WO2009068324A1

WO2009068324A1 - Optical sampling

Info

Publication number: WO2009068324A1
Application number: PCT/EP2008/050706
Authority: WO
Inventors: Antonella Bogoni; Luca Poti; Antonio Malacarne; Francesco Fresi
Original assignee: Telefonaktiebolaget LM Ericsson AB
Current assignee: Telefonaktiebolaget LM Ericsson AB
Priority date: 2007-11-28
Filing date: 2008-01-22
Publication date: 2009-06-04
Anticipated expiration: 2010-05-28

Abstract

The present invention relates to an optical sampler, to a systemincluding such a sampler, and to a method of sampling an optical signal comprising a pattern having a frequency. The method comprises providing the optical signal to a nonlinear medium, generating an optical sampling signal comprising a series of optical pulses, providing the optical sampling signal to the nonlinear medium to interact with the optical signal and thereby induce a nonlinear effect; and obtaining a sample result by making a measurement indicative of said induced nonlinear effect. The step of generating the optical sampling signal comprises controlling the frequency of said series ofoptical pulses using a controlsignal dependent upon the difference between a predetermined reference signal frequency and a predetermined function of the frequency of the pattern.

Description

Optical Sampling

TECHNICAL FIELD

The present invention relates to an optical sampler, to a system including such a sampler, and to a method of sampling an optical signal. Embodiments of the present invention are particularly suitable for use in the monitoring (and subsequent control) of optical communications signals

BACKGROUND

The need for direct monitoring of both ultra-short pulses and long bit sequences in the time domain is rapidly increasing, being of interest in a large number of applications such as ultra- fast communications, bio -photonics, sensing, large systems synchronization, testing and dynamic characterization of new materials. For example, the trend in optical telecommunications systems is to transmit the optical signals at faster bit-rates.

A number of different optical sampling techniques have been proposed in order to correctly visualize the behaviour of ultra-fast optical signals, but each technique introduces some limitations in terms of resolution, stability, complexity, elaboration time and scanning time interval.

Synchronous sampling methods can typically reach sub-picosecond resolution, but are able to analyze limited time intervals, for example only allowing the eye-diagram of a data signal to be resolved.

Asynchronous optical sampling methods can analyze longer time intervals (e.g. greater than lμs), but with longer refresh times due to the need to perform post-processing of the data. Moreover, asynchronous optical sampling introduces a higher jitter, intrinsic in the asynchronous operation.

Quasi-asynchronous (QA) sampling techniques are known, which attempt to combine the advantages of both synchronous and asynchronous sampling techniques, namely analysis of longer time intervals of optical signal, with high resolution. For example, the articles "High sensitivity waveform measurement with optical sampling using quasi-phasematched mixing in LiNbO3 waveguide" by S. Kawanishi et al., Electron. Lett., vol. 37, no. 13, pp. 842-843, June 2001, and M. Shirane et al., "A compact optical sampling measurement system using mode- locked laser-diode modules", by M. Shirane et al., IEEE Photon. Technol. Lett., vol. 12, no. 11, pp. 1537-1539, Nov. 2000, both describe quasi-asynchronous sampling techniques.

SUMMARY

It is an aim of embodiments of the present invention to provide a method of sampling an optical signal that substantially addresses one or more problems of the prior art, whether referred to herein or otherwise. It is an aim of particular embodiments of the present invention to provide a method of sampling an optical signal in the optical domain with quite high accuracy together with the possibility of visualizing relatively long time-intervals.

In a first aspect, the present invention provides a method of sampling an optical signal comprising a pattern having a frequency. The method comprises providing the optical signal to a nonlinear medium, generating an optical sampling signal comprising a series of optical pulses, providing the optical sampling signal to the nonlinear medium to interact with the optical signal and thereby induce a nonlinear effect; and obtaining a sample result by making a measurement indicative of said induced nonlinear effect. The step of generating the optical sampling signal comprises controlling the frequency of said series of optical pulses using a control signal dependent upon the difference between a predetermined reference signal frequency and a predetermined function of the frequency of the pattern.

Controlling the frequency of the optical sampling signal using a predetermined reference signal in combination with a frequency of the optical signal being sampled, allows the provision of an optical sampling signal that is relatively stable ie with low timing jitter. Consequently, due to the low time jitter, it is possible to sample relatively long bit sequences of the optical signal with relatively high accuracy, avoiding any data post-processing. The predetermined function of the frequency of the pattern may include a component indicative of the frequency of said series of optical pulses.

The predetermined function of the frequency of the pattern may be fs/N-fc, where fs/N is the frequency of the pattern, N is a predetermined integer, and fc is the frequency of the series of pulses.

N may be the pattern bit number.

The predetermined reference signal frequency may be determined based upon a desired optical sampling resolution.

