoring. The arrangement of circula-

tors, one at each side of the ampliﬁcation setup, al-

lows for the incoming light pulse to be routed through

the ampliﬁcation, and the backscattered light to by-

pass the ampliﬁcation. In the following sub-sections,

speciﬁc parts of the topology are explained in detail,

namely the pulse detection and conditioning, polar-

ization control, and ampliﬁcation in the remote node.

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

140

2.1 Ampliﬁcation System Overview

The over-saturation ampliﬁcation scheme, depicted as

a single unit Fig. 1, has been decomposed into its

most fundamental blocks, namely a polarizing beam

splitter (PBS), an optical delay line, an SOA and a

polarization controller (PC), and is presented in detail

in Fig. 2. In it, a pulsed light signal, polarized such

that an equal ratio division takes place in the PBS,

enters the system; without loss of generality, the PBS

axis is deﬁned as the rectilinear one, i.e., spanned by

the horizontal (H) and vertical (V) polarization states.

SOA

Input

Output

Figure 2: Ampliﬁcation system architecture. Time-

polarization multiplexing is enforced by the optical delay

line connected between the t (transmitted - H) and r (re-

ﬂected - V) arms of the PBS.

Upon splitting, two pulses carrying half of the in-

cident optical power traverse the same looped path in

opposite directions. Time-polarization multiplexing

is enforced by the optical delay line and allows the

SOA to amplify each pulse individually. Finally, the

PC acts on the propagating pulse’s polarization so that

coherent recombination takes place in the PBS and

the ampliﬁed pulse exits through the remaining port.

Fig. 3 shows the ampliﬁcation curves of the single

SOA and the proposed over-saturation ampliﬁcation

system for different input peak powers.

Figure 3: Saturation curves for a single SOA and for the

over-saturation ampliﬁcation scheme. Symbols correspond

to experimental data, and solid lines correspond to curve

ﬁtting.

A 1.96 dB over-saturation gain could be achieved

experimentally according to Fig. 3. Note that the pro-

posed ampliﬁcation system becomes advantageous

for input powers higher than 1.3 dBm, and as the input

power increases above this threshold value, the over-

saturation gain also increases, limited by 2 dB due of

the PBS’ insertion loss.

Another important result from (Resende et al.,

2017) is that this system can be scaled to achieve

even higher over-saturation ampliﬁcation gains with

the addition of extra PBS’s and delay lines, where

each PBS added contributes with ∼ 2 dB of ampli-

ﬁcation gain. Fig. 4 shows the scalability feature of

this system with the use of N PBS’, what leads to a

N×2 dB gain. For instance, in case N = 2, to obtain ∼

4 dB of ampliﬁcation gain it must be guaranteed that

the input pulse peak power would still be high enough

to saturate the SOA as the number of pulses traversing

the SOA will be doubled and its power halved. Also,

all the polarization alignment constraints should be

met so that the coherent recombination takes place in

the PBS, so extra PCs are necessary.

PCN_1

PCN_2ODN

PBSN

Input

Output

PCin

PCN-1_1

PCN-1_2ODN-1

PBSN-1

SOA

PC1

PBS1

OD1

PC2_1

PC2_2OD2

PBS2

Figure 4: Block diagram showing the scalability feature

of the ampliﬁcation system architecture. One optical de-

lay line (OD) and a pair of polarization controllers (PC) is

needed for each PBS added.

2.2 Pulse Detection and Conditioning

The core concept of this work is to develop a low-cost

remote ampliﬁcation unit which surpasses the satura-

tion limits of an SOA working in the pulsed regime,

the block diagram of which is presented in Fig. 5.

Due to the pulsed nature of the application, the electri-

cal pulses sent to the ampliﬁer must be synchronized

to the arrival of the monitoring pulses for optimal per-

formance. The pulse emission rates and instants may

differ based on the scenario (link length and whether

the monitoring is performed at the same time as the

data transmission), so a synchronization mechanism

is implemented in an embedded electronic structure,

a Field Programmable Gate Array (FPGA). As it will

become clear in the next section, the FPGA is also

responsible for part of the polarization stabilization,

so the two functionalities are delegated to the same

processing unit.

Remote Over-saturation Ampliﬁcation of Optical ﬁber Monitoring Pulses

141

SOA

polarizer

A/D

FPGA

combiner

D/A

Input

Output

feedbacksignal

PC PC

Over-saturation

AmplificationSetup

Polarizationand

Synchronization

Unit

Figure 5: Block diagram of Remote Ampliﬁcation Unit.

