Philippe Marchegay

Non coherent spectral analysis of ADC using filter bank

The spectral analysis of ADC digital data has traditionally been done with discrete Fourier transform. This method imposes restrictions to optimize the results: one of these is coherent sampling. Recently, some filter structures have been used for the spectral analysis of sinusoidal signal corrupted by harmonics and noise. In this paper, we present a filter bank structure used for decomposition of a signal into his main spectral components. The main application examined is ADC spectral parameters extraction (SINAD, SNR, THD) with non coherent sampling. Computer simulations are used to demonstrate the performance of the proposed filter structure.

Signal generation using single bit sigma delta techniques

In this paper, we present a digital oscillator, based on the single-bit /spl Sigma//spl Delta/ modulation principle. With the exception of a front end analog filter, the proposed circuit is entirely digital. The oscillator is achieved by the use of a /spl Sigma//spl Delta/ modulator in the digital resonator loop. A novel topology for constructing stable high-order 1-bit modulators is described. A fourth-order low-pass and eighth-order bandpass oscillators are implemented in field-programmable gate array prototype. The experimental results demonstrate stability and feasibility, achieving more than 110-dB dynamic range for the generated signals.

Hybrid Newton/Gauss-Newton algorithm for time-domain analysis of A/D converters

This brief describes a maximum-likelihood estimator for a sum of harmonically related sinewaves for characterizing analog-to-digital (A/D) converters. The frequency is estimated by means of the Newton method. For the iterations where singularity problems occur, the algorithm switches to the Gauss-Newton method combined with the singular value decomposition. The initial frequency accuracy required by this algorithm is determined and the algorithm is modified in order to converge to the global minimum, whatever the initial frequency uncertainty. Thus, one can use the frequencies of the signal generators to provide the starting value. Once the frequency is determined, the linear parameters of the model are estimated by solving a linear set of equations. It is shown through simulation and experimental results that the proposed algorithm is fast, accurate, and robust, with respect to the initial guess of the frequency.

Optimization and Implementation on Fpga of the DCT/IDCT Algorithm

In this paper, we present a comparison between two methods, the modified Loeffler algorithm (11 MUL and 29 ADD) and distributed arithmetic, to implement the DCT/IDCT algorithm for MPEG or H.26x video compression using VHDL description language. The implementation has been achieved on Altera Stratix EP1S10 FPGA which provides a dedicated DSP blocks required for common signal processing functions. A new solution based on this DSP blocks used to implement multipliers for the modified Loeffler algorithm in order to optimize speed and area

A 3.3 V switched-current second order sigma-delta modulator for audio applications

This paper presents a low pass sigma-delta modulator (/spl Sigma//spl Delta/) for Analog-to-Digital conversion which uses the switched-current technique. The growth of this technique is limited by the clock feedthrough which induces an error on the output signal value. In order to cancel the effect of this phenomena a new method based on a current feedback loop is used. The /spl Sigma//spl Delta/ modulator implemented is a conventional second order and the whole circuit is designed with fully differential current mode building blocks. The circuit, designed in 0.8 /spl mu/m AMS single poly CMOS process, shows 55 dB dynamic range at 2 MHz sampling rate with 8 kHz input frequency bandwidth.

The initial frequency uncertainty tolerated by the Levenberg-Marquardt method applied to dynamic testing of A/D converters

This paper examines in some detail the convergence of the Levenberg-Marquardt method applied to A/D converter testing with sinusoidal stimulus. Firstly, we have determined the required initial guess accuracy in order to obtain the global convergence. Then, this result is used in order to develop a globally convergent modified Levenberg-Marquardt algorithm whatever the initial guess uncertainty. Simulation and experimental results are presented in order to show the effectiveness of this algorithm.

Dynamic testing of A/D converters: how many samples for a given precision?

The statistical analysis used for an ADC performance evaluation necessitates a finite number of samples N. The subject of our work is to optimize this value for a precision /spl beta/ with a given confidence interval /spl psi/. Two different approaches were treated in the literature (deterministic or probabilistic). We consider in our case an intermediate approach based on the following concept: the jitter introduced by the ADC and the test bench, /spl sigma/, influences the distribution of sampling. Consequently, the evaluation of N has to be reconsidered and will depend on the ratio defined by /spl sigma//T/sub re/ where T/sub re/ is the equivalent sampling period for a reconstructed signal. We also defined a generalized relation giving N as function of this ratio whatever the periodic input signal wave form. Simulations corroborate this study.

Jitter measurement of an ADC by statistical analysis

Abstract A method for measuring the jitter of an ADC based on the ‘lcked histogram’ test is presented, and the conditions for validation of the test as a function of the sampling position is studied. As all test bench instruments have their own independent jitter a mathematical model is used to extract the jitter of the ADC. Some experimental results are given.

[ADC Characterization in Time Domain] Frequency Estimation to Linearize Time-Domain Analysis of A/D Converters

Time-domain analysis is a powerful tool for testing A/D converters (ADCs), particularly by incoherent sampling. This sampling is encountered for instance in real application and systems where the clock is generated by an on-board crystal oscillator. In the case of the sine-wave model for ADC output samples, fitting a sine-wave function to the data record is nonlinear with respect to the frequency. To overcome the convergence problem, the authors propose some alternative methods based on frequency estimation by means of fast Fourier transform and eigenanalysis of the autocorrelation matrix. This leads to a simple linear problem. The effectiveness of these methods is proven on both simulation and experimental results

Non Coherent Spectral Analysis of ADC using FFT Windows: an Alternative Approach

This article presents an alternative method for testing Analog-to-Digital Converters (ADC) by non-coherent spectral analysis using weighting windows and focusing on the maxima of the window spectrum instead of summing the power inside the main lobe of the window spectrum, as it is always done. Comparing to the existing method, theoretical laws for computing fundamental, harmonics and noise power spectral densities are derived and software simulations are used to validate them.