Robin Bornoff

Simulation-based design optimization methodologies applied to CFD

Finding the optimal physical design for an electronic system is extremely time-consuming. In this paper, we describe a sequential global optimization methodology that can lead to better designs in less time, and illustrate its use by optimizing the design of a heat sink for a simple system. The results show the need for a global approach, the insights that can be gained through automated design optimization, and illustrate the efficiency of the reported methodology in finding the optimum design.

Measurement based compact thermal model creation - accurate approach to neglect inaccurate TIM conductivity data

In this paper two possible ways are investigated to create accurate thermal models without having validated information on the thermal properties of the applied thermal interface materials. One way is the calibration of a detailed numerical thermal model based on the physical information which can be derived from experimental structure functions. In the paper we show a complete calibration procedure using a TO-220 package as an example. Another approach is the generation of dynamic compact models based on real measurements. In order to apply this approach one has to identify the junction-to-case thermal resistance of the tested package using the JEDEC JESD 51-14 standard.

An additive design heatsink geometry topology identification and optimisation algorithm

Established heatsink manufacturing processes such as extrusion and casting impose constraints on the methods used to design the heatsink. These affect both allowable geometry topologies and absolute sizes. The advent of 3D printing (additive manufacture) may remove many of these constraints, forcing us to reconsider the approach taken during design. This paper proposes and explores a new approach to heatsink design where the geometry topology is not defined a priori, but allowed to develop as part of an additive design process involving a number of sequential simulations.

Creating multi-port thermal network models of LED luminaires for application in system level multi-domain simulation using spice-like solvers

The multi-domain operation of LEDs characterized by the tight coupling between their electrical, thermal and optical properties manifests not only on chip and package level, but has to be tracked on LED products of higher integration level such as LED modules or LED luminaires. At present the system level numerical analysis of such integrated LED products is typically performed by the different design teams dealing with the electrical, thermal and optical design separately. One reason of this practice is the lack of a unified modeling approach in which the right models are used for the chip, package, module and luminaire (system) level. Spice-like LED models can address this issue. The present paper describes how thermal network models of luminaires can be created in an automated way to allow system level multi-domain simulations of complete LED luminaires.

Heat sink design optimization using the thermal bottleneck concept

Calculation and display of a thermal bottleneck scalar field as an integrated part of a CFD simulation enables a practitioner to interact with and understand the physical mechanisms by which heat is removed from an electronics system. By applying the characteristics of this thermal bottleneck scalar to heat sink design aspects, one can identify near optimal solutions with a minimal number of simulations. This work will detail the principles of using thermal bottleneck information to optimize fin thickness distribution and copper slug design and compare the results to that obtained by more traditional Design of Experiments and numerical optimization techniques.

Novel MOR approach for extracting dynamic compact thermal models with massive numbers of heat sources

A novel Model Order Reduction approach for the construction of Dynamic Compact Thermal Models is presented. With respect to previous approaches, this methodology allows reducing the complexity of the constructed models, from quadratically to linearly dependent on the number of independent heat sources. In such a way, the approach allows constructing Dynamic Compact Thermal Models practically without limitations on the number of heat sources. The proposed methodology is validated through the application to two state-of-the-art electronic systems.

A detailed IC package numerical model calibration methodology

Experimentally derived structure functions can be used to provide insights into the thermal resistances and capacitances heat experiences as it travels from a die through and beyond an IC package. A 3D “detailed” numerical model of the package that purports to explicitly represent the internal construction of the package requires material properties and geometric sizes to be accurately specified. A structure function derived from simulating the detailed numerical model can itself be compared to the experimentally derived reference example. Deviations between experimental and numerical SFs indicate error sites within the detailed model and an indication of whether the thermal resistances or thermal capacitances of the numerical would need to be increased or decreased to match the experimentally observed values. Iterative modifications of the detailed model, based on successive structure function comparisons, will achieve a fully calibrated detailed numerical package model.

Delphi4LED — from measurements to standardized multi-domain compact models of LED: A new European R&D project for predictive and efficient multi-domain modeling and simulation of LEDs at all integration levels along the SSL supply chain

There are a few bottlenecks hampering efficient design of products on different integration levels of the SSL supply chain. One major issue is that data sheet information provided about packaged LEDs is usually insufficient and inconsistent among different LED vendors. Many data such as temperature sensitivity of different light output properties are provided to a limited extent only and usually by means of plots. Also, reported light output properties are typically rated for a junction temperature of 25 °C, which is obviously much below the junction temperature expected under real operating conditions. Even if “hot lumens” measured at a junction temperature of 85 °C this is not the actual operating temperature and there is little information about how such “hot lumen” tests are performed. The gap between and reported LED test data and actual operating conditions can be bridged by proper simulation models of LEDs and their environments. Such models should be accurate, hence capable of proper prediction of LED operation but simple enough to assure fast numerical simulations. However, LED integration do not get access to detailed LED information to perform those simulation at system level, thus perform reverse engineering which is time and cost consuming. A bridge, in the form of standardization, has to be established between the semiconductor industry and the LED component integrators. In order to achieve this, the following tools have to be provided: · Generic, multi-domain model of LED chips · Compact thermal model of the LED chips `environment (including the package and module assembly) · Modeling interface towards the luminaire The goal of the project is to develop a standardized method to create multi-domain LED compact models from testing data.

Transient thermal characterization of a fcBGA-H device

In this paper, we describe a study in which the thermal performance data of Theta jc (Rth-JC), Theta ja (Rth-JA), and structure functions of a flip-chip ball grid array device with heat spreader, fcBGA-H, was measured. For Rth-JC, various boundary conditions for the thermal resistance modeling were considered and are discussed here. A transient measurement method was used to obtain the temperature responses of the diodes. The structure functions of the diodes were measured; and the thermal resistances were calculated. Furthermore, the effect of power map on the structure functions was studied, and a thermal simulation was conducted to match the simulated structure functions with the experimental structure functions. The matched simulation structure functions provides the most accurate thermal resistor network for system level thermal evaluation.

Lifetime estimation of power electronics modules considering the target application

The design of power electronics modules and power packages is heavily influenced by thermal concerns. New substrate materials, thinner and thermally more conductive attachment materials are used to decrease the thermal resistance of a given module. When a new material or technology is applied, its reliability has to be tested thoroughly before the module can be considered for production. The reliability or the expected lifetime of the module can be expressed by a number of temperature or power cycles the system can withstand. These numbers however make more sense if they can be linked to the foreseen lifetime of the application where the power module will be applied. In our changing industrial environment the design of power modules is not aimed at maximum lifetime anymore, but it is determined by the application itself. In this article we will summarize the steps which take the designer from the application requirements (mission profile) to the expected lifetime of the application considering the reliability of the power modules used.