A polarization and synchronization unit is responsible for

pulse detection, maintaining the polarization stable at |45

◦

and triggering the SOA for pulse ampliﬁcation. The over-

saturation ampliﬁcation setup (Resende et al., 2017) per-

forms pulse ampliﬁcation over the saturation limit of the

SOA. A/D: Analog-to-Digital converter; BS: arbitrary split-

ting ratio ﬁbred beam splitter; D/A: Digital-to-Analog con-

verter; FPGA: Field Programmable Gate Array; PC: po-

larization controller; PT: Polarization Tracker; SOA: Semi-

conductor Optical Ampliﬁer.

The synchronization between light pulses and

electrical pulses at the remote node is performed as

follows. A ﬁbred beamsplitter with very high split-

ting ratio (99:1) is used to collect a small portion of

the incoming light signal, which passes through a po-

larizing ﬁbred device (more details in the next sec-

tion) and is detected in a photodector. The analog

signal from the photodetector is converted to digital

in an analog-to-digital converter (ADC) and sent to

the FPGA, which implements a discrete derivative al-

gorithm that allows to transform the detected pulses

into narrow pulses that indicate the start and end of

a pulse based on whether its polarity is positive or

negative, respectively. In case of a positive transi-

tion, the FPGA imposes different time delays inter-

nally and generates two electrical pulses at its output

which are calibrated to match to the optical delay ex-

perienced by the incoming optical pulse. The delays

created by the FPGA must be different in order to trig-

ger the SOA at the exact moments when each of the

optical pulses pass through it.

In a real case scenario, where noise contribu-

tions from different sources affect the measured data,

thresholds for a transition in the discrete derivative al-

gorithm must be deﬁned. Otherwise, false transitions

will be identiﬁed and the system will trigger the SOA

in times uncorrelated to the propagation of a monitor-

ing pulse, which translates to increased noise in the

ﬁber probing measurement. Therefore, the sensitivity

of pulse detection can be controlled in advance when

the speciﬁcations of devices involved in the detection

of the light pulse, the link length and losses, and the

input pulse power are known a priori.

Mathematical methods to implement an adap-

tive identiﬁcation of transition thresholds, although

present in the literature (Harrison, 2003), have not

been implemented in this work but are a subject of

future study. Furthermore, it is interesting to notice

that the discrete derivative is DC-proof, i.e. it identi-

ﬁes a transition due to the arrival of a pulse regardless

of the voltage value of its amplitude. This feature is

related to the fact that, when the signal is subtracted

by a one-cycle delayed version of itself (as it is done

in the discrete derivative), the DC component is elim-

inated.

2.3 Polarization Control

Operation in the pulsed regime has, as commented in

previous sections, a number of desired effects both

for time-domain optical ﬁber supervision and over-

saturation ampliﬁcation; the former would be ex-

tremely complex in the continuous-wave case (in-

cluding a de-convolution algorithm) while the for-

mer would be impossible. Polarization control, on

the other hand, does not beneﬁt from pulsed opera-

tion and becomes more complex: measuring the po-

larization state of the incoming light pulse can only

be performed during said pulse duration. Further-

more, in order to make use of available components,

a Polarization Tracker module (PT) (General Photon-

ics, 2013) is employed in the experimental realization

of the ampliﬁcation remote node, the system within

which does not operate under the pulsed regime.

Making use of the synchronization scheme pre-

sented in the last section, it was possible to close the

feedback loop and ensure polarization stability. Based

on the arrival of an incoming light pulse, heralded by a

positive transition of discrete derivative, an embedded

electronic structure initializes acquisition of the pulse

amplitude until this signal outputs a negative transi-

tion, heralding the end of the light pulse; the ampli-

tude within the pulse duty cycle is thus acquired and

its average is stored in a memory structure. The out-

put is then calculated based on a convex combination

between the instantaneous acquired pulse amplitude

and the averaged value over the last N pulse detection

events, i.e.,

y(k) = α · u(k) +

(1 − α)

∑

i=1

u(k − i), (1)

where y (k) denotes the output amplitude, u (k) de-

notes the input average amplitude within the pulse

duty cycle, and α is the relative weight between the

instantaneous and accumulated acquired values.

The output value translates the evolution of the

pulse amplitude as the polarization is changed, and

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

142

can then be fed into the PT closing the feedback loop.

