Time-Frequency Encyclopedia

Focus on time and frequency, precise and stable.

04

2024

-

07

Time‑synchronization system frequency source specifications

The primary performance metrics for frequency sources are frequency accuracy and frequency stability, with frequency stability further categorized into long-term stability and short-term stability.

  The primary performance metrics for frequency sources are frequency accuracy and frequency stability, with frequency stability further categorized into long-term stability and short-term stability.

  Frequency stability refers to the variation in a frequency source’s frequency accuracy over time, and it characterizes the source’s ability to maintain a constant output frequency. Frequency accuracy describes the deviation between the actual output frequency of a frequency source and its nominal frequency. In practical applications, a frequency source may be subject to both internal factors and external environmental conditions, leading to some degree of frequency drift. Frequency accuracy is the metric that quantifies this deviation and is typically expressed as either absolute frequency accuracy or relative frequency accuracy.

  During actual measurements, to improve accuracy, multiple readings can be taken and averaged. Frequency stability refers to the degree of frequency variation—caused by internal device noise, component aging, or external environmental changes—of an oscillator’s output frequency relative to its nominal frequency over a specified time interval.

  Currently, frequency stability is characterized in two ways: in the time domain, it is quantified by the internationally recognized Allan variance, which reflects the random fluctuations of the frequency mean; in the frequency domain, it is typically expressed as the power spectral density of relative frequency fluctuations.

  When measuring frequency stability, it is important to consider the potential impact of the test environment on the results. Ambient temperature, noise, and other factors can all introduce inaccuracies. Therefore, when testing high‑stability crystal oscillators, stringent requirements must be imposed on the test environment. In time‑frequency measurements of frequency stability, the Allan variance—widely recognized internationally as a metric for relative frequency fluctuations—is currently the standard method of expression.

  Here, it is required that there be no time interval between two consecutive measurements; in other words, the result of the previous sampling serves as the starting point for the next. In practical measurements, the number of samples cannot be infinite, so only an estimated value can be obtained. The greater the number of samples, the closer the estimate will be to the true value.