Buck Average Mode Control - LC Double pole in the current loop?

[Magnethicc]

Does anyone have experience with average mode control and could help me clarify why do I see a double pole of the LC in the current loop?
When modeling and simulating an average mode controlled buck/boost converter the voltage loop does not see any double pole but when using an FRA (or AC Analysis) in the current loop there is a double pole.

As simulation shows, an LC double pole at ~1.8kHz is present in the current loop only, which is the same resonant frequency is L1 and C6 in the simulation.
I didn't find any resources about this and I would be thankful to anyone who could contribute
1. why is there LC double pole in the current loop? i though that if you control the current in the choke you damp the LC.
2. Why there is no double pole in the voltage loop also?
3. how to damp the current closed loop double pole? is it necessary to damp it?

Thank you!

[youngda9]

What you're seeing in the second plot is the resonance of the output LC (500uH inductor and 16.6uF capacitor).  That calculates to a corner frequency of 1.75KHz =1/(2*pi*sqrt(L*C)).

You can start by adding realistic impedances into the inductor and capacitor model.  ESR of the output capacitor is important here.  This should help to damp it out.

Does it need additional damping...It doesn't really matter if you are getting the loop gain performance and transient response that you desire.  It will have some effect on the output impedance over frequency as well, if that is a concern.

Suggestion:  When doing Bode plots...for loop gain.  Set 0db in the middle of the plot on the left side.  Set 0deg in the middle of the plot on the right side.  Set both scales so that they have the same number of divisions.  This way it is easy to quickly read off the zero-crossing, PM, and GM.  The way that you have it now, it takes a lot more work.  Example attached.

[Magnethicc]

Thank you for your answer!
I found a video about average current mode control by dragan maksimovic (in a coursing site) and in his simulation the double pole of the LC is present in the current loop and not present in his voltage loop - so its very similar to what my simulation shows.
so I guess this is just an inherent characteristic of the current loop in the average current mode control that the LC double pole is present.

adding capacitor and inductor ESR did not do much (added as much as 0.5ohm in each).

My issue with the undamped double pole is that the double-pole present a significant drop in phase margin and I was wondering if the phase near the double-pole drops to 180degrees than oscillations are imminent? (because at the LC double-pole frequency the gain is still very high, assuming the loop bandwidth is higher than the double-pole frequency).

thank you for your suggestion, will make sure to impliment it!

[youngda9]

My issue with the undamped double pole is that the double-pole present a significant drop in phase margin

Your first plot is your control loop, where phase margin is measured.  At the resonant frequency the plot is smooth and the phase continues rising linearly.  This resonance isn't having any effect on your phase margin.

[Magnethicc]

My issue with the undamped double pole is that the double-pole present a significant drop in phase margin

Your first plot is your control loop, where phase margin is measured.  At the resonant frequency the plot is smooth and the phase continues rising linearly.  This resonance isn't having any effect on your phase margin.

In the inner current loop the LC double pole exist and effect the phase margin at near the resonance frequency.

[youngda9]

In the inner current loop the LC double pole exist and effect the phase margin at near the resonance frequency.

The upper plot, where you calculate phase margin, looks linear at the resonant frequency.