3 years ago

Essais & Simulations n°127

  • Text
  • Ainsi
  • Projet
  • Solutions
  • Mesures
  • Conception
  • Composites
  • Simulations
  • Simulation
  • Mesure
  • Essais
spécial jec World : prendre la mesure des essais dans les composites

mesures Transparent 3D

mesures Transparent 3D view of our board model. Watch the drills underneath the package. component modeling The focus of TRM is to calculate the thermal behavior of the printed board. Components are treated as simple as possible, also because no information about the complex inner structure is available. Common practice is to create a rectangular volume and to assign a bulk conductivity together with a thermal power value. As a material definition for this type of QFN package we chose the entry “comp_diel_loc$TRM” from the TRM material database. The power dissipation for the experiment in Fig. 4 can be estimated from the data sheet and the text for Fig. 1 in Taranovich [4]: “The first test was as follows: Vin=12V, Vout=1V, Fsw=500kHz, Iout=40A with the board configured in a two-phase parallel single output mode (Mode 5A). Thermal imaging test results showed a maximum temperature of 99.6 C. See Figure 1.” When we look into the Data Sheet [1] we find on p.27 the graph shown in Fig. 7. Unfortunately the combination Vout=1V, Fsw=500kHz cannot be found. From other graphs in the Data Sheet [1] we infer that the efficiency at 500 kHz must be close to that at 300kHz. We decide to use a power loss of 8 Watts on the blue line at 40 A. Environment conditions Any thermal calculation needs as a reference temperature the external air temperature plus the cooling conditions. The Taranovich report [4] says: “Tests were performed using the evaluation board at room temperature with zero air flow.” Unfortunately he does not tell the mounting situation precisely: was the board placed vertically or horizontally and held within free air flow or did it lie on Graph for power dissipation from Data Sheet of ISL8240 [1] the table, etc. nor he gave the value of the room temperature (we wouldn’t believe that is was 39.6 C Fig. 1). In our thermal calculation we just assume an air room temperature of 20 C plus a heat transfer coefficient h=13W/m²K. To get this value is quite complicated. There are two fundamentals: first, semi-empirical formulae that describe the heat transfer by free convection air flow over the surface of a plate and, second, the Stefan-Boltzmann law of radiation. Both laws together lead to an equilibrium temperature of a homogeneously heated plate and from this the effective total heat exchange coefficient can be deduced. The procedure is coded in the TRM Heat Transfer Assistant module. basic result Fig. 8 shows the simulation result for a 2,1,1,2 oz board with an 8 Watt component at 20 C air temperature in free convection in steady-state. The maximum value in the calculated temperature field is around 99 C in a spot somewhere in the center of the component model. More or less this temperature can be interpreted as the surface temperature because of the choice of the bulk conductivity, with some uncertainties. Component. Maximum is 99.5 C Top layer. Maximum is 99 C 40 IESSAIS & SIMULATIONS • N° 126 • octobre - novembre 2016

mesures Bottom layer, viewed from top. Maximum is 95 C Simulated temperature of ISL8240MEVAL4Z with our basic parameters. Apart from the maximum temperature also features in shape and look of the infrared image are nicely reproduced. Edges and corners are caused by the artwork of the layers, which is the same in simulation and experiment. However, the infrared image is that of an uncoated board i.e. metallic surfaces have low emission and appear blue. TRM assumes coated board (emissivity 0.9) and component surfaces in order to be able to compare simulation with infrared imaging. This makes a strict quantitative comparison difficult. However, we can adapt the same color bar as was used in the thermograph. In TRM we can show the layout as a background picture. This was done (right) using the top layer. We may wonder why the yellow isotherms end abruptly at coordinate x=30mm (s. Fig. 8). This is caused by the insulation in the inner layer L2. Variations There are some major uncertainties in the data: • The unknown room temperature and heat transfer • The ambiguous thickness of the inner layers (1 oz or 2 oz?) • The unknown power dissipation of the component • The unknown case-to-board thermal resistance of the component • The unknown and non-uniform emissivity of the infrared image. How does the temperature change if we change parameter values? Table 1 gives values for the maximum temperature relative to the ambient air temperature. Table 1: Parametric study. Temperature above ambient in K. Parameter 1 Parameter 2 2,1,1,2 oz 6 oz 2,2,2,2 oz 8 oz = 6+30% 15W/m²K 6 oz 8W 80K (Fig. 8) 72K – 10.5W (8+30%) 105K 94K – 6W (8-30%) 62K 57K – 8W 80K (Fig. 8) – 74 • The maximum temperature relative to ambient is proportional to dissipated power. This implies that the thermal resistance to ambient does not change much with power. • Increasing the total thickness of layers by 30% leads to a decrease of relative temperature of -10% only. This is not predictable in a back-of-the-envelope calculation and a genuine 3D simulation result. • As in the theory of point like heat sources the maximum temperature relative to ambient is proportional to the square root of the heat exchange coefficient. ● J. Adam, Adam-Research, Leimen, Germany Acknowledgement We thank Intersil Corp. for permission to use the information and files found on their website for this article and on our websites. References [1] Data Sheet: Intersil/documents/isl8/isl8240m.pdf [2] Intersil Corp. reference-designs/isl8240mevalxz.html [3] Johnston, J.: “Carrying the Heat Away from Power Module PCB Designs” in http://www. (2015) and similar material in dam/Intersil/whitepapers/power-module/carryheat-away-from-power-module.pdf [4] Taranovich, S.: “Unique Intersil thermal design removes heat from encapsulated, compact 50A power modules” in electronics-products/electronic-product-reviews/ other/4439182/Unique-Intersil-thermal-designremoves-heat-from-encapsulated--compact-50Apower-modules (2015) [5] User Guide: Intersil/documents/an19/an1922.pdf ESSAIS & SIMULATIONS • N° 127 • Janvier-Février 2017 I41

Copied successfully!

Essais et Simulations - Découvrez la revue en version digitale

Pour vous abonner, joindre la rédaction, communiquer dans la revue ou simplement avoir plus d'informations sur votre métier, rendez-vous sur notre site internet :