The present invention relates to communications systems and in particular to a system and method for monitoring a fibre optic transmission system.
In long haul fibre optic communications systems it is important to be able to monitor the transmission properties of the system. Monitoring can be used to locate faults in the fibre or in repeaters or amplifiers. It is also desirable to be able to carry out monitoring whilst the communications system is in service. Various schemes have been proposed.
GB2268017 describes a number of techniques based on the provision of a supervisory signal and a high loss loopback system in which a small proportion of the composite traffic and supervisory signal is looped back at every repeater in the transmission link. The supervisory signal may comprise a pseudo random signal, a reference signal with frequency sweep or a short pulse, and is transmitted using a sub-carrier over a single wavelength traffic signal. Information from the loopback signal is then obtained by detecting correlations, beat frequencies or simple time delays, respectively.
One problem with the techniques described in GB2268017 is the very low signal-to-noise ratio (SNR) associated with the returning signal, wherein the loopback supervisory component must be extracted from the background traffic signal and noise. An associated problem is the slow speed at which information on faults in the transmission system can be collated. For example, in the case of the short pulse supervisory signal, detection relies on a photodiode followed by a simple electrical bandpass filter for discrimination and the filtered signal is simply displayed on an oscilloscope.
U.S. Pat. No. 5,825,515 also discloses a method based on a high loss loopbacK system with a pseudo random supervisory signal. A low bit rate supervisory Pseudo Random Bit Sequence (PRBS) signal, which is Binary Phase Shift Keyed (BPSK) encoded and optically Intensity Modulation (IM) encoded is transmitted using a sub-carrier over a single wavelength traffic signal. A small proportion of the composite traffic and supervisory signal is looped back at every repeater and the supervisory signal is deconvolved from the noise by integration and correlation to yield a supervisory signal level from each repeater. However, using this method in a typical system means an in-service measurement time of the order of 4 hours.
The reason for the long measurement time is again the very low signal to noise ratio of the returned pulses. As a result, long integration times are required to extract the signal with any degree of confidence. Although a data cancellation scheme is employed in the system of U.S. Pat. No. 5,825,515, it is not possible to eliminate much of the noise generated in the PIN diode in the receiver. This problem is addressed U.S. Pat. No. 5,969,833, in which the supervisory signal is transmitted over a separate wavelength. This scheme also reduces the traffic signal penalty due to the supervisory modulation. However, both patents advocate the use of PRBS sequences, which must be de-convolved at the supervisory receiver in order to measure the loop loss to a given repeater. A separate de-convolution process is required for every repeater, which is expensive when many repeaters are present in a system.
The present invention aims to provide a simple, low-cost solution for the in-service monitoring of fibre optic transmission systems, and moreover a solution which delivers the required information in a satisfactory time period.
According to first aspect of the present invention, an optical transmission system comprises:
a first optical fibre extending between a first terminal and a second terminal for carrying optical signal traffic from the first terminal to the second terminal;
a second optical fibre extending between the first terminal and the second terminal for carrying optical signal traffic from the second terminal to the first terminal; and
a plurality of optical repeaters coupled to the first and second optical fibres between the first terminal and the second terminal, each repeater having a permanently connected passive high loss loop back circuit between the first and second fibres;
the first terminal including a transmitter and a receiver, wherein in use, the transmitter launches a pulsed supervisory signal on a dedicated supervisory wavelength on the first optical fibre, and the receiver detects a portion of the supervisory signal looped back from each repeater into the second optical fibre in order to identify the existence and location of faults in the transmission system, and wherein the pulsed supervisory signal is of sufficiently short duration such that portions of the signal returned from each repeater do not overlap with one another.
By employing a pulsed supervisory signal the present invention permits the use of relatively cheap and simple detection circuitry. Moreover, interference with counter-propagating traffic is avoided in the present invention by utilising a dedicated supervisory wavelength, thereby improving the SNR.
In addition, cross-talk between supervisory signals returned from repeaters is avoided by transmitting a pulse that is sufficiently short that when looped back from each repeater there is no overlap. Clearly, the maximum duration of the pulsed supervisory signal is dependent on the spacing of the repeaters in the system, and in particular on the smallest spacing between any two repeaters, and so the transmitter must be tuned to suit the system to which it is coupled.
Preferably, the supervisory wavelength is different to that of any optical signal traffic on the first optical fibre. Preferably, the pulsed supervisory signal is periodically repeated, the period of repetition such that each pulse is returned to the receiver from the furthest repeater before the immediately subsequent pulse is returned from the closest repeater. This is dependent on the time it takes the supervisory signal to be looped back from the furthest repeater.
Preferably, the supervisory signal is modulated to prevent transient effects arising from amplifiers in the transmission system.
In order to measure wavelength dependent losses a plurality of supervisory wavelengths may be used. The supervisory signals may be generated using a single tunable signal source or a plurality of separate signal sources.
A key aspect affecting the cost of the system and the speed of data collection is the architecture of the receiver.
Preferably, the receiver comprises a heterodyne receiver.
