Costs, Logistics And Best Practice – Thermal Design

Recent innovations in efficient, easy-to-use thermal interface materials and insulated metal substrates are helping companies reduce component count, bills-of-materials, assembly processes and overall product dimensions. Convenient thermal pads and fillers are also letting manufacturers reduce stockholding
and simplify inventory management. These are the claims of
The Bergquist Companys European marketing manager,
Nico Bruijnis: read on to discover more. Keeping cool

Heat is generated in all semiconductor components (notably processors and power devices) due to internal conduction and switching losses. Among other component groups, power resistors also generate large quantities of internal heat proportional to the square of the conducted current. This heat can destroy the device if it isnt removed sufficiently quickly to prevent the die, or resistive element, from overheating. In semiconductors, the temperature at the silicon junction is the critical parameter determining the devices survival. However, manufacturers usually recommend a maximum case temperature, since this is more easily monitored when the device is in-circuit. The maximum permitted case temperature is calculated knowing the thermal gradient from the die to the case surface.

Heat from the surface of the device may be transferred directly into the surrounding air. However, if the device is being driven hard, or is particularly small, the available surface area may not be sufficient to dissipate the thermal energy at the same rate it is being generated. Thus, the temperature of the die will steadily increase until the device is destroyed. A heatsink extracts thermal energy from the die and dissipates it into the surrounding atmosphere. In some cases, a heatsink alone may suffice. On the other hand, forced air cooling may also be necessary.

Reducing component count

Heat generated in power semiconductors such as thyristors, power MOSFETs and IGBTs often forces these devices to be operated below their maximum ratings, to enhance reliability and longevity. Several devices may be operated in parallel to distribute the load current, which is responsible for die heating. Optimal thermal design, removing heat quickly and efficiently from the device junction, lets engineers use the fewest possible devices to share the specified load current.

Eliminating still air from between heat-generating electronic devices and heat-dissipating components is a cost-effective way to optimise thermal performance, since still air is one of natures most effective thermal insulators.

Simplifying assembly

To achieve this, a thermally conductive material is generally inserted into the interface between a semiconductor device and an attached radiator. This ensures the two surfaces are in thermally efficient contact across the entire surface area. Thermal grease is a traditional solution, but a thin, uniform layer is required to reduce thermal resistance: around 65m. This can only be achieved via automation which introduces additional complexity and demands the purchase and maintenance of capital equipment.

Phase-change materials (such as Bergquist Hi-Flow) are non-tacky at room temperature and thus easier to handle than grease. As the temperature of the component reaches the phase-change temperature, typically 55C, the phase-change material flows to wet-out the entire surface. This eliminates any discontinuities in the thermal path. Hi-Flow products are available as standard or custom pre-cut pads in a range of thicknesses.

Smaller, simpler

Heatsinks are expensive, particularly in relation to the cost of thermal pads and gap fillers that significantly improve the overall thermal performance of an electronic product. Typically, if insufficient attention is paid to minimising the thermal resistance from the case to the heatsink, then an over-sized or complex heatsink must be fitted to compensate. Also, poor thermal design, requiring an unnecessarily large heatsink, may stop a product meeting its target form factor or external dimensions.
A thermally conductive adhesive, such as Bond-Ply, can replace the use of screws to attach and thermally couple devices such as power transistors and heatsinks. This also saves the manufacturer from arranging a constant feed of fixings to the factory floor, either from inventory or on a Kanban basis.

Low thermal resistance interface materials or a high thermal capacity substrate, such as IMS (Insulated Metal Substrate), are valuable in reducing heatsink size. In some cases an IMS (such as Bergquist Thermal Clad) can completely replace a conventional heatsink and let power components operate closer to their rated power.

In one application, revising a forklift truck motor controller using Thermal Clad IMS eliminated 66 through hole power MOSFETs, 15 high profile capacitors and 11 bus bars, in favour of just 48 surface mount MOSFETs, five low profile capacitors and four bus bars. This trimmed over 1kg from the assemblys total weight.
Heatsink designers use techniques including material selection, shape, surface area, special features, orientation and surface texture to improve thermal capacity and dissipation. However, high-performance heatsinks are expensive. If lower cost techniques (such as eliminating air gaps with thermal interface materials) can be applied, heatsink selection criteria can be relaxed, leading to BoM savings.

Sound thermal design can also avoid or reduce forced air cooling. Cooling fans are not designed-in lightly: they add to the BoM, increase mounting requirements and add to the products power budget. Optimising the thermal path from component case to ambient, while making best use of heatsinks via phase-change material or a thermally conductive electrical insulator (such as Bergquist Sil-Pad) reduce fan size and cost. In borderline situations, this may even eliminate the fan.

If the products enclosure can be used to dissipate heat, it can reduce the designers reliance on discrete heatsinks, thus reducing overall size and cost. Likewise, if a large metallic surface is close, it too could be used to sink heat. However, a low thermal resistance path must be established between the enclosure/heatsink and devices cases. Thermally conductive gap filling materials (such as Bergquist Gap Pad) can eliminate air gaps between a device and nearby cool surface.

Liquid gap filling materials, compatible with dispensing processes, are also available. These are useful for eliminating air gaps relating to: fragile components; devices with high topography; or applications where large tolerances in stack-up height prevent designers calculating the optimum pad thickness.
While cooling fans and complex heatsinks are eye-catching and suggest attention to detail, eliminating still air remains the single most effective goal of best practice thermal design.