Does the mounting direction of the SMPS inductor affect radiation?

The amount of copper at the top-level SW node should certainly be minimized to limit antenna size. With a monolithic switching regulator (power switch inside the IC), the SW node goes from the IC all the way to the inductor, leaving a short trace on the top level. By using a controller (power switch external to the switching controller IC), the SW node can be independent of the switch and away from the IC. The SW node copper is connected to one side of the inductor in both buck and boost switching topologies. The layout of the Layer 1 SW node in the XY plane of the PCB or on an inner layer is tricky due to the numerous performance parameters involved (see Figure 1).

Figure 1. Highlighted SW node in the XY plane of layer 1 on the DC3008A LT8386 low EMI LED driver

Figure 2. The white stripe on the Coilcraft XAL inductor is a marking for the short coil lead because the coil leads are not visible. It indicates the orientation of the terminals and short leads. Connect here for high dv/dt for lowest EMI.

Of course, the SW node also extends vertically (in the Z plane) when the inductor terminals are considered. The vertical orientation of the inductor terminals may increase the antenna effect and radiation of the SW node. In addition, the internal inductor windings may not be symmetrical. Even though the symmetrical terminals of the inductor indicate a symmetrical structure hidden in the package, the polarity indication on the top of the component says otherwise. Figure 2 shows the internal winding structure of the Coilcraft XAL series inductor. The flat wire winding starts at the bottom of the component and ends at the top, so in the Z plane, one terminal ends up being much shorter than the other.

Additionally, an inductor with exposed SW nodes on the side may be worse than an inductor with shielded vertical metal, as shown in Figure 3. Board designers can choose an inductor with the fewest vertical exposed terminals to reduce EMI, but what about the orientation of the two inductor terminals and their relative impact on radiation?

The low radiated performance of the board under test is a combination of the IC radiated performance and layout considerations. Even with low radiated monolithic ICs, careful layout must be approached, taking into account the placement of critical radiating components. To demonstrate this, the effect of the orientation of the main inductor L1 of the LT8386 demonstration circuit on the board was examined (see Figure 4). In this case, the inductor manufacturer Coilcraft specifies the short terminal of the XAL6060 series inductor as marked with a white line above the component. Standard CISPR 25 conducted emissions (CE) and radiated emissions (RE) testing in an EMI chamber shows that the orientation of this inductor (see Figure 5) can seriously affect performance.

Figure 3. Pay attention not only to orientation but also to the type of inductor terminals on EMI-sensitive designs

Figure 4. Highlighted SW node in the DC3008A LT8386 low EMI LED driver schematic. Compare the full radiation results by placing the short-side terminals in Direction 1 and Direction 2.

Figure 5. Directional radiation test of Coilcraft XAL6060-223MEB inductor with DC3008A LT8386 LED driver. L1 Direction 1 (left), short terminal at SW node; L1 Direction 2 (right), long terminal at SW node. Radiation results are shown in Figures 6 to 8.

Figures 6, 7, and 8 show that the radiated performance of the DC3008A is directly affected by the orientation of L1 on the demo circuit, with no changes to other components. Specifically, for orientation 1, with the short-side terminal placed on the SW node, low-frequency RE (150 kHz to 150 MHz) and FM band CE (70 MHz to 108 MHz) have lower EMI. The 17 dBµV/m to 20 dBµV/m difference in the AM band cannot be ignored.

Not all inductors are created equal. Winding orientation, terminal shape, shape of terminal connections, and even core material may vary. The strength of magnetic and electric fields from different core materials and constructions may play a role in changing the inductor's radiation. However, this case study reveals an area of ​​concern that we can turn to our advantage.

Figure 6. Radiated emissions shows that the orientation of the inductor on the DC3008A has a significant impact on the results. The short side terminals are attached to the SW node to minimize the SW antenna (red), and the radiated emissions (RE) are significantly improved.

Figure 7. With the short side of the inductor attached to the switch node, the current probe method shows improved conducted emissions (CE) compared to the other polarity (>3 MHz)

Figure 8. With the short side of the inductor attached to the switch node, the voltage-based conducted emissions (CE) are improved compared to the other polarity (>3 MHz).