Preferably, the heterodyne receiver comprises a local oscillator coupled to a first mixer for mixing the received loopback signal portion with a signal from the local oscillator. Preferably, the heterodyne receiver further comprises a second mixer coupled to both the local oscillator and a transmit oscillator in the transmitter for mixing signals from the two oscillators to provide a reference frequency for synchronous demodulation of the received signal.
It is preferred that the heterodyne receiver comprises an IQ demodulator.
The integration of each return pulse is done sequentially so that a set of parallel detectors is not required. The need for a PRBS signal, a set of parallel correlators at the receiver and a data cancellation unit is also removed. Using the present invention in a typical system, the measurement time is reduced to about 1 minute.
According to a second aspect of the present invention, a method of monitoring the performance of an optical transmission system comprising a first optical fibre extending between a first terminal and a second terminal for carrying optical signal traffic from the first terminal to the second terminal, a second optical fibre extending between the first terminal and the second terminal for carrying optical signal traffic from the second terminal to the first terminal, and at least one optical repeater coupled to the first and second optical fibres between the first terminal and the second terminal, each repeater having a permanently connected passive high loss loop back circuit between the first and second fibres, comprises the steps of:
launching a pulsed supervisory signal on a dedicated supervisory wavelength on the first optical fibre, and
detecting a portion of the supervisory signal looped back from each repeater into the second fibre in order to identify the existence and location of faults in the transmission system, wherein the pulsed supervisory signal is of sufficiently short duration such that portions of the signal returned from each repeater do not overlap with one another.
Preferably, the supervisory wavelength is different to that of any optical signal traffic on the first optical fibre. Preferably, the pulsed supervisory signal is periodically repeated, the period of repetition such that each pulse is returned to the receiver from the furthest repeater before the immediately subsequent pulse is returned from the closest repeater. This is dependent on the time it takes the supervisory signal to be looped back from the furthest repeater.
Preferably, the method further includes the step of modulating the supervisory signal, in order to prevent transient effects from amplifiers in the transmission system.
In order to improve the SNR in the received signal, the method preferably includes the further step of optically filtering the loopback signal portion.
Preferably, the method further includes the step of heterodyne reception of the detected loopback signal portion.
Preferably, the method further includes the step of synchronous demodulation of the received signal.
To improve the accuracy of the measured signal still further, it is preferred that the method further comprises the step of averaging the loopback signal portion of a supervisory pulse returned from a repeater. More preferably, the method comprises the step of averaging the loopback signal portions of a plurality of supervisory pulses returned from a repeater.
According to a third aspect of the present invention, an optical transceiver comprises:
means to produce a pulsed supervisory signal on a dedicated supervisory wavelength; and
means to detect a portion of the supervisory signal looped back to the transceiver from repeaters in an optical transmission system coupled to the transceiver, in order to identify the existence and location of faults in the transmission system,
wherein the pulsed supervisory signal is of sufficiently short duration such that portions of the signal returned from each repeater do not overlap with one another.
Preferably, the supervisory wavelength can be selected to be different to that of any optical signal traffic on the optical transmission system. Preferably, the means to produce the pulsed supervisory signal includes means to periodically repeat the supervisory signal, the period of repetition such that each pulse is returned to the receiver from the furthest repeater before the immediately subsequent pulse is returned from the closest repeater.
Preferably, the transceiver further includes means to modulate the supervisory signal to prevent transient effects from amplifiers in the transmission system.
Preferably, a transmitter section of the transceiver comprises a pulsed laser.
In order to improve the SNR in the received signal, it is preferred that a receiver section of the transceiver comprises a tunable optical filter locked onto the supervisory wavelength.
Preferably, a receiver section of the transceiver comprises a heterodyne receiver. It is preferred that the heterodyne receiver comprises a local oscillator coupled to a first mixer for mixing the received loopback signal portion with a signal from the local oscillator. Preferably, the heterodyne receiver further comprises a second mixer coupled to both the local oscillator and a transmit oscillator in a transmitter of the transceiver for mixing signals from the two oscillators to provide a reference frequency for synchronous demodulation of the received signal.
Preferably, the heterodyne receiver comprises an IQ demodulator.
Thus, the present invention provides a simple cost effective technique for monitoring the performance of components such as repeaters in an optical transmission system. A sequence of short supervisory pulses is launched into the system on a dedicated wavelength and a high-loss loop-back arrangement provides a sample of each supervisory pulse returned from each repeater. Modulation of the pulses mitigates against non-linear effects in system components. Fast collection of high accuracy data with low SNR is obtained by a combination of targeted optical filtering, pulse averaging and heterodyne reception of the loopback signal portions. Simple calibration of the system can be performed during its commissioning with pulses of sufficiently high power for a good SNR.
Examples of the present invention will now be described in detail with reference to the drawings, in which:
The transmission system comprises a first optical fibre 10 which carries optical signals from a first terminal to a second terminal, as indicated by the arrows in
Each repeater also includes a lossy loopback circuit. The loopback circuit comprises an optical coupler 14, typically a tap coupler, at the output of the amplifier on each fibre. The optical coupler couples a portion of the signal in the fibre to the input of a pad attenuator 15. This portion of the signal which is tapped off is then coupled into the other fibre, i.e. to the return traffic path. The amplifier 13 compensates for the attenuation of not only the fibre itself but also of the coupler on each repeater.
