When examining conducted and radiated emissions in a given system, the emissions are broken down into two types: differential-mode (DM) noise and common-mode (CM) noise. By examining and understanding the pathways these noise types can take, techniques can be developed to help mitigate or remove noise from the TMCS11xx.

Differential-mode noise has an opposite polarity when looking into the pins of the device. The aforementioned conducted emissions are seldom common-mode noise and are typically differential-mode where potentially one signal trace on a PCB is the noise source with respect to a clean analog ground signal. In most cases for the TMCS112x or TMCS113x, the reference can be GND pin of the device, which is independent of the earth or chassis GND (but can be referenced to one or the other based on the application). Differential-mode noise is recognized by most engineers, as this is the typical noise most engineers think of when observing an output signal from an IC on an oscilloscope.

Common-mode noise is most often created from a radiating aggressor emitting electric fields or magnetic fields and coupling via parasitic capacitance and/or parasitic inductance. Therefore, a distinct path to GND is not necessary to propagate this type of noise in the system. Once coupled into a nearby victim circuit, the common-mode currents travel into the pins of a victim device with the same polarity and terminate into an earth or chassis ground. As a result, this type of noise is more difficult to measure in a system. Figure 2-2 provides a visualization of these noise types and how the noise types can flow into the TMCS112x/3x. With CM and DM noise present on both L1 and L2, probing the differential voltage between these two nodes can disregard common-mode noise while effectively measuring differential-mode noise.

Figure 2-2 Differential-Mode (DM) vs. Common-Mode (CM) Noise Paths Through an IC

The challenge then becomes providing low-impedance pathways for these unwanted signals to traverse while simultaneously preserving the integrity of the desired signal chain. The classical way to achieve this in power systems is through the use of reactive components, as losses in these components are typically minimal, but come with the tradeoff of resonant peaking that must be addressed to make sure a successful design. In fact, most power supply designs today start at the inputs with some form of an EMI filter. Figure 2-3 shows a visualization of the relevant filter networks that must be developed to effectively control these noise sources away from the TMCS112x/3x. Note that the purpose of these networks is to provide a designed for, low-impedance path for these high-frequency signals to return upon, and failure to design an impedance path lower than that of the part does not result in a proper filter. However, this is typically manifested in the frequency response of the filter, and therefore design of these filters can be concentrated in the area of frequency shaping.

Figure 2-3 Filtering Techniques to Shape Noise Paths Away From TMCS112x/3x