After a sample result has been obtained by making a measurement indicative of said induced nonlinear effect, the value of the predetermined reference signal frequency may be altered so as to thereby alter the frequency of said series of optical pulses, the method further comprising: obtaining a further sample result by making a further measurement indicative of a nonlinear effect induced by the interaction of the optical sampling signal with the optical signal.

The step of generating the optical sampling signal may comprise operating a mode-locked laser having a laser cavity to generate the series of optical pulses, the length of the laser cavity being controlled in dependence upon said control signal.

The mode-locked laser may be an active mode-locked laser controlled to have a timing jitter less than 200 fs.

The optical sampling signal may be polarised.

The method may further comprise the step of polarising the optical signal.

The method may further comprise the step of controlling the polarisation of the optical sampling signal to maximise the induced nonlinear effect. The optical sampling signal may be provided to the nonlinear medium to induce the nonlinear effect of cross-phase modulation of the optical signal; and the step of obtaining a sample result may comprise making a measurement of a portion of the optical signal rotated by the cross- phase modulation.

The optical sampling signal may be provided to the nonlinear medium to induce the nonlinear effect of four- wave mixing with the optical signal to produce a resultant optical signal; and the step of obtaining a sample result may comprise making a measurement of at least a portion of said resultant optical signal.

The method may further comprise the step of controlling a parameter of the optical signal in dependence upon the obtained sample result.

The non- linear medium may comprise non- linear optical fibre.

In a second aspect, the present invention provides a sampler for sampling an optical signal comprising a pattern having a frequency. The sampler comprises: a nonlinear medium comprising an input arranged to receive the optical signal; an optical sampling signal generator arranged to generate an optical sampling signal comprising a series of optical pulses; an input to the nonlinear medium arranged to provide the optical sampling signal to the medium to interact with the optical signal and thereby induce a nonlinear effect; and a measurement device arranged to obtain a sample result by making a measurement indicative of said induced nonlinear effect. The sampler further comprises: a controller arranged to control a frequency of said series of optical pulses using a control signal dependent upon the difference between a predetermined reference signal frequency and a predetermined function of the frequency of the pattern.

The sampler may further comprise an oscillator arranged to generate said predetermined reference signal frequency.

The sampler may further comprise a frequency meter arranged to determine the frequency of the pattern of the optical signal.

The optical sampling signal generator may comprise a tuneable optical filter for selecting the optical wavelength of the optical sampling signal.

In a third aspect, the present invention provides a system comprising the above sampler.

BRIEF DESCRIPTIONS OF THE DRAWINGS

Preferred embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 is a schematic diagram of an optical sampler using Cross Phase Modulation in accordance with an embodiment of the present invention;

Figure 2 is a schematic diagram of the actively mode-locked fibre laser within the embodiment of Figure 1 ;

Figure 3 is a schematic diagram indicating the relative timings of the optical signal being sampled and the optical sampling signal during the operation of the quasi-asynchronous sampler of Figure 1;

Figure 4 shows two graphs of the results obtained from sampling of a first optical signal using respectively (a) a commercial 53 GHz oscilloscope, and (b) the optical sampler of Figure 1;

Figure 5 shows a graph indicating the corresponding pulse autocorrelation of the first sampled optical signal obtained from a commercial autocorrelator (thicker line) and the optical sampler of Figure 1, with the inset figure showing the acquired pulse trace; Figure 6 is a graph of the bit sequence as a function of time obtained from sampling of a

160Gbit/s second optical signal comprising 2ps return-to -zero pluses with a bit time of 6.25ps, using the optical sampler of Figure 1;

Figure 7 shows a graph indicating the corresponding pulse autocorrelation of the second sampled optical signal obtained from a commercial autocorrelator and the optical sampler (dotted, smoother line) of Figure 1;

Figure 8 shows the eye-diagram evaluation for a 160Gbit/s optical time division multiplexed signal as obtained by the optical sampler of Figure 1; Figure 9 is a schematic diagram of an optical sampler using Four-wave mixing in accordance with a further embodiment of the present invention;

Figure 10 is a graph of the bit sequence as a function of time obtained from sampling of a 640 Gbit/s third optical signal comprising 550 fs return-to-zero pulses with a bit time of 1.56 ps using the optical sampler of Figure 9;

Figure 11 shows a graph indicating the corresponding pulse autocorrelation of the sampled optical signal of Figure 10 obtained from a commercial autocorrelator and the optical sampler (smoother line) of Figure 9; and

Figure 12 shows the eye-diagram evaluation for a 640 Gbit/s optical time division multiplexed signal as obtained by the optical sampler of Figure 9.

DETAILED DESCRIPTION

Figure 1 is a schematic diagram illustrating the optical sampler 100 and its method of operation, with Figure 2 illustrating further details of the optical sampling signal generator 300 of the optical sampler 100.

