E. Rubiola

Correlation-based phase noise measurements

In the characterization of the phase noise of a component, it is common practice to measure the cross-spectrum density at the output of two phase detectors that simultaneously compare the component output signal to a common reference. This technique, which is based on correlation and averaging, allows the rejection of the phase detector noise. On the other hand, it is known that the interferometer exhibits lower noise floor and higher conversion gain than other phase detectors suitable to radio-frequency and microwave bands. Thus, we experimented on an improved instrument in which the phase noise of a component is measured by correlating and averaging the output of two interferometers. The measurement sensitivity, given in terms of noise floor, turns out to be limited by the temperature uniformity of the instrument, instead of the absolute temperature T. This feature makes the instrument suitable to investigate the spectrum Sφ(f) of phase fluctuations below kBT/Po, i.e., the thermal energy kBT referred to...

Flicker noise measurement of HF quartz resonators

Frequency flicker of quartz resonators can be derived from the measurement of S/sub /spl phi// (f), i.e., the power spectrum density of phase fluctuations /spl phi/. The interferometric method appears to be the best choice to measure the phase fluctuations of the quartz resonators because of its high sensitivity in the low power conditions, which is required for this type of resonator. Combining these two ideas, we built an instrument suitable to measure the frequency flicker floor of the quartz resonators, and we measured the stability of some 10-MHB high performance resonators as a function of the dissipated power. The stability limit of our instrument, described in terms of Allan deviation /spl sigma//sub y/(/spl tau/), is of some 10/sup -14/.

Very high frequency and microwave interferometric phase and amplitude noise measurements

The interferometric technique allows close-to-the-carrier measurements of both phase and amplitude noise, improving the instrument noise floor by 10–25 dB as compared to the traditional method based on a saturated mixer. Principles and basic equations describing the noise measurement system are given, together with design strategies suitable to microwave and very high frequency bands. Two prototypes, operating at 9 GHz and 100 MHz are discussed in detail. The relevant features of these prototypes are the capability to operate in a wide power range, below 0 dBm and above 20 dBm, and low noise floor. The latter is about −180u2002dBu200arad2/Hz (white) and 150u2002dBu200arad2/Hz (flicker) at 1 Hz Fourier frequency, at carrier power from 9 to 15 dBm.

Improved interferometric method to measure near-carrier AM and PM noise

An improved version of the interferometric method to measure near-carrier AM and PM noise is being presented. The main feature of this scheme is the capability to reduce the instrument noise by correlating the output of two equal interferometers built around the same device to be tested, thus enhancing the sensitivity. Two double interferometers are described, operating in the microwave and VHF bands. The latter shows a noise floor of -194 dB rad/sup 2//Hz when the signal power is +8 dBm. The sensitivity of both the instruments turns out to be significantly higher than the ratio of the thermal noise divided by the carrier power.

High stability current control in the 10 A range

This paper reports the design and the evaluation of an high stability current source developed for energizing the C-field magnet of the high field cesium standard experiment, currently in progress at Politecnico di Torino. This source, devoted to C-field uniformity measurements, supplies about 1 kW dc power with a current stability of about 10/sup -7/. >

Long-term behavior of operational amplifiers

The voltage and current offsets of two typical precision operational amplifiers (OPAs) with BJT and FET input, respectively, were continuously measured for two years. The paper presents the experiment, explains the method of data analysis, and discusses the results. The long-term stability turns out to be limited mainly by random walk processes.

Phase Noise Metrology

As a result of a major technological trend towards high speed digital communications and circuits, phase noise turns out to be a relevant concern for scientists and engineers. This paper describes methods and instruments to measure the phase noise of oscillators, components and more complex devices in the radiofrequency and microwave bands, from approximately 100 kHz to 30-40 GHz, and even beyond. After a brief introduction, two sections deal with basic definitions and traditional methods, and one section presents a set of schemes that cover most actual needs. Then a new approach— known as the interferometric method— is discussed in detail, providing design strategies and examples; this method exhibits the highest sensitivity in real time, which can alse be exploited to dynamically correct the phase noise of amplifiers and oscillators. The last section deals with an improved version of the interferometric method, in which correlation is used to remove the instrument noise of two equal interferometers that simultaneously measure the same device. This scheme enables the measurement of low noise processes, even below the thermal floor, and therefore it represents the state of the art in the high sensitivity phase noise metrology.

Noise induced time interval measurement biases

Time jitter in the output fronts of hysteretic triggers is not the only effect of additive noise. An average leading time bias error can be observed when large bandwidth noise is added to a poor slope signal. Time jitter is shown to be a possible limitation to the accuracy of high precision time interval measurements, as the average switching time is always before the zero crossing of the ideal noiseless analog signal. The resulting bias in the start-stop type of measurements, in which a noisy signal is compared to a reference, is shown to depend on S/N ratio and on noise bandwidth. An experimental evaluation of this effect is reported, based on time interval measurements in which noise is added to a sinewave signal.<<ETX>>

A measurement of the period stability of a free pendulum in vacuum

High period stabilities can be expected of a fiber suspended light pendulum swinging freely in vacuum, mainly because of the extremely high Q factors (>10/sup 5/) obtainable in this almost frictionless mechanical oscillator. In this paper, the modelization and the projected stability limitations of such a system are discussed. Both the physical arrangements and the electronic data collection solutions adopted in the experimental realization of a prototype pendulum apparatus are described, and sample series of period measurements are reported, analyzed and discussed. Mechanical instrumental noise (no isolation against vibrations was provided) appears to limit the Allan deviation of post modelization residuals with a white noise of a few /spl mu/s per period, which masks all projected period measurement stability limitations. It is argued that this observed noise is generated by instability in the position of the period detector, and possible solutions to this problem are proposed.

Oscillators and the Characterization of Frequency Stability: an Introduction

This paper provides an introduction to basic concepts commonly used in time and frequency metrology, and is addressed to other scientific communities. Thus no attempt is made to provide an exhaustive review. Instead, attention is focused on the most important tools used by physicists and engineers involved in time and frequency metrology. We first explain the principles of the oscillator through an example. Then we introduce the concepts of frequency reference, oscillating loop, frequency stability and accuracy. Finally, we define the power spectrum density of frequency fluctuations and the Allan variance as means to characterize the stability of frequency standards.