Frequently Asked Questions (FAQ) on Micromechanical Resonators:
(Answers provided by Clark T.-C. Nguyen: . Email me if you must, but be warned:
Question: What is the frequency limit for micro- or nano-mechanical devices?
Answer: There is no vibrational frequency limit for mechanical devices. As long as a given mechanical structure can be scaled to an appropriately small size or made to vibrate with an appropriate mode shape, that structure can be made to resonate at any frequency, without limit. The better question to ask, here, is: “At what frequency will the Q of a vibrating mechanical resonator become uselessly small?” It’s one thing to vibrate at 10 GHz; it’s another to be able to vibrate at 10 GHz while retaining a Q >1,000. Q and what happens to it as frequencies rise is the main reason why frequency extension research is so interesting.
Question: Why should I care about the high Q devices in a wireless transceiver? After all, aren’t the latest transceiver architectures based on direct-conversion “doing away” with the need for off-chip, high-Q passives?
Answer: Direct-conversion architectures and similar approaches aimed at minimizing the number of off-chip, high-Q passives used in a given implementation have actually only successfully eliminated one off-chip passive: the IF filter, and this at the cost of performance. Many other off-chip components remain, including the RF filters (pre-select and image-reject, or something similar to image-reject for further noise suppression), the oscillator frequency reference (e.g., quartz crystal), and inductors and capacitors for biasing and medium-Q tank applications (e.g., for the VCO in the synthesizer). In addition, without the IF filter to passively remove out-of-channel interferers, the transistors in a direct-conversion receiver must work harder to suppress such interferers, expending more power in the process, i.e., higher power is needed to attain higher dynamic range transistor circuit stages. This last point is an example of the Q versus power trade-off that so often arises in communication circuit design and determines the battery lifetime of portable wireless applications. In particular, the higher the Q of a given communication function (e.g., the pre-select filter), the lower the power consumption
Question: Why would anyone want a Q >10,000? Isn’t a Q on the order of 1,000 to 3,000 good enough for most RF front-end applications?
Answer:If one confines thinking to only today’s communication architectures, then it’s true that Q’s from 500 (for the pre-select filter) to 3,000 (for the image-reject filter) are all that are required for today’s conventional super-heterodyne communication architectures. the pre-select filter that immediately follows the antenna need not have