The sampler 100 acts to sample an optical signal (i.e. the signal under test), which is received at the optical sampler input 110. The optical signal comprises a pattern of frequency fs/N (and corresponding period T), and optical wavelength λs. For example, the optical signal could be a bit pattern, with T=NTb (Tb being the bit time i.e. Tb = l/f_s where f_s is the bit frequency) and N being an integer (i.e. the bit sequence number).

The optical sampler 100 exploits a nonlinear interaction between the signal under test and the optical sampling signal 124. The optical sampling signal comprises a series of optical pulses of frequency fc (and corresponding period Tc). Preferably, so as to provide a high resolution, the pulses are relatively short e.g. 2ps or less, or even less than lps e.g. 500fs.

A frequency mismatch is imposed and maintained between the repetition rates of the sampling signal and the signal under test. The period of the sampling signal is Tc = T + Δt, where Δt is the desired temporal resolution. The pulsewidth of the sampling signal is preferably relatively short, so as to obtain the desired temporal resolution. If the signal under test is periodic with T=NTb (with Tb being the bit time and N the sequence bit number), the period of the sampling signal can be chosen as Tc = NTb + Δt. Thus, samples can be collected consecutively, so no post processing is required, as will be described further below with reference to Figure 3.

In this particular embodiment, the sampler exploits the ultra-fast response time (<50fs) of the Kerr effect, with the sampling signal inducing Cross Phase Modulation (XPM) on the signal under test. The XPM results in an induced phase shift, which leads to a rotation of the polarization of the signal under test. The optical sampler operates by performing a measurement indicative of the optical power of the resulting rotated portion of the signal under test. As XPM is utilised, it is desirable to maximise its efficiency. Hence, the polarization state of the optical sampling signal 124 is optimized in order to maximize the XPM effect induced on the signal under test.

The signal under test may already be polarized. If not, then the sampler input 110 can include a polarizer, to polarize the signal under test for subsequent sampling.

The polarized signal under test 112 is provided to an input 120 of a nonlinear medium 122. The polarized sampling signal 124 is also provided to an input of the nonlinear medium 122. In this instance, both inputs are provided by a coupler 120 e.g. a 50:50 coupler. The nonlinear medium is a nonlinear optical fibre e.g. a HNLF (Highly Nonlinear Fibre) such as a fibre having a nonlinear coefficient > 10 km^"1 W^"1.

Preferably, the power of the signal under test 112 is maintained at a level sufficiently low to avoid Self Phase Modulation e.g. the mean power of the signal can be 0 dBm or less. Preferably, to reduce spectral crosstalk, the wavelength of the sampling signal is separated by a predetermined range from the wavelength λs of the signal under test. For example, the wavelength separation of the signals can be between 10 and 50 nm e.g. 20nm. The wavelength of the sampling signal is tuneable, to allow an optimal separation to be selected for different signals under test.

An APC (automatic polarization controller) 126 is used to control the angle of polarization of the sampling signal 124 provided to the input 120. To maximise the XPM, the polarization state of the sampling signal is preferably controlled so as to be at an angle of 45° to the polarization state of the signal under test.

Within the nonlinear medium, the nonlinear effect of XPM leads to the signals 112 & 124 interacting. A change in the phase of the signal under test results, as each pulse within the sampling signal interacts with the signal under test. Figure 3 illustrates the sampling effect provided by this interaction.

The sampling signal 124 comprises a series of relatively short pulses, repeated at a period of Tc. As the period Tc of the sampling signal is different from (here, longer than) the period T of the signal under test, each subsequent pulse from the sampling signal 124 will interact with a different portion of the signal under test 112. Each interaction leads to a rotated portion of the signal under test, with the optical power of the rotated portion being indicative of the optical power of the portion of the signal under test that interacted with the relevant pulse from the sampling signal. Thus, the XPM effectively results in a series of rotated portions (or pulses) of the signal under test, the power of each of which can be measured to provide a series of sample results that correspond to the profile of the signal under test (as shown at the right hand side of figure 3, with the sample result being indicated by to, t_{l s}... I₁). The optical sampler makes a measurement indicative of the optical power of each of the rotated portions, as will now be described.

The output from the nonlinear medium 122 will comprise components at the wavelengths of both the sampling signal and the optical signal under test. The output is thus passed through a bandpass filter 130, having a pass band arranged to pass the wavelength λs of the signal under test, and to remove (not pass) the component(s) of the sampling signal and any out-of-band noise

To isolate the rotated portions of the signal under test, the (filtered) signal is then passed through a polariser 136, which is arranged to isolate the rotated portions of the signal under test for subsequent detection/ measurement e.g. to pass the rotated portions of the signal under test, and to not pass the unrotated portions. In this embodiment, the polarizer 136 is a fixed polarizer implemented as a Polarizing Beam Splitter. An APC 132 is thus located between the nonlinear medium 122 and the polarizer 136, to control the polarization of the signal output from the nonlinear medium. The APC is controlled in dependence upon a feedback signal from the polarizer 136, indicative of the optical power of the portion of signal still in its original polarization state passed by the polarizer. The APC 134 is operated to adjust the polarization of the optical signal output from the nonlinear medium, to maximise the rotated portion passed by the polarizer 136.

