window averaging

Copy of Window Averaging - on Wed, 07/22/2020 - 18:23

MASA
Designer208

MASA

Member for

9 years 1 month

575

designs

8

groups

Description

**************

Continuous PWM waveforms combine an averaged response with variations within the switching cycles. These switching cycle details can be removed by averaging with a sliding window, i.e. by replacing the value at each time point with an average taken over the previous switching period.

This design demonstrates important characteristics of window averaging using a copy of the Inductor current waveform from the buck converter at

https://www.systemvision.com/design/managing-pwm-waveforms-buck-converter

Specific window averaging characteristics demonstrated here include the following:

1) Window averaging is actually a linear filter function. This design implements two instances of that filter using continuous function blocks.

2) The filter will virtually eliminate the switching cycle details if the window width is set equal to the period of that switching cycle.

This design demonstrates items 1 and 2 at the upper left of the schematic.

3) Window averaging introduces a delay equal to half the window width. The schematic demonstrates this by comparing the averaged waveform to a delayed version of the original, full switching waveform.

4) At the upper right of the schematic there's an inverse filter that rebuilds the unsmoothed waveform. The rebuilt waveform is a virtually perfect match to the unsmoothed original.

5) At the bottom of the schematic a copy of the filter is driven with a narrow pulse by which the 50us window is generated at the filter output. A frequency domain simulation outputs the corresponding sin(x)/x transfer function. These waveforms confirm that the filter realizes the window averaging operation.

About text formats

Norm
Designer43361

Norm

Member for

7 years 11 months

328

designs

2

groups

Description

**************Continuous PWM waveforms combine an averaged response with variations within the switching cycles. These switching cycle details can be removed by averaging with a sliding window, i.e. by replacing the value at each time point with an average taken over the previous switching period.This design demonstrates important characteristics of window averaging using a copy of the Inductor current waveform from the buck converter athttps://www.systemvision.com/design/managing-pwm-waveforms-buck-converterSpecific window averaging characteristics demonstrated here include the following:1) Window averaging is actually a linear filter function. This design implements two instances of that filter using continuous function blocks.2) The filter will virtually eliminate the switching cycle details if the window width is set equal to the period of that switching cycle. This design demonstrates items 1 and 2 at the upper left of the schematic.3) Window averaging introduces a delay equal to half the window width. The schematic demonstrates this by comparing the averaged waveform to a delayed version of the original, full switching waveform.4) At the upper right of the schematic there's an inverse filter that rebuilds the unsmoothed waveform. The rebuilt waveform is a virtually perfect match to the unsmoothed original.5) At the bottom of the schematic a copy of the filter is driven with a narrow pulse by which the 50us window is generated at the filter output. A frequency domain simulation outputs the corresponding sin(x)/x transfer function. These waveforms confirm that the filter realizes the window averaging operation.

About text formats