White Papers
Thermal Transient Characterization of Single and Stacked-Die
This paper proposes an unambiguous definition for the junction-to-case thermal resistance, based on a transient measurement technique with much higher repeatability for very low thermal resistances compared to existing two-point measurement methods. The technique is illustrated on thermal transient measurements of power MOSFETs. The concept is extended to multi-chip and stacked-chip structures, where transfer impedances have to be introduced.
More White Papers
Ten Good Reasons Why Thermal Measurements are Important to Your Design
The Mentor Graphics MicReD T3Ster, a best-in-class thermal measurement system, is fully equipped to deliver the efficiency, repeatability, flexibility, and ease of use needed for this application. There are at least ten good reasons to include thermal measurements as a routine step in any electronic component or system design process.
Amid all the promotion of solid-state superlatives ranging from data rate to feature size to LED light output, one characteristic is never touted: junction temperature. That's because junction temperature (TJ) is an undesired but unavoidable side-effect of high currents and/or switching speeds. A p-n junction, whether it is one of millions on a CPU chip or the only one within a power LED, generates heat. In the past two decades the industry has seen heat dissipation increase by orders of magnitude. Faster is better, but faster is also hotter. This trend is not without consequences. A 10° increase in TJ can cause a 50% reduction in a semiconductor device's life expectancy. Differing thermal expansion responses where dissimilar packaging materials meet can degrade and ultimately destroy devices.
In LEDs, both brightness and color can suffer as TJ increases. And of course the twin issues of safety and cooling can impact the design of an entire system, not just the semiconductor device producing the heat. All these facts point toward the need for a thorough grasp of thermal behaviors at the chip level, and beyond that to thermal interface materials (TIM) and even heat sinks. True understanding comes with physical measurements performed on actual devices.
Thermal Characterization Confirms Real-World LED Performance
LED thermal characteristics, beginning with the PN junction and extending to the ambient environment, must be well understood in order to ensure a safe, reliable design and satisfactory performance. There may be multiple thermal interfaces such as die attach or glue layer in the heat flow path, and their thickness and resistance can be difficult to control in manufacturing. Moreover the thermal interface between the LED package and the luminaire acts as a heat-sink from the LED’s perspective, further complicating the design challenge. Thermal resistance must be understood as early as possible in the prototype phase.
Thermal Analysis of Memory Module Using Transient Testing Method
The electrical transient testing method has become popular as a useful thermal analysis tool because of its accuracy, high repeatability and rich information content compared to the use of traditional steady state thermal characterization techniques. This paper presents a thermal study of a 16-chip memory module using transient testing. The two variables in this study are the thermal boundary conditions of and the power distribution within the module. By applying the method of network identification by deconvolution (NID) to a transient temperature measurement, we can identify the structure function, which is the dynamic thermal resistance versus capacitance along a particular heat flow path for a given boundary condition and power distribution. Comparisons of the structure functions reveal differences in the heat flow paths for the cases of one chip and multiple chips dissipating heat. We have successfully used transient testing on a three-dimensional memory module, and determined the contributions to the overall dynamic thermal resistance by each of the components including the heat spreader (HS), socket and even thermal interface material (TIM). This information about a 3-D assembly is often difficult to obtain using steady state techniques. Thermal engineers can use such information to differentiate the relative merit of materials and heat transfer mechanisms in a cooling solution to optimize the overall thermal budget.
When Designing with Power LEDs, Consider Their Real Thermal Resistance
As of today, there exist no widely accepted international standard for high power LED testing. Many vendors follow the JEDEC JESD51 series of standards for obtaining thermal metrics of their LEDs, but in calculating the thermal resistance they do not count with the emitted light energy. On the other hand, when the light output of an LED is measured following the recommendations of the CIE 127-2007 document, thermal aspects are rarely considered properly. To overcome these problems the best practice would be to combine the recommendations of the above international standards. This way the real thermal resistance of power LEDs can be identified. Knowing this, light output metrics of power LEDs can also be obtained as function of the real operating temperature of LED chips' pn-junctions. This way design for the LEDs' actual 'hot lumens' becomes possible...
Dynamic Cooling Mount Compact Models for Board-Level Design
This paper tries to extend the DELPHI and PROFIT compact thermal modeling methodologies to allow complete dynamic models of device packages with compact models of cooling assemblies for the same purpose: co-simulation with a detailed board model. A few case studies are presented showing how such models can be constructed using structure functions and transient model fitting tools.
Electric and Thermal Transient Effects in Optical Devices
By studying electrical and thermal transients in high power LEDs this paper proposes a multi-domain "compact" model suitable for single devices and LED arrays in a board-level simulation environment. The thermal part of the model has been identified from structure functions extracted from measured thermal transients. Measurements were also carried out in a combined photometric/thermal measurement setup.
In case of opto-electronic devices the power applied on the device leaves in a parallel heat and light transport. For this reason the interpretation of Rth is not obvious. By studying electrical and thermal transients in high power LEDs this paper proposes a multi-domain "compact" model suitable for correct simulation of single devices as well as LED arrays in a board-level simulation environment. The thermal part of the model has been identified from structure functions extracted from measured thermal transients. Several measurements were carried out in a combined photometric/thermal measurement setup, which is proposed for the characterization of power LEDs. Transient simulation results compared to measured transients are also presented.
Non-Linearity Issues in the Dynamic Compact Model Generation
In this paper we show how to generate dynamic nonlinear compact thermal models. Algorithms with which network simulators can simulate nonlinear Rth and Cth elements are also given. Various experiments are presented, in which the magnitude of the error caused by neglecting the temperature dependence of the Rth and Cth elements of the dynamic compact thermal models is checked.
Thermal Issues in Stacked Die Packages
This paper discusses two major subjects: the qualification of die attach in stacked die structures and compact thermal modeling. An overview of the current techniques for die attach qualification in stacked structures for failure analysis is given. Finally, the state-of-the-art and the major issues in compact thermal modeling of stacked die packages is presented.
Thermal Measurement and Modeling of Multi-Die Packages
The paper gives an overview of thermal measurement and modeling techniques and results for stacked and MCM structures. Results for an opto-coupler device with 4 chips mounted in a combined lateral and vertical arrangement are presented. The paper also shows how to derive junction-to-pin thermal resistances directly from transient measurement results using structure functions.