CN-121981054-A - MEMS-CMOS collaborative design method and device based on model

CN121981054ACN 121981054 ACN121981054 ACN 121981054ACN-121981054-A

Abstract

The invention provides a MEMS-CMOS collaborative design method and device based on a model, which comprises the steps of constructing a parameterized collaborative model library, driving an MEMS simulator and a CMOS simulator to perform joint simulation through an automatic script engine based on the collaborative model library, automatically extracting a system-level key performance index from a simulation result, taking the system-level key performance index as an optimization target, adopting a multi-target optimization algorithm to simultaneously perform collaborative optimization on design parameters in an MEMS parameterized multi-physical-field model and a CMOS parameterized circuit macro model, generating a pareto optimal solution set, performing Monte Carlo analysis on a design selected from the pareto optimal solution set, and evaluating the yield of the design. The collaborative design method provided by the invention can find the optimal solution of the system level, avoid the problem of reduced overall performance due to local optimization, improve the success rate of streaming, and improve the design efficiency and the design precision of the whole microsystem.

Inventors

CAO JING
ZHANG MENG
HU XIAOYAN
WANG WEIPING
WANG ZIXIN
XU JIUZHI

Assignees

中国电子科技集团公司信息科学研究院

Dates

Publication Date: 20260505
Application Date: 20251204

Claims (10)

1. The MEMS-CMOS collaborative design method based on the model is characterized by comprising the following steps of: firstly, constructing a parameterized collaborative model library; the collaborative model library comprises an MEMS parameterized multi-physical field model, a CMOS parameterized circuit macro model and an interconnection interface coupling model for representing the interconnection interface coupling effect between MEMS and CMOS; Based on the collaborative model library, driving the MEMS simulator and the CMOS simulator to perform joint simulation through an automatic script engine, and automatically extracting a system-level key performance index from a simulation result; Taking the system-level key performance index as an optimization target, and adopting a multi-target optimization algorithm to simultaneously perform collaborative optimization on design parameters in the MEMS parameterized multi-physical-field model and the CMOS parameterized circuit macro model to generate a pareto optimal solution set; Step four, based on a process deviation model, carrying out Monte Carlo analysis on the design selected from the pareto optimal solution set, and evaluating the yield of the design; the first to fourth steps form a closed loop iteration process until a design scheme meeting the requirements of system performance indexes and yield is obtained.
2. The method of claim 1, wherein the design parameters of the MEMS parameterized multi-physical field model include at least one of mass, stiffness, damping, initial gap, and electrode area; The design parameters of the CMOS parameterized circuit macro model at least comprise one of transistor width-to-length ratio, bias current, compensation capacitance and switching frequency.
3. The method of claim 1 or 2, wherein the interconnect interface coupling model comprises: A physical interface model for characterizing bond wire or through silicon via parasitics; A noise coupling model for characterizing the effects of superposition of CMOS circuit noise and MEMS thermo-mechanical noise; and the reverse coupling model is used for representing the feedback effect of electrostatic force generated by the operation of the CMOS circuit on the MEMS movable structure.
4. The method of claim 1, wherein the system level key performance indicators comprise at least one of signal to noise ratio, dynamic range, bandwidth, total power consumption, and nonlinearity.
5. The method of claim 1, wherein the multi-objective optimization algorithm is a pareto-ranking-based multi-objective evolutionary algorithm.
6. The method of claim 5, wherein the pareto-ranking-based multi-objective evolutionary algorithm is the NSGA-II algorithm.
7. The method of claim 1, wherein the step three comprises: Combining MEMS design parameters and CMOS design parameters into a combined design variable vector, taking the system-level key performance index as an optimization target, and defining design constraint conditions; And (3) carrying out iterative optimization on the joint design variable vector by adopting a multi-objective optimization algorithm, automatically calling the joint simulation and performance index extraction flow of the step two in each iteration to evaluate objective function values, and finally outputting the pareto optimal solution set.
8. The method of claim 1, wherein the process bias model comprises a statistical variation model of a MEMS fabrication process and an angular model and a mismatch model of a CMOS fabrication process.
9. A model-based MEMS-CMOS co-design apparatus, comprising: the construction module is used for constructing a parameterized collaborative model library; The simulation module is used for driving the MEMS simulator and the CMOS simulator to perform joint simulation through an automatic script engine based on the collaborative model library, and automatically extracting system-level key performance indexes from simulation results; The optimization module is used for taking the system-level key performance index as an optimization target, adopting a multi-target optimization algorithm to simultaneously carry out collaborative optimization on design parameters in the MEMS parameterized multi-physical-field model and the CMOS parameterized circuit macro model, and generating a pareto optimal solution set; and the evaluation module is used for carrying out Monte Carlo analysis on the design selected from the pareto optimal solution set based on the process deviation model and evaluating the yield of the design.
10. The apparatus of claim 9, wherein the optimization module is further configured to combine the MEMS design parameters and the CMOS design parameters to form a joint design variable vector, take the system level key performance indicators as optimization targets and define design constraints, perform iterative optimization on the joint design variable vector using a multi-objective optimization algorithm, and automatically invoke a joint simulation and performance indicator extraction procedure of the simulation module in each iteration to evaluate objective function values, and finally output the pareto optimal solution set.

