Hello all,

I am trying to extract the loop gain and phase of a switched mode power supply that I am designing, using this by LTSpice documentation; https://www.analog.com/en/technical...-steps-in-generating-a-bode-plot-of-smps.html.

When you are running a transient analysis, it is important to process measure commands for efficiency when the converter has reached steady state. How does it work with AC analysis? I am trying to compensate my power supply properly, because the current control system is oscillating around the set-point. At what time is the correct point to begin measuring loop gain and phase for AC analysis?

 

My experience with switchmode power supplies is that they have at least two different modes of operation, startup and steady state. Both capacitors and inductors have different effects, charged versus not charged capacitors, and inductors with or without changing magnetic fields. So you will need to use two different sets of initial conditions. THAT is part of why I don't design switching power supplies.

 

How does it work with AC analysis?

AC analysis only works for AC linear models of all the devices, so won't directly work for a switch-mode supply.
You have to substitute a linear model for the PWM modulator input feedback to output voltage transfer function, i.e. feedback volts in to PWM voltage out.
Determining that gain requires knowledge of exactly how the PWM modulator works for the particular design you have.

I did that for a switching regulator I was designing, and it worked well to optimize the loop compensation values, especially since it simulates a lot faster than the transient analysis.
I then checked the stability of the loop using the transient analysis of the switching circuit by injecting a small plus and minus pulse into the feedback loop, and looking for any significant overshoot or ringing.

 

AC analysis only works for AC linear models of all the devices, so won't directly work for a switch-mode supply.
You have to substitute a linear model for the PWM modulator input feedback to output voltage transfer function, i.e. feedback volts in to PWM voltage out.
Determining that gain requires knowledge of exactly how the PWM modulator works for the particular design you have.

I did that for a switching regulator I was designing, and it worked well to optimize the loop compensation values, especially since it simulates a lot faster than the transient analysis.
I then checked the stability of the loop using the transient analysis of the switching circuit by injecting a small plus and minus pulse into the feedback loop, and looking for any significant overshoot or ringing.

So in a different method than is included in the document I have attached which claims to find the loop gain? Have you possibly got a reference to an external source that I could study on how to do what you describe, please?

 

So in a different method than is included in the document I have attached which claims to find the loop gain?

Yes, I believe that does it using the Transient analysis.

Have you possibly got a reference to an external source that I could study on how to do what you describe, please?

Not that I'm aware of.
The main task is to determine the PWM linear transfer function, which is the change in PWM duty-cycle versus the input error signal multiplied by the change in output voltage versus the change in duty-cycle.
This is fairly straight-forward if you know the PWM modulator design.

 

How does it work with AC analysis? I am trying to compensate my power supply properly, because the current control system is oscillating around the set-point. At what time is the correct point to begin measuring loop gain and phase for AC analysis?

The AC analysis is ALWAYS (per definition) performed under steady state conditions. You have no other choice and you cannot determine any time the analysis begins. It is a small-signal analysis - of course, without any transient effects (which are described in the time domain).