Adjusting the parameters α and N is no trivial task,

and the results of the automatic polarization control

for different values of α and a ﬁxed value of N = 32

are presented in Fig. 6. Based on empirical result

analysis, N = 32 and α = 0.6 have been set as stan-

dard experimental parameters.

Figure 6: Polarization stabilization signal evolution as the

number of received probing pulses increases for different

values of α. Distinct characteristics in the curves’ be-

haviours can be identiﬁed: when the instantaneous mea-

sured value represents a higher contribution to the output

signal (α = 1), the system reaches saturation faster but with

sharper oscillations; when the accumulated measured val-

ues contribute more (α = 0), the system takes longer to

reach convergence but with a steadier evolution.

The facts that the full control of the polarization

state is not necessary (note that any polarization state

at the equator of the Poincaré Sphere that yields an

equal division of the input optical power between the

arms of the PBS can be used) and that the polarization

drift is expected to be a low-frequency process (ther-

mal effects that lead to change in the ﬁber birefrin-

gency) allow the polarization stabilization to be im-

plemented with state-of-the-art technology speciﬁc to

this application, which can reduce the cost of the pro-

posed system. Furthermore, the fact that the FPGA is

an integral part of the design and that FPGA-based

algorithms for polarization tracking have been pro-

posed (Garcia and Amaral, 2016) and implemented

in the literature (Lima et al., 2017) indicates that this

possibility is within the technological reach; this de-

velopment, however, is left as a future point of in-

vestigation and the proof of principle implementation

presented here is conducted with manufactured com-

ponents (General Photonics, 2013), which are avail-

able experimentally.

2.4 Power Consumption and Autonomy

The implementation of a remote ampliﬁcation unit

would not be complete without the assessment of

its power consumption, and the reason for that is

twofold. Firstly, it enables one to determine which

type of supply is going to be necessary, i.e., a con-

nection to the power grid, or a simple battery. This

point is relevant in the case that the remote node is

placed in a location where access to the power grid

is costly. Secondly it enables the network operator to

balance the cost added by the addition of such remote

ampliﬁcation unit and the advantages introduced by

its employment. This last point can indeed become

the most relevant one since it is not scarcely that the

operation expenditures (OPEX), the cost associated

to keeping the system running – to which the power

supply costs would be added –, overcome the capi-

tal expenditures (CAPEX), the cost of implementing

the system. In Table 1, the power consumption of the

proposed system, stratiﬁed with respect to each indi-

vidual component, is presented.

Table 1: Power consumption of each individual device that

composes the proposed remote ampliﬁcation unit.

Power supply (mW)

ADS805 (ADC) 300

DAC7541A (DAC) 300

Polarization Tracker 7800

Spartan-3E (FPGA) 154

SOA and Driver 4100

New Port 2011 (Detector) 360

Total 13014

When applicable, the data on Table 1 was ac-

quired through the devices datasheet (ADS805 (Texas

Instruments, 2002), DAC7541A (Farnell, 1993),

Polarization Tracker (General Photonics, 2013),

SOA1013SXS (ThorLabs, 2011), and New Port 2011

(New Port, 2009)) based on a pessimistic estimate,

where the devices are fully operational (driving the

maximum nominal current) at all times. In the case of

the FPGA, the Xilinx Power Estimator was employed,

and, for the combination of the SOA and its electronic

driver, the power consumption was acquired experi-

mentally. Among the devices listed, the PT exhibits

the by far higher power consumption since it inte-

grates many functionalities that, as commented in the

previous Section, might not be necessary for the im-

plementation of the proposed system. It is within the

future project’s goal of developing an FPGA-based

polarization tracker to substantially reduce its power

consumption.

Considering the pessimistic estimate, an off-the-

Remote Over-saturation Ampliﬁcation of Optical ﬁber Monitoring Pulses

143

shelf 12V battery (Expert Power, 2012) with 18000

mAh capacity could supply the system for around 16

days, an extremely limited autonomy. If one consid-

ers, however, that the CO is not arguing the ﬁbre con-

stantly and does so four times per day during the span

of an hour, which constitutes a conservative rate, the

autonomy of the system would reach 100 days, i.e.,

a mobile team would only be scheduled every three

months to recharge the battery. It is interesting to note

that this calculation is still pessimistic in the sense

that, in general, the devices are not operating at their

full capacity and, thus, the practical power consump-

tion could be expected to reach much lower values.

