Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Generation and performance Investigation of 40GHz phase stable and pulse width-tunable optical time window based on a DPMZM

Open Access Open Access

Abstract

We experimental demonstrate stable, simple and pulse width-tunable 40 GHz short time window generation using a dual parallel Mach-Zehnder modulator (DPMZM) driven simply by an electrical clock. The pulse widths are measured to be tunable continuously from 5.6ps to 12.6ps by simply adjusting the DC bias voltage of DPMZM. The timing jitter, extinction ratio (ER), optical signal-to-noise ratio (OSNR) and insertion loss (IL) of the generated width-tunable time window are measured to be better than 50fs, 18dB, 50dB and 16dB respectively.

©2012 Optical Society of America

1. Introduction

Stable short optical time window of low timing jitter, short pulse-width, phase preserving characteristics, low insertion loss at high repetition rate has attracted a lot of interest due to its important application in all-optical signal processing, optical sampling, return-to-zero (RZ) data, fiber-optical sensing, generating sub-picosecond pulse sources assisted by pulse compression and ultrahigh optical time-division-multiplexing (OTDM) with high level modulation formats. Moreover, in nearly all optical systems, the ultimate performance can be critically dependent on matching the optical pulse-width and wavelength to the optimal overall system parameters. For example, the performance of an OTDM transmission link can vary significantly depending on the pulse width and wavelength, even for small changes in fiber characteristics [1]. Short pulses providing pulse-width and wavelength tunability would be highly desirable in order to optimize the overall system performance.

As has been reported, high repetition rate up to 40 GHz time windows can be generated using electro-optic devices such as electro-absorption modulator (EAM) [2] or fiber looped amplitude modulator [3] or fiber looped polarization modulator (FL-PM) [4] driven simply by an electrical clock or cascaded Mach–Zehnder intensity modulator (MZM) [5] or single dual-drive MZM [6]. However, the EAM and cascaded MZM based techniques suffer from large insertion loss, the FL-MZM and FL-PM is unstable due to the non-integrated fiber loop and dual-drive MZM based method suffers from large RF clock power. In ref [7], we proposed a stable time switch based on a dual parallel Mach-Zehnder modulator driven simply by an electrical clock.

At the same time, several methods have been reported for generation of width-tunable pulses such as: by adjusting both the chirp applied to the pulse and the dispersion value of a tunable dispersion element [8]; by using four-wave mixing (FWM) in highly-nonlinear fiber (HNLF) [9,10] or semiconductor optical amplifier (SOA) [11] to combine two parallel optical pulse trains on different wavelengths and varying the delay between the two pulse trains; by driving a LiNbO3 modulator with differential electrical clock signal and adjusting the phase shift between the driving clock signals [12]; or by changing the gain of distributed Raman amplifier (DRA) based adiabatic soliton compressor [13]. All the above-mentioned methods suffer from complicated adjustment method or non-integrated long fiber used which makes the time window unstable.

In this paper, we report for the first time, a pulse width-tunable, power effective, phase preserving and stable 40 GHz time switch by using a dual parallel Mach-Zehnder modulator driven simply by an electrical clock. The power of the driving clock is less than 22 dBm. Experimental results show the full-width at half-maximum (FWHM) pulse widths can be tuned continuously from 5.6 ps to 12.6 ps by simply adjusting the DC bias voltage of DPMZM. The timing jitter, extinction ratio (ER), optical signal-to-noise ratio (OSNR) and insertion loss (IL) are measured to be better than 50 fs, 18 dB, 50 dB and 16 dB respectively.

2. Concept and experimental setup

Figure 1 shows a schematic illustration on how to generate a short time window using a single DPMZM driven by one electrical clock. A DPMZM comprises two child Mach-Zehnder modulators (MZM1 and MZM2) nested within a third parent Mach-Zehnder modulator (MZM3). There are three independent DC bias voltages and two RF inputs. Let VRF1 and VRF2 represent the RF modulating electrical voltage of MZM1 and MZM2, VDC1, VDC2 and VDC3 represent the DC bias voltage applied to MZM1, MZM2 and MZM3, when chirp-free operation of each child MZM is assumed, the electric fields at point (a) and (b) can be described as:

 figure: Fig. 1

Fig. 1 Schematic of DPMZM based short time window generation.

