Vout pkpk ripple of Boost PFC?

[zenerbjt]

Hi,
AND8123 page 16 (equation 88) tells that the Vout ripple peak value of a 200W Boost PFC with VOUT(AV) = 400V and Cout = 22uF would be 33.2V (ie 66.4V pkpk).

However, the actual simulation of a 200W Boost PFC with Vout(Av) = 400V and Cout = 22uF gives the Vout ripple as 72Vpkpk. This simulation is attached, and also the PDF picture of it.

Equation 88 is derived by assuming that the average output current through the Boost diode is a train of Half sinusoids. However, the attached LTspice simulation shows that if the current really were a train of half sinusoids, then the Vout pkpk ripple would be 49V.

So how do you really assess what is the Vout(pkpk) ripple of a Boost PFC? Clearly there is an element of dead-time at the mains zero crossing which means that the analogy with the train of half sinusoids is not fully accurate, would you agree?

I would say that the LTspice simulation is the most accurate means. Would you agree? Equation 88 clearly does not take into account the dead-time at the mains zero cross.

AND8123

https://www.onsemi.com/pub/Collateral/AND8123-D.PDF

 

[Easy peasy]

any deviation from true sine wave in will give worse ripple - ( unless you draw a square wave off the mains with a dip in the middle - this will give you less ripple than the sine ... )

 

[FvM]

Ther assumptions of the Vout ripple equivalent circuit are wrong. An ideal single phase PFC sinks a sinusoidal current from the mains, resulting in an instantaneous power swing between 0 an 2*Pavg. The current into the boost converter is ideally a rectified sine, the output current into the bulk capacitor isn't. It's scaled by the instantaneous boost voltage ratio, resulting in a sine waveform, see Ic in AN figure 14.

Not sure how you arrived at 66Vpp, I get the same value of 72Vpp when applying eq. 91. A small deviation from the sinusoidal input current waveform near the zero crossing, e.g. due to switcher deadtime doesn't significantly affect the power waveform and the ripple voltage.

 

[Easy peasy]

The ripple on the boost caps in in fact a sine^2 function @ 2x the line freq

We have found - empirically - that lack of very high power factor - makes the ripple worse ... except for the exception discussed above ...