How to optimize PCB layout to improve EMI performance

Power supply design engineers often use some DC / DC buck converters in automotive systems to support multiple power rails. However, there are several factors to consider when choosing these types of buck converters. For example, on the one hand, a high switching frequency DC / DC converter (working frequency higher than 2 MHz) needs to be selected for the car infotainment system / host unit to avoid interference with the radio AM band; on the other hand, it is also necessary to select a relatively small Inductors to reduce solution size. In addition, high switching frequency DC / DC buck converters can also help reduce input current ripple, thereby optimizing the size of the input electromagnetic interference (EMI) filter.

However, for large automotive original design manufacturers (ODMs) trying to create the latest automotive systems, meeting the required EMI standards is critical. These requirements are very strict and manufacturers must comply with standards such as the International Special Committee on Radio Interference (CISPR) 25 standard. In many cases, if the manufacturer does not meet the standard, the car manufacturer cannot accept the corresponding design.

Therefore, for the EMI performance improvement of DC / DC step-down converter, PCB layout is very important. To obtain good EMI performance, it is critical to optimize the high-current power loop and reduce the influence of parasitic parameters on the loop.

Take the two-output DC / DC step-down converter composed of LMR14030-Q1 as an example, as shown in Figures 1 and 2 of two different printed circuit board (PCB) layouts. The red line shows how the power circuit flows in the layout. The flow direction of the power circuit in FIG. 1 is U-shaped, while the flow direction in FIG. 2 is I-shaped. These two layouts are the most common layouts in automotive and industrial applications. So, which layout is better?



Conducted EMI is divided into two types: differential mode and common mode. Differential mode noise originates from the rate of change of current (di / dt), while common mode noise originates from the rate of change of voltage (dv / dt). And whether it is di / dt or dv / dt, the key point of EMI performance is how to minimize the parasitic inductance.

Figure 3 is an equivalent circuit of a buck converter. Most designers know how to minimize the parasitic inductance of Lp1, Lp3, Lp4, and Lp5 in high-frequency loops, but ignore Lp2 and Lp6. For two different layouts, U-type and I-type, the parasitic inductance on Lp2 and Lp6 of the U-type layout is smaller than that of the I-type layout. In a U-shaped layout, reducing the power loop when the switch Q1 is turned on will also help improve EMI performance.


In order to verify the best layout, measuring EMI data becomes critical. Figures 4 and 5 compare the conducted EMI of a two-output converter. At the same time, the circuit uses phase shift control to reduce the input current ripple, thereby optimizing the input filter. It can be seen from the test results that the EMI performance of the U-type layout is better than the EMI performance of the I-type layout, especially in the high-frequency part.


Adding EMI filters can effectively improve EMI performance. Figure 6 shows a simplified version of an EMI filter that includes a common mode (CM) filter and a differential mode (DM) filter. Generally speaking, the noise of the differential mode filter is less than 30MHz, and the noise of the common mode filter ranges from 30MHz to 100MHz. Both filters affect the entire frequency band that EMI needs to limit. Figures 7 and 8 compare the conducted EMI with common-mode and differential-mode filters, respectively. U-shaped layouts can meet CISPR 25 Class 3 standards, while I-shaped layouts do not.