The optical power of the optical output signal from the polarizer 136 (i.e. comprising the rotated portions of the signal under test) is subsequently measured by optical detector 144, as a function of time. For example, detector 144 can be formed as a photodiode, arranged to provide an output to an oscilloscope 146, with the oscilloscope subsequently displaying the sample results. The optical detector can also be configured to provide an output to a control unit 150. The control unit 150 could then use the sample results from the sampler to control one or more parameters of the signal under test e.g. the sample results can act as a feedback signal, for controlling the generation and/or modulation of the signal under test, for optimisation of that signal.

In the particular embodiment shown, a number of further elements (138, 140, 142) are provided between the polarizer 136 and the detector 144. An optical amplifier 140 is provided, to amplify the power of the optical signal output from the polarizer 136, to allow convenient detection. For example, the optical amplifier could be an EDFA (Erbium Doped Fibre Amplifier). A band-pass filter 142 is provided at the output of the amplifier 140, to pass the wavelength of the signal under test and to inhibit other wavelengths e.g. to minimise the out-of- band noise.

A splitting device 138 is coupled to the output of the polarizer 136, and arranged to pass the majority (e.g. 90%) of the optical signal output from the polarizer 136 towards the detector 144. The remaining portion is used as a feedback signal to the APC 126, to allow the APC 126 to control the polarization of the sampling signal 124, so as to maximise the resulting XPM.

The polarized sampling signal 320 input to the APC 126 is generated by the optical sampling signal generator 300, details of which are shown in Figure 2. In this particular example, the generator 300 comprises a mode-locked laser, and in particular an actively mode locked laser. The generator 300 is arranged to generate the sampling signal 320 consisting of a series of pulses of frequency fc.

The laser cavity of the generator comprises an Optical Tunable Filter 306 (e.g. a tunable Optical Band Pass Filter that selects the wavelength of the sampling signal). An Optical Isolator 304 ensures the optical signal only pass around the ring cavity in a single direction. A tunable optical delay line (ODL) 302 allows the length of the cavity to be varied. An optical amplifier 308 e.g. an Erbium Doped Fiber Amplifier (EDFA), provides the gain medium. A modulator 310 (e.g. a phase or amplitude modulator), driven by an appropriate control signal 316 is used to lock the phase of the cavity modes. The modulator used here is a Mach Zehnder (MZ) modulator, driven by a sinusoidal RF wave, whose frequency is equal to a multiple of the cavity natural frequency.

Optical splitter 314 is used to split of a portion of the optical signal with the cavity, to act as the sampling signal 320 and also a feedback signal 322 indicative of the frequency fc. These individual elements 302-310 & 314 are connected by optical fibre 312. The laser cavity is composed of all Polarization Maintaining (PM) fibers and components to eliminate polarization instabilities and to reduce the vibration sensitivity.

Mode-locking (ML) refers to the situation in which the cavity modes are made to oscillate with comparable amplitude and with locked phase. Considering 2n+l longitudinal optical modes, the phase-locked condition requires a constant phase difference between all the longitudinal optical modes circulating inside a laser cavity. This property of ML modes allowed the generation of a train of ultra-short evenly spaced light pulses. The pulsed signal period depends on the round-trip time of photons along the laser cavity (τp) according with:

τ_P =y_c ⁽ⁱ⁾ where L is the length of the cavity and c is the speed of light into the laser cavity. Therefore the repetition rate fc of the generated pulses is given by: fc =f = Nf_cav (2) with N an integer number (N is in the range 2000-4000 for a 10 GHz source and it depends on the length of the cavity), and fcav=c/L the natural (fundamental) cavity oscillation frequency.

Thus, the frequency fc of the sampling signal is a function of the cavity length, and hence varying the length of the cavity (by the ODL 302) controls the frequency fc of the sampling signal. In this particular example, the ODL 302 comprises a piezoelectric controller, with a control signal 212 (typically a voltage signal) applied to the controller controlling the ODL length, and hence the frequency fc of the sampling signal. The generator 300 can thus be regarded as an optical voltage controlled oscillator, as the frequency fc depends upon the control signal 212.

