Heat transfer: Electronic design
Overview
The aim of this test case is to validate the following:
- Heat transfer in electronics
- Transient analysis
The simulation results of SimScale are compared with the results presented in [Bruce]. The mesh used was created using first order tetrahedral elements with local mesh refinement algorithm on the SimScale platform.
Open the validation project into workspace
Geometry
The current analysis is carried out on a high-power IC package that is attached between the heatsink base and the PCB substrate as shown in figure below. The components explicitly represented in the model are the die, TIM1, lid, TIM2 and the heat sink base (where TIM = Thermal Interface Material). To proceed with the simulation, only a portion of the geometry is considered with width and length of all components equal to that of the die.
The simulation ready CAD model with its specific dimensions are as shown below.
| Component | Length | Width | Thickness |
|---|---|---|---|
| Die | 13 mm | 13 mm | 0.50 mm |
| TIM1 | 13 mm | 13 mm | 0.10 mm |
| Lid | 13 mm | 13 mm | 0.50 mm |
| TIM2 | 13 mm | 13 mm | 0.05 mm |
| Heat sink base | 13 mm | 13 mm | 6.00 mm |
Analysis type and Domain
Tool Type : Code_Aster
Analysis Type : Transient heat transfer
Mesh and Element types :
| Mesh algorithm | No. of nodes | No. of 3D elements | Solver |
|---|---|---|---|
| Tetrahedral with local refinements | 327326 | 1045594 | Code_Aster |
A tet-dominated mesh elements with fine local mesh refinements at the interfaces between the components are as shown in the figure below.
Simulation Setup
Material:
| Component | Material | Thermal conductivity (W/mK) | Density (kg/m3) | Specific heat (J/kgK) |
|---|---|---|---|---|
| Die | Silicon | 111 | 2330 | 668 |
| TIM1 | Ag-Epoxy | 2.0 | 4400 | 400 |
| Lid | Copper | 390 | 8890 | 385 |
| TIM2 | Grease Aluminium filler particle | 1.0 | 2500 | 900 |
| Heat sink base | Copper | 390 | 8890 | 385 |
Initial Condition:
- Temperature = 273.15K
Heat Flux Loads:
A power of 1 W is applied to the top surface of the die which is in contact with the TIM. Therefore, the surface heat flux that is applied on the top surface of the die is calculated by dividing the power with respect to the die surface area.
- Surface heat flux = 5917.1598W/m2 (top of the die)
Convective heat flux:
The cooling effect of the heatsink fins are collectively represented through a heat transfer coefficient that is directly applied to the heatsink base, top surface. Therefore, a convective heat flux boundary condition is used to represent the thermal resistance offered by the heatsink to the surrounding air.
- Convective heat flux = 20000W/m2K (heat sink to air top surface)
Results
Most of the temperature rise occurs at the TIM surfaces. The temperature distribution of the IC package is as shown below.
Comparison of the SimScale results with that of [Bruce] is as shown below.
Temperature change w.r.t ambient temperature versus time plot comparison with [Bruce] and SimScale simulation results for 3 different location of the model, Die-top, Lid-top and heatsink-bottom
A slight deviation in temperature graphs from [Bruce] with SimScale is due to the approximation error caused during the extraction of temperature values from the digital plots.
References
| [Bruce] | (1, 2, 3, 4) Bruce Guenin, “Calculation Corner: Transient Thermal Modeling of a High-Power IC Package, Part 1” Calculation Corner, Computer, IT Products, Number 4, Software/Modeling, Volume 17, December 22, 2011 |