Application Note

Submarine Cable Testing

MW90010A
Coherent OTDR

By Stuart Whitehead

Background
With the submarine optical cable industry continuing to grow at a steady rate and the ever increase reliance on transcontinental
data traffic, the importance of minimizing optical network downtime has become more critical than ever.

Required data rates will continue to increase but more importantly, the reliance on connectivity is becoming more and more the
expected norm. This has caused several of the latest submarine cable cuts to become not only news within the industry, but
also within the wider public arena.

C-OTDR Overview
What is it?
The Coherent Optical Time Domain Reflectometer (C-OTDR) is the only way to accurately measure and characterize the
optical submarine network allowing accurate fault location to within 10 meters.

How does it work?

The C-OTDR works utilizing the Rayleigh backscatter caused by the impurities inherent in optical fiber to reflect light back to
the source much the same as a standard OTDR. As the submarine network or “submarine portion” of the network consists of
not only the optical fiber, but also EDFA’s (erbium-doped fiber amplifier), standard OTDR technology isn’t a viable option.
EDFA’s only amplify in the forward direction and employ components which are unidirectional therefore the backscattered
light is not able to return via its original path. The majority of installed and planned systems include an optical feedback path
within the EDFA enclosure, this feedback path allows the backscatter light to be returned to the C-OTDR on the second fiber
of the pair.

The Network
Network overview
The submarine optical network is made up of several key components:
- LTE (Line Terminal Equipment) - this can be provided from many different vendors, normally 10Gig interfaces
- EDFA Section - lying on the seabed at depths of up to 3+km and containing several different components
- EDFA (Erbium-Doped Fiber Amplifier) - optical amplifier used to increase the optical transmission signal
- Optical Feedback Path - part of the EDFA used during testing of the network
- Fiber - submarine cable normally consisting of only several fiber pairs, power feed and strengthening material.
Side A Side B

LTE LTE

EDFA Section EDFA’s Optical Feedback Path Fiber (40km or more)

Fig 1; Major components of a Submarine System
©
Anritsu MW90010A-E-F-1-(2.00)
 C-OTDR
Working within the Network
Each submarine repeater has a path allowing the backscatter signal to pass backwards along the link, this path allows for
monitoring the submarine portion of the optical fiber using the OTDR method. The distance between two repeater sections is
joined by approximately 40 to 90 kilometers of optical fiber.
A submarine system uses multiple DWDM channels to increase the amount of total traffic between two locations (these
DWDM signals are today normally at 10Gbps line rate while some companies are trialing 40Gbps). The C-OTDR probe pulse
is assigned a DWDM channel.
The C-OTDR probe pulse and dummy pulse are normally placed as far away from the active traffic as possible to minimize
any chance of interference. The dummy pulse occupies a second channel commonly adjacent to the probe pulse.

C-OTDR probe pulse Submarine Repeater

Dummy probe pulse

Uplink

λ
Communication signal

Downlink

Fig 2; Submarine Repeater system showing C-OTDR return path.

The dummy pulse is required due to the automatic gain control system of the EDFA’s. In a live system, the input to an EDFA
is at a constant power level across multiple channels while C-OTDR testing is often completed on an unlit system (no traffic).
When completing testing on an unlit system, the EDFA gain control is not able to maintain a stable output due to the pulsing
power nature the C-OTDR presents to it. In order to avoid this problem, the C-OTDR is able to output on two channels which
ensures a constant input level to the EDFA. The test pulse is being generated for a short period while the load pulse is on for
the remainder of
Probe time while the
Time ratio between the
Dummy two is determined
by the testing PW
selected on the C-
Fig 3; C-OTDR Probe and Dummy light OTDR.

The C-OTDR works on the same basic principles of an OTDR which is by transmitting a light into the fiber then looking at the
reflections (or backscatter) from the fiber under test. It also has the added ability to transmit on an adjustable narrow
wavelength which allows the unit to be used in a live network alongside customer traffic within the DWDM network. On the
receiver side of the OTDR there are several enhancements over a standard OTDR. The first major difference is the input of the
C-OTDR is filtered to remove the active DWDM channels as well as extra noise. While the second and most important is the
coherent detection. Coherent detection is a way of re-injecting the original transmitted wavelength allowing the resultant to
show only information at exactly that wavelength. This method removes all other noise allowing for a large signal to noise
improvement including reconstruction of data from well below the normal noise floor.
A submarine network is made up of many optical amplifiers to increase the power level to the DWDM wavelengths but this
also increases the Amplified Spontaneous Noise (ASE) level. As each amplifier increases the ASE level, the coherent
detection method enables the C-OTDR to see signals which would normally be considered hidden within or below this noise.

©
Anritsu MW90010A-E-F-1-(2.00)
The block diagram Figure 4 offers an overview of the key areas of a C-OTDR.

