The Research For Microwave Signal Optical Analog-to-digital Conversion

The optical analog-to-digital conversion is the most promising solutions for the digital signal-processing (DSP) systems to digitize high-speed analogue signal and to overcome the bottlenecks caused by limited sampling and switching speeds in electronic signal processing circuits. Although there has been more than thirty years for the investigation of optical analog-to-digital converter (ADC) in the world, and many kinds of optical ADCs have been proposed, there has no significant breakthrough either in the functional unit, such as sampling process and quantizing process, or in the whole optical ADC operation system. It is still a challenge for a high sampling rate and high resolution quantization ADC. Therefore, the optical analog-to-digital conversion for microwave signal has been investigated in this dissertation. After the introduction of R&D of optical ADC at home and abroad, a novel optical analog-to-digital conversion for high-speed signal was investigated and built up, based on optical phase sampling, optical polarization sampling and optical wavelength sampling, respectively.Firstly, using the basic concepts of optical ADC, a function relationship of sampling rate and quantization resolution induced by noise has been obtained. Taking the quantization noise as criterion and assuming the other optical components to be ideal, the effective numbers of bit (ENOB) has been lucubrated, which affected by the CNR of an analogue optical system, the amplitude fluctuation of sampling light pulses, the average light power of light pulses, the timing jitter and the finite pulse width, respectively. The results have shown that the CNR of analogue optical system, the amplitude fluctuation and average power of sampling light pulses only decay the ENOB of optical ADC, moreover, the timing jitter and the finite pulse width decline the ENOB of optical ADC and the sampling rate. The state-of-the-art mode-locked laser sources will limit optical ADCs to less than 10 bits of ENOB at the sampling rate of 10 GS/s.Secondly, the principle of the Taylor's phase-encoded optical sampling ADC was studied and found a theoretical limit for the sampling rate and the ENOB of optical ADC by the transit time of light pulse. Using the theory limit, the best performance of Taylor's phase-encoded optical sampling ADC was limit to around 4 bits of ENOB at the sampling rate of 1 GS/s. An improved phase-encoded optical sampling method is presented by us to avoid the shortcoming of Taylor's, which can be realized through doubling the product of both the modulating voltage and the length of electrode of M-Z modulator. The improved method could increase about 2 bits of the ENOB than Taylor's method at the same sampling rate. To overcome the difficult in experiment resulted by the highly half-wave voltage of the existing LiNbO3 M-Z modulators, a biased M-Z modulators method was utilized in the experiment of a phase-encoded optical sampling ADC, and resulted in 2 GS/s @2 bits.Thirdly, a novel high-speed polarization sampling method using nonlinear polarization rotation (NPR) in an SOA is proposed in this dissertation. Using the carrier rate equation in an SOA for the two polarization orients, a model is proposed to describe the relationship between the polarization rotating angles of sampling light pulse and the analogue signal light power. Utilizing the model, an optimized transfer curve for polarization sampling has been obtained through adjusting the initial polarization angle of the sampling light pulse, as well as the injected current of the SOA. Meantime, the high-speed polarization sampling process was also simulated. The theory analyses results indicate that the polarization sampling has promoting potential to improve the sampling rate at hundreds GS/s, but the quantization resolution is only few bits.Fourthly, since the light wavelength is almost immune from the noise and disturbance in propagation, a wavelength sampling and wavelength quantizing (WSWQ) method is also proposed in this dissertation for a high-speed optical ADC. The wavelength sampling can be realized by using a narrow pass-band EO high-speed tunable optical filter (TOF), and the wavelength quantizing function is realized by using an arrayed-waveguide grating (AWG). Taking the quantization noise as criterion, the performance of our WSWQ optical ADC is analyzed in theory and four results are obtained as follows: a) The sampling rate is equal to the light pulse repetition rate and the quantization resolution is limited by the jitter of light pulse, as well as the pulse width; b) The WSWQ optical ADC is immune to the intensity noise of light pulse and the CNR of analogue optical system; c) The state-of-the-art mode-locked laser sources will limit the WSWQ optical ADCs to less than 10 bits of ENOB at the sampling rate of 10 GS/s. d) Owing to the nonideal performance of the practical TOF, the quantization resolution of our WSWQ optical ADC is mostly limited by the narrow wavelength tuning range of the TOF.The high-speed TOF is the key component in our WSWQ optical ADC scheme, which tuning speed is crucial for wavelength sampling rate, and which tuning rang and narrow pass-band is crucial for quantization resolution. Therefore, we firstly design a narrow pass-band EO TOF in this dissertation, which consists of two paralleled and identical waveguides with long period waveguide gratings (LPWGs) thereon. Using the coupled mode theory, the coupling equations of the two symmetry waveguides with LPWGs have been derived and solved by Laplace Transform & Inverse Laplace Transform method. The solutions give out the output light power and the coupling coefficient of the two symmetry waveguides. The filtering principle of our filter is clarified by using the phase-matching condition of the LPWGs.Using electro-optical polymer material to fabricate its waveguide cores, the EO TOF is designed and simulated in this dissertation. Compared with the conventional TOF, the EO TOF of ours can achieve fast tuning speed in nanosecond order with widely tuning range and narrow pass-band, and can output a single band-passed result and band-rejected result, simultaneously. A model of the EO TOF is presented to analysis the output wavelength, the 3-dB pass-band width and the coupling efficiency, which is affected by the period of LPWGs, the whole LPWGs length, the gratings etching depth and the coupling coefficient, respectively. As a result, the weakly coupling coefficient method to compress the pass-band is presented. The simulation results are shown that the TOF can realize continuously high-speed linearity tuning in wavelength across 30 nm range with a narrow pass band of <0.8nm.For accurate calculation of the value of the coupling coefficient, the full-vector finite difference mode solver proposed by A. B. Fallahkhair is used in this dissertation to analysis and to compute the modes located in the cladding complex waveguide, which can not be obtained by the existing waveguide analytic softwares.