UCC28780: an argument on single stage efficiency and two stage efficiency -- ( for hundred watts ac-dc conversion )

Hello Norman,

Thank you for your interest in the UCC28780 ACF controller.

In most conversion applications over 75 W input power, PFC is usually required. At lower power levels, single-stage conversion is often considered, but much less so at levels such as 300 W or higher.  Efficiency is not the only major consideration.

I think there are two main reasons for this (although there can be more):

a. most applications require a tightly-regulated, high-bandwidth output while PFC needs a loose, low-bandwidth control, and

b. the apparent simplicity of a flyback converter is overshadowed by the huge peak currents necessary at such power levels.

In a low-power single-stage converter, the energy storage needed to maintain output regulation during the input line zero-crossings must come from the output capacitance. The size of this capacitor becomes very large even at low power.  Since capacitive energy varies with the square of the voltage, it is much more efficient to store energy at high voltage to keep the capacitance value low.  The boost-PFC stage does this. Then, the output converter can regulate Vout with fast transient response and small ripple voltage even with a low output cap.  The PFC controller is optimized to shape the input current and the DC-DC controller is optimized to shape the output voltage.  Combining them into one stage de-optimizes both functions.  This all applies for the usual low voltage outputs.

Since your application output is at 300Vdc, the situation is not as clear. It is almost like a PFC output already.   However, you have not indicated what kind of regulation is necessary, how much low frequency ripple is allowable, and how much input harmonic content is allowed.  Using a flyback topology will require a specialized control method to obtain high-pF, low-THD such as average current mode control.  Peak current mode control will need slope compensation and constant on-time (which works well for Transition Mode) produces a lot of input distortion (because the flyback current is discontinuous).  But even at 300-V out, any fast transient response must be handled by the output cap because the loop bandwidth must be < 10Hz for PFC purposes.

Also, at 300W the huge peak flyback currents will need very large MOSFETs (or other semiconductor) which will incur heavy switching losses from the Coss. This is one of the reasons why LLC is preferred over flyback at power levels > 150W or even lower.  Active-clamp should theoretically eliminate switching losses with recycled leakage energy and ZVS, but the conduction losses may negate much of the expected gains. 

These are all qualitative arguments, and it is difficult to assign real numbers to each situation. One can’t assume that since the ACF achieves 92%+ in the usual <75W AC-DC applications that it will be able to do so at 300W. I think you’ll have to design and construct a breadboard to try it.   The UCC28780 ACF controller has an external feedback input which should allow it to follow a single-stage PFC control law, but this IC’s design is optimized for the typical off-line bulk voltage range of  70~375Vdc.  It’s not clear how it will behave around the zero-crossings. And the feedback path will need a regulator that manages the input current as well as the output voltage, not the usual TL431-type regulator. 

 I hope this helps your technical considerations.

Regards,

Ulrich

Thank you so much ... Ulrich,

our internal discussion focus on full load efficiency for hundreds wattage without emphasis on output voltage regulation ; yes, I thanks and agree on your points.

In general cases,  for hundreds wattage AC-DC with PFC , boost pfc + LLC are still the  best in full load efficiency ; someone just can't simply use the simple logic that  one stage conversion   will be always good. furthermore , ACF will be no more 92%+ efficiency for hundreds wattage. 

Best regards, 

Norman