Motor Controller Heat-Sink Fin Optimisation

The problem

Motor controllers lose a few per cent of their power as heat in the switching devices. If the heat sink is too small, junction temperatures rise, the controller derates or fails. Fin design is a classic trade-off: more fins add area but block airflow. A parametric CFD study lets students find the best fin geometry instead of copying one.

Objectives

• Estimate device losses and set a temperature target for the heat sink base.
• Model a parametric extruded heat sink with a design table for fin count, thickness and height.
• Run natural convection and forced-air studies in Flow Simulation with conjugate heat transfer.
• Find a fin geometry that meets the target with minimum mass and check it against a hand calculation.

What you need

Key concepts

Thermal resistance chain

Junction temperature equals ambient plus losses times the sum of junction-to-case, case-to-sink (thermal pad) and sink-to-air resistances. CFD gives the sink-to-air part.

Optimum fin spacing

In natural convection, fins too close together choke the buoyant flow and fins too far apart waste area. Correlations such as Elenbaas give a starting spacing to check against CFD.

Contact resistance

The thermal pad or grease between device and sink adds resistance. Flow Simulation can model it as a contact resistance condition on the interface.

Conduction and switching losses

Conduction loss is roughly I²·RDS(on) for a MOSFET; switching loss depends on switching frequency and the energy per transition from the datasheet.

Step-by-step guide

1. 1

   Estimate losses and set targets

   Take an example 48 V, 3 kW scooter controller with six MOSFETs. From the datasheet, estimate conduction loss using an example RMS phase current and the hot RDS(on), and switching loss from the switching frequency and transition energy. Sum to a total, for example 60 to 90 W, and record it as an estimate. Set an example target: device case temperature under 90 °C at 40 °C ambient, leaving margin below the junction limit after the junction-to-case step.

   Checkpoint: A loss table per device with total loss, and a case-temperature target that includes the junction-to-case drop.

2. 2

   Model the parametric heat sink

   Sketch the heat sink profile: base plate plus one fin, with dimensions for base thickness, fin thickness, fin height and pitch. Extrude to length and create the fins with a Linear Pattern whose instance count is linked to a dimension. Add a mounting pad area and tapped holes for the devices. Set the material to an aluminium alloy such as 6063 from the SOLIDWORKS material library, the common extrusion grade.

   Checkpoint: Changing fin count or thickness rebuilds the part cleanly and Mass Properties updates.

3. 3

   Build a design table

   Insert a Design Table with columns for fin count, fin thickness, fin height and base thickness. Add rows for a spread of options, for example 8 to 20 fins, 1.5 to 3 mm thick and 25 to 45 mm tall, keeping the overall width fixed. Check that every row has a fin gap above a manufacturable minimum for an extrusion, which you should confirm with an extruder's guideline, and that no row produces fins thinner than the die allows.

   Checkpoint: Every configuration in the table builds without errors and has a positive fin gap.

4. 4

   Assemble devices and pad

   Make simple TO-247 or module-shaped device bodies and mate them onto the mounting pad. Between device and sink, either model a thin pad solid with its own conductivity or apply a Contact Resistance condition using the pad's thermal resistance per area from its datasheet. Add a simple PCB plate above the devices only if it affects airflow. Keep the controller housing out of the first study.

   Checkpoint: Interference Detection shows the devices sitting flush on the pad with no overlap.

5. 5

   Set up the natural convection study

   Run the Wizard as an External analysis with Heat Conduction in Solids and gravity on, pointing in the real mounting direction because fin orientation matters in natural convection. Use air at 40 °C ambient. Apply a Surface Source or Volume Source equal to each device's loss. Make the domain large enough above the fins for the plume to develop, for example three times the sink height. Add Surface Goals for maximum device temperature.

   Checkpoint: The baseline run converges and a velocity cut plot shows a rising plume between the fins.

6. 6

   Set up the forced-air study

   Clone the project and add a fan or duct inlet. For a ducted case, model a shroud and use Internal analysis with a fan curve on the inlet and Environment Pressure on the outlet. For a simple case, use External analysis with a free-stream velocity, for example 3 m/s, along the fins. Add goals for device temperature and pressure drop across the sink. Keep the same heat sources and contact resistance as the natural case.

   Checkpoint: The forced-air run converges with device temperatures well below the natural-convection case.

7. 7

   Check mesh and hand calculation

   Refine the global mesh and add a Local Mesh over the fins so at least three cells span each fin gap. Confirm device temperature changes by less than 1 °C. Then do a hand check: compute sink-to-air thermal resistance from CFD as (sink base temperature minus ambient) divided by total loss, and compare with a correlation-based estimate for a finned plate. Expect agreement within about 20 to 30 per cent, given the correlation's limits.

   Checkpoint: Mesh change under 1 °C and a CFD thermal resistance in the same range as the hand estimate.

8. 8

   Run the parametric study

   Use the Flow Simulation Parametric Study to run each design-table configuration for both natural and forced cases, or run batch jobs with Batch Run. Collect maximum device temperature, pressure drop and mass. Plot temperature against fin count for each fin thickness. For natural convection you should see a minimum at an intermediate fin count; for forced air, more fins usually help until pressure drop limits the flow.

   Checkpoint: Plots showing an optimum fin count for natural convection and the pressure-drop penalty for forced air.

9. 9

   Choose the design and check junction temperature

   Pick the lightest configuration that meets the case-temperature target in the worst case you must cover. Add the junction-to-case resistance from the datasheet to get junction temperature and confirm margin below the device limit. Check the heat sink fits the controller housing and that fin tips do not clash with connectors. Run Interference Detection on the controller assembly, and leave room for a screwdriver to reach every device mounting screw.

   Checkpoint: A chosen design with junction temperature below the datasheet maximum with a stated margin.

10. 10

    Drawings and report

    Make a drawing of the extrusion profile with critical dimensions and ISO 2768 general tolerances, plus the machined mounting holes and flatness callout on the device pad. Write up losses, set-up, mesh check, hand calculation, parametric plots and the chosen design. Recommend a thermocouple test on a powered controller to validate the prediction, and note that losses are estimates that should be refined from measured currents.

    Checkpoint: A drawing ready for an extruder quote and a report with plots, chosen design and test plan.

Testing & validation

• Mesh refinement changes device temperature by less than 1 °C.
• CFD sink-to-air thermal resistance agrees with a correlation estimate within the correlation's accuracy.
• Energy leaving the sink surfaces matches the total device loss.
• Natural convection results show the expected optimum fin spacing.

Expected results

A parametric heat sink with a design table, natural and forced convection results across fin geometries, a chosen design that meets the temperature target with the least mass, and an extrusion drawing.

Take it further

• Add a transient study for a short peak-power event using device thermal capacitance.
• Compare a pin-fin or skived-fin design with the extruded fins.
• Model the full controller housing and its vents.
• Instrument a real controller with thermocouples and compare with the prediction.

Viva and review questions

Learn it at DIYguru

These DIYguru courses and programmes teach what this project needs.