I have designed a LM25085 based Zeta converter with the "minimum ripple configuration" circuit. The circuit work fine until the converter is in continuous current mode.
Under light load, the output show a big ripple (about 800 mV) at low frequency (6 millisec period); the average voltage is 0.4V lower than that in CCM. In this situation, the behaviour is the following: the regulator generates a burst of switching cycles (for 400 microsec) that make raise the output voltage; then the switching stop and the output voltage begins to decrease until the next burst.
I think that is the topology of ripple injection circuit itself that produce this behaviour, because it can not distinguish between CCM and DCM.
Has anyone found a way to solve this drawback ?
