Voltage mode flyback has LC resonant frequency too close to feedback loop frequency?

[treez]

Hello,
The below voltage mode flyback has an output LC filter pole at 618Hz.
The feedback loop crossover frequency is 1150Hz (this can be seen in the simulation by the frequency of the ringing on the "COMP" pin)

So, the feedback loop crossover frequency is less than double the LC filter pole frequency.
To ensure stability, the output LC filter pole frequency should be at least three times greater than, or three times less than the feedback loop crossover frequency.

So howcome this voltage mode flyback is actually stable?..it should be unstable(?)

Schematic and LTspice simulation attached

 

[mtwieg]

Hello,
The below voltage mode flyback has an output LC filter pole at 618Hz.
The feedback loop crossover frequency is 1150Hz (this can be seen in the simulation by the frequency of the ringing on the "COMP" pin)

So, the feedback loop crossover frequency is less than double the LC filter pole frequency.
To ensure stability, the output LC filter pole frequency should be at least three times greater than, or three times less than the feedback loop crossover frequency.

So howcome this voltage mode flyback is actually stable?..it should be unstable(?)

Schematic and LTspice simulation attached

There's no reason the crossover frequency can't be at or above the LC frequency, so long as the compensation is well designed.

 

[treez]

There's no reason the crossover frequency can't be at or above the LC frequency, so long as the compensation is well designed.

..thanks but I don't think I explained myself enough...I am not speaking about an output LC filter...I am speaking about the LC which comprises the flyback secondary and the output capacitor.
The crossover frequency cannot be "at" this frequency. -to design a compensation network like that would surely be like shackling ones legs prior to climbing a mountain?
Pages 216-218 of Marty Browns cookbook shows the crossover freq being well above the LC pole frequency...ok Brown speaks of a bridge converter, but its the same situation with voltage mode flyback.

 

[mtwieg]

..thanks but I don't think I explained myself enough...I am not speaking about an output LC filter... I am speaking about the LC which comprises the flyback secondary and the output capacitor.

They're effectively the same thing for the purpose of transfer function analysis.

The crossover frequency cannot be "at" this frequency.

Of course it can.

-to design a compensation network like that would surely be like shackling ones legs prior to climbing a mountain?

No, it's quite feasible. Above resonance, the converter contributes 180 degrees of phase shift, and a type III compensation scheme (maybe combined with some phase boost from output capacitor ESR) can provide enough phase boost to give good phase margin despite that. The RHPZ frequency is the effective limit, not the LC resonance.

 

[treez]

yes but surely if you do a proper type 3 compensator, then you will place 2 zeros at the output double pole filter frequency, and you will then place a pole at the esr zero frequency, and if you do this, you will then have a crossover frequency which is at least three times above the LC frequency.

Its not possible to do a type 3 compensator properly and end up with the crossover frequency at the LC frequency.

 

[mtwieg]

Its not possible to do a type 3 compensator properly and end up with the crossover frequency at the LC frequency.

There's no exactly "proper" way to do type III compensation. All it means is you have DC pole, and two pole/zero pairs. They can be placed anywhere to get whatever crossover frequency.

 

[FvM]

Why do you think that the converter is working in voltage mode?

 

[treez]

because the ramp is created entirely by a ramp generator, and not by the current sense ramp.

 

[treez]

Mtwieg:

There's no reason the crossover frequency can't be at or above the LC frequency, so long as the compensation is well designed.

Page 47 of "power supply design, volume 1:control" by Dr Ridley states that for a voltage mode flyback.."The crossover frequency should be at least twice the resonant frequency".......by "resonant frequency" , Dr Ridley means resonance of Cout and L(sec)

So I conclude, from Dr Ridley, that the crossover frequency cannot be at the resonant frequency?

 

[mtwieg]

I don't see how you could possibly conclude that. "should be" is not the same as "must be."

 

[treez]

Pages 47-51 of the Ridley book “power supply design, volume 1:control” states how poles and zeros should be placed in voltage mode flybacks.

The book states that the phase boost of a type 3 compensator is needed to offset the sharp phase drop that occurs after the resonant frequency of the power stage. –consequently the compensator zeros are placed closely around the power stage resonant frequency.
In fact page 48 shows graphs of Voltage mode flyback power stage, as well as the type 3 compensator gain, and the overall resultant loop gain……it does appear from looking at that diagram, that indeed the placement of compensator poles and zeros “must be” done in a specific way in order that the compensator plot and the power stage plot add together to form the overall loop plot effectively.

Im looking right at this diagram now, on page 48, and I really don’t see any better way of compensating that power stage……I can well see that the crossover frequency “should best be” at least twice the resonant frequency.
Looking at the diagram , I’d say that it would be particularly unwise to put the crossover frequency “at” the LC resonant frequency.

The best way forward is, as the book says…
First pole of compensator at origin
Compensator zeros placed around the LC resonant frequency
Second pole of compensator at esr zero freq
3rd pole of compensator at rhpz freq

The graph shows the power stage gain plummeting at –40db/dec after the LC freq…..clearly the compensator poles need to be placed closely around the LC freq so that the overall loop gain can be picked up.

Ridley states that for some converters, the rhpz is at such a low freq that it is not possible to cross the loop over above the resonant freq, and the converter cannot be properly controlled…this issue is fixed, he says, by doing current mode control instead.