Search

Digital signal processing (DSP) with both linear equalization and nonlinear compensations are studied in this paper for the single-carrier 400G system based on 65-GBaud 16-quadrature amplitude modulation (QAM) signals. The 16-QAM signals are generated and pre-processed with pre-equalization (Pre-EQ) and Look-up-Table (LUT) based pre-distortion (Pre-DT) at the transmitter (Tx)-side. The implementation principle of training-based equalization and pre-distortion are presented here in this paper with experimental studies. At the receiver (Rx)-side, fiber-nonlinearity compensation based on digital backward propagation (DBP) are also utilized to further improve the transmission performances. With joint LUT-based Pre-DT and DBP-based post-compensation to mitigate the opto-electronic components and fiber nonlinearity impairments, we demonstrate the unrepeatered transmission of 1.6Tb/s based on 4-lane 400G single-carrier PDM-16QAM over 205-km SSMF without distributed amplifier. [ABSTRACT FROM AUTHOR]

This paper presents a sub-2-dB Noise Figure (NF) wideband differential Common Gate (CG) Low Noise Amplifier (LNA) in 0.18 µm CMOS technology. The circuit benefits from several new techniques to improve the overall performance. A new current bleeding scheme is proposed in which not only enhances the gain of the circuit but also improves transconductance and noise performance simultaneously. In addition, double positive feedback is employed with only an auxiliary NMOS transistor to break the trade-off between noise and input matching condition of the CG structure. Moreover, the added transistor is biased in moderate inversion region to form pre-distortion performance and sink third order nonlinearity of main transistors at the input. Thus, low noise figure is obtained simultaneous with good linearity. Post-layout simulation results in TSMC 0.18 µm RF-CMOS show a voltage gain of 20.1 dB with a −3 dB bandwidth of 0.38–2.7 GHz, while S 11 is below −10 dB over the frequency band. The minimum NF equals to 1.88 dB and an average NF of 2.04 dB is achieved over the frequency band. Third Input Intercept Point (IIP3) of the proposed circuit is +6 dBm. The circuit occupies an area of 0.055 mm 2 while consuming 2.55 mA from a 1.2-V supply voltage. [ABSTRACT FROM AUTHOR]

This paper presents an Ultra Wide-Band (UWB) high linear low noise amplifier. The linearity of Common Gate (CG) structure is improved based on pre-distortion technique. An auxiliary transistor is used at the input to sink the nonlinear terms of source current, resulting linearity improvement. Furthermore, an inductor is used in the gate of the main amplifying transistor, which efficiently improves gain, input matching and noise performance at higher frequencies. Detailed mathematical analysis show the effectiveness of both linearity improvement and bandwidth extension techniques. Post-layout simulation results of the proposed LNA in TSMC 0.18 µm RF-CMOS process show a gain of 13.7 dB with −3 dB bandwidth of 0.8–10.4 GHz and minimum noise figure (NF) of 3 dB. Input Third Intercept Point (IIP3) of 10.3–13 dBm is achieved which shows 8 dB improvement compared to conventional common gate structure. The core circuit occupies an area of 0.19 mm 2 including bond pads, while consuming 4 mA from a 1.8-V supply. [ABSTRACT FROM AUTHOR]

A novel technique to improve the linearity of Low Noise Amplifier (LNA) based on pre-distortion in the input matching circuit is proposed. The linearization technique is applied to a Capacitive Cross Coupled Common Gate (CCC-CG) structure. An auxiliary transistor operating in weak inversion region is used to adjust the input matching circuit nonlinearity. It has been shown that the nonlinearity of input matching circuit can efficiently compensate the nonlinearity of transconductance ( G m ) of the amplifier, and improves the third Input Intercept Point (IIP3). Furthermore, the added transistor is used in a feedback path which provides a noise canceling scheme. Thus, linearity improvement is achieved in addition to low Noise Figure (NF). Post-layout simulation results in TSMC 0.18 µm CMOS technology, show a gain of 18.55 dB with an upper −3 dB frequency of 2.96 GHz, the minimum NF equals to 2.63 dB and IIP3 of 11 dBm, while consuming 2.77 mA from a 1.5 V supply. [ABSTRACT FROM AUTHOR]

Abstract: In this paper a novel all-optical scheme is proposed to reduce fiber nonlinear distortions in high-speed dense wavelength division multiplexed (DWDM) optical fiber communication systems by applying time-domain fractional Fourier transform (FRFT) based on the time–space duality theory. The modulated optical pulses will be pre-distorted by an FRFT module before being launched into fiber links. It has been validated that the pre-distorted signal performs better fiber nonlinearity tolerance compared to that without pre-distortion. Simulation results show that the received bit error rate (BER) is improved by several orders of magnitude for a 40Gbit/s carrier-suppressed return-to-zero on-off keying (CSRZ-OOK) DWDM system, which can be used in a current 40Gbit/s system easily to improve the nonlinearity effect tolerance. [Copyright &y& Elsevier]

Abstract: An optical waveform pre-distortion method is proposed in reducing the temporal broadening of pulse based on the optical implementation of time domain fractional Fourier transform (FRFT). Moreover, a new analysis of the dispersion and self-phase modulation effects is investigated. The simulation results show that prechirp coefficient and proper fractional orders of FRFT can be beneficial in mitigating pulse broadening. [ABSTRACT FROM AUTHOR]

In this work, the nonlinear behavior on pulse-amplitude-modulation (PAM) and their impact on the performance of RGB light-emitting-diodes (LED) based visible light communications (VLC) is presented. Particularly, how the limited source bandwidth affects the nonlinearity of fully modulated RGB LEDs is investigated. It is found that when symbol-rate is higher than the channel bandwidth the nonlinearity gradually diminishes. It causes the symbol-rate dependent nonlinearity. When applying PAM in VLC links, a pre-distortion method is often applied to compensate for the nonlinearity. Due to this symbol-rate dependency, however, the conventional pre-distortion method has a limitation on linearizing the non-linear VLC channel, especially in the case of applying various symbol-rates according to the channel condition. Although it can be adaptively compensated by nonlinear equalizers, their complexity is considerably high. In this work, the characteristics of multi-level PAM scheme in a band-limited nonlinear channel is presented. Also, the feasibility of a low complex but efficient pre-distortion method derived from monotonically decreasing characteristics of the frequency dependent nonlinearity is investigated. [ABSTRACT FROM AUTHOR]