Polarization
CW Scramble
Output
LD SCR EDFA A/O
Communication Signals
Pulse signal
Local output
Optical path Signal
λ ASE
Electrical path
λ
Test Signal
Optical Input
Filter
Optical
interference
0.2 Coherent detection

0
P (t)

0.2
0 20 40 60 80 100

Distance (km)

A/D Display

Fig 4; Basic internal working of a C-OTDR

Testing a severed network
A submarine network consists of pairs of fibers, direction A to B and B to A linked by the Optical Feedback Path at each
repeater. The feedback path makes it possible for the C-OTDR to receive backscatter from the cable transmitting in the
opposite direction. When testing a cable which has been totally severed (thus a single cut in a single location), the task is
relatively straight forward however this can become more complicated in either of the following situations:

Single direction if the cable is severed (i.e. A to B but not B to A):

- The distance length will appear different depending on which end of the network you are testing from. The true fault
location will only be shown when testing in the same direction as the transmission link fiber. Testing from the
receiver end side, the fault will show the end location of the repeater directly following the cut location which could
be inaccurate by as much as the distance of the repeater sections (up to 90km).

The cable is totally severed in two locations:

- A major cause of fiber cuts is due to seabed movement. This movement can cover a large geographical area and
affect a large section of cable. A simultaneous cable cut in two locations can adjust the response and repair actions to
be taken thus fully understanding this situation is very important.

Figure 5 shows testing from different ends of the network (not completed simultaneously):
- From direction A, the correct fault location is able to be identified while testing from direction B is incorrect.
By understanding how the OTDR technology and the optical feedback path work together, the true fault location can be
quickly identified.
In a situation of two cuts within close proximity of each other careful evaluation reduces the risk of badly deploying expensive
resources.

©
Anritsu MW90010A-E-F-1-(2.00)
 Side A Side B
down link
C-OTDR C-OTDR

up link
COTDR Signal from Side A
Backscattered light from the fiber colored yellow

Side A Side B
up link
C-OTDR C-OTDR

down link
COTDR Signal from Side B
Side A to Backscattered light from the fiber colored orange
Side B

Side B to
Side A

Note: Trace depicted inverted to simplify drawing

Fig 5; C-OTDR showing the repeater and break location

C-OTDR resolution
The importance of a C-OTDR comes from enabling accurate fault location on any length of submarine network. The data
point resolution of many traditional OTDR with 50,000 data points Range / data points = Resolution
OTDR’s is normally determined by
10km Range 0.2m resolution 10,000 / 50,000 = 0.2
the km range setting of the OTDR.
100km Range 2m resolution 100,000 / 50,000 = 2
For example;
Table 1; OTDR Data Point resolution
- An OTDR with only 50,000
data points will be affected by the range setting.
This becomes an even more critical issue with submarine networks as the distance of the submarine portion is several orders of
magnitude larger than terrestrial networks. Due to this the latest C-OTDR’s are designed with 1.2 Million data points and
automatically reduces the number of points depending on the distance range setting which has several advantages:
- Maintain the same 10m resolution irrelative of the km range setting
 10m resolution was selected as it ensures the C-OTDR is not the weakest point when locating a fault
 10m accuracy is well within the requirements for fault location in a Submarine Network.
- Reduce the processing time of the C-OTDR while calculating the trace
 The C-OTDR can take approximately 8 samples per second Note 1 each sample consisting of up to 1.2 Million data
points. These samples are then averaged over time before the trace is displayed.
 This processing time reduction is due to being able to reduce the number of data points being averaged when a range
setting of less than 12,000km is selected.

©
Anritsu MW90010A-E-F-1-(2.00)
Note 1
The number of samples per minute is dependent on several factors such as, speed of light, Range and Resolution of the OTDR.
To complete 1.2 million data points (each 10m apart), the C-OTDR would be set to a range of 12,000km.
The time taken for light to travel 12,000km (there and back) would be approximately 120mS,
Thus 1second / 120mS = 8.3 scans per second.

Using a C-OTDR with less data points can

become a large issue if the user was required Data point of low
to measure longer links. resolution tester
- An example would be if the C-OTDR Event slopes are
had a maximum of 10,000 data points sharper with higher
and the range was set to 8,000km, Data points
 8,000km / 10,000 = 800m (data
point inaccuracy). See Fig 6,
Data point of high
 This inaccuracy could cause
resolution tester
extended delays in locating
the end fiber fault, in turn
delaying network restoration.

Data point
inaccuracy error

Fig 6; Data Point resolution of different C-OTDR’s

Conclusion
Why the C-OTDR technology
The C-OTDR offers the best technology for testing submarine fiber optic cables. The newer generation of C-OTDR’s allows
for not only extremely accurate distance measurements, but also full characterization of optical events. The combination of the
C-OTDR’s coherent technology and submarine cable optical feedback path ensures thousands of kilometers of fiber can be
characterized quickly and efficiently. With some basic knowledge and a simple interface, even an inexperienced engineer is
able to ensure the expensive task of fault restoration is completed as quickly and efficiently as possible.

©
Anritsu MW90010A-E-F-1-(2.00)
Anritsu Optical Solutions
MS9740A Optical Spectrum Analyzer
The MS9740A is a benchtop optical spectrum analyzer covering 600 nm to 1750 nm. The
MS9740A offers high-performance optical resolution and high-speed measurements to
support the needs of customers.
CMA5000a Multi-Layer Network Test Platform
CMA5000 transmission test modules can support field installers and maintenance
engineers who require a single tool for data network testing form n x 64 BER and physical
interfaces of 1.5 Mbit/s to 10 Gbps for SDH/SONET. OTN networks on both 2.66 Gbps
and 10.7 Gbps are also supported. Ethernet interfaces from 10 Mbit/s to 10 Gbps (LAN-PHY
and WAN-PHY) allowing full data network testing abilities.

©
Anritsu MW90010A-E-F-1-(2.00)