To ensure that sampling operation has a high accuracy, it is important that the frequency fc of the sampling signal is accurately controlled, and hence it is important that the signal 212 (used to control fc) represents a good control signal. This is achieved in the embodiment by using a double PLL (phase locked loop) architecture 208, 210 in combination with a signal of frequency fiχ_> generated by an oscillator 202, to allow accurate generation of the control signal 212. In particular, the controller 200 generates signal 212 (which controls fc) so as to lock the difference between the pattern frequency fs /N and fc to fLo-

That function can be implemented as follows. The controller 200 receives a signal 114 indicative of the frequency fs i.e. the frequency of the signal under test. This signal 114 could be provided by the same source that provides the signal under test. Alternatively, input 110 can comprises a frequency meter arranged to measure the frequency of the signal under test, and provide a signal indicative of this frequency to the controller 200.

The signal 114 indicative of fs received by the controller 200 is then passed to a divider 204 (e.g. a frequency divider), to calculate fs/N. The resulting signal indicative of fs/N is then passed to mixer 208. A signal 206 (indicative of the frequency fc of the sampling signal) is also passed to mixer 208, which outputs a signal indicative of fs/N-fc to mixer 210. Oscillator 202 also passes a signal of frequency fiχ_> to mixer 210, which generates the control signal 212 in dependence upon the difference between the predetermined reference signal frequency (frequency fLo) and a predetermined function of the frequency of the pattern (i.e. fs/N-fc).

Thus, using a local oscillator 202, the difference between the Nth sub-multiple of the signal frequency fs and the sampling frequency (i.e. fLo=fs/N-fc) is maintained constant, allowing accurate control of the sampler signal generator 300, and hence accurate generation of the sampling signal.

The frequency mismatch imposed using the local oscillator 202 is determined by the desired sampling resolution (Δt), and can be calculated according to the formula:

For example, if the integer N=128, it is possible to resolve every bit pattern with a period of 2ⁿ bit and n < 7. When fs = 10GHz, it is possible to obtain a 100 fs temporal resolution choosing a repetition rate fc = 78.124MHz for the sampling signal and a frequency mismatch fiχ_> ~ 610Hz. This solution provides a refresh time of - 1.6 ms. Moreover, tuning the repetition rate of the sampling signal by altering the repetition rate of the sampling signal (i.e. altering fc), it is possible to also resolve bit patterns with a standard length of 2^{n l} bit.

The sampling signal frequency fc can be adjusted by altering the value of fLo. Thus, if desirable, the signal under test could be sampled at a first value of fc to obtain a first sampling result, then f_Lo altered, and the signal under test sampled at a second value of fc so as to obtain a further sample result, at a different resolution.

Experimental Results

The results of experimental measurements obtained using a particular implementation of the above optical sampler will now be described.

In the experimental measurement, the value N=8 was used. The first signal to be sampled was obtained from a 10GHz Actively Mode-Locked Fibre Laser (AMLFL) producing 4-ps pulses (as measured with a commercial autocorrelator). The pulse train was modulated by a Mach Zehnder modulator using an 8 bit pattern. The sampling signal was obtained using a second AMLFL acting as an optical VCO (Voltage Controlled Oscillator), followed by a pulse compression stage to reach a pulsewidth of 500 fs. A temporal resolution Δt of 500 fs was selected by setting a frequency mismatch (ie the oscillator frequency value) to fLo=781.25 KHz. The sampling pulses were coupled with the signal under test into the nonlinear medium, a HNLF (Highly Nonlinear Fibre). The mean power of the sampling signal was set at about 6dBm, while the signal power was maintained low (<0dBm), to avoid Self Phase Modulation.

The sampling signal optical wavelength was a set predetermined distance (20nm) from the wavelength of the signal being sampled. Such a separation does not affect XPM and reduces the spectral crosstalk. After output of the rotated signal from the polarizer, the optical samples were then amplified, filtered, photo-detected and viewed on a 600 MHz real-time oscilloscope.

Figure 4 shows a comparison between the bit sequence evaluated on a 53GHz commercial oscilloscope (a) and the one reconstructed by means of the present sampler (b). Due to the bandwidth limitations, the 53GHz oscilloscope is not able to measure the real pulsewidth; on the other hand, the optical sampling oscilloscope presents a higher resolution.

In order to confirm the effectiveness and the accuracy of the sampler, the obtained sampled trace was compared with the sample trace obtained from a commercial autocorrelator. The inset in Figure 5 shows the optical sampled pulse shape, whose measured pulsewidth is 4.2 ps, with the main portion of Figure 5 showing good conformity between the autocorrelation supplied by the commercial autocorrelator and the one evaluated from the sampled pulse.

Further measurements were also carried out using a 160Gbit/s OTDM (Optical Time Division Multiplexed) signal as a second, different signal under test. The OTDM signal was obtained by time-multiplexing the previous 10Gbit/s signal (after a pulse compression stage to reduce pulsewidth down to 2 ps). When the oscilloscope was directly triggered by the frequency signal from the oscillator (i.e. by the signal fLo), the bit pattern was visualized (Figure 6).

