SOLID-STATE SIGNAL GENERATION at millimeter-wave frequencies is limited by the cutoff frequency and breakdown voltage of active devices. The low quality factor (Q) of the associated passive components in an oscillatordegraded by ohmic and substrate lossesalso comes into play. There is a question, however, about whether oscillators have exploited the full capacity of active devices in terms of output power and frequency. In any given process, it is essential to find the maximum frequency of oscillation (fmax) for a given device and circuit topology, considering the Q of the passive components. For a fixed-frequency oscillator, it also is important to determine the topology that results in maximum output power.
At Cornell University, Omeed Momeni and Ehsan Afshari investigated how oscillator topology and passive component Q impacted oscillation frequency, using the activity condition of the transistors. The researchers developed a systematic approach to designing high-frequency and high-power oscillators using device activity condition. This method finds the best topology for achieving frequencies close to the fmax of the transistors. It also determines the fmax for a fixed circuit topology while considering the Q of the passive components.
In a 0.13-m CMOS process, the engineers designed and implemented 121- and 104-GHz fundamental-frequency oscillators with output power of -3.5 and -2.7 dBm, respectively. Using a novel triple-push structure, Momeni and Afshari then introduced 256- and 482-GHz oscillators. The 256-GHz oscillator, which was implemented in 0.13- m CMOS, provided output power of -17 dBm. In contrast, the 482-GHz oscillator generated -7.9 dBm in a 65-nm process. See "High Power Terahertz and Millimeter-Wave Oscillator Design: A Systematic Approach," IEEE Journal Of Solid-State Circuits, March 2011, p. 583.