Two SMPS with front end PFC connected to same AC voltage source/grid 50Hz.. PFC(digital control) is operating at 24kHz in both SMPS. I am observing a 0.2Hz envelop on PFC current which is changing if the difference in switching frequency is increased. The low frequency ripple in current is high at lower voltages (high current). Instantaneous currents are in opposite magnitude at switching frequency level. There is a CM (pi) filter at front end. Voltage sense for PFC is after EMI filter. 0.2Hz envelop is also seen in voltage after EMI filter. How to avoid this?

Thanks for ur time.

Synchronise both PFC clocks and see.This is due to the coupling of switching frequencies via line due to the currents drawn.

Hi Gowri,

these are two different units having different control cards. It is difficult to synchronise both these clocks. Is there any other solution?

Seems a little unlikely that unsynchronized converters would be that close in frequency. Are you sure it is beat frequencies of the clocks?

How much ripple current/voltage are you actually observing?

Hi Ray,

We actually changed the switching frequency of one unit by 12Hz, then input current has 12 Hz beat frequency.

For single unit: Unfortunately, i observed more ripple(almost double in magnitude) in current at mains input than at EMI filter output. With function generator, for EMI(CM pi filter) i could see signal amplifying below 50kHz and starts attenuation thereafter. Probably need to increase capacitor and see.

Ripple current in the DM choke of EMI filter can certainly be higher than the ripple present in the PFC choke (or EMI filter output), depending on how far the resonant frequency of DM filter is from the PFC switching frequency. Consequently mains input current may also manifest higher ripple. Bode plot measurement of DM section can give good insight as to how much Q is pressent in the LC circuit, and help in quantifying the "almost double ripple" observation.One can always play with C or L value to move the resonant frequency far away from Fsw.I would suggest to reduce L value or remove it altogether, if it does not affect CE and RE test results.

Oh Murali, from your comment I reckon you have only CM (PI) filter. Still, the resonant frequency theory applies here for the differential-mode equivalent ckt of the CM filter. And bode plot measurement is all the more helpful.

Can we see DM noise as current source? If yes, do we need to provide the path for ripple current through DM caps? Does it heat the cap??