Download Full Size | PDF

Ea(t)=Ein(t)2cos[π2Vpi(VRF1(t)+VDC1)]ejπVDC12VpiEb(t)=Ein(t)2cos[π2Vpi(VRF2(t)+VDC2)]ejπVDC22Vpi

Where Ein(t)is the input signal. Vpi is the driving voltage for swithing the output power of the modulators from zero to its maximum. The power spliting ratio of arm two for the input Y-branch waveguide is assumed to be 0.5. The output electric field Eout(t)at point (c) given by the interference of Ea(t) and Eb(t) is:

Eout(t)=Ea(t)+ejπVDC3VPIEb(t)

LetVRF1=V0sin(2πf0t) and VRF2 = 0, the output optical amplitude comes to be:

Eout(t)=Ein(t)2{[cosπ2Vpi(V0sin(2πf0t)+VDC1)]+ej2πVDC3+πVDC2πVDC12VPIcosπVDC22Vpi}ejπVDC12Vpi

Where V0 and f0 are the peak to peak amplitude and frequency of RF modulation clock.

Figure 2 illustrates the concept diagram of the proposed pulse-width tunable time window generation. The blue lines in Fig. 2(a) show the modulation curves of a typical Mach-Zehnder modulator (MZM), the solid blue line and dashed blue line represent optical power and electric field separately. Typical optical pulses with a full-width at half maximum (FWHM) of 50% of the bit duration can be realized by biasing the Mach-Zehnder modulator at quadrature point (0.5 Vpi) as indicated by solid circle. Three cases are discussed here for generating three kinds of tunable pulses.

 figure: Fig. 2

Fig. 2 Concept diagram of the DPMZM based optical switch. (a) Modulation curve of MZM; (b ~d) pulse waveform and phase output from MZM1and MZM2; (e ~g) time window waveform output from DPMZM

Download Full Size | PDF

Firstly, when Child-MZM1 is biased at a point between quadrature point and peak point (VDC1 = 0~0.5Vpi) as depicted by open circle “Bias-A” in Fig. 2(a), inverted pulses with continuous phase is generated after Child-MZM1 as shown in Fig. 2(b); By adjusting VDC2, a CW light with average optical power around peak of the inverted pulse at the output of child-MZM2 is generated as depicted by red line in Fig. 2(b); Then, by adjusting VDC3 to make sure the output from the two child MZMs have a phase difference of π, low duty cycle short pulses as indicated in Fig. 2(e) can be obtained according to the interference of the inverted pulses and the CW light. Since the pulse-width of the generated time window is proportional to that of the inverted pulses, and the inverted pulses broaden when VDC1 changes from 0 to 0.5Vpi, the emitted pulses from DPMZM is pulse-width tunable by adjusting the DC bias voltage.

Secondly, when Child-MZM1 is biased at a point between quadrature point and null point (VDC1 = 0.5Vpi ~Vpi) as depicted by open circle “Bias-B” in Fig. 2(a), pulses with periodic phase information as illustrated in Fig. 2(c) are generated at the output of Child-MZM1, the optical power and phase are illustrated by solid black line and dotted blue line respectively. The phase difference of the pulses and their pedestals is π. Therefore, by adjusting VDC2 to make sure the optical power of the CW light from Child-MZM2 equal to the maximum optical power of pulses from Child-MZM1, and by carefully adjusting VDC3, pulses with higher duty cycle, as illustrated in Fig. 2(f), can be generated according to the constructive and destructive interference of the light from two child MZMs. The pulse-width is still tunable through adjusting the DC bias voltage of the three MZMs.

Lastly, when Child-MZM1 is biased at a point between null point and second quadrature point (VDC1 = Vpi ~1.5Vpi) as depicted by open circle “Bias-C” in Fig. 2(a). Pulses as illustrated in Fig. 2(d) are generated at the output of Child-MZM1, Solid black line and dotted spot line in Fig. 2(d) indicate the power and phase of the output pulses. Comparing the pulses in Fig. 2(d) with that in Fig. 2(c), we can find the two pulses sequences have similar shape and phase with a time delay of half-bit duration; When the power of the CW light from Child-MZM2 is adjusted to be equal to the second peak power of Child-MZM1 pulses, the pedestals of Child-MZM1 pulses are cancelled according to destructive interference by carefully adjusting VDC3, broad pulses without pedestals as shown in Fig. 2(g) are realized.

3. Simulation and experimental results

The experimental setup is similar as shown in Fig. 1. A CW light at 1550.12 nm with output power of 12 dBm is injected into a chirp free DPMZM with alpha parameter of less than ± 0.2 and 3 dB optical bandwidth of 39.5 GHz and 35.3 GHz at two RF port separately. The on/off extinction ratio at low frequency for parent MZI and child MZI are all better than 30 dB, the total insertion loss of the DPMZM is 4.1 dB. The child MZM with 39.5 GHz bandwidth is driven by a 40 GHz sinusoidal clock with tunable RF power between 10 dBm to 30 dBm (peak to peak voltage between 2 V and 20 V). The output pulses are amplified by an EDFA with noise figure of 4.5 dB and saturation output power of 14 dBm, and then fed into a 500 GHz optical sampling oscilloscope (OSO) and an optical spectrum analyzer (OSA). In simulation process, all the parameters are set to be equal to the experimental setup for the convenience of comparative analysis.

