Integrated 18th-order pseudo-N-path filter in VIS-SC technique

An integrated 18th-order frequency symmetrical switched-capacitor (SC) bandpass filter is described that was realized by the use of voltage inverter switches (VISs) in the form of recharging devices. The filter was implemented in a standard CMOS technology. It was conceived for use as a `channel filter' in modulation schemes for PCM/FDM and audio/FDM conversion. Starting from a high-pass filter as reference filter, the SC filter is designed as a pseudo-N-path filter, with N=2. The application of the pseudo-N-path principle leads to a substantial saving in hardware.     

Narrow-bandpass filtering with modulation

To overcome the problems of carrier leak etc. experienced with N-path filters incorporating active lowpass filters, it has been suggested that N-path filters incorporating bandpass filters (which must be well matched) be used to produce a pair of passbands away from the carrier frequency, with another filter at the required final frequency to discriminate between the wanted and the image passbands. An alternative and practical solution is described here in which neither the well matched low-frequency filters nor the additional high-frequency filter are required; instead a single low-frequency bandpass filter and two pairs of elementary phase-shift networks are used. The circuit is well suited to realisation as a thin-film or hybrid integrated circuit, since all the frequency-selective circuits operate at the low frequency.     

An integrated CMOS switched-capacitor bandpass filter based on N-path and frequency-sampling principles

A new combined switched-capacitor (SC) frequency-sampling N-path filter is presented, which allows the implementation of very narrow bandpass filters. The included frequency-sampling (FS) filter suppresses undesirable passbands of the SC N-path filter. The center frequency f/SUB m/ of the bandpass filter is identical to the circuit clock frequency f/SUB c/. Experimental results are presented for a CMOS SC frequency-sampling four-path filter with second-order filter cells, a center frequency of 1 kHz, and -3-dB passband bandwidth of 11.5 Hz.     

Alternative approach to the design of switched-capacitor filters

An alternative switched-capacitor filter design is shown to be possible using an economic form of c.c.d. A simple example of a biquadratic digital-filter section is included to illustrate the point. This leads to comment on the relationship between switched-capacitor filters, the N-path filter and the digital filter.     

具备谐波抑制的高阶有源N路径带通滤波器

采用三级2阶N路径滤波单元设计了一种带谐波抑制功能的高阶有源N路径带通滤波器。在第二、三级之间插入负电阻和回转器,可提高滤波器的Q值、带宽和线性度;在末级的串联型2阶N路径滤波单元中采用有相位差的时钟信号进行控制,可有效地抑制三次谐波。基于0.18μm CMOS工艺仿真。结果表明,该滤波器的最高增益为20.12 dB,中心频率调谐范围为0.1~1 GHz,带外抑制高达50.2 dB@700 MHz,三次谐波抑制大于40 dB,噪声系数为4.71~6.9 dB,带外输入3阶交调点(IIP3)为16.3 dBm@50 MHz。     

Delayed N-path structure for high speed adaptive FIR digital filtering

A structure and its algorithm for high speed adaptive FIR digital filtering based on the delayed N-path concept is presented. Using N/sup 2/ processors, the throughput rate of the proposed filter can be N/sup 2/ times faster than that of the same filter realised directly using one such processor.<>     

Pseudo-N-path switched capacitor filters with out-of-band noise peaks

Novel pseudo-N-path (PNP) switched-capacitor (SC) integrator structures which avoid noise peaks in the main passband are described. They combine the simplicity of the previously proposed RAM-type PNP SC circuits with the desirable properties of hybrid-type circuits. Simulation results are presented for filters employing these structures to verify their usefulness.<>     

Finite-gain-insensitive circulating-delay type pseudo-N-path filter

A novel circulating-delay switched-capacitor (SC) pseudo-N-path (PNP) filter is described which has low sensitivity to finite amplifier gain. The circuit exploits the narrowband characteristic of the PNP filter to predict its output voltage a number a cycles in advance and hence give accurate error compensation.<>     

Improved systolic allpass digital filters for very high-speed applications

An improved systolic realisation of an arbitrary-order allpass digital filter for delayed N-path digital filtering is presented. Using this method, the sampling rate at the input and output of a delayed N-path digital filter can be reduced to (T/sub m/+2T/sub a/)/(N(N-1)) (where N>or=2, and T/sub m/ and T/sub a/, respectively, represent the times for 2-input real multiplication and 2-input real addition), which is attractive for very high-speed applications.<>     

Variable-bandwidth filters using switching

It is shown that, if a switching technique is used to scale the bandwidth of a lowpass filter, the resultant circuit has some of the characteristics of an N-path filter. The consequence of this is that the signal frequency must be bandlimited to avoid spurious responses at the switching frequency and its harmonics.     

