This video demonstrates how to account for thermal contact resistance in your designs and the importance of doing so. The model is a thyristor bridge output module from an industrial motor control.The presentation includes a comparison of temperature results for with and without contact resistance. It also compares the results with and without electrical insulators installed between the semiconductors and heat sink.

Video Narration Script (click to view)

 

0:05	This is one of three output modules belonging to a full-wave, half-controlled, thyristor bridge – specifically, one designed to drive a 20 horsepower, 1770 RPM, 460 VAC, 3-phase electric motor. We will be performing a Mechanical-Thermal analysis to determine the steady-state operating temperatures when driving the motor at 15% overload.
0:28	In this sectioned view, you can see the internal objects comprising the components. The semiconductor dies are the heat source.
0:36	To maximize thermal conduction to the heat sink, the components are mounted without insulators, so the heat sink is electrically hot and must be insulated from the cabinet.
0:46	For time and convenience, several operations have already been performed, as follows:
0:51	1. Appropriate materials have been assigned to all objects.
0:56	2. The objects in contact with the heat sink have been imprinted on it for the purpose of isolating contact areas from the exposed portion of the front and back mounting faces.
1:07	Later, we will assign convection boundaries to the exposed faces of the heat sink. Currently, user-assigned contact areas are not automatically excluded from convection boundaries, as they are for default bonded contact areas. So, the mounting faces had to be split manually. Our convection assignment will exclude these contact faces.
1:29	To keep the objects used as imprinting tool, the option, “Clone tool objects before operation” was selected before completing the Imprint command.
1:39	3. A total of four face selections have been created to accommodate the selection of assignment faces for contact and for convection boundaries.
1:48	We’ll define different thermal resistances for the contact faces on the front and back sides of the heat sink. Thermal grease is applied between the component bodies and the front contact faces to reduce thermal resistance. These are the primary heat flow faces since heat generation occurs on the front side of the module. No heat sink compound is applied to the stud and nut threads, between the washer or terminal and the nut, or between the washer or terminal and the heat sink. Accordingly, we’ll define a higher thermal contact resistance for the back contacts.
2:22	We will also be assigning two different convection boundaries to account for the fins having grooved faces on their left and right sides.
2:29	V-grooves increase the surface area for heat dissipation by about 41% relative to smooth faces. Rather than complicating the geometry with many additional and very thin faces, we can account for the increased surface area by raising the convection film coefficient while leaving the model’s faces smooth.
2:48	4. A length-based mesh operation has been assigned inside all objects, specifying a maximum length of 4 mm.
3:02	5. The ambient temperature variable has been increased from its default value to 30 Celsius.
3:09	6. An analysis setup has been added using default settings.
3:15	To obtain realistic simulation results, you must consider the thermal resistance that exists between adjacent objects. Roughness along faces of manufactured parts, exaggerated in this image for clarity, reduces the area of true metal-to-metal contact to an amount much less than the apparent contact area. Thermal grease applied to the parts fills in the surface irregularities and improves heat transfer. Nonetheless, even with thermal grease applied, contact resistance is significant. In electronic assemblies, temperature step-changes occur across component-to-heat sink interfaces.
3:50	If you don’t assign contacts, adjacent objects are bonded by default and there is no thermal resistance between them. That is, the thermal conductivity from one part to the next is perfect, and the temperatures are identical at both sides of the contact face. Therefore, you could significantly underestimate the temperature of the electronic components, especially if an insulator is used. In that case, the heat must flow through two contact interfaces and through the thickness of the insulator.
4:19	Let’s select the front contact faces and assign contact. Note that, where objects meet, you only select the contact face of one of the two objects. The target face on the adjacent object is found automatically by the solver. For front contact, we are choosing the two imprinted circular faces on the front of the heat sink.
4:39	Specify Contact_Front as the Name.
4:43	For the Resistance Type, choose Thermal Impedance. This option is a distributed value for which the total resistance varies depending on the contact area. It is the inverse of the thermal conductivity per unit area. Specify 25 Celsius mm squared per watt, which is a reasonable assumption for smooth surfaces with thermal grease applied and with a relatively high clamping force.
5:08	Repeat the procedure for the back contact faces. For this assignment, we are choosing the front and back faces of the washer and terminal (for heat sink and nut contact, respectively) and the surface of the hole in each nut (for nut-to-stud contact).
5:23	This time, specify Contact_Back as the Name and 150 Celsius mm squared per watt as the Thermal Impedance. We are assuming that the contact resistance of a smooth, bare metal-to-metal interface is approximately six times greater than that of a similar interface with thermal grease applied.
5:44	The remaining object interfaces will all be treated as bonded contact without thermal resistance, which includes internal contact between all objects comprising the diode and SCR.
5:55	Next, select the smooth convection faces and assign a convection boundary.
6:00	Specify a uniform Film Coefficient of 10 watts per meter squared Celsius.
6:09	Repeat this procedure for the grooved convection faces but this time, specify 14.1 watts per meter squared Celsius as the Film Coefficient. As mentioned previously, this value is 41% greater than the other convection boundary to account for the increased surface area of grooved faces.
6:28	Next, select the SCR Die and assign a Heat Generation with a Total Power of 35 watts. The wattage is based on a current of 28 amps and a forward voltage drop of 1.25V through the SCR.
6:51	Repeat this procedure for the Diode Die, but this time, specify a Total Power of 21 watts, because the forward drop of the diode is about 0.75 volts, and the current is the same.
7:03	We’re ready to run the analysis, which takes about 3 minutes to complete on a current high-end workstation.
7:15	Select all objects and overlay the temperature results.
7:26	It takes a few moments to read in the values. Press F6 to display the geometry as a wireframe, suppressing the model shading for a cleaner contour plot.
7:31	Press F6 to display the geometry as a wireframe, suppressing the model shading for a cleaner contour plot.
7:39	The minimum temperature is approximately 63 Celsius, putting the coldest part of the heat sink 33 degrees above ambient. The maximum temperature occurs at the SCR Die and is about 113.4 Celsius, which is a typical junction operating temperature for power semiconductors.
7:58	Hide the temperature overlay, select all objects again, and create a mesh plot.
8:12	As you can see, the mesh uniformity and quality appear to be good.
8:18	Next, let’s create a Fields Summary listing the minimum, mean, and maximum temperatures for each of the objects in the model.
8:37	You can see, for example, that the temperatures within the heat sink range from about 63 to 75.8 Celsius and the temperatures in the SCR Die range from about 104 to 113.4 Celsius.
8:53	Finally, to better quantify the importance of considering thermal contact resistance, two variants of the original model were created. This design shows the results if thermal contact is not assigned, and all objects are in perfect bonded contact. The maximum temperature of 106.8 C is about 6.6 degrees cooler than the version with thermal contact resistance considered.
9:17	And this design shows the results if 3 mil thick mica insulators are added to electrically isolate the SCR and Diode from the heat sink. The Thermal Impedance was increased to include the effects of two contact interfaces (at the front and back of each insulator) plus the impedance of the mica itself, which has a much lower thermal conductivity than metals. The resultant temperature of 148 C is 34.6 degrees hotter than the thermal contact case without insulators and 41.2 degrees hotter than the case with perfect contact throughout.
9:52	Thermal contact resistance is easy to account for in Mechanical–Thermal simulations, and we hope you appreciate the importance of doing so. Use this tool to ensure the reliability of your circuit and system designs. Thank you for watching this video.