This one specifically deals with a LLC, but many of the general considerations also apply to other topologies.
https://www.edn.com/power-tips-103-llc-design-considerations-for-audio-amplifiers/
As the requirements of an audio-oriented switching supply seem to be widely misunderstood in this forum, this piece is a must-read for anyone trying to design such a supply from scratch, or to reuse a "normal" switcher for audio applications
https://www.edn.com/power-tips-103-llc-design-considerations-for-audio-amplifiers/
As the requirements of an audio-oriented switching supply seem to be widely misunderstood in this forum, this piece is a must-read for anyone trying to design such a supply from scratch, or to reuse a "normal" switcher for audio applications
Cuk would also be a good SMPS for audio due to the very low ripple inherent in the design but like most SMPS it does hammer the crap out of the capacitors.
Selecting the controller (PFC or DC link supply to load):
* fixed / variable frequency switching based on load
* multi-mode support for conditions such as no-load, regulated load, high load (>80%, inrush etc).
Seems to me the basis of a PFC with a fixed switching frequency of ~100-300KHz then have enough capacitance to cope with gap between peak load and the next AC wave form plus charging time of the capacitance. A step down often run in the Mhz switching for example - faster switching the more loss..
Once you get above 500W then you're really looking at active rectification to minimise power loss, with parallel stages etc and it's all about not loosing 50W+ on high speed switching.
SMPS design is driven by the need for magic numbers to show the regulatory compliance and to wave the flag of efficiency in a 'just in time' on demand. So if you need peaks then go with a big stick (big power supply with large PFC stage) or use a bank of caps..
I don't think the design decisions have changed. I think that it's almost impossible to find a specifically designed 'audio' SMPS that is as cheap. Given the components and complexity of the designs to support that, you're really back to a off the shelf cheap small SMPS and a big cap with a nod to the odd sag caused by some passages..
Selecting the controller (PFC or DC link supply to load):
* fixed / variable frequency switching based on load
* multi-mode support for conditions such as no-load, regulated load, high load (>80%, inrush etc).
Seems to me the basis of a PFC with a fixed switching frequency of ~100-300KHz then have enough capacitance to cope with gap between peak load and the next AC wave form plus charging time of the capacitance. A step down often run in the Mhz switching for example - faster switching the more loss..
Once you get above 500W then you're really looking at active rectification to minimise power loss, with parallel stages etc and it's all about not loosing 50W+ on high speed switching.
SMPS design is driven by the need for magic numbers to show the regulatory compliance and to wave the flag of efficiency in a 'just in time' on demand. So if you need peaks then go with a big stick (big power supply with large PFC stage) or use a bank of caps..
I don't think the design decisions have changed. I think that it's almost impossible to find a specifically designed 'audio' SMPS that is as cheap. Given the components and complexity of the designs to support that, you're really back to a off the shelf cheap small SMPS and a big cap with a nod to the odd sag caused by some passages..
One disadvantage of the LC/LLC types is their unidirectional operation. A bidirectional converter with PFC would inherently support anything between no-load and over-load conditions. Besides, they always run in CCM irrespective of loading, simplifying design calculations.
Good point. An audio supply should run in CCM to minimise ripple but also reduce the harmonic content caused by the zero crossover of the more energy efficient DCM.