The power consumption analysis of the proposed

remote unit yields, therefore, positive results with re-

spect to the previously mentioned ﬁrst point, i.e., its

autonomy. With respect to the second point, it is in-

teresting to analyze, for a more complete balance be-

tween cost and advantage, the gain in monitoring ef-

ﬁciency when the ampliﬁcation is present. For this, it

is useful to consider the monitoring time spent by the

supervision technique – where we consider, as in all

the experimental results, the photon-counting OTDR

of (Herrera et al., 2016) – to achieve a given mea-

surement accuracy in a speciﬁc scenario. In order to

maintain consistency with respect to the experimental

results of the next Section, a 24 km ﬁber with a 6 dB

power drop – that can be compensated by the imple-

mentation of the proposed remote over-saturation am-

pliﬁcation – at kilometer 12 is considered as the sce-

nario; furthermore, an accurate measurement is de-

ﬁned as a minimum of 10 dB signal-to-noise ratio

(SNR), which, for the ﬁber measurement, translates

into achieving a 10 dB SNR in the last position of the

ﬁber.

Under these conditions, it is possible to estimate

the total monitoring time in case the proposed system

is implemented and when no ampliﬁcation is present.

Following the calculations of (Amaral, 2014), i.e.,

with a 500 ns separation between 20 ns gates in the

photon-counting OTDR, an accurate ﬁber measure-

ment with no ampliﬁcation in the remote node would

take about 45 minutes. In great contrast with this

result, the elapsed time of an accurate measurement

in presence of remote node ampliﬁcation would only

take 5 minutes, a nine times gain in measurement

time. With this margin, the network operator can

either supervise more ﬁbers in less time, or activate

the monitoring system during a much shorter time; in

both cases, the implementation of the proposed sys-

tem is advantageous with respect both to the power

consumption in the CO side, which ﬁgures as a mi-

nor advantage, and in the data transmission downtime

necessary to supervise the ﬁber (in case in-service

monitoring is not possible), therefore reducing the

OPEX of the network operator.

3 OTDR MEASUREMENT

RESULTS

As shown in section 2.1, the proposed ampliﬁca-

tion system will over-perform an ampliﬁcation sys-

tem consisting of a single SOA for pulses with in-

put peak power above ∼ 1.3 dBm. In the remote

over-saturation scheme, however, this limit is ∼ 1 dB

higher because of extra insertion losses imposed by

the PT and the beam splitter (99:1). This 2.3 dBm

limit represents the turning point where the additional

losses introduced by extra devices are compensated

for the multiplexed ampliﬁcation of the pulses. There-

fore, for values below this limit, the over-saturation

ampliﬁcation system will cause power losses. At the

same time, if the input power is extremely high, even

the over-saturation system will reach saturation. En-

suring the correct input power balance is a crucial as-

pect of the network design using the proposed sys-

tem, a feature which is pictorially presented in Fig. 7,

where OTDR traces of the same ﬁber with different

input peak powers and making use of the ampliﬁca-

tion system are depicted.

Figure 7: OTDR traces of the same optical ﬁber link com-

posed of a ∼ 12 km and a ∼ 4 km ﬁbers connected in be-

tween to the ampliﬁcation system. The peak power levels

depicted are measured at the CO’s output. Measurement

times for each trace were 6 minutes.

The black trace corresponds to a condition where

the input peak power saturates the over-saturation am-

pliﬁcation system and the power after the remote node

drops slightly. The red and blue traces, on the other

hand, are within the input peak power level region that

allows for over-saturation ampliﬁcation gain. Finally,

the green trace corresponds to a power level below

the over-saturation limit and a slight power drop is

observed. Even though the over-saturation ampliﬁca-

tion system can indeed saturate, this limit can be in-

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

144

creased by scaling the topology as suggested in (Re-

sende et al., 2017).

In order to provide a fair and practical compara-

tive analysis of the capabilities of the presented sys-

tem, ﬁber proﬁles acquired using a photon-counting

OTDR system are presented. The test for which the

results are depicted in Fig. 2 simulate a simple WDM

network composed of a ∼ 12 km feeder ﬁber, a remote

splitting node, and a single ∼ 12 km ﬁber directed to

the user. The remote node includes an AWG with an

experimentally measured ∼ 6 dB insertion loss. The

three traces correspond to: the ﬁber proﬁle without

ampliﬁcation in the remote node; the ﬁber proﬁle with

an SOA as remote node ampliﬁcation; and the over-

saturation ampliﬁcation system as remote node am-

pliﬁcation.

