Detection of Chaos in the Noise of Electronic Oscillators by Time Series Analysis Methods
Frequency fluctuations of an electronic oscillator are studied by time series analysis methods in order to detect an eventual underlying attractor. Measurements are performed by a period counting technique with a reference signal of 10 MHz and various frequencies of the local oscillator. Three different behaviours of the Allan variance are observed, depending on the mean frequency of the beat signal. The results of time series analyses clearly show that two of these behaviours are associated with the presence of a chaotic process, whereas the third is more intricate. When applied to data computed from continued fraction expansions of real numbers, these methods lead to the same conclusions. For two kinds of truncation, the data present a chaotic behaviour, which is not obvious for the third kind. A correspondence between the behaviour of the frequency fluctuations and the way of truncating the continued fraction expansions can thus be proposed.
KeywordsCorrelation Dimension Beat Frequency Local Slope Continue Fraction Expansion Allan Variance
Unable to display preview. Download preview PDF.
- 2.Perera A.G.U., Matsik S.G. (1995) Chaotic to Periodic Spontaneous Pulsing in Current Driven Silicon p-i-n Structures. Physica D84:615Google Scholar
- 16.see the companion paper by Planat M.Google Scholar
- 17.Planat M., Dos Santos S., Ratier N., Cresson J., Perrine S. (1999) Close to Resonance Interaction of Radiofrequency Waves in a Schottky Diode Mixer: 1/f Noise and Number Theory. In: Handel P., Chung A. (Eds.) Quantum 1/f Noise and other Low Frequency Fluctuations in Electronic Devices. AIP Proceedings 466:177Google Scholar
- 24.Eckert C., Miehé J.A. (1996) Test of the Deterministic Conjecture for the Instantaneous Time Phase Fluctuations of an Oscillator. Ann Telecommun 51:351Google Scholar