[Paper Review] Characterization of the individual short-term frequency stability of Cryogenic Sapphire Oscillators at the 1e-16 level

Christophe Fluhr, S. Grop+4 more|arXiv (Cornell University)|Dec 10, 2015

Advanced Frequency and Time Standards17 references3 citations

TL;DR

This paper characterizes the individual short-term frequency stability of three Cryogenic Sapphire Oscillators (CSOs) using the three-cornered-hat method with commercial components, achieving sub-10⁻¹⁶ stability. The best CSO exhibits fractional frequency stability below 5×10⁻¹⁶ at 1 s and below 2×10⁻¹⁶ for integration times up to 5,000 s, revealing that residual thermal fluctuations and discriminator noise limit performance despite ultra-low intrinsic noise.

ABSTRACT

We present the characterisation of three Cryogenic Sapphire Oscillators using the three-corner-hat method. Easily implemented with commercial components and instruments, this method reveals itself very useful to analyse the frequency stability limitations of these state-of-the-art ultra-stable oscillators. The best unit presents a fractional frequency stability better than 5e-16 at 1 s and below 2e-16 for integration times less than 5,000s.

Motivation & Objective

• To measure the individual short-term frequency stability of three Cryogenic Sapphire Oscillators (CSOs) at the 10⁻¹⁶ level.
• To identify technical noise sources degrading stability beyond intrinsic oscillator noise.
• To evaluate the effectiveness of the three-cornered-hat method for characterizing ultra-stable oscillators using off-the-shelf components.
• To optimize thermal and control systems by identifying instability sources such as temperature modulation and mechanical perturbations.
• To determine the dominant noise sources limiting performance in state-of-the-art CSOs.

Proposed method

• The three-cornered-hat method was applied to extract individual Allan deviation (ADEV) from three CSOs by cross-comparing their beat notes using commercial frequency counters and signal processing software.
• A 70 kHz phase modulation was applied via a Voltage-Controlled Phase Shifter (VCPS) to enable Pound servo feedback for frequency stabilization.
• The frequency discriminator gain D was varied by adjusting the microwave power injected into the sapphire resonator, allowing noise analysis of the discriminator stage.
• The voltage spectral density en of the demodulator output was computed from the ADEV data using the relation σy(τ) = en / (√2 × D × ν₀) × τ⁻¹/² to estimate intrinsic noise.
• Thermal filtering was implemented using copper braids and thermal ballast to reduce mechanical and temperature fluctuations from the pulse-tube cryocooler.
• The turnover temperature T₀ was stabilized to minimize thermal sensitivity, with T₀ values between 5–8 K depending on sapphire crystal quality.

Experimental results

Research questions

• RQ1 What is the individual short-term frequency stability of each of three Cryogenic Sapphire Oscillators at the 10⁻¹⁶ level?
• RQ2 What technical noise sources degrade the stability of CSOs beyond their intrinsic oscillator noise?
• RQ3 How effective is the three-cornered-hat method when implemented with commercial components for characterizing ultra-stable oscillators?
• RQ4 Why do two of the three CSOs exhibit a hump in their ADEV at short integration times, and what causes this instability?
• RQ5 What is the dominant noise source limiting the performance of the best-performing CSO?

Key findings

• The best-performing CSO achieves a fractional frequency stability of 4.6×10⁻¹⁶ at 1 s and remains below 2×10⁻¹⁶ for integration times up to 5,000 s.
• CSO-3 exhibits no hump in its ADEV and demonstrates a flicker floor below 2×10⁻¹⁶, indicating superior thermal and control stability.
• The voltage spectral density en of the Pound discriminator is estimated at approximately 9 nV/√Hz, consistent with expected noise from the lock-in amplifier and diode detector.
• The ADEV of CSO-1 and CSO-2 is limited by residual temperature fluctuations, likely due to mechanical perturbations from prior transport and thermal anchoring changes.
• Despite identical design, the three CSOs show different stabilities due to variations in thermal configuration, resonator quality, and post-assembly handling.
• The three-cornered-hat method successfully isolates individual oscillator noise using only commercial instruments, enabling stability characterization at the 10⁻¹⁶ level.