Figure 8: OTDR traces acquired using a photon-counting

OTDR monitoring system (Herrera et al., 2016). The link

follows the diagram of Fig. 1, with an AWG right after the

ampliﬁcation system. Both the feeder and user ﬁbers are

∼ 12 km-long. The over-saturation system achieved higher

AWG loss compensation than the one achieved with a single

SOA, conﬁrming the system’s application proposal. Mea-

surement times for each trace were 30 minutes.

It is important to note that ampliﬁcation with the

SOA also requires pulse synchronization but not po-

larization control (the SOA presents negligible polar-

ization dependent loss (PDL)), therefore, the excess

loss of the polarization tracker is not present in this

measurement. The bypassing structure, however, is

unnecessary in both cases since ampliﬁcation is uni-

directional. Even though, in the results of Fig. 8,

a ∼ 0.5 dB difference is measured between the two

traces, this represents a ∼ 1.0 dB over-saturation gain,

which is halved by the round-trip factor of the OTDR.

It is also interesting to note that, apart from peaks

and troughs in the resulting trace in the vicinities of

the ampliﬁcation setup, the proﬁle of the ﬁbers is un-

altered, showcasing the transparent adaptation of the

system to monitoring applications. These peaks and

troughs are associated to reﬂections present in the am-

pliﬁcation setup that can be neglected while analyzing

the resulting proﬁle.

Extending the reach of OTDR measurement have

been given a considerable amount of attention in the

past decade, and the Coded OTDR (Park et al., 2007)

has been, without a doubt, one of the most successful

propositions. However, the implementation complex-

ity of such technique is considerably higher and the

impact on the transmitted data in case of in-service

monitoring can be higher than for the single-pulse

OTDR, which has been shown to induce negligible

noise (Amaral et al., 2014). Furthermore, the present

limitation of 0.6dB over a single SOA is expected to

be extended in a scalable design.

4 CONCLUSION

Losses in optical ﬁber links are extremely detrimen-

tal to network operation and single-ended link super-

vision ﬁgures as a solution for robust and centralized

network management. Within this context, ﬁber faults

can be located and repaired, but intrinsic losses due to

passive elements necessary for the network operation,

such as splitters, must be otherwise compensated for.

In this work, a remote automated ampliﬁcation unit

capable of operating at over-saturation is presented as

a mean of compensating the loss of such splitters. The

structure of the remote ampliﬁcation node, including

optical and electronic parts, is discussed and experi-

mentally veriﬁed.

The results from photon-counting OTDR monitor-

ing systems are consistent with respect to the dynamic

range gain factor, i.e., ∼ 0.5 dB when compared to a

regular ampliﬁer. Even though this is still a small gain

factor, the scalability of the over-saturation ampliﬁ-

cation would allow for even higher gain factors, and

the present work paves the way towards such imple-

mentation. When compared to the case where no re-

mote ampliﬁcation is present, a considerable increase

in dynamic range is observed, enforcing the hypothe-

sis that the strategical placement of ampliﬁcation sys-

tems in the vicinities of high a priori known power

losses is advantageous. Not only that, but this par-

ticular network design allows for centralized and re-

mote ampliﬁcation of the probing pulses sent from the

Central Ofﬁce before they are split into different user

ﬁbers. At the same time, as the results of (Amaral

et al., 2016; Resende et al., 2017) show, the gain of

over-saturation ampliﬁcation systems decreases dra-

matically when the input power is not sufﬁcient to sat-

urate the ampliﬁer, so planning the system’s location

is crucial for its optimal performance.

Remote Over-saturation Ampliﬁcation of Optical ﬁber Monitoring Pulses

145

ACKNOWLEDGMENT

The authors would like to thank brazilian agencies

CNPq, Capes and FAPERJ for ﬁnancial support.

REFERENCES

Amaral, G. C. (2014). FPGA applications on single photon

detection systems. MsC Thesis.

Amaral, G. C., Herrera, L. E., Vitoreti, D., Temporão, G. P.,

Urban, P. J., and der von Weid, J. P. (2014). Wdm-pon

monitoring with tunable photon counting otdr. IEEE

Photonics Technology Letters, 26(13):1279–1282.

