Typical switching frequencies of 50 - 300kHz can be emitted by both differential and common-mode mechanisms. Lower frequencies are more prone to differential-mode emission while higher frequencies are usually worse in common mode. This section considers aspects of the design of the converter itself that can impact on the emissions profile.

The supply filter is a necessary part of all switching converter circuits, in order to attenuate the high frequency emissions that are inevitably present on the supply lines. Virtually all mains-powered apparatus is subject to some kind of limitation on these emissions, but even DC-powered switching supplies will often need some degree of supply filtering.

Capacitive coupling from high dv/dt

High dv/dt at the switching point (the collector or drain of the switching transistor) will couple capacitively to ground and create common-mode interference currents. The solution is to minimize dv/dt, and minimize coupling capacitance or provide a preferential route for the capacitive currents.

dv/dt is reduced by a snubber and by keeping a low transformer leakage inductance and di/dt. These objectives are also desirable to minimize stress on the switching device and inductor/transformer end winding, although they increase power losses. The snubber acts to damp ringing on the switching waveform, which is both an extra emissions source and an over-voltage threat to the switching device; the resistor should be sized to give critical damping to the LC circuit formed by the transformer leakage inductance L_lkg and the switching device output capacitance C₀, R = √(L_lkg/C₀).

Capacitive coupling is reduced by providing appropriate electrostatic screens, particularly in the transformer and on the device heatsink. Note the proper connection of the screen, to return circulating currents to their source, not to the chassis or external ground. Physical separation of parts carrying high dv/dt is desirable, or the offending component(s) can be given extra screening. Placing a safety-rated capacitor C1 across the transformer from primary to secondary DC nodes provides another means of local attenuation of the coupling due to the inter-winding capacitance C_TX. Acting as a capacitive voltage divider, this will attenuate transformer-coupled interference by a factor C1/C_TX, as long as the parasitic lead and track impedance connecting C1 across the transformer terminals is kept to a minimum. But a large value of C1 leads to large leakage current, therefore the capacitance value is limited due to the leakage current requirement (safety critical).

Single layer transformer windings need a proper electrostatic screen – a layer of copper or aluminium foil, broken in the winding direction to avoid a shorted turn – between the windings. Even if the transformer is not screened, its construction can either aid or prevent capacitive coupling from primary to secondary. Separating the windings onto different bobbins reduces their capacitance but increases leakage inductance. Coupling is greatest between nodes of high dv/dt; so the end of the winding which is connected to VCC or ground can partially screen the rest of the winding in a multi-layer design. 

Screening of the whole unit will ensure that capacitive coupling from individual nodes is referred to the screen rather than to the external environment.

As semiconductor switches become faster and semiconductor device packaging improves over the years, it is crucial to consider the switching device package when designing a circuit. Generally, through-hole devices like TO-247 have lower costs and are easier to design thermal heatsinks for. However, they often exhibit drawbacks from an EMC perspective.This is primarily due to the extra-long leads of through-hole devices, which inevitably introduce higher inductance compared to surface-mounted package types such as the Dpak. The figure illustrates simulated ringing and overshoot when the inductance of the device is reduced.

In cases where changing the devices is not an option at a late design stage, incorporating a small lossy ferrite ring core on the switching device, as shown in the picture, can help dampen the ringing and reduce noise. Though used in the 1990s, but it still finds its way into today’s designs: The core increases inductance, but the ring is damped. Therefore, the frequency and amplitude of the resonance go down. Also, the square loop material delays the current rise, making the switching softer.

Magnetic coupling

Magnetic field radiation from a loop which is carrying a high di/dt can be minimized by reducing the loop area or by reducing di/dt. Loop area is a function of layout and physical component dimensions. di/dt is a tradeoff against switching frequency and power losses in the switch. It can to some extent be controlled by slowing the rate-of-rise of the drive waveform to the switch, although the trend towards minimizing power losses and increasing frequencies goes directly against this requirement. If the output circuit is at a lower voltage than the input, it has to carry a higher di/dt; it is therefore as important, if not more so, to pay attention to the output as much as the input circuit loop area.

The leakage flux from the transformer or inductor can also be significant and is affected by the placement, orientation and construction of the core. No sensitive components or tracks should be routed near to this component. It is also important that the filter choke cores should not couple magnetically with the transformer leakage flux, as this will bypass the filtering components.

