Research On The Fractional Analyzing Method And Software Of Noise In Electronic Device

With the enlargement of the scale of the integrated circuit, the devices of the semiconductor have already reached deep sub-micrometer level. Because the noise of the device is inversely proportional to the area of the device, so the noise is concerned by people extremely, and become the subject for research which can't be ignored in miniaturization. Among them inside noise (especially the noise of low frequency represented by 1/f noise), it is a key index of restricting the sensitivity of the device and measuring the precision, signify the quality of the device and an important sensitive parameter at the same time. So it receives to be quite extensive research and application already in the world. Electronical noise is the most worth studying content.Because electronical noises are not entire stochastic. They exhibit complex behavior characterized by long-range power-law correlations. Traditional approaches such as the power-spectrum are used to quantify long-range correlations in these signals. The method primary focused on extract of FFT spectral characteristic parameter (mainly some power spectrums andγindexes ).It is a sign of the whole singularity to the noise signal, but has neglected some local characteristics. And power-spectrum are not suited to accurately quantify long-range correlations in non-stationary signals. It required a new analysis method for electronical noises. Fractal analysis is one of the most promising new approaches for extracting hidden information from electronical noise.Fractal analysis primary deal with nonregularity object. Electronical noise is nonregularity and singularity at everywhere, so fractal analysis method is suitable to used for noises. In this paper, we emphases on some typical method such as fractional dimension estimated, detrended fluctuation analysis , singularity spectrum and local h(?)lder exponent determined . We will realized their algorithm program and verification, and applied those method for electronical noise and emulational noise.Apply Boxdimension analysis, we found noise's fractional dimension can be viewed as an indicator of the "roughness" of the original time series. Fractional dimension is a parameter absolute differ fromγindex. Fractional dimension estimated from four kinds model noise with identicalγindex are different. Application of DFA method in AL interconnection electromigration signal shows evidences for a crossover pjenomenon associated with changes in short and long-range correlation exponent .This method may be of use in distinguishing devices condition based on differences in the correlation property. And we attempt to measure the singularity spectrum changes versus time during the accelerated lifetime tests. The result shows h(?)lder exponent corresponding to the summit are valid in characterizing elecromigration process and predicting interconnection. And through the local h(?)lder exponent measure from optocoupler'signal., we find the mean h(?)lder exponent nearly to zero can be proved burst noise in the signal.Fractional analysis method is so effective that can be further used in the research and qualification of noise and model. And it provide another simply and conveniently means for screening too. There are better application prospects .