If the inductor manufacturer indicates the difference in internal terminal size with a silkscreen front marking or dots, it is easy to determine the orientation. If one of these inductors is selected for a design, it is wise to mark it on the PCB silkscreen, on the installation drawing, and even on the schematic. Unfortunately, some inductors do not have polarization or short terminal indications. The internal winding structure may be close to symmetrical, or there may be known structural differences. There is no malicious intent here - the manufacturer may not be aware of this specific mounting orientation distinction inherent in its product. Regardless, we recommend evaluating the selected inductor's radiation in both directions in a certified chamber to ensure repeatable high-performance measurements.

Sometimes, without external markings, the mounting orientation of an inductor is inevitably arbitrary, but the inductor is still required because of other parameters. For example, Würth Elektronik's WE-MAPI metal alloy power inductors are very small and efficient. Their terminals are only located on the bottom of the case. Each component has a dot on the top near the WE logo, but the data sheet does not specify this point as the starting point of the winding indication (see Figure 9). Although this may cause some confusion initially, the component has a fairly symmetrical internal winding structure and performance should be the same in both mounting orientations. Therefore, the dot on the top of the IC does not need to be indicated on the mounting silkscreen. However, if used in a circuit where EMI is critical, it is wise to test in both directions to confirm performance.

We tested the DC3008A with a high-performance Würth inductor, with the point on the top of the package and the start of its windings indicated in the data sheet (see Figure 10). For the form factor and current requirements of the LT8386, the 74439346150 15μH inductor was a good fit. Again, for comparison with the Coilcraft, we mounted this inductor in both orientations for radiated testing (see Figure 11).

The results (see Figure 12) are similar to those of the Coilcraft inductor. The radiation results show that the orientation of the inductor has a significant effect on the radiation. In this case, orientation 1 in Figure 11 is clearly the best orientation with the lowest radiation. Orientation 1 clearly radiates better in the lower frequency AM band (RE) and FM band (CE).

Figure 9. The WE-MAPI inductor data sheet does not give the winding start point, but there is a winding start point on the top logo of the component. These inductors may not have direction-dependent radiation effects, but this should be confirmed through testing.

Figure 10. The top marking on the WE-XHMI series inductors indicates where the winding starts.

Figure 11. Directional radiation test of Würth 74439346150 ("WE 150") inductor with DC3008A LT8386 LED driver. L1 Direction 1 (left), the short terminal start point of the winding is at the SW node; L1 Direction 2 (right), the long terminal is at the SW node. The radiation results are shown in Figure 12, indicating that the winding start point should be connected to the SW node for best results.

It is obvious that the inductor orientation has an impact on the radiation in a single switch node boost LED driver. We can assume that the SW node of the boost regulator has the same characteristic radiation because the power conversion and switching elements are the same in the voltage regulator and LED driver circuit.

We can also assume that the buck regulator has similar SW node design priorities to minimize the antenna effect of the inductor terminals. However, since the SW node of the buck regulator is closer to the input side of the converter, follow-up work may help determine whether the impact of the inductor orientation in the RE and CE regions is the same as for the boost regulator.

For dual switch node buck-boost converters, there is a bit of a dilemma. Common buck-boost converters, such as those in the LT8390 60 V synchronous 4-switch buck-boost controller family, have important low EMI features such as SSFM and a small hot-loop architecture. Single-inductor designs do not clearly reveal the effect of inductor orientation on radiation. If the short terminal is placed on one SW node, the long terminal on the other SW node will act as an antenna. In these designs, which orientation is best? What happens when all four switches are switching in the 4-switch operating region (V _IN is close to V _{OUT )?}

We will explore this in a future article - Testing EMI of a 4-switch buck-boost controller with two SW nodes at different inductor orientations. Just something to think about: maybe there are more than two options for this topology, 180° apart?

Figure 12. Radiated and conducted emissions show that the mounting orientation of the Würth 74439346150 high-performance inductor has a significant impact on the radiation results.

The orientation of the inductor in a switching regulator is important. When measuring emissions, pay attention to the inductor orientation and its repeatability—know how the selected inductor differs in these respects, test in both orientations, and clearly communicate possible mounting pitfalls to the board production department if the orientation is uncertain. It may be possible to improve emissions simply by rotating the inductor 180°.