The characteristics of DPMZM generated time window in various case of Child-MZM1 DC bias with fixed RF clock power are investigated theoretically and experimentally as shown in Fig. 3 and Fig. 4 . The RF power is set to be 1.5Vpi for simplification; the voltages of VDC2 and VDC3 are properly set to optimize the DPMZM generated pulses in both experiment and simulation. According to the simulated time window waveform in linear and logarithm coordinates as illustrated in Fig. 3(a) and (b), the pulses-width get wider as VDC1 increases from 0.3Vpi to 1.5Vpi. the insertion loss get lower and the extinction ratio increases as VDC1 changes from 0.3Vpi to Vpi, then the insertion loss get larger and the ER decreases as VDC1 changes from Vpi to 1.5.We can also observe that, too small Child-MZM1 DC bias voltage lead to pedestal pulses with lower ER, and too large bias voltage cause the pulses depress at top. Therefore, with given RF clock power, the pulse-width tunable range is limited to guarantee the quality of generated time windows. The frequency chirp of the time window is determined by chirp of the DPMZM. For chirp free DPMZM the frequency chirp of the generated time window can be neglected as shown by the dotted lines in Fig. 3(a). A chirp free time window would be a better choice for RZ pulse carver or OTDM de-multiplexer, it can also be used as shor pulse source by following pulse compression stages.

 figure: Fig. 3

Fig. 3 Simulated time window waveform and frequency chirp in linear (a) and logarithm (b) coordinates.

Download Full Size | PDF

 figure: Fig. 4

Fig. 4 Experimental measured FWHM pulse width (a), Insertion loss and ER (b) and optical spectra (c~f) of the DPMZM generated time window versus VDC1.

Download Full Size | PDF

Figure 4(a) and Fig. 4(b) give the experimental measured FWHM pulse width, Insertion loss, and ER of the DPMZM generated time window versus VDC1. When VDC1 is adjusted from 0.2Vpi to 1.4Vpi, the pulse width ranges from 5.3 ps to 12 ps, which agree well with the simulated data as depicted by triangle-line in Fig. 4(a). When VDC1 is adjusted from 0.2 Vpi to 1.4 Vpi, the insertion loss decreases from 13.8 dB to 4.6 dB, then increases to 5.6 dB, the lower IL appears near the bias of Vpi. When VDC1 becomes larger, the ER increases firstly and then decrease. The measurement error of ER is mainly caused by the ASE noise from EDFA, which makes the OSO cannot measure an ER higher than 20 dB accurately.

Insets in Fig. 4(a) depict the measured waveform of DPMZM generated optical time window using a 500 GHz OSO by setting VDC1 at 0.26Vpi, 0.35Vpi, 0.86Vpi, and 1.32Vpi respectively. Shorter time windows show higher peak level due to the saturated amplification of EDFA, pedestals are observed in the case of VDC1 = 0.26Vpi as shown in Fig. 4(a); with the increasing of VDC1, the pedestals become small and can be ignored. As shown in the optical spectra of Fig. 4(c ~f), OSNR of 52 dB, 56 dB, 58 dB and 61 dB for the four cases are measured by using an OSA with resolution of 0.02 nm. Since higher VDC1 lead to lower insertion loss, higher OSNR can be achieved after EDFA amplification.

To choose a proper RF clock power to realize tunable pulses with large tunable range, low insertion loss and high ER, we investigate the time window characteristics dependence on RF clock power as shown in Fig. 5 and Fig. 6 . As is described in Fig. 3, if the RF clock power is fixed, there is a tradeoff between the minimum pulse-width and the ER of the generated time window. Figure 5 shows the minimum available FWHM pulse width and the corresponding insertion loss for different RF clock power in the case of ER<18 dB. It can be observed that, as RF clock power increasing, the available minimum pulse-width becomes larger and the corresponding IL becomes lower. The experimental results agree well with the simulation results. The insets in Fig. 5 give experimentally measured waveforms with minimum pulse-width for different RF clock power. The pulse pedestals are very small and can be ignored. OSNR of 50 dB, 56 dB and 60 dB for VRF of 1.15Vpi, 1.45Vpi and 1.8Vpi are measured as shown in the optical spectra of Fig. 5(b),(c) and (d).

 figure: Fig. 5

Fig. 5 Minimum available pulse-width and the corresponding insertion loss (a) and the optical spectras (b~d) for different RF clock power.

