Harley: What accounts for the sonic differences among switching amplifiers that use the same module? It seems as though some companies have a “house sound” in their switching amplifiers that parallels the house sound in their linear designs.
Putzeys: That should not really be a surprise. I prefer to be absolutely agnostic about the technology that’s used to obtain a certain sonic result. These days I can make a linear amplifier and a switching amplifier that sound absolutely identical. There are only a few correct basic topologies for building a linear amplifier. Of course there are many people trying out other ideas, and usually these have more or less obvious flaws, but the basic foundation of all good linear amplifiers is roughly similar throughout.
And yet, linear amplifiers that are built along very similar lines but by different people have their particular “house sound,” as you call it. A similar thing happens when different people start with the same Class D circuit. Part of that has to do with how one translates sonic requirements into technical criteria. You can say you want a particular kind of sound, but in the end you have to get your hands dirty, pick up the soldering iron, and do something technical that corresponds to what you want to do sonically.
So that sets some parameters in the design. After that—and much to my personal frustration as an “audio objectivist”—there’s still a lot of influence from the kinds of components you use, such as capacitors and resistors.
You use these things to tune toward a certain kind of sound. Everyone has a different conception of what a good amplifier should sound like, so it’s quite clear that different designers strive to achieve their own interpretation.
D’Agostino: I take exception to your comment that people “design” switching amplifiers, because most of them are based on somebody else’s chipset. What you’re really doing is designing the output filter in a way that expresses your idea of what an amplifier should sound like, because as everyone knows, there is no ideal filter for switching amps yet. We’ve had some great results with switching power supplies. We built a six-kilowatt switcher that is a pretty amazing thing, and using that with a linear output stage seems to work really well. I have not had very good results with any of the Class D chipsets that I’ve seen out there, so again, my opinion is still that this technology isn’t finished. I think it has a lot of maturing left to do.
Putzeys: When I talk about how I design amplifiers it’s probably, indeed, very different from what the majority of people do. Most designers just look at what is available in terms of modules, but I actually make original circuits from the ground up, using discrete transistors and so on. Superficially speaking, they are very simple discrete circuits where the performance really hinges on the math behind it. So far I’ve not yet come across any integrated circuits that allow me to build an amplifier of the performance level that I want to achieve.
There are too many switching amplifiers around currently that do not fulfill even the most basic requirements of an amplifier. The basic requirement of an amplifier is that it has a specific voltage gain, that it has low distortion at all audio frequencies and that the output impedance is so low that the loudspeaker can really not influence the amplifier’s behavior. Almost none of my competitors have anything remotely resembling this, and certainly in the integrated-circuit domain I’m not seeing people even trying to achieve this.
Top | Previous | Next
Advertisement
All documents notede with file size are in PDF format.
Download Acrobat Reader
