top of page

Exploring Bandwidth, Gain Margin and Phase Margin in Non-Invasive Measurement

Summary Among the most frequently asked questions regarding non-invasive stability measurement are “How do I determine the bandwidth, phase margin and gain margin from the non-invasive measurement”. The answer is simple. These terms are all fragments of our imagination, despite the fact that we have come to know and love them. They were created to simplify the process of defining and describing stability by defining three points along the curve. The proper assessment of stability comes down to the proximity of the transfer curve to the singular unstable point at 1,0 (or -1,0 depending on where you include the negative feedback term). There can be one or many points that come close to the singular unstable point, and these may or may not be the phase margin and gain margin.

Keywords Signal injector, current injector, high speed, POL, PDN, line injector, output impedance, input impedance, bode plot, PSRR, transient response

Introduction It has become more difficult to assess the stability of linear regulators, voltage references, switching regulators and audio amplifiers as these devices become more integrated, eliminating the circuit access needed for the traditional Bode measurement to evaluate stability. While we have written numerous articles on the subject of non-invasive stability measurement and the limitations of Bode analysis in general, we continue to receive many questions about the specific assessment of bandwidth, phase margin and gain margin from the non-invasive measurement data.

These terms, are related to the Bode plot, and I would argue have little significance in a stability assessment. It is shown in reference ii that it is possible for a control loop to have good phase margin and gain margin results, while at the same time presenting poor stability. Since the control loop is designed to optimize the closed loop circuit performance, this article will look at both open loop and closed loop characteristics.

13 views0 comments


bottom of page