Accurate Loop Gain Measurements, Tips & Tricks

by Florian Hämmerle - OMICRON Lab
Loop Gain measurements are a powerful tool when designing a compensator of a switching power supply or to verify the small-signal stability of voltage regulators or switchers having a feedback circuit for a stabilized output voltage.
Since the availability of price effective vector network analyzers like the Bode 100, loop gain measurements have gained in popularity enabling the systematic stabilization and control design for simple as well as complex topologies in power electronic circuits.
In this presentation, the basics of the loop gain measurement method using voltage injection are reviewed and common pitfalls, like noise at low frequency, are explained in detail. Choosing the right measurement signal size is not only crucial to fight noise but also to get accurate results. The presentation includes tips and tricks for better measurements and demonstrates some of the most common challenges in a real-life demonstration.

[@ankitpratihasta1556]

Sir, please could you explain how to measure the loop gain if we use the hall sensors for the voltage measurement?

[@OMICRONLabTutorials]

The principles explained in this video apply to any control loop / feedback where the information flows in form of a voltage. Please contact [e-mail] for specific questions.

[@zktec1]

13:15 Is the additional +180 of phase on in the Analyzer Suite true for only the loop gain measurement? In other words, if I switch the probes around to measure the compensator as in 43:60, will its phase also be shown with an additional 180 degrees of phase? Based on actual measurements with my Bode 100, i'm seeing a low-freq phase of a Type-II OTA compensator as around +90 degree. But in a spice simulation it starts out at -90 degree. Just wanting to make sure I'm interpreting this correctly. This is a super-informative video by the way. Thank you Florian.

[@OMICRONLabTutorials]

The "additional 180° phase shift" is actually not "additional" but it is there in the real physical system whereas it is not there in a theoretical open-loop analysis of a negative feedbacks system. Bode 100 will always measure the real physical system. In case of an inverting error amplifier with integral part, there is a -180° phase shift from the inverting configuration and a -90° phase shift from the integrator part. In total this is -270° which equals +90° phase. Did you use an inverting configuration in your spice simulation?

[@zktec1]

@@OMICRONLabTutorials You are exactly right. I went back to my simulation and sure enough did not have it configured to inverting. Even in my own derivation of the compensator I didn't propagate negative sign through. Now everything is reading +90 on the compensator's phase at low-freq. Thank you Florian for point this out.

[@paulpaulzadeh6172]

Nice Florian , just wonder if you have current feedback amplifier , could we use the same method , requirements about output impedance vs inputs impedance, what it be then ?!

[@OMICRONLabTutorials]

Dear Paul, the requirement for the impedance condition comes from the fact that we inject a voltage and measure two voltages. Only if the impedance condition is satisfied, it can be ensured that the information flow at the injection point is in form of voltages. If not, the information flow could be in form of a current and voltage combination which makes the measurement hard. If you have the opposing situation where impedance backwards is large and forward is small, then you could inject a current and measure two currents to get the loop gain. Hope that helps! Otherwise, please contact us at [e-mail]

[@biswajit681]

When you are going to upload this year videos?

[@OMICRONLabTutorials]

As soon as we have prepared the recordings. It can take some days.

[@biswajit681]

@@OMICRONLabTutorials Thanks for the response... please keep priority of Dr. Ali Shrishabar and Basso video