Computational Efficient Models
IntelliSense introduces a new class of extraction tools for dramatically reducing the computational time to perform accurate analysis of MEMS. Based upon cutting edge, Krylov subspace model reduction algorithms, very large nonlinear thermoelectromechanical FEM models can be reduced into computationally efficient behavioral models. These behavioral models can be used in system simulators and analog mixed-signal workflows.
IntelliSuite behavioral models fully capture the complex nonlinear dynamics inherent in MEMS due to electrostatic forces, residual stresses, squeeze film damping, packaging effects and multiple mechanical degrees of freedom. These models fully capture all of the harmonic modes of the MEMS so that you can fully account for sub-harmonic contributions while developing readout or control electronics.
download download download
3D N-DOF System Models (click picture to play movie)
IntelliSuite extracted system models are full 3D models. As a result system simulators can provide full 3D output of the system response. As a result the response of a device to a complex mechanical or electrical input can be captured accurately. (a) Extracted beam model responding to a 50 kHz voltage input (b) the same beam responding to a 3 MHz voltage input (c) a micro-mirror response showing multiple modes being actuated.
Orders of magnitude faster
Extracted models match FE simulations in accuracy but are 100-1000 times faster than traditional FE based simulations.
Coupled Fluidic, Electrostatic, Mechanical analysis of a 3x3 micromirror array; effect of one mirror being actuated on surrounding mirrors. Extraction techniques can speed up computation by a few orders of magnitude.
Extraction is performed in 3 stages. (a) The total strain and capacitance energy and fluid dissipation is captured (b) a Lagrangian Look Up Table (LUT) formulation is used to generate a description of the device and (c) the model is output into various HDLs for incorporation into EDA tools of your choice.
Lagrangian Formulation
System model extraction is based upon capturing the total energy within a MEMS device. Strain energy, electrostatic energy, fluidic dissipation and other damping terms are captured for each mode of interest. Sub-modeling techniques can be used to capture the interaction of the packaging related stresses on the device as well. Process related aspects such as residual stresses and strain gradients in films, orthotropic or anisotropic material properties, electrostatic levitation effects, thermal or packaging effects are fully captured in the resulting model. The extracted model is an arbitrary N-DOF 3D Model. User can use the 3D model to investigate dynamics of the MEMS device either locally or in full 3D .
Packaging analysis can now capture fluidic damping and spring force for complex MEMS structures. Our macromodel techniques now extend to fully capture fludic dissipation and automatically include ambient damping into system simulations. Fluidic damping can be captured for each of the modes of interest.
As MEMS make their successful entry into mobile devices, the demand for slim profile packages becomes paramount. MEMS devices have started to leverage IC stacking technologies such as wafer bumping, die-on-wafer assembly, through-silicon vias (TSV) and towards low-cost Wafer Level Chip Scale Packages (WLCSP). As the MEMS subsumes the package into the structure, the need for efficient package modeling and compensation of packaging effects at the die level become important.
One of the most innovative features is the ability to include packaging effects on the device performance into the system models. Package extraction and sub-modeling techniques allow you to estimate the impact of the package stresses and temperature effects on the system performance.
Feature Highlights
General features
Tightly integrated into IntelliSuites Thermo-Electro-Mechanical analysis tools
Ability to define areas or points of interest within a model
Automated or semi-automated behavioral model extraction.
Three levels of behavioral models of varying complexity can be generated.
Level 1 models. Models based on rigid body approximations. These capture up to 6 mechanical DOFs and multiple electrical DOFs. Useful for designing readout electronics
Level 2 models: Accurately capture second order effects such as parasitics, non-linear deformation, temperature effects and package related effects. Useful for designing compensation electronics
Level 3 models: Fully capture harmonic and sub-harmonic response of the devices. Useful for designing active controls of MEMS.
Behavioral modeling features
Support for arbitrary electrical & mechanical DOFs. Capture moments of inertia (Ixx, Iyy, Izz, Ixy Iyz, Ixz), masses, 3D squeeze-film damping effects, non-linear spring constants (fully accounting for residual stresses and stress gradients)
Automated extraction of non-linear electrostatic forces (even for fringe field dominated devices)
Automatic calculation of modal contribution factors and modal energies. Fully capture of static, dynamic and harmonic response of devices
Behavioral model error bound estimation
Automatic formulation of device dynamics (Lagrangian formulation)
Model output
Model output to SYNPLE, Verilog-A, VHDL, PSPICE, HSPICE, Simulink (with purchase of EDA Linker)
Slideshows, Publications, Application Notes