As in the previous case, to confirm the accuracy of the acquired trace a comparison between the autocorrelation supplied by a commercial autocorrelator and the one measured from the optical sampler trace was carried out, with the results shown in Figure 7. The good agreement between the two traces confirms the effectiveness of the sampler for both shape and pulsewidth measurement; the measured pulsewidth is 2 ps.

Eye-diagram evaluation (Figure 8) was also performed by inserting a frequency multiplier to multiply the clock frequency provided by the local oscillator, prior to input to the oscilloscope.

The results show the successful implementation of a QA all-optical sampling oscilloscope based on XPM-induced polarization rotation in fibre, able to benefit from the advantages of both synchronous and asynchronous methods.

In particular, the results show that the sampler was able to resolve long bit sequences (ie bit- sequences of tens of nanoseconds) with sub-ps-resolution, and with a low refresh time (< 1.6ms). The optical sampling of a bit sequence at 10Gbit/s (4-ps pulse) and of a bit sequence 160Gbit/s (2-ps pulse) has been experimentally demonstrated, showing high accuracy, as confirmed by comparisons with commercial instruments. The capability of performing eye- diagram evaluation has also been experimentally demonstrated.

Thus, the optical sampler 100 provides a technique for QA sub-ps-resolution optical sampling oscilloscope based on polarization rotation induced by XPM (Cross Phase Modulation). The operation of the particular embodiment described herein is based on the non- linear interaction in optical fibre between the signal to be resolved (the signal under test) and a sampling ultra- short pulse train whose frequencies are locked to a fixed difference. In this way long bit sequences can be displayed with high accuracy, avoiding any data post-processing and consequently with a very low refresh time (<1.6ms). A double PLL scheme able to maintain a fixed frequency mismatching is described. Finally, sub-ps-resolution has been demonstrated through a commercial autocorrelator. Further embodiment

Whilst the above particular implementation has been described with respect to an optical sampler 100 that exploits the nonlinear interaction of XPM, it should be appreciated that other nonlinear effects maybe utilised.

Figure 9 is a schematic diagram of an optical sampler 400 that utilises the nonlinear effect of four- wave mixing (FWM).

FWM is a nonlinear effect rising from a third-order optical nonlinearity. FWM can occur if at least two different wavelength components propagate together in a nonlinear medium, e.g. FINLF. Assuming just two input optical wavelength components X₁ X₂ (with X₂>λi), a refractive index modulation at the difference wavelength occurs, which creates sidebands for each of the input waves. In effect, two new wavelength components (λ₃ and X₄) are generated: λ₃=λi-(λ₂-λi)=2λi-λ₂, and λ₄=λ₂+(λ₂-λ₁)=2λ₂-λ₁.

Thus, FWM can be utilised in sampling in a similar manner to XPM, by providing the signal under test at a first optical wavelength (e.g. X₁), and the sampling signal at a second optical wavelength (e.g. X₂). A measurement indicative of the induced nonlinear effect of FWM can be obtained by making a measurement of the resultant optical signal at wavelength X₃, and/or a measurement of the resultant optical signal at wavelength X₄. Assuming Pi is the power of the Xi component and P₂ is the power of the X₂ component, theory indicates that the power P₃ of the X₃ component is proportional to Pi². P₂, whereas the power P₄ of the X₄ component is proportional to P₂ ²P₁. It is thus preferable to measure the power of the component that is proportional to the power of the signal under test.

For example, if Xi is the wavelength of the signal under test, to obtain a sampled signal proportional to the power of the signal under test, it is preferable to make a measurement indicative of the power of the resultant optical signal at wavelength X₄.

Figure 9 is a schematic diagram of the optical sampler 400 that utilises the nonlinear effect of FWM. It would be seen that the optical sampler 400 is generally similar to the optical sampler 100 described with reference to Figure 1, with identical reference numerals being used to represent similar elements. As the majority of the sampler 400 operates in a similar manner to the sampler 100, only the elements of the sampler 400 that operate in a different manner will be described in detail.

It would be noted that the sampler 400 does not include an APC 134 or a polariser 136. Such an APC/polariser combination is not required, as the resultant optical signal from the FWM nonlinear effect is of a different wavelength (X_FWM) than either the sampling signal or the signal being sampled (i.e. the signal under test).

Within the optical sampler 400, the controller 200 and the optical sampling signal generator 300 operate in the same manner as within the sampler 100.

