Measurement and Characterization Technique for Real-Time Die Temperature Prediction of MOSFET-Based Power Electronics

This paper presents a technique to predict the die temperature of a MOSFET based on an empirical model derived following an offline thermal characterization. First, a method for the near-simultaneous measurement of die temperature during controlled power dissipation is presented. The method uses a l...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on power electronics 2016-06, Vol.31 (6), p.4378-4388
Main Authors: Davidson, Jonathan N., Stone, David A., Foster, Martin P., Gladwin, Daniel T.
Format: Article
Language:English
Subjects:
Controllers
Cross coupling
Current measurement
Diodes
Disengaging
Electronics
Energy dissipation
IIR digital filters
Mathematical models
MOSFETs
power control
Power dissipation
Power measurement
power MOSFETs
prediction methods
Real time
Semiconductor device measurement
Temperature effects
Temperature measurement
Temperature sensors
Transistors
Voltage measurement
Waveform analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper presents a technique to predict the die temperature of a MOSFET based on an empirical model derived following an offline thermal characterization. First, a method for the near-simultaneous measurement of die temperature during controlled power dissipation is presented. The method uses a linear arbitrary waveform power controller which is momentarily disconnected at regular intervals to allow the forward voltage drop of the MOSFET's antiparallel diode to be measured. Careful timing ensures the power dissipation is not significantly affected by the repeated disconnection of the power controller. Second, a pseudorandom binary sequence-based system identification approach is used to determine the thermal transfer impedance, or cross coupling between the dice of two devices on shared cooling using the near-simultaneous measurement and control method. A set of infinite impulse response digital filters are fitted to the cross-coupling characteristics and used to form a temperature predictor. Experimental verification shows excellent agreement between measured and predicted temperature responses to power dissipation. Results confirm the usefulness of the technique for predicting die temperatures in real time without the need for on-die sensors.
ISSN:0885-8993
1941-0107
DOI:10.1109/TPEL.2015.2476557