Description

MEMS-CMOS collaborative design method and device based on model Technical Field The disclosure relates to the field of microsystem design, in particular to a MEMS-CMOS collaborative design method and device based on a model. Background Typical MEMS-CMOS microsystems integrate MEMS structural function modules, voltage drive control or weak signal readout CMOS function modules and the like, relate to multiple disciplines such as electricity, material mechanics, heat and the like, and have the characteristic of obvious multi-discipline intersection. In order to shorten the design period of the micro-system and reduce the design iteration times, the collaborative design method based on the model is used for researching, collaborative design is realized for each functional unit of the MEMS-CMOS micro-system based on the existing model, and the overall behavior response of the simulation system is simulated to achieve the performance index matched with the design requirement of the system. Specifically, the development process of the microsystem comprises the processes of demand analysis demonstration, system function architecture design, behavior level design and simulation, physical level design and simulation, process implementation and the like, the development process fully shows multi-level characteristics, the behavior level design and the physical level design and the simulation stage also intensively show the characteristics of multi-specialty and multi-tool interaction of MEMS-CMOS microsystem design, frequent interaction among simulation tools of different disciplines often brings a large amount of repeated software linking workload, the design efficiency and the design accuracy are seriously influenced, and meanwhile, the technical requirements on designers are increased. Therefore, a new collaborative design mode is needed, each sub design is subjected to real-time automatic data interaction, design and collaborative simulation links such as multi-physical field simulation, link simulation, heterogeneous integration simulation and the like are uniformly controlled, and finally complete MEMS-CMOS cross-device cross-specialty collaborative design and simulation closed loop are realized, so that the overall performance of the system is ensured to realize high-confidence design optimization before process manufacturing. Disclosure of Invention The embodiment of the disclosure provides a MEMS-CMOS collaborative design method and device based on a model, which are used for solving the problem that the design efficiency and the design accuracy are affected by multi-specialty multi-tool cross in microsystem design. Based on the above-mentioned problems, in a first aspect, an embodiment of the present disclosure provides a method for collaborative design of a MEMS-CMOS based on a model, including the steps of: firstly, constructing a parameterized collaborative model library; the collaborative model library comprises an MEMS parameterized multi-physical field model, a CMOS parameterized circuit macro model and an interconnection interface coupling model for representing the interconnection interface coupling effect between MEMS and CMOS; Based on the collaborative model library, driving the MEMS simulator and the CMOS simulator to perform joint simulation through an automatic script engine, and automatically extracting a system-level key performance index from a simulation result; Taking the system-level key performance index as an optimization target, and adopting a multi-target optimization algorithm to simultaneously perform collaborative optimization on design parameters in the MEMS parameterized multi-physical-field model and the CMOS parameterized circuit macro model to generate a pareto optimal solution set; Step four, based on a process deviation model, carrying out Monte Carlo analysis on the design selected from the pareto optimal solution set, and evaluating the yield of the design; the first to fourth steps form a closed loop iteration process until a design scheme meeting the requirements of system performance indexes and yield is obtained. With reference to the first aspect, in one possible implementation manner, the design parameters of the MEMS parameterized multi-physical field model at least comprise one of mass, stiffness, damping, initial gap and electrode area; The design parameters of the CMOS parameterized circuit macro model at least comprise one of transistor width-to-length ratio, bias current, compensation capacitance and switching frequency. With reference to the first aspect, in one possible implementation manner, the interconnection interface coupling model includes: A physical interface model for characterizing bond wire or through silicon via parasitics; A noise coupling model for characterizing the effects of superposition of CMOS circuit noise and MEMS thermo-mechanical noise; and the reverse coupling model is used for representing the feedback effect of electrostatic force g