Amaral, G. C., Herrera, L. E. Y., Resende, M. M., Tem-

porão, G. P., Urban, P. J., and von der Weid, J. P.

(2016). Time-polarization multiplexing for increased

output power of Semiconductor Optical Ampliﬁers in

the pulsed regime. Appl. Opt., 55(28):7878–7884.

Bergman, A., Davidi, R., Shalev, A. I., Ovadia, L., Langer,

T., and Tur, M. (2016). Increasing the measurement

dynamic range of Rayleigh-based OFDR interrogator

using an amplifying add-on module. IEEE Photonics

Technology Letters, 28(22):2621.

Expert Power (2012). EXP12180. https://cdn.shopify.

com/s/ﬁles/1/2364/9089/ﬁles/EXP12180-Specs.pdf?1

0653271711015005675.

Farnell (1993). DAC7541A. http://www.farnell.com/data

sheets/673292.pdf.

Garcia, J. D. and Amaral, G. C. (2016). An optimal polar-

ization tracking algorithm for lithium-niobate-based

polarization controllers. In Sensor Array and Mul-

tichannel Signal Processing Workshop (SAM), 2016

IEEE, pages 1–5. IEEE.

General Photonics (2013). Polarization Tracker. http://

www.generalphotonics.com/downloads/manuals/POS

-002-Manual-V2-1-9-27-13.pdf.

Govind, P. A. (2002). Fiber-Optic Communication Systems.

John Wiley, New York.

Harrison, R. R. (2003). A low-power integrated circuit for

adaptive detection of action potentials in noisy sig-

nals. In Engineering in Medicine and Biology Society,

2003. Proceedings of the 25th Annual International

Conference of the IEEE, volume 4, pages 3325–3328.

IEEE.

Herrera, L. Y., Calliari, F., Garcia, J. D., Amaral, G. C., and

von der Weid, J. P. (2016). High Resolution Auto-

matic Fault Detection in a Fiber Optic Link via Photon

Counting OTDR. In Optical Fiber Communication

Conference, page M3F.4. Optical Society of America.

Lai, Y. W., Chen, Y. K., and Way, W. I. (1994). Novel

supervisory technique using Wavelength-Division-

Multiplexed OTDR in EDFA repeatered transmis-

sion systems. IEEE Photonics Technology Letters,

6(3):446–449.

Lima, V. M., Amaral, G. C., Calliari, F., Temporão, G. P.,

von der Weid, J. P., Garcia, J. D., and Garcia, J. P.

(2017). Fast polarization switch for polarization-

based quantum communication. In Proceedings of

the 5th International Conference on Photonics, Op-

tics and Laser Technology - Volume 1: PHOTOPTICS,

pages 288–293. SciTePress.

New Port (2009). Model 2011. https://www.newport.

com/medias/sys_master/images/images/hc5/hb2/8796

990341150/2001-2011-User-Manual-RevB.pdf.

Park, N., Lee, J., Park, J., Shim, J. G., Yoon, H., Kim,

J. H., Kim, K., Byun, J.-O., Bolognini, G., Lee, D.,

et al. (2007). Coded optical time domain reﬂectom-

etry: Principle and applications. In Passive Compo-

nents and Fiber-based Devices IV, volume 6781, page

678129. International Society for Optics and Photon-

ics.

Resende, M. M., Tovar, P., Amaral, G. C., and von der Weid,

J. P. (2017). Overcoming the maximum ampliﬁcation

limit of coherent optical pulses in Semiconductor Op-

tical Ampliﬁers with time-polarization multiplexing.

Optical Engineering, 56(11):110502.

Simmons, J. M. (2005). On determining the optimal optical

reach for a long-haul network. Journal of Lightwave

Technology, 23(3):1039.

Texas Instruments (2002). ADS805. http://www.ti.com/lit/

ds/symlink/ads805.pdf.

ThorLabs (2011). C-band Semiconductor Optical Ampli-

ﬁer. Accessed: 2010-09-30.

Urban, P. J., Vall-Llosera, G., Medeiros, E., and Dahlfort,

S. (2013). Fiber plant manager: An OTDR-and OTM-

based PON monitoring system. IEEE Communica-

tions Magazine, 51(2):S9–S15.

Yu, J. and Zhang, J. (2016). Recent progress on high-

speed optical transmission. Digital Communications

and Networks, 2(2):65 – 76.

PHOTOPTICS 2019 - 7th International Conference on Photonics, Optics and Laser Technology

146