Download Full Size | PDF

 figure: Fig. 6

Fig. 6 Maximum available pulse-width and the corresponding insertion loss (a) and the optical spectras (b~d) for different RF clock power.

Download Full Size | PDF

Figure 6 shows the maximum available FWHM pulse-width and the corresponding IL of generated pulses without depressed top for different RF clock power. It can be observed that, as RF clock power becomes larger, the available maximum pulse-width decreasing and the corresponding IL become lower. The experimental results agree well with the simulation results. The insets in Fig. 6 give experimentally measured optical pulse waveforms with maximum pulse-width and there optical spectras for different RF clock power. OSNR of 60 dB, 61.5 dB and 64.5 dB for VRF of 1.15Vpi, 1.45Vpi and 1.8Vpi are measured as shown in Fig. 6(b~d).

Therefore, to choice proper RF clock power, we have to make a tradeoff between the insertion loss and the pulse-width tunable range. Lower RF clock power can be used for the applications need wider pulse-width tunable range without regard to the insertion loss; higher RF clock power can be used for the applications need lower insertion loss irrespective of tunable range. Table 1 compares the experimentally measured Characteristics of tunable optical pulses for different RF clock power. For VRF = 1.15Vpi, the pulse-width can be adjusted from 5.6 ps to 12.6 ps with OSNR better than 50 dB; For VRF = 1.8Vpi, the pulse-width can be adjusted from 6.3 ps to 11.4 ps with OSNR better than 60 dB. For all the cases the measured timing jitter is lower than 50 fs. It should be point out that the stability of the time switch is mainly affected by three biasing voltage and one RF drive voltage, DPMZM bias control circuit and power control circuit for RF driver are required for production quality applications

Tables Icon

Table 1. Characteristics of Tunable Optical Pulses for Different RF Clock Voltage

4. Conclusions

In this paper, we generate and investigate a pulse width-tunable, phase-preserved and stable 40GHz time switch by using a dual parallel Mach-Zehnder modulator driven simply by an electrical clock. The time window performance with respect to the DC bias voltage and the RF clock power is researched experimentally. Experimental results show the pulse-width can be tuned continuously from 5.6 ps to 12.6 ps by simply adjusting the DC bias voltage of DPMZM; the timing jitter, extinction ratio, OSNR and insertion loss are measured to be better than 50 fs, 18 dB, 50 dB and 16 dB respectively. The proposed pulse-width time switch can be used in QPSK modulated OTDM system as de-multiplexer and pulse source assisted by pulse compression.

Acknowledgment

This work was partly supported by “NSFC” program 61001121, 60932004, 61006041, “863” program 2012AA011303, “973” program 2011CB301702 and the Fundamental Research Funds for the Central Universities.

References and links

1. C. Yu, L. S. Yan, T. Luo, Y. Wang, Z. Pan, and A. E. Willner, “Width-tunable optical RZ pulse train generation based on four-wave mixing in highly nonlinear fiber,” IEEE Photon. Technol. Lett. 17(3), 636–638 (2005). [CrossRef]  

2. Y. Ji, Y. Li, J. Wu, F. Z. Zhang, K. Xu, W. Li, X. B. Hong, and J. T. Lin, “A Phase Stable Short Pulses Generator Using an EAM and Phase Modulators for Application in 160-GBaud DQPSK Systems,” IEEE Photon. Technol. Lett. 24(1), 64–66 (2012). [CrossRef]  

3. Y. Ji, Y. Li, J. Wu, B. Luo, Y. T. Dai, K. Xu, W. Li, X. B. Hong, and J. T. Lin, “160-GBaud DQPSK Optical Time-Division Demultiplexing Using a Polarization Modulator,” IEEE Photon. Technol. Lett. 24(9), 772–774 (2012). [CrossRef]  

4. A. O. J. Wiberg, C. S. Bres, B. P. P. Kuo, E. Myslivets, and S. Radic, “Cavity-less 40 GHz pulse source tunable over 95 nm,” in European Conference and Exposition on Optical Communications, OSA Technical Digest (CD) (Optical Society of America, 2009), paper 5.2.3.