In the sampler 400, the polarised sampling signal 320 output from the optical sampling signal generator 300 is passed through a polarisation controller 426. The polarisation controller 426, which is here implemented as an APC, is arranged to control the polarisation of the polarised sampling signal, so as to ensure that the sampling signal 424 provided to input 120 is in the same polarisation state as the signal under test (112). Having the two optical signals 112 and

424 at the same polarisation maximises the FWM efficiency. Within the nonlinear medium 122 (e.g. 250m of HNLF), the optical sampling signal 424 and the signal under test 112 will undergo FWM, which will produce two new optical frequency (i.e. optical wavelength) components. Either or both of these two resultant optical signals could theoretically be measured to obtain a sample result. However, in this particular embodiment, only the resultant optical signal (at wavelength X_FWM) is measured (i.e. the component is measured that has a power proportional to the power of the signal under test).

Accordingly, the output of the nonlinear medium 122 is passed through a bandpass filter 430, which has a passband arranged to pass the wavelength λpwM of the resultant optical signal, and to remove (not pass) the wavelengths of the sampling signal and the signal under test. The output from the bandpass filter 430 is then passed to a splitter 138, which as previously splits a portion of the signal (e.g. 10%) to provide as feedback to the polarisation control 426. The polarisation controller 426 can thus control the polarisation of the sampling signal 424, so as to maximise the power of the optical signal received from splitter 138 i.e. to thus maximise the FWM efficiency between the sampling signal and the signal under test.

The majority of the resultant optical signal is passed, via the splitter 138, through an optical amplifier to amplify the power of the resultant optical signal, to allow convenient detection. A bandpass filter 442 is provided at the output of the amplifier 140, to pass the wavelength of the resultant signal (X_FWM) and to inhibit other wavelengths, e.g. to minimise the out of band noise.

The optical power of the resultant optical signal is measured by optical detector 144. As previously, the output of detector 144 is conveniently attached to an oscilloscope 146, with the oscilloscope subsequently displaying the sample results.

In the sampler 100 the oscillator 202 was shown as providing a trigger direct to the oscilloscope 146. It will be seen that in the sampler 400, a slightly different arrangement is in place to provide the trigger signal to the oscilloscope 146.

In particular, a switch 446 is provided on the trigger input to the oscilloscope 146. The switch 446 can switch between receiving a signal direct from the oscillator 202 (i.e. at frequency fLo), and receiving a N multiplied version of the oscillator frequency. If the input trigger is taken as the oscillator frequency fLo then the bit pattern can be directly visualised (e.g. the oscilloscope output will be of the type shown in Figure 4) whilst if the switch 446 is actuated to provide an input of f_Lo*N, then the eye-diagram can be visualised (e.g. an output of the type shown in Figure 8 would be shown by the oscilloscope).

Either sampler 100 or sampler 400 can be used within a larger system. In Figure 9, sampler 400 is shown as being included within a larger system 500. Optical detector 144 is configured to provide an output to a control unit 450. The control unit processes the sample results from the sampler to control one or more parameters of the signal under test i.e. the sample results act as a feedback signal, for controlling the generation and/or modulation of the signal under test, for optimisation of that signal.

For example, the system 500 can be a telecommunications system or a node within a telecommunications system, with the signal under test being an optical signal used for transmission of information by the telecommunications system. The telecommunications system or the node can include an optical signal generator for generating the signal for transmission of information, or an optical signal modifier arrange to modify (e.g. condition, or alter the properties of) the optical signal. The optical signal generator or modifier can include control unit 450, for controlling the generation or modification of the optical signal used for transmission of information. The telecommunications system or node 500 can also include a medium for transmission of the generated optical signal, and a receiver for receiving the optical signal. The samplers 100 and 400 can of course also be used in other systems or portions of systems.

Further experimental results

The results of experimental measurements obtained using a particular implementation of the optical sampler illustrated in Figure 9, will now be described.

In the experimental measurements described here, the signal under test was a 640 Gb/s Return- to-Zero (RZ) data signal having 550 fs RZ pulses, with a bit time of 1.56 ps. A sampling signal of pulse-width of 500 fs was utilised to exploit the FWM effect.

The signal under test was generated using a 10GHz actively mode-locked fibre laser producing 4 ps pulses. The pulse train was then compressed down to 500 fs using a multiple stage compressor based on alternate spans of FINLF and single mode fibre. The 500 fs pulses were then multiplexed using a split-and-delay multiplexer in order to achieve the 640 Gb/s bit pattern. Both the signal under test and the sampling signal were coupled into a 50m length of

FINLF as the nonlinear medium 122. A 3dB coupler was used as the input 120, with the mean power of both the sampling signal and the signal under test provided to the input 120 each being respectively controlled by a separate optical amplifier, in order to maximise the effect of nonlinearities within the FINLF.

