Quartz Crystal Oscillators

Figure 4

Figure 4 Quartz crystal oscillator frequency stability versus temperature.

Quartz crystals satisfy the medium- to long-term stability requirements of most microwave systems.6 Quartz is a bulk electroacoustic element. While crystal oscillators operate at 10 to 100 MHz, they resonate acoustically. Two properties make quartz an advantageous material: it has high Q values, typically between 10,000 in small low-cost variations to about a million in very expensive versions and quartz has a flat frequency versus temperature plateau, as shown in Figure 4.

The quartz crystal resonator is cut as either AT or stress-compensated (SC). AT is the room temperature cut used in temperature-compensated crystal oscillators (TCXOs). This type of oscillator has very little frequency change versus temperature around 25°C. Additional compensation circuitry can extend this useful plateau significantly. The SC cut has a higher Q and a flat frequency-to-temperature plateau around 90°C. Instead of temperature compensation, the SC crystal is used in an OCXO device, held at a constant temperature in an oven, to achieve even greater stability than is achievable in a TCXO.

A disadvantage of quartz-based crystal oscillators is their performance degrades at frequencies above about 100 MHz. To utilize a quartz-based oscillator in X-Band requires multiplication and this approach adds a 20log10(N) noise multiplier, where N is the frequency multiplication factor. Figure 5 shows a 10 MHz and 100 MHz high-quality OCXO and their noise after ideally multiplying to 8 GHz. The 10 MHz device will almost always outperform a 100 MHz device at 1 to 100 Hz offsets. However, the 100 MHz oscillator will perform better at higher frequencies because it has been optimized for a better noise floor farther from the carrier. High performance systems may integrate and phase-lock both oscillators for the optimum phase noise and stability performance.


Figure 5

Figure 5 Comparison of 10 MHz and 100 MHz OCXOs at 8 GHz.

Figure 6

Figure 6 8 GHz oscillator phase noise comparison to Saetta Labs’ SLCO.

A SAW oscillator is a surface acoustic device with a good Q-factor and small size. They operate from around 300 MHz to 2 GHz and serve as an intermediate oscillator to lock while multiplying OCXOs higher in frequency. These devices can sometimes replace a 100 MHz OCXO or a DRO. The performance of the SAW oscillator sits between the DRO and the OCXO in frequency and the material has significant size advantages. SAW oscillators require frequency multiplication for operation at X-Band and above.


Oscillators operating fundamentally at X-Band will have better phase noise than a quartz oscillator at offsets far from the carrier. Figure 6 is a phase noise plot showing the quantitative performance of OCXO, voltage-controlled oscillator (VCO), DRO and SLCO designs at 8 GHz. As discussed, the Q-factor is the main performance difference in these oscillators. The Q-factors range from about 10 to 100,000, depending on the resonator technology. Size and cost are usually related to the Q-factor.


A VCO is the smallest, lowest cost and most prevalent of all the X-Band oscillators. They are typically realized as a lumped-element oscillator, either discrete or integrated on-chip. Some are embedded directly into the PLL IC for the highest level of integration. The advantages of this approach are wide tuning range, fast acquisition, low power consumption and small size. The disadvantage is a relatively low Q-factor, typically between 10 and 100. While this option has the lowest performance, it offers the best SWaP-C solution and the performance is often “good enough.”


Yttrium iron garnet (YIG) material enables wide tuning, high Q, magnetically-tuned oscillators. The material resonates in the 3 to 10 GHz range and this range can be made wider if the material is doped. The resonance is directly proportional to a magnetic field. These oscillators have higher Q than VCOs with wide, but slow tuning. The large magnetic field results in significant power consumption and heat dissipation. These devices may be the best option when applications require a wide tuning range and better phase noise than VCOs can provide. Q-factors are in the 500 to 1000 range. YIG oscillators are significantly bigger and more expensive than VCOs and they are typically seen in higher performance systems like test equipment requiring a wide tuning range and low phase noise.


DROs are single-frequency oscillators that operate with a TM or TE electromagnetic resonance in a cylinder of high dielectric material. In a coaxial resonator oscillator (CRO), this high dielectric material is in a small coaxial cylinder. The material options range from larger dielectric pucks with high Q-factors to smaller, lower Q pucks for different applications and temperature stabilities. Unlike VCOs, they are single-frequency devices that are typically coarse-tuned mechanically and fine-tuned electrically for phase locking. DRO-based devices are about the size of a YIG but provide better phase noise performance at significantly lower operating currents. DROs are a good high performance option for architectures where fixed LOs drive ADCs and DACs. Q-factors can be around 1000, although phase noise varies greatly between manufacturers and models.

Sapphire Oscillators

Sapphire oscillators are relatively new to commercial applications. They operate as an electromagnetic device, like a DRO, but they use a different material and mode of resonance. The resonator material is sapphire and the resonance is called a whispering-gallery mode. The mode was named for the acoustic resonance in St. Paul’s Cathedral in England, where the acoustic waves from a whisper travel around the perimeter of the circular hall with almost no attenuation. In a circular shape, a whispering-gallery mode can travel on the inside of two dielectric boundaries with near-perfect reflection. This removes the metallic losses inherent in DROs and cavity oscillators, improving the Q-factor from 10× to 100×. The only loss mechanism is the dielectric loss of the sapphire, which is extremely low. Sapphire oscillators have Qs around 100,000, which are like quartz but operate fundamentally at X-Band. These microwave oscillators have the lowest available noise performance but they are the largest solution and they do require power for thermal control like an OCXO. These constraints make these oscillators more expensive.


Direct microwave conversion capabilities in the X-Band and above frequency range are expanding system design capabilities. The low phase noise performance of the new generation of DACs and ADCs spotlights the need for increasingly stringent clock performance requirements. In addition to improvements in materials, careful design considerations are needed to achieve the highest performance possible while still being able to accommodate SWaP-C requirements. This article has discussed oscillator phase noise and stability concerns along with the advantages and disadvantages of quartz, IC, VCO, YIG, DRO and newer SLCO technologies for designing and manufacturing oscillators that will become essential as direct conversion techniques move higher in frequency.


Thank you to David Guidry at Texas Instruments for providing data and helpful conversations on driving high frequency DACs and ADCs.


  1. T. Galla, “Selecting Phase-Locked Oscillators for Frequency Synthesis,” Microwave Journal, Vol. 62, No. 9 Supplement, September 2019.
  2. D. Leeson, “Oscillator Phase Noise: A 50-Year Review,” IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, August 2016.
  3. RF, RFIC and Microwave Theory, Design, RFIC, Web: www.ieee.li/pdf/essay/phase_noise_basics.pdf.
  4. L. Geller, “Characterization of Additive Phase Noise in Microwave Amplifiers,” Microwave Journal, Vol. 64, No. 1, January 2021.
  5. J. Breitbarth and J. Koebel, “Additive (Residual) Phase Noise Measurement of Amplifiers, Frequency Dividers and Frequency Multipliers,” Microwave Journal, Vol. 51, No. 6, June 2008.
  6. J. Emmerich and H. Rudolph, “The Importance of Crystal Oscillators with Low Phase Noise,” Microwave Journal, Vol. 66, No. 4, April 2023.