Hi all,

Just wondering if there are any power supply designers on this board who may be able to help me out. I have a converter which I have measured the frequency response and I have a very similar response to the following.

I am curious to know what causes the "blip" at around 8kHz that causes the response to go up and then back down. At first, I thought that perhaps it may be a ESR zero, but thinking more closely, a single zero would not cause the response to rise like that. Furthermore, the response shows another pole afterwards, and I can't quite see where its coming from either. Anybody experience this or seen the cause of this before?

Thanks for your time.

 

Do you have the schematic you can post?

 

Absolutely. The LLC Resonant Converter

 

I have to ask the obvious question: What is the resonant frequency?

 

From where to where did you measure this response?

 

Ron_H: the resonant frequency is in the hundreds of kHz range, so it is far away from this measured response.

crutschow: the measurement is the control-to output measurement, so the output voltage and the switching frequency of the gates.

Not sure if that information really helps, but I guess what I'm looking for is just general suggestions as to what is causing that blip in the frequency. I've sort of figured out the poles may be attributed to delays in the circuit, but I have absolutely no idea what might be causing the positive dB gain there.

EDIT: I'm doing a bit of reading, and then I remembered in boost converters that there exists a right hand plane zero due to the charging and releasing of energy in the inductors (although my memory of this is shady). Could there be any chance the same type of thing is happening here? Although I can't find any literature regarding this phenomenon for this topology, I think this might be a good starting point for discussion.

 

You must have a control loop that determines how these switches are driven. How can we answer your question without seeing that?

 

The response is for an open loop response. I just used an oscillator circuit to provide the gate signals.

 

What does the frequency scale represent?
How is gain defined?

 

I repeat, where and how are the measurements taken. Without a diagram of this with component values we can't help.

 

If those plots are loop gain, it will oscillate. The phase shift drops to -180 while the loop still has positive gain.

If the peak in gain is true (not measurement noise), it can only be caused by a zero from some place.

 

The response is for an open loop response. I just used an oscillator circuit to provide the gate signals.

That doesn't make a lot of sense unless I've misunderstood your explanation. I would think that one would run the switches in pwm mode at the nominal switching frequency and then modulate the pwm around a mean modulation index whilst sweeping the audio modulation frequency range to elicit the frequency response. Maybe that's what you did anyway???