Semiconductors
Our SEMICONDUCTOR design and fab EXPERIENCE:
FABLESS
Fabless players have to address a unique set of challenges, beyond reference designs, wafer supply, and relationships with foundries. We help companies optimize the software systems that manage tapeouts, specifications, fab cards, test automation, wafer tracking, and many other information system changes that are often necessary with every new design and process.
We are one of few small software companies with the semiconductor industry knowledge and expertise to be able to step in and very quickly deliver custom solutions for this industry.
EDA
Our experience with Electronic Design Automation includes the development of custom modules and integrations for Ansys HFSS, COMSOL and Cadence / AWR. We have worked on highly-specialized Method-of-Moments EM solvers for thin-film SAW design, and implemented hybrid-cloud distributed simulation deployments across captive HPC data centers and the public AWS and Azure clouds. Our work includes rewriting research and experimental MATLAB code to tuned C++ code that takes advantage of CUDA GPU 64 bit floating point numerical calculation capabilities.
Foundry
Foundries operate at the edge of what is physically possible and face challenges with cost-effective manufacturing. The unprecedented complexity of modern semiconductors makes it a constant challenge to keep cost, quality, and manufacturing cycle times under control.
We have worked with clients to develop optimized process control systems that dramatically increase the ability of a fab to start up new processes, pursue yield improvement and perform efficient process control during mass production.
Watch the Video here:
Use Cases and Projects
SNP / Touchstone Database
Our client needed that ability to store and retrieve both simulation results as well as measurement data products in s-parameter and y-parameter formats.
We developed a high-performance web front-end system backed by an application server and a database schema, that uses a combination of modern SQL modeling techniques and JSON data types to allow 1- to 16- pot SNP data to be efficiently stored, collected with robotic wafer probes, and automated-sweep VNA equipment, indexed, and searched. The ability to retrieve multiport frequency response data by design, turn, PDK, wafer, shot and response parameters enabled faster turns and improved designer efficiency.
 |  |  |
|---|---|---|
 |
Process Monitoring and Control
Our client developed proprietary machine-learning techniques and algorithms that are not sensitive to measurement parasitics, to allow the use of simple wafer-probe measurements of PCMs to deliver Critical to Quality (CTQ) data including dimensional information across an entire wafer, and to replace slow, expensive, and destructive FIB-SEM or TEM imaging that can only measure a few specific locations.
oPMTx algorithms are not sensitive to measurement parasitics and unlike SEM/TEM PMTx is a non-destructive process.
We developed the system wrappers around the core algorithms to allow multitenant use of the system, parallel execution, data storage, validation, and visualization of results.
The efficiencies fabs gained from the technology include cutting 6 to 12 months from time to market and revenue generation when bootstrapping each new FAB Process, bringing new products to market approximately 6 months faster, cutting yield optimization time between ramp and mass production by 2 to 3 months and saving 3% yield/year by eliminating excursions and out of control incidents.
Need more details? Contact us.
We are here to assist. Contact us by phone, email or via our Social Media channels.
