Tunable Oscillators Aim At Reduced Phase Noise

Tunable oscillators are instrumental in the operation of many systems, from commercial communications to military radars. And, while many characteristics define the performance of a tunable RF/microwave oscillator, one of the more difficult parameters to optimize is phase noise, which is critical to the performance of so many systems. For that reason, the design team at Synergy Microwave (www.synergymwave.com) has targeted phase noise as one of the key specifications for its innovative line of compact coupled planar resonator (CCPR) oscillators.

These high-quality-factor (high-Q) tunable oscillators currently can be designed for frequencies from about 50 MHz through 14 GHz, using low-loss substrate material from Rogers Corp. (www.rogerscorporation.com). Oscillators are currently in production and have been produced and tested through 14 GHz. Introduced earlier this year (see Microwaves & RF, June 2008, p. 92 and July 2008, p. 90), the tunable oscillators are based on the use of self-injection-locked CCPRs to produce stable output frequencies with low phase noise and low jitter. Because this is a new technology, it has been compared to existing oscillator technologies, such as dielectric-resonator oscillators (DROs), ceramic resonator oscillator (CRO), surface-acoustic-wave (SAW) resonator oscillators, YIG oscillator, and even traditional voltage-controlled oscillators (VCOs). And the industry has been inquisitive about the novel oscillator technology, based on the feedback received so far from inquiries made at the Microwaves & RF website (www.mwrf.com).

Some of these questions about the technology have been as simple as "does it really work?" During a recent visit with Ulrich L. Rohde, Chairman of Synergy Microwave, and Ajay K. Poddar, Chief Scientist at Synergy Microwave, the authors of the original articles on the CCPR oscillators, some of the issues raised by readers were addressed. From the measurement results, it is clear that not only has the Synergy design team built a number of these oscillators, but has also thoroughly tested them for such characteristics as output power, frequency stability with temperature and time, and phase noise. Based on the Rogers substrate material, the team has succeeded in reaching to 14 GHz with the technology, with excellent stability over temperature and time.

As Rohde explains, the CCPR oscillator design provides outstanding stability: "We have had several inquiries about the frequency stability of these oscillators. Stability is an issue where the frequency either drifts because of temperature changes or because of aging. These are VCOs and they are not intended to serve as frequency standards, but to function as part of a phase-locked loop (PLL) in a synthesizer circuit. We have designed them to provide the lowest phase noise possible. Any question of stability does not really apply to these oscillators in their normal application, in which they will be locked in frequency to a reference source as part of a PLL frequency synthesizer."

In PLL circuits built for use with the CCPR oscillators, the Synergy design team has used sampling discriminators in the PLL based on harmonic sampling of the reference oscillator. Using a sampling divider provides advantages in phase-noise performance of large offsets from the carrier. However, as Poddar points out, "our technology will also work with fundamental dividers as well. These oscillators can reach 15 GHz with fundamental dividers and the substrate materials we are currently using, and the phase noise performance is exceptionally good for a given class and toplogy." The company has spent countless hours in their test laboratory to characterize these oscillators for phase noise close in and further out from the carrier, using the latest spectrum analyzers from Rohde & Schwarz (FSUP 20Hz -50 GHz) and Agilent Technologies (10 MHz -7 GHz).

In addition to the barrage of measurements, the design team has developed accurate electromagnetic (EM) models for the CCPR oscillators. According to Ulrich Rohde, "We performed 2.5 or 3-D modeling of the resonant structure and incorporated this model into a model for the nonlinear oscillator circuit. The nonlinear circuit contains the oscillator's active device, with S-parameters representing the passive circuit part (resonator)." This partitioning of the oscillator into its modeled component parts apparently worked quite well since, as Rohde, notes: "the combination of the S-parameters and the nonlinear circuit model agree closely with the measured data from the circuits we have built." Ajay Poddar adds that "the S-parameters that we use for these transistors are large-signal S-parameters," so the firm is able to understand the behavior of the oscillator active devices under quasi-linear (low signal drive level) and nonlinear (large signal drive level) conditions, which improves the optimization cycles using harmonic balance simulators (Serenade/Ansoft/Nexxim/Ansys and ADS 2008) to the limit of the physics.

In spite of the firm's success with its discrete-component versions of the CCPR oscillators, it is seeking a partnership with the right IC house in order to develop IC versions of the oscillators suitable for larger-scale, lower-cost markets. In order to realize the full potential of the technology in IC form, the right partner must be willing to experiment with multiple masks and wafer runs, be conversant with the Cadence design environment in order to optimize different oscillator characteristics on materials such as GaAs, SiGe, and BiCMOS that are considerably more lossy than the Rogers substrates used for the discrete-component CCPR oscillators, and be willing to invest in a technology that should offer significant benefits in IC form. Among those benefits are considerably smaller size than current VCOs on the market, with lower phase noise and much lower power consumption. The power-efficient version of evanescent mode CCPR oscillators that can be right candidate for transition into integrated circuit (IC) typically require +2-3V VDC bias and consume only about 8-10 mA current.

Two-way radio companies now look to cover the spectrum from 100 MHz to 2000 MHz with better performance than currently available due to tighter systems specifications and much less power consumption requirements. This new patented approach, currently and transferred to a hybrid IC can accommodate the present and future requirement of software defined radio (SDR), which needs faster tuning sensitivity to cover the wideband operations without compromising the phase noise performances.

Poddar volunteers that the firm is working on broadband solutions with the technology: "We are working at pushing the technology to 18 GHz in first phase and the "know how" can be utilized to extend the frequency in Ka band using N-Push techniques for which Synergy owns the patent. The technology can go as low as 500 MHz, and even down as far as 50 MHz, but some fundamental changes must be made to the basic oscillator design." He notes that the company has built 500-MHz oscillators with extremely low phase noise and measuring a mere 0.3 x 0.3 in. "We are also working on a similar research effort in the area of crystal oscillators, around VHF," says Rohde. He offers, "we would like to work with DARPA or other military organizations to help them improve their oscillator technology." For example, in SATCOM and radar system, the lower phase noise of these oscillators could dramatically improve the range of the system. Synergy Microwave Corp.; 201 McLean Blvd., Paterson, NJ 07504; (973) 881-8800; Fax: (973) 881-8361, E-mail: sales@synergymwave.com, Internet: www.synergymwave.com.