5. M. D. Pelusi, “160-Gb/s Optical Time-Division Demultiplexing Using a Mach-Zehnder Modulator in a Fiber Loop,” IEEE Photon. Technol. Lett. 20(12), 1060–1062 (2008). [CrossRef]  

6. T. Sakamoto, T. Kawanishi, and M. Tsuchiya, “10 GHz, 2.4 ps pulse generation using a single-stage dual-drive Mach-Zehnder modulator,” Opt. Lett. 33(8), 890–892 (2008). [CrossRef]   [PubMed]  

7. Y. Li, J. Wu, Y. Ji, D. M. Kong, K. Xu, W. Li, X. B. Hong, H. X. Guo, Y. T. Dai, Y. Zuo, and J. T. Lin, “160Gbaud/s to 40Gbaud/s OT DM-DQPSK De-multiplex Based on a Dual Parallel Mach-Zehnder Modulator ”, in Optical Fiber Communication Conference, OSA Technical Digest (Optical Society of America, 2012), paper OWI1.2.

8. S. M. R. M. Nezam and A. B. Sahin, J. E. MeGeehan, T. Luo, Q. Yu, and A.E. Willner, “Enhanced Robustness of RZ WDM Systems Using Tunable Pulse-Width Management at the Transmitter” in European Conference and Exposition on Optical Communications, OSA Technical Digest (CD) (Optical Society of America, 2002), pape 10.6.2.

9. C. Yu, L.-S. Yan, T. Luo, Y. Wang, Z. Pan, and A. E. Willner, “Width-Tunable Optical Pulse Generation Based on Four-Wave Mixing in Highly-Nonlinear Fiber”, Conference on Lasers and Electro-Optics, OSA Technical Digest (CD) (Optical Society of America, 2002), paper CTuN4.

10. A. Zhang, H. L. Liu, M. S. Demokan, and H. Y. Tam, “Width and Wavelength-Tunable Optical Pulse Train Generation Based on Four-Wave Mixing in Highly Nonlinear Photonic Crystal Fiber,” IEEE Photon. Technol. Lett. 17(12), 2664–2666 (2005). [CrossRef]  

11. M. A. Ummy, N. Madamopoulos, and R. Dorsinville, “SOA-based loop mirror for tunable pulse-width generation,” Opt. Commun. 282(13), 2608–2614 (2009). [CrossRef]  

12. Y. Dong, Z. H. Li, J. Mo, Y. Wang, C. Lu, and T. H. Cheng, “Pulsewidth-Tunable CS-RZ Signal Format With Better Tolerance to Dispersion and Nonlinear Degradation in Optical Transmission System,” IEEE Photon. Technol. Lett. 16(5), 1409–1411 (2004). [CrossRef]  

13. Q. Nguyen-The, H. N. Tan, M. Matsuura, and N. Kishi, “Generation of multi-wavelength picosecond pulses with tunable pulsewidth and channel spacing using a Raman amplification-based adiabatic soliton compressor,” Opt. Express 20(2), 1230–1236 (2012). [CrossRef]   [PubMed]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (6)

Fig. 1
Fig. 1 Schematic of DPMZM based short time window generation.
Fig. 2
Fig. 2 Concept diagram of the DPMZM based optical switch. (a) Modulation curve of MZM; (b ~d) pulse waveform and phase output from MZM1and MZM2; (e ~g) time window waveform output from DPMZM
Fig. 3
Fig. 3 Simulated time window waveform and frequency chirp in linear (a) and logarithm (b) coordinates.
Fig. 4
Fig. 4 Experimental measured FWHM pulse width (a), Insertion loss and ER (b) and optical spectra (c~f) of the DPMZM generated time window versus VDC1.
Fig. 5
Fig. 5 Minimum available pulse-width and the corresponding insertion loss (a) and the optical spectras (b~d) for different RF clock power.
Fig. 6
Fig. 6 Maximum available pulse-width and the corresponding insertion loss (a) and the optical spectras (b~d) for different RF clock power.

Tables (1)

Tables Icon

Table 1 Characteristics of Tunable Optical Pulses for Different RF Clock Voltage

Equations (3)

Equations on this page are rendered with MathJax. Learn more.

E a (t)= E in (t) 2 cos[ π 2 V pi ( V RF1 (t)+ V DC1 )] e j π V DC1 2 V pi E b (t)= E in (t) 2 cos[ π 2 V pi ( V RF2 (t)+ V DC2 )] e j π V DC2 2 V pi
E out (t)= E a (t)+ e j π V DC3 V PI E b (t)
E out (t)= E in (t) 2 {[cos π 2 V pi ( V 0 sin(2π f 0 t)+ V DC1 )]+ e j 2π V DC3 +π V DC2 π V DC1 2 V PI cos π V DC2 2 V pi } e j π V DC1 2 V pi
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.