Class D - output LPF

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
What I like about post-filter phase-shift self oscillation is that no RC dampers are required. and that the switching frequency drops nicely as the output approaches the rails and current increases, to compensate for the increased switching losses ;)
 
..the smooth change of the switching frequency is for sure nice if we look to the switching losses.

In clocked designs you may also get lower switching frequencies when aproaching the rails. Carrier aliasing - which is not just a good luck / bad luck question, but a reliable effect that can be considered during design. But it changes the switching frequency in a step by factor 2 (first step), which looks ugly on the scope. The resulting increase of distorsions is IMHO acceptable.

IMHO the self resonant topologies offer a charming combination of simplicity and performance. The only two points which pushed me away from selfresonant were:
- Not synchronizable (...still unclear for my if this feature is really important or not...)
- Lock up effects, which should be no issue at all after proper design - but can increase the chocolate consumption during design phase.
 
My preferred modulator for sinewave inverters and the like uses the instantaneous duty cycle to produce the slopes of the trianglewave. The slopes are changed linearly with duty cycle, one increased and the other decreased, and vice-versa, being equal for 50%. This works nice for switching losses too, the amount of slope change (and thus, of frequency drop) can be selected for almost constant carrier residual.

I fear using post-filter self oscillation in sinewave inverters because the load can be really anything during normal operation.

As a bonus, the error amplifier is stable with more gain :D:D:D particularly in average current mode.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.