Thermal Via Array Design for Power Devices

A surface mount power device dissipates heat into a pad that is only a few square millimetres. Without a path through the board, that heat has nowhere to go except laterally through thin copper and down through still air. A thermal via array is the structure that carries it to the other side, and its design determines the thermal resistance of the whole path.

How Heat Moves Through the Board

Heat enters the pad, spreads laterally through the copper, and then has to cross the laminate to reach a cooler surface. The laminate is a poor conductor, so the crossing is the bottleneck. A via filled with plated copper provides a much better path through the same thickness.

An array of such vias behaves as a composite conductor whose effectiveness depends on the copper cross section of the vias, their number and how well the copper at each end spreads the heat into them. A single via is nearly useless; an array with inadequate copper at the ends is not much better.

Choosing Via Count and Spacing

More vias lower the thermal resistance, but with diminishing returns, because the limiting factor eventually becomes the spreading resistance in the pad above them. The useful range is usually a spacing of about one to one and a half millimetres on a grid, which for a typical power pad means a dozen or more vias.

The vias should be placed under the device footprint rather than around it, since heat flow follows the area of the heat source. Placing vias outside the footprint adds thermal mass without providing a direct path.

Copper Area at Each End

A via array needs a copper area to collect heat at the top and to spread it at the bottom. On the top side that area is the device pad, which is fixed. On the bottom side it is a copper plane or a dedicated pad, and its size determines how effectively the heat is dispersed.

A bottom side plane that is much smaller than the array is a common design flaw, because it re-concentrates the heat that the array worked to spread. The plane should be at least as large as the array, and preferably considerably larger. The wider thermal strategy is described in our article on thermal management design.

Via Construction and Fill

Vias under a pad are usually tented, plugged or filled, depending on whether paste must be printed over them. An open via in a pad draws paste away during reflow and produces a starved joint, so the array should be filled and capped if the pad is to be soldered normally.

Fill material affects the thermal path as well. A resin plug conducts far less heat than a copper filled via, so thermal arrays on high power boards are often copper filled. The manufacturing detail and the surface requirements are described in our article on via filling and plugging.

Thermal Resistance and Verification

The thermal resistance of the array can be estimated from the copper cross section, the board thickness and the number of vias, and the result can be compared with the target for the device junction temperature. A calculation is enough to choose the array; a measurement confirms it.

Measurement means running the device at a known power and observing the junction temperature or the board surface temperature with an infrared camera. The difference between the calculated and measured values indicates whether spreading or the interface dominates, and the measurement approach follows the principles in our note on thermal measurement accuracy.

Assembly Considerations

An array of vias under a pad changes how the pad heats during reflow. The copper of the vias conducts heat away from the joint, so a pad with a large array needs more energy to reach soldering temperature than one without.

The usual remedy is a preheat that allows the array to reach temperature gradually, or a longer soak. Where the device is sensitive to the total heat it sees, that constraint has to be balanced against the thermal requirement, and both should be considered in the same design review.

Alternative Substrates

Where the thermal load exceeds what a via array can carry, the board itself has to change. A metal core substrate conducts heat directly and can remove the array from the calculation, and a ceramic substrate does the same with better electrical isolation. Both change the assembly process.

The comparison is worth making early, because a design that pushes a conventional board to its thermal limit may cost less overall on a metal core. The trade-offs are set out in our article on metal core and ceramic substrates.

Checks Before Release

The checks that matter are the ones performed on the product rather than on a sample kept for the purpose, because a coupon that travels with the panel is the only evidence about that panel. Where an operation cannot be verified afterwards, it has to be controlled during the operation, and that control has to be visible in the record.

Consumables have a life measured in cycles, and the replacement point should come from the measurement rather than from a failure. The sequence of operations is part of the specification, because a different order produces a different result from the same steps.

Where the process window is narrow, the measurement resolution has to be better than the window, or the data cannot distinguish a good part from a marginal one.

FAQ

How many thermal vias are enough? Enough that the array is no longer the dominant resistance. Beyond that the spreading resistance in the pad sets the limit.

Do thermal vias need to be filled? If paste is printed onto the pad, yes. An open via wicks alloy away from the joint and leaves it starved.

Does the via array help if the bottom side is insulated? Much less. The heat has to reach a surface that can dissipate it, so the bottom side copper and any airflow matter as much as the vias.