More About
SPAYN:

Brochure (PDF) 1.5 MB

Release Notes

Manual

Product Roadmaps
90 Day | 1 Year | 3 Year

Competitors

Application Notes

Presentation Materials

Training Materials

Software Download

See Also:

VWF
Virtual Wafer Fab

UTMOST III
Spice Modeling Software

UTMOST IV
Optimization Module for Compact / Macro-Modeling

Spice Modeling Services

SPAYN

Statistical Parameter and Yield Analysis

SPAYN is a statistical modeling tool for analyzing variances from model parameter extraction sequences, electrical test routines, and circuit test measurements. SPAYN helps to identify the relationship between device or circuit performance variations and the process fluctuations.

Key Features

• Automatic generation of worst-case and corner SPICE models
• Advanced principal factor / component analysis
• Identification of inter-relationship among groups of parameters
• Statistical process control, process monitor, and yield analysis
• Advanced Wafer Map displaying wafer-to-wafer and die-to-die variances
• Seamless integration with fast SmartSpice API for statistical circuit design
• Flexible data input, output, append, merge, and split options with histogram, 2D & 3D scatter plot, and response surface model display options
  
• Silvaco's strong encryption is available to protect valuable customer and third party intellectual property.

SPAYN Objective

SPAYN is a statistical analysis software package tailored for the semiconductor industry. It is applied for two main fields:

• Characterization
• Process Control / Analysis

Spread in performance of analogue or digital circuits is largely due to inherent fluctuation in the anufacturing process. Achieving statistical control is fundamental to product quality.

SPAYN is the ideal statistical modeling tool for analyzing variances from model parameter extraction sequences, electrical test routines, and circuit test measurements

SPAYN performs statistical analyses on parametric data, and helps determine the complex relationships between circuit design and process fabrication parameters.

SPAYN Users

• Parameter Extraction and Circuit Simulator Support Personnel
• Device Modeling Engineers
• Circuit Designers
• Process Development / Integration Engineers
  
  

SPAYN Applications

Used standalone, or linked to a circuit simulator, SPAYN is a user-friendly tool requiring little detailed knowledge of statistics and no user-created programs.

Users range from SPICE parameter extraction, process development and integration, product, and yield engineers, to circuit designers.

SPAYN is a simple, powerful and affordable tool for statistical modeling and process control.

Spreadsheet data format is used to view combined data from various sources.

Database Filtering

SPAYN is able to manage very large number of Attributes and Parameters and Data Sets. Limitations are only provided by working station memory and process capability.

Even with spreadsheet table, it can be tedious and numbing to work with all of them. To address this, SPAYN provides filters to precisely select sample and reduce number of Data sets in use. Filtering can be done at Attributes or at Parameters level. First, “Filter by Attributes”, defines database identification selection key. Second, “Filter by Parameters”, defines values limits to reduce number of sets that satisfy criteria.

Flexible site selection from many samples using independent Filter by Attributes. Filter could be specified by ‘Match’, ‘Values’, or ‘Tables’.	Subset of dataset definition using powerful Filter by Parameters. Some tools such as ‘compute limits’, ‘histogram plot’, Automatic Filter’,... help building the Filter.

Statistical Analysis, Data Management and Visualization

SPAYN accepts UTMOST-generated SPICE model library formats, Virtual Wafer Fab format, and other industry formats (RS/1TM, CSV, MS EXCELTM). SPAYN offers unique and extremely fast relational database search, merge, append, and split capabilities. Many built-in general purpose statistical analysis features aid in the examination of parametric data: Gaussian, exponential, log-normal, and Gamma distributions.

Scatter plots analyze the relationships between parameters by using the least squares method to fit linear, logarithmic, parabolic, reciprocal, hyperbolic, exponential, power, root, or 3rd order polynomial. The resulting error residuals, ANOVA information and correlation coefficients are generated. The “Golden Device feature” allows the user to calculate which observation or observations in a particular database most closely match the mean, based on some dissimilarity measure.

Response surface model of regression are displayed using TonyPlot	Histogram plot with attributes for selected parameters highlighted

The contour plot displays ellipses representing +s, +2s, and + 3s distribution of parameters	The relationship among any three selected parameters can be viewed in 3D

Multiple linear regression capabilities generate equations that relate one selected parameter to other sets of user-defined parameters. SPICE models card can be generated for any point in a scattergram Plot. The SmartSpice link and the user Netlist feature allow running simulation for each SPICE models card.

A correlation matrix of parameters selected for analysis.
Significant correlations are highlighted in red.

Advanced PCA/PFA – Principal Component and Factor Analysis

SPAYN identifies dominant parameters and/or factors of variations. The data is then automatically grouped so that each parameter in a specific group is controlled by the same source of variation. Analysis of each parameter group then identifies relationships between process input, process monitoring, device, circuit and manufacturing parameters within the group.

The following analysis features are available: Advanced Principal Component and Factor Analysis, Dominant parameter identification, VARIMAX and QUARTIMAX rotations, Histograms or scatter plots using components/ factors, PCA parameter weights and sensitivity analysis, and Unique user-defined dominant parameter selection and equations.

PCA / PFA capabilities allow the user to define dominant parameters or identify them automatically, and then generate linear or non-linear equations	Powerful factor analysis, coupled with built-in SPICE model equations, delivers unmatched statistical modeling capabilities

Worst-Case and Corner SPICE Model Generation

SPAYN generates equations relating each parameter under analysis to the uncorrelated dominant factors or dominant parameters. These independent variables are interactively or automatically perturbed to produce realistic “corner” and “worst-case” models using built-in SPICE model equations.

User-defined, Monte Carlo (left figure), corner (right figure), or worst-case simulations can be performed using fast SmartSpice API. Single device or very complex circuit can be simulated and any kings of analysis are usable.

Statistical Process Control and Yield Analysis

SPAYN can be used to determine the underlying statistical nature of IC manufacturing processes and their impact on device performance.

SPAYN creates statistical process control (SPC) and process monitor (PM) charts to track the variation of dominant parameters or factors over selected wafers or lots. SPC charts plot and store Shewhart mean and range/sigma data with the parameters grouped in terms of their attributes.

SPAYN helps to identify an efficient process monitoring strategy by identifying the minimum set of dominant factors that must be monitored in production in order to control the yield. Process monitoring charts are then used to view these parameters.

SPICE models are generated using built-in SPICE. Models can then be used to run simulations directly from PM charts.	Wafer Map displays parameters, or any other measured data together with their distributions over die locations

Data points lying outside process control limits (outlier) are marked with a square.

Golden Device feature calculate which observation or observations most closely match the mean, based on some dissimilarity measure. i.e., the measured device that best describes the given database.

Three neighbors of Golden Device
computed and Attributes displayed.

SPAYN in Silvaco Inductor PDK Flow

Rev. 101410_08