Screening the transformer will reduce the associated electric field, although it will have little effect on the magnetic field. A flux-cancelling band of copper tape around the outside of the transformer, in the same sense as the windings, reduces some of the leakage flux by its magnetic cancellation effect, due to the induced current in the band. The transformer (or inductor) core should be in the form of a closed magnetic circuit in order to restrict magnetic radiation from this source. A toroid is the optimum from this point of view; if you use a gapped core such as the popular E-core type, the gap should be directly underneath the windings, not in the outside legs, since the greatest magnetic leakage flux is to be found around the gap.

Differential-mode input and output noise

Differential-mode interference is caused by the voltage developed across the finite impedance of the reservoir or DC link capacitor at high di/dt. Careful choice of the capacitor type can reduce the value of its internal impedance but its physical size will mean that there is always significant inductance across which high frequency interference components will appear. In high-power switching converter applications, the preferred choice is a laminated DC bus bar with a capacitor bank, as it minimizes inductance in the DC link (see an example in the picture).

Using multiple capacitors in parallel

In many cases, achieving lower EMI in a switched converter can be accomplished by connecting numerous capacitors in parallel on the DC link, yielding significant benefits. This configuration often involves the inclusion of several electrolytic capacitors as well as a number of non-electrolytic, high-value (e.g., 22uF) surface-mounted multi-layer ceramic capacitors (MLCCs). This approach is also commonly applied in the design of high-frequency components such as GHz+ microprocessors and video systems. So, why not simply opt for a single, large-value electrolytic capacitor, considering it is a much cheaper solution?

There are two perspectives to consider. First, let's examine the design from a circuit point of view. Large capacitance value electrolytic capacitors tend to be physically bulky, with higher equivalent series resistance and inductance (ESR and ESL). The construction of electrolytic capacitors involves using a long, thin, rolled-up strip of aluminium foil submerged in an electrolyte fluid, resulting in relatively high internal resistance and inductance. By employing several smaller capacitors in parallel, we can effectively treat them as multiple ESR and ESL components in parallel, significantly reducing the total ESL and ESR compared to a single large capacitor. Large-value MLCCs are even more advantageous, as they typically exhibit much lower ESL and ESR (they are made of interleaved layers of deposited metal between layers of specialized ceramic) when compared to electrolytic capacitors. Nevertheless, electrolytic capacitors are still necessary in most applications for two key reasons. First, they can store substantially more energy compared to MLCCs. Second, multiple MLCCs in parallel may potentially introduce resonance issues, as the circuit may behave like an L-C tank circuit without sufficient damping components (ESR).

Another perspective is to consider that ultimately, energy must be delivered from the source to the load through the switching devices (IGBTs, MOSFETs, or SiC). The rate at which energy is delivered depends on how quickly the switches turn on and off (i.e., their rise and fall times). With switches being switched faster, it is preferable to have multiple energy sources supplying energy simultaneously. These energy sources must also be capable of delivering energy rapidly, making MLCCs an attractive choice due to their ability to supply energy at a faster rate.

The case study presented in this article and this presentation demonstrates that adding capacitors can effectively reduce EMI.

Filters

Extra series inductance and parallel capacitance on the input side will attenuate the voltage passed to the input terminals. Series inductors of more than a few tens of microhenries are difficult to realize at high DC currents (remembering that the inductor must not saturate at the peak ripple current, which is much higher than the DC average current), and multiple sections with smaller inductors will be more effective than a single section since their self resonance frequencies will be higher.

Switching spikes are a feature of the DC output of all switching supplies, partly because of the finite impedance of the output reservoir and partly because of capacitive coupling through the transformer. Such spikes are conducted out of the unit on the output lines in both differential and common mode, and may re-radiate onto other leads or be coupled to the ground connection and generate common-mode interference. A low-ESL reservoir capacitor is preferable, but extra suppression in differential mode can be obtained, as with the input, with a high frequency L-section filter.

VHF diode noise

The abrupt reverse recovery characteristic of the output rectifier diode(s) can create extra high frequency ringing and transients, due to the very fast di/dt at turn-off exciting the resonant circuits of the output transformer and filters. The spectrum extends well into the VHF range and it often appears as broadband noise, apparently unrelated to the switching frequency. These can be attenuated by using soft recovery diodes or by paralleling the diodes with a capacitor or an RC or snubber, although this will have an impact on the power efficiency.