The combined loss of the two tap couplers and the intermediate pad is typically 45 dB. Therefore only a small proportion of the optical signal on one fibre is looped back onto the other fibre from each repeater. It should be noted in this regard that the looped back circuit of
The optical signals transmitted from the first terminal will comprise a wavelength division multiplex (WDM) traffic comb over a set of wavelengths. A supervisory wavelength is also included and is coupled into the aggregate traffic signals. It is the supervisory signal which is used to measure losses at each repeater along the link. The supervisory-wavelength can be at either end of the WDM comb or between traffic channels. The supervisory wavelength is chosen to be between two adjacent traffic wavelengths of the counter propagating traffic. In other words, the supervisory wavelength for the first fibre 10 is chosen to lie between two traffic wavelengths of the traffic propagating along fibre 11. No absolute restrictions are placed on the location of the supervisory wavelength with respect to the co-propagating traffic, but performance is improved if it can also be placed between co-propagating traffic wavelengths.
Wavelength dependent loop back losses are measurable by inserting supervisory wavelengths as at a number of points along the WDM comb. If measurement time permits, this can be accomplished using a single tunable laser and a corresponding tunable optical filter at the receiver.
The supervisory signal is generated by a pulsed laser. The pulse duration is chosen to be as long as possible, but sufficiently short so that return signals from each repeater do not overlap. The launched 20 and returned pulses 21 are shown schematically in
Referring to
and the path difference Δzi between repeaters i-l and i is
Δzi=zi−zi-l=2li (2)
The time required for a signal to return from repeater is
where nr≈1.5 is refractive index of the transmission fibre. The difference in loopback time between adjacent repeaters i-l and i is then
To prevent the overlap of the return signal from adjacent repeaters, the pulse duration must be no greater than ΔTi|min, i=1 . . . n, where the min subscript refers to the minimum repeater spacing lmin. To allow for switching time between returned pulses, a guard band should also be provided of duration TG. The pulse duration is then shortened according to:
where
is the proportion of Tp+TG reserved for the guard band.
After a pulse has been launched, a second cannot be transmitted until all the loopback pulses have arrived. From (1) and (3), the time for the leading edge of the pulse to return from the nth repeater is
where
is the length of the system, excluding the last, (n+1)th, span.
The trailing edge of the pulse arrives after a time Tn+Tp and the pulse repetition time TR, allowing for the guard band, is then
where Tn+Tp is substituted from (5).
The transmitter section operates to produce a stream of supervisory pulses. Gated pulses from oscillator 30 are amplified by amplifier 32 and applied to the bias pin of laser diode 31. The resulting optical pulse from laser diode 31 has burst duration Tp, repetition rate Tr and modulation frequency fs. Prior to application to the laser diode 31, the pulses are shaped by pulse shaping network 33 in order to ensure that the pulses are of the desired profile, in this example a square shape.
The receiver section detects and processes the incoming signals to isolate the looped back supervisory wavelength and determine the amplitude of the returned pulses. As it is necessary to ensure that the receiver is tuned to the correct frequency for the returned pulses, the receiver includes a narrowband optical filter (not shown) prior to the optical detector 34, which can be tuned prior to operation of the system to correspond to the wavelength of the received pulses. A dither technique may be used, for example, so that the maximum SNR is obtained.
The receiver shown in
In operation relatively low power pulses must be used for the supervisory signal in order to avoid four wave mixing with data signals. There is also typically a large amount of noise on the returned signals owing to attenuation and amplification of the pulses in the repeaters. This leads to a low signal to noise ratio (SNR), which makes it necessary to average more than one measurement for each repeater in order to ensure a meaningful result. The received signals are detected as a sequence of samples and averaging is performed over a plurality of the samples. The averaging is performed across each pulse and may be repeated for a plurality of pulses from each repeater. Averaging the voltages in this way increases the SNR linearly with the number of samples whereas averaging power would increase the SNR as the square root of the number of samples.
In order to calibrate the system, it is necessary to determine when the returned pulses will be received from each repeater. This can be done using high power pulses during commissioning of the system. The returned high power pulses can be rapidly detected with minimal averaging because the SNR is sufficiently large.
Number | Date | Country | Kind |
---|---|---|---|
0409005.6 | Apr 2004 | GB | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/GB2005/001508 | 4/21/2005 | WO | 00 | 12/30/2006 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2005/104406 | 11/3/2005 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
5510925 | Suzuki et al. | Apr 1996 | A |
5825515 | Anderson | Oct 1998 | A |
6708004 | Homsey | Mar 2004 | B1 |
20040047629 | Evangelides et al. | Mar 2004 | A1 |
Number | Date | Country |
---|---|---|
0 883 254 | Dec 1998 | EP |
2 294 373 | Apr 1996 | GB |
Number | Date | Country | |
---|---|---|---|
20090154919 A1 | Jun 2009 | US |