Semiconductor Delay Calculation Using Input Pin Capacitance Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional delay calculation methods in semiconductor integrated circuits face inaccuracies due to oversimplification of input pin capacitance, which depends on both input slew and drive load capacitance, leading to excessive margin requirements and deviations in delay calculation results, especially as manufacturing processes become more precise.
Innovation Solution
A method that calculates input pin capacitance by expressing it as a function of input slew and drive load capacitance, allowing for more accurate representation and minimization of errors in delay calculations, incorporating statistical processing and specific capacitance calculations to reflect actual circuit conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the input pin capacitance is represented by one value or a range (maximum and minimum values) in the delay library, then the delay calculation can be performed with simple data structure, but the accuracy of delay calculation deteriorates because the input pin capacitance actually varies depending on input transition and drive load capacitance
Solution Approach 1:
The patent applies parameter changes by representing input pin capacitance not as a fixed value or simple range, but as a function of multiple parameters including input slew and drive load capacitance. This allows the capacitance value to dynamically adapt to different circuit conditions, significantly improving delay calculation accuracy while managing complexity through systematic parameterization.
Solution Approach 2:
The patent implements dynamics by making the input pin capacitance value variable rather than static. The capacitance is expressed as a function that changes based on input slew and drive load capacitance conditions, allowing the delay model to reflect actual circuit behavior under varying operating conditions rather than relying on fixed or simplified capacitance values.
2Productivity
If the input pin capacitance is estimated to be relatively larger or smaller using conventional methods, then the delay calculation can be performed with simple coefficient multiplication, but excessive margin is demanded which reduces the number of obtained chips
Solution Approach 1:
The patent changes the approach from using fixed coefficients for capacitance estimation to using a functional representation that incorporates input slew and drive load capacitance as variables. This enables more accurate capacitance values to be calculated for each specific circuit instance, reducing the need for excessive design margins and thereby increasing the number of obtainable chips while maintaining reliability.
3Measurement precision
If the coefficient table with input slew as index is used to calculate input pin capacitance, then the calculation can be performed with simple interpolation, but error is generated because the input pin capacitance depends on both input slew and drive load capacitance
Solution Approach 1:
The patent extends the conventional single-parameter (input slew) coefficient table approach to a multi-parameter model that includes both input slew and drive load capacitance. This functional representation with multiple parameters captures the actual dependencies of input pin capacitance more accurately, improving calculation precision while managing the increased complexity through systematic modeling.
4Measurement precision
If the number of points of input slew is increased to reduce interpolation error, then the accuracy of input pin capacitance calculation is improved, but the coefficient calculation becomes impractical because it is obtained simultaneously with delay value characterization
Solution Approach 1:
The patent changes from using a large number of discrete input slew points to using a continuous functional representation of input pin capacitance in terms of input slew and drive load capacitance. This functional approach achieves high accuracy without requiring an excessive number of characterization points, making the coefficient calculation process practical while maintaining precision.
Data Source
AI summary
An input pin capacitance of a cell is obtained in advance in a function expression, and a delay is calculated in such manner that the input pin capacitance is calculated in functions of an input slew and a drive load capacitance in each instance. In a cell characterizing process, a total volume of a current running into an input terminal before a voltage value of the input terminal reaches a reference voltage is obtained so that a value approximate to a real input pin capacitance can be obtained.