2-phase circulating shift register for switched-capacitor N-path filters

A novel circulating shift register for switched-capacitor N-path filters is presented. The configuration affords a means of increasing the allowable discharge time for the shift register capacitors and eliminates clock feedthrough at the signal frequency. The shift register employs a simple 2-phase clock signal for any number of paths.     

New form of delayed N-path recursive digital filters

A new form of delay N-path recursive digital filter with transfer function expressed as a summation of N-delayed parallel recursive subfilters (each with a distinct denominator) is presented. This new filter is suitable for efficient systolic realisation and does not contain uncontrollable don't care bands. Use of the proposed filter for system modeling is demonstrated.<>     

Passive N-path-filter realisation

Through the intermediary of state-variable equations, an equivalence is obtained which shows how passive RC circuits incorporating time-variable gyrators can be used to realise N-path filters. The effects of gyrator losses in the realisation of time-invariant transfer functions are briefly discussed to indicate the practical significance of the result.     

Combined switched-capacitor FIR N-path filter using only grounded capacitors

A simple technique is presented that allows the unwanted sidebands of an N-path filter to be suppressed. The circuit realising this technique uses only grounded capacitors.     

High-Qfactor shunt-switched N-path filters

At its fundamental resonant frequency, a shunt-switched N-path filter is similar to an LCR resonant circuit, except that the selectivity or Q factor achievable, which is a direct function of the passive components employed, may be very much better than normally available in tuned circuits. The limit on the maximum attainable Q factor is shown to depend on the impedance of the shunt arm switches and filter load, as well as the speed of the switches.     

Some comments on notch capability of N-path filters

The notch capability of the N-path filter is investigated experimentally for the simple bipolar-switching case. Verification of the predicted notch for multilevel switching is also presented.     

Method of reducing the f.m. threshold using a tracking N-path filter

A self-synchronised N-path filter is used to track the instantaneous incoming f.m. carrier. As a result of the narrow bandwidth, a marked reduction of the f.m. threshold, for all f.m. modulation indices, is achieved. The system is described briefly.     

Globally convergent adaptive IIR filters based on fixed polelocations

A new class of adaptive filters, dubbed fixed pole adaptive filters (FPAF's), is introduced. These adaptive filters have infinite impulse responses, yet their adaptation exhibits provable global convergence. Good filter performance with a relatively small number of adapted parameters is permitted by the new filter structure, thus reducing the computational burden needed to implement adaptive filters. The implementation and computational complexity of the FPAF is described, and its modeling capabilities are determined. Excitation conditions on the filter input are established that guarantee global convergence of a standard set of adaptive algorithms. Some methods are described for selecting the fixed pole locations based on a priori information regarding the operating environment of the adaptive filter. The FPAF is tailored to applications by such a procedure, enabling improved performance. In examples, the FPAF is shown to achieve a smaller minimum mean square error, given an equal number of adapted parameters, in comparison with adaptive FIR filters and adaptive filters based on Laguerre and Kautz models     

Gaussian particle filtering

Sequential Bayesian estimation for nonlinear dynamic state-space models involves recursive estimation of filtering and predictive distributions of unobserved time varying signals based on noisy observations. This paper introduces a new filter called the Gaussian particle filter. It is based on the particle filtering concept, and it approximates the posterior distributions by single Gaussians, similar to Gaussian filters like the extended Kalman filter and its variants. It is shown that under the Gaussianity assumption, the Gaussian particle filter is asymptotically optimal in the number of particles and, hence, has much-improved performance and versatility over other Gaussian filters, especially when nontrivial nonlinearities are present. Simulation results are presented to demonstrate the versatility and improved performance of the Gaussian particle filter over conventional Gaussian filters and the lower complexity than known particle filters.     

A 28-MHz wideband switched-capacitor bandpass filter with transmission zeros for high attenuation

A 28-MHz wideband switched-capacitor (SC) bandpass filter employs an N-path technique and implements transmission zeros in the transfer function to achieve high attenuation. A modified SC biquadratic filter architecture is proposed to achieve high-speed operation. Implemented in a 0.35-/spl mu/m CMOS process, the bandpass filter operates at 28-MHz center frequency with a 3.84-MHz bandwidth and adjacent-channel attenuation of more than 35 dB. At 3-V supply, the filter measures a dynamic range of 37 dB at the 1% THD3 point while dissipating 19.6-mW per pole and occupying a chip area of 1.65 mm/sup 2/.