At the beginning of this project a thorough comparison of existing modelling methodologies has been carried out. An empirical approximation of the ambipolar diffusion equation was chosen as the common basis for the time-dependent description of the charge stored in the lightly doped parts of power devices. The common approach proved to be successful for the establishment of compact models for all power devices.
For all power devices Levels 1, 2 and 4 models have been developed. Level 5 models including a simplified description of the dominant failure mechanisms have been created for the diode, bipolar transistor and MOSFET.
Parameter extraction schemes could be established for all parameters.
Six benchmark circuits as close as possible to important operational conditions of the different devices have been extracted, designed and set up. These circuits have been used for a comprehensive validation of all models developed. Except for the GTO, this process could be finalised as scheduled. All assessment activities together with the guidelines of using the models are collected in a validation library.
A multitude of industrial processes and products use power electronic circuits to synthesize the special supply voltages and/or waveforms required for their operation. Examples include motor controls (for robotics, factory automation, traction systems etc.), switch mode power supplies (for low-energy lighting, consumer electronics etc.), uninterruptable power supplies (used to provide back-up supplies for data processing systems etc.) and driver stages (used in deflection systems for televisions, monitors etc.). However, despite their importance, these power sub-systems are still designed through a series of prototyping cycles, even though CAD techniques have been shown to be effective in reducing both the time-to-market and the overall design costs of other electronic products. This difference in design methodology arises because the computer models, on which the CAD approach is based, do not reproduce the behaviour of the power components used in these sub-systems correctly. The first goal of the project is to extend the existing benefits of CAD to these power conditioning sub-systems by developing new computer models for a range of important power devices. The second goal is to provide a new design capability, that of reliability simulation and optimisation, by further enhancing these models so that they are able to predict their own failure. The use of computer-aided design is seen as essential to reducing significantly the time taken to introduce new products to the marketplace, a strategic goal of many European companies who wish to improve their competitive edge. Pressure for faster innovation also motivates the need for computer-aided reliability analysis as existing techniques are often time-consuming and yet provide no real understanding on the cause of failure.
Funding SchemeCSC - Cost-sharing contracts
RH1 5HA Redhill
SA2 8PP Swansea