Magnetics Placement and Layout on Switching PCBs

Magnetics are the largest components on most power boards and the ones that cause the most layout trouble. An inductor radiates a magnetic field that couples into any loop it can reach, a transformer couples between its own windings, and both dissipate enough heat to change the behaviour of whatever is placed next to them. Placement is therefore decided by field and thermal geometry, not by where the component fits. This article covers the rules that keep a switching design working after it is built.

What Magnetics Do to a Layout

Magnetics placement therefore has two separate goals that are often in conflict: keeping the field away from sensitive circuits, and keeping heat away from temperature sensitive components. Solving one by adding distance usually helps the other, which is why distance and orientation are the first tools to reach for and why a compact layout needs the most careful review.

An inductor stores energy in a field that extends well beyond the component body. That field induces a voltage in any conductor loop inside it, and the magnitude depends on the loop area and on the orientation of the loop relative to the field. A sense trace routed near an inductor can therefore carry a switching frequency signal that has nothing to do with the quantity it is supposed to measure.

The second effect is thermal. A magnetic component with a significant loss runs hotter than the surrounding board, and heat spreads by conduction through the copper and by convection through the air. A component placed downstream of that heat runs at a higher temperature than the design calculation assumed, which is a common cause of a converter that meets its specification on the bench and derates in the enclosure.

Stray Field Coupling and Loop Area

Stray field coupling is addressed by distance, orientation and loop area. Distance reduces field strength quickly, so moving a sensitive circuit a few millimetres away from an inductor has a large effect. Orientation matters because a loop whose plane is parallel to the field lines couples almost nothing, while one that is perpendicular couples strongly.

Loop area is the part of the problem the designer controls directly. A feedback or sense path that runs out to a divider and back forms a loop, and that loop acts as an antenna. Routing the forward and return paths together so that they enclose the smallest possible area is the standard remedy, and the general rules are set out in switching regulator layout.

Shielded and Unshielded Components

A shielded inductor contains most of its field, which reduces coupling to neighbours and to the ground plane. An unshielded one does not, so it needs more clearance and is a poor choice next to a sensitive analogue circuit. Shielding also changes the thermal behaviour, because a shield reduces convection from the winding.

The choice should follow the application rather than the price. Where the board is dense and the switching frequency is high, the reduced coupling of a shielded part is usually worth the cost and the temperature penalty. Where space is generous and the circuit is not sensitive, an unshielded part may be adequate if the clearance rules are respected.

Thermal Placement and Copper Design

Thermal placement starts with identifying which components are the heat sources and which are sensitive to temperature. An electrolytic capacitor has a limited life that halves for every ten degrees of additional temperature, so placing one next to an inductor is a direct reliability decision rather than a layout preference. A temperature sensitive reference or a crystal has the same problem in a different form.

Copper design determines where the heat goes. A large copper area under the component spreads heat into the board, which lowers the component temperature but raises the board temperature everywhere nearby. Thermal vias move heat to the other side, where it can be spread over a larger area. The trade offs are the same as those in PCB thermal management.

Keep Out Areas and Component Orientation

A keep out area around each magnetic component prevents the placement of sensitive parts and the routing of sensitive traces where the field is strongest. It should be defined in the library footprint rather than remembered by the designer, so that it is enforced automatically at placement time.

Orientation is the second control. Two adjacent inductors can be rotated so that their fields oppose rather than reinforce, which reduces the external field of the pair. A feedback trace should be routed so that the field crosses it at the least sensitive angle, and where that is not possible, the trace should be rerouted rather than left in place.

Sense Traces and Grounding

A sense trace carries a small signal and is vulnerable to any field it crosses. It should be routed as a differential pair where the measurement allows, kept short, and referenced to a ground that does not carry switching current. A sense trace that runs alongside a switch node will pick up capacitive coupling regardless of its orientation, so separation and shielding matter alongside loop area.

Grounding determines whether the coupling has a return path. A ground plane that is split between power and analogue sections, with a single defined connection point, keeps switching currents out of the sense reference. The principles that govern that split are covered in power integrity, and the connection point should be placed where the return current actually flows rather than where it is convenient on the drawing.

Verification and Common Mistakes

Verification is a combination of measurement and inspection. A near field probe moved over the board shows where the field is strongest and whether the sensitive circuits sit inside it. Measuring the sense signal with the converter loaded and unloaded shows whether the coupling changes with operating point, which is a strong indicator of a field problem rather than a component problem.

The gopcb engineering team reviews every power layout against a short list of mistakes: a sense trace inside an inductor keep out area, a feedback loop routed under the component, an electrolytic capacitor placed downstream of a hot magnetic, and a ground plane split that forces return current through a sensitive region. Each of these is cheap to correct at layout and expensive to correct after the board is built.

Layout Review Checklist

A short checklist catches most magnetics placement errors before the board is released. Confirm that every magnetic component has a keep out area in its footprint, that no sense or feedback trace enters it, that the loop area of every sensing path is minimised, and that no electrolytic capacitor sits downstream of a hot component in the airflow. Then confirm that the ground split forces return current away from the sensitive region.

The checklist should be applied at the layout review rather than at the prototype stage, because the corrections are nearly free before the board is fabricated. Where a design is dense enough that the clearances cannot be met, the answer is usually to change the switching frequency or the component selection rather than to accept the coupling, and that decision belongs to the circuit designer working with the layout engineer.

FAQ

Can a shielded inductor be placed next to a sensitive circuit? With a small clearance, usually yes. The field is reduced but not eliminated, so a keep out area is still required.

Does rotating an inductor help? Yes, particularly when two magnetic components are close together. The relative orientation changes the field that reaches the sensitive area.

How much clearance should be reserved? It depends on the component and the sensitivity of the circuit. Measure the field with a near field probe rather than relying on a general figure.