SIMBA Conf 2025 #4: Case Study for Single Phase Interleaved Totem Pole Inverter

Real-world circuits operate with various non-idealities. First are the non-idealities in the power loop circuit, i.e., parasitic components (parasitic resistance, capacitance, inductance). Non-linear inductance (saturation) and capacitance are some of the most common examples; these non-idealities are generally well understood and are commonly added to the simulation.

Second is measured noise on the circuit control loop feedback: (2a) picked-up noise in signal transmission and (2b) sensor self-generated noise. Noise (2a) generally does not affect the transient behaviour since it occurs at very high frequency and can be easily filtered out by using a simple low-pass filter (note: the filter must be fully modelled to ensure accurate transient response). Anyway, (2b) is more complicated since the noise characteristics are mostly either flicker noise or white noise, making such noise more difficult to filter.

Among the sensors commonly used for power converter feedback, the open-loop Hall-effect current sensor is known to be the noisiest and will be the focus of this presentation. For the case study, the interleaved totem pole inverter topology is selected since this topology is particularly sensitive to noise, especially for the low-frequency leg control. The following points will be discussed to help the audience understand the significance of noise modelling:

Impact of the noise on the zero-crossing control of the interleaved totem pole inverter
Physical Hall-effect-based current sensor noise characterization
Noise modelling in the SIMBA environment
Controller modification to mitigate the sensor noise issue

www.simba.io