Performance-on-demand mems (podmems): Electrical control of effective mass, damping, and stiffness

We propose the use of electrostatic force feedback to control the stiffness, damping, or mass of MEMS. If feedback forces are proportional to sensed displacement, velocity, or acceleration of a MEMS proof mass, then feedback can be used to increase or decrease the apparent stiffness, damping, and or mass of the MEMS. Such feedback can be used to compensate for process variations, packaging stress, thermal drift, viscous damping, etc. Prior efforts by others include position or velocity based feedback for modifying frequency, bandwidth, quality factor, or sensitivity of resonators. We present a means of quantitative control of stiffness, damping, and mass of MEMS to achieve performance on demand, which we call Performance-on-Demand MEMS (PODMEMS). Our comprehensive control on effective parameters may enable two devices with different geometry to behave identically. This technology might enable a single PODMEMS to adjust its dynamic response depending on an application\u27s requirements. We derive and study both steady-state and transient PODMEMS models that include feedback forces, circuit delay, and noise. We compare transient and steady-state results for verification. There exists cross-talk among effective parameters. Cross-talk in effective damping from electrical mass and stiffness can decrease/increase the damping. The effective damping may become negative due to cross-talk, making the system unstable. The delayed feedback forces develop hysteresis in displacement, velocity, and acceleration and the width of hysteresis loop increases as the delay increases. Due to delayed feedback forces, the potential/kinetic energy show late or early minima/maxima and there are two minima/maxima throughout a cycle. Although potential energy and kinetic energy are affected by hysteresis of the feedback forces, the total energy is constant throughout a cycle. We have also simulated a test structure and found that frequency shift due to temperature variation can be reduced by a factor of ~2600 for our test case