Semiconductor Switch Failure Prediction Without Temperature Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor monitoring systems face challenges in increased sensor complexity, high costs, and inefficiency in predicting device failures, particularly for SiC semiconductor devices, where certain parameters like Rdson are temperature-dependent, making it difficult to accurately predict device health and remaining useful life.
Innovation Solution
A semiconductor monitoring system that includes sensors to measure steady-state electrical characteristics, a control system to estimate junction temperature without a temperature sensor, and a method to predict device failure by calculating remaining useful life using lookup tables for channel and body diode resistance values, simplifying circuitry and eliminating the need for temperature measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If temperature sensors and high bandwidth circuitry are used to measure fast changing electrical variables, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent extracts the temperature measurement function from dedicated temperature sensors and high bandwidth circuitry by using standard voltage and current measurements during normal device operation. The junction temperature is derived from electrical characteristics (Rdson, body diode voltage) that are already being measured, eliminating the need for separate temperature sensing hardware.
Solution Approach 2:
The patent makes standard voltage and current measurement circuitry perform multiple functions: they measure both the electrical characteristics needed for power device control and the temperature-dependent parameters needed for failure prediction. This multi-functionality eliminates dedicated temperature sensors and reduces overall system complexity.
2Ease of operation
If temperature-dependent parameters like Rdson are used for failure prediction, then ease of operation is improved, but reliability deteriorates due to temperature compensation requirements
Solution Approach 1:
The patent changes the approach from using absolute parameter values to using parameter ratios and trends. By monitoring changes in Rdson and body diode voltage relative to their expected values at measured junction temperatures, the system achieves temperature-compensated failure prediction without complex compensation algorithms.
Solution Approach 2:
The patent implements feedback by continuously measuring junction temperature through electrical characteristics and using this information to adjust failure prediction thresholds. The system compares measured parameters against temperature-corrected reference values, enabling reliable failure prediction across varying operating conditions.
3Reliability
If multiple sensors and complex circuitry are deployed to monitor semiconductor health, then reliability is improved, but manufacturing cost increases
Solution Approach 1:
The patent enables the power semiconductor device to monitor its own health status using its inherent electrical characteristics during normal operation. The device's own voltage and current waveforms provide the information needed for failure prediction, eliminating the need for external monitoring hardware and reducing manufacturing complexity.
Solution Approach 2:
The patent recovers useful temperature and health information from electrical measurements that would otherwise be discarded or used only for basic control functions. By extracting temperature-dependent parameter information from standard voltage and current measurements, the system achieves reliable failure prediction without additional sensors.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system effectively predicts near-term semiconductor device failures, reducing downtime and costs by simplifying sensor complexity and providing accurate health monitoring without requiring temperature sensors, thus enhancing reliability and reducing sensor complexity.
Implementation Method 1
a semiconductor switch structured to allow electric current to flow in forward conduction from a source to a drain in a channel while the switch is turned on, in reverse conduction from a drain to a source in the channel while the switch is turned on, and in reverse conduction from the drain to the source by way of an intrinsic body diode while the switch is turned off
Implementation Method 2
measuring a second voltage of the body diode after turning off the semiconductor device; determining a resistance value using the second voltage measurement
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Unique systems, methods, techniques and apparatuses of semiconductor failure prognostication. One exemplary embodiment is a power converter (123) comprising a semiconductor switch (110) and a converter control system (121). The converter control system (121) is configured to turn on the semiconductor switch (110), measure a first voltage and a current during reverse conduction, estimate junction temperature of the semiconductor device (110), turn off the semiconductor device (110), measure a second voltage after turning off the semiconductor device (110), determine a resistance value using the second voltage measurement, determine an expected resistance value, predict a failure of the semiconductor device (110) using the resistance value and the expected resistance value, and transmit a semiconductor device failure warning.