The wavelength of the sampling signal was selected to be 14nm separate from the wavelength of the signal under test. A frequency f_Lo=200kHz was selected, with the optical samples then being viewed on a 600MHz real-time oscilloscope. Figures 10, 11 and 12 show respectively the bit sequence, the corresponding pulse autocorrelation, and the eye-diagram evaluation of the sample signals obtained using this apparatus. Figures 10, 11 and 12 are experimental results using a sampler 400 exploiting FWM, and can be regarded as corresponding figures to Figures 6, 7 and 8 for the sampler exploiting XPM.

Experimental results indicate that the sampler 400 exploiting FWM provided a better (higher) extinction ratio and a lower noise floor than the experimental results obtained using a sampler exploiting XPM.

From the foregoing description, various alternatives will be apparent to the skilled person as falling within the scope of the present invention, as defined by the appended claims.

Claims

1. A method of sampling an optical signal comprising a pattern having a frequency, the method comprising: providing the optical signal to a nonlinear medium; generating an optical sampling signal comprising a series of optical pulses; providing the optical sampling signal to the nonlinear medium to interact with the optical signal and thereby induce a nonlinear effect; and obtaining a sample result by making a measurement indicative of said induced nonlinear effect, wherein the step of generating the optical sampling signal comprises: controlling the frequency of said series of optical pulses using a control signal dependent upon the difference between a predetermined reference signal frequency and a predetermined function of the frequency of the pattern.

2. A method as claimed in claim 1, wherein the predetermined function of the frequency of the pattern includes a component indicative of the frequency of said series of optical pulses.

3. A method as claimed in claim 1 or claim 2, wherein the predetermined function of the frequency of the pattern is fs/N-fc, where fs/N is the frequency of the pattern, N is a predetermined integer, and fc is the frequency of the series of pulses.

4. A method as claimed in claim 3, wherein N is the pattern bit number.

5. A method as claimed in any one of the above claims, wherein the predetermined reference signal frequency is determined based upon a desired optical sampling resolution.

6. A method as claimed in any one of the above claims, wherein after a sample result has been obtained by making a measurement indicative of said induced nonlinear effect, the value of the predetermined reference signal frequency is altered so as to thereby alter the frequency of said series of optical pulses, the method further comprising: obtaining a further sample result by making a further measurement indicative of a nonlinear effect induced by the interaction of the optical sampling signal with the optical signal.

7. A method as claimed in any one of the above claims, wherein the step of generating the optical sampling signal comprises operating a mode-locked laser having a laser cavity to generate the series of optical pulses, the length of the laser cavity being controlled in dependence upon said control signal.

8. A method as claimed in claim 7, wherein the mode-locked laser is an active mode- locked laser controlled to have a timing jitter less than 200 fs.

9. A method as claimed in any one of the above claims, wherein the optical sampling signal is polarised.

10. A method as claimed in claim 9, further comprising the step of polarising the optical signal.

11. A method as claimed in claim 9 or claim 10, further comprising the step of controlling the polarisation of the optical sampling signal to maximise the induced nonlinear effect.

12. A method as claimed in any one of claims 9 to 11, wherein: the optical sampling signal is provided to the nonlinear medium to induce the nonlinear effect of cross-phase modulation of the optical signal; and the step of obtaining a sample result comprises making a measurement of a portion of the optical signal rotated by the cross-phase modulation.

13. A method as claimed in any one of the above claims, wherein: the optical sampling signal is provided to the nonlinear medium to induce the nonlinear effect of four- wave mixing with the optical signal to produce a resultant optical signal; and the step of obtaining a sample result comprises making a measurement of at least a portion of said resultant optical signal.

14. A method as claimed in any one of the above claims, further comprising of the step of controlling a parameter of the optical signal in dependence upon the obtained sample result.

15. A method as claimed in any one of the above claims, wherein the nonlinear medium comprises highly nonlinear optical fibre.

16. A sampler for sampling an optical signal comprising a pattern having a frequency, the sampler comprising: a nonlinear medium comprising an input arranged to receive the optical signal; an optical sampling signal generator arranged to generate an optical sampling signal comprising a series of optical pulses; an input to the nonlinear medium arranged to provide the optical sampling signal to the medium to interact with the optical signal and thereby induce a nonlinear effect; a measurement device arranged to obtain a sample result by making a measurement indicative of said induced nonlinear effect, wherein the sampler further comprises: a controller arranged to control a frequency of said series of optical pulses using a control signal dependent upon the difference between a predetermined reference signal frequency and a predetermined function of the frequency of the pattern.

17. A sampler as claimed in claim 16, further comprising an oscillator arranged to generate said predetermined reference signal frequency.

18. A sampler as claimed in claim 16 or claim 17, further comprising a frequency meter arranged to determine the frequency of the pattern of the optical signal.

19. A sampler as claimed in any one of claims 16 to 18, wherein the optical sampling signal generator comprises a tuneable optical filter for selecting the optical wavelength of the optical sampling signal.

20. A system comprising a sampler as claimed in any one of claims 16 to 19.