Amplifier linear zone and class D

Hey gang, I’m trying to understand if overhead is as important in a class D amp as in a class A/B. IN a few of his videos @Paul has stated that an amplifier’s linear zone is around 20% of rated power. Is this due to coming out of the A biased region? Is it due to power supply capabilities? Is this also the case for class D and, if so, why (no A bias region, relatively tiny power supplies)?

No amp is perfectly linear across its entire operating rage, and any non-linearity means distortion, even with a perfect power supply. Modern Class AB amps are quite good at minimizing the effects of supply voltage variation, but they do not eliminate them completely. Class D amps work entirely differently by rapid pulsing of current from the supply, and any PS voltage variation is going to be reflected in the output. Class D needs a good power supply more than AB does.


I have some stupidly expensive class D amps. If you look inside the amp, the case is filled 7/8 with power supply and 1/8 with Class A driven Class D amplifier module. What I own has insanely good power supplies. And guess what, it sounds wonderful.

There are a lot of people who see “Class D” and they immediately reject the possibly that it can be done well. I am here to tell you it can indeed be done well. If I could invent a need for the PS Audio M700 amps I would buy them without fear of the Class D subject.

1 Like

Oh it’s nothing to do with sound quality. I know that a well designed and voiced amp will sound well regardless of class. My question is more to do with overhead. So for example: let’s say I’ve calculated that to get 80db @3 meters from the speakers with headroom of 15-20db I need 100 watts. So for an A/B amp to be in its linear region (~20% of full power per Paul) I would need a 500 watt amp. I’m trying to understand if it’s the same case with a class D (~20% of full power) or the percentage/range is different.

You have included quite a lot of headroom in your calculation already. When I do similar calculations I allow for 23db between average level and peak, which is the broadcasting standard, although I have never measured more than 19db when I have been experimenting. Frankly, for me, I wouldn’t worry too much about having the amp still in its sweet region for the most extreme transients and peaks, These are occasional and very brief events and, providing the amp doesn’t clip or distort grossly at that level then you would need to be a very extreme audiophile to detect any difference.

jedi - This doesn’t actually address your question - more of a personal experience that MAY have something to do with your question.

I’d been using a 250wpc nCore amp to power my Harbeth SHL5+ Speakers, that “need” at most, 100w. They might actually use 30w of a given amp’s power at most. Recently, while trying new amps, I had an ARC Ref 7, which as I recall is about 60wpc. Absolutely gorgeous 60w, and great for jazz, but lacking “balls” on the Bottom for pop and rock sorts of music.

At that same time, I had the prototypes for Darren’s new M1200 Class D mono’s here for around a month. They flabbergasted me with respect to showing what it was possible for the Harbys to do. Just didn’t think them capable of the effortless dynamics and “grip” on the bottom they displayed until then. I rolled some NOS RCA’s in, in place of the PSVanes that will likely come with them, and was very pleased by how they took to tube-rolling.

I’ve ended up with a Vitus RS-100, which is a 300wpc A/B piece, and is a great compromise between the finesse of the ARC’s sound and the clout and dynamics of the 1200’s. At any rate, if you’d told me it would be worth considering an amp rated at 10x the speaker’s rated needs, and 30 or 40x what they actually use, I’d have thought you nuts.

So - not sure if this is actually related to your headroom question or not. On one level it is always about “how Good are those bottom 30 watts?” I do wonder if it has something to do with what you’re asking though.

Which ones?

NBS Universal 2

They were a gift to me from the man who owns the company.

Thanks for the responses. Perhaps I got too wordy with my question so let me try again:

@Paul has mentioned that amps stay in their sweet spot up to about 20% of power rating - is this specifically for A/B amps or does it include class D as well?

Follow up: if class D has a different sweet spot range, what is it?

From memory, a plot for a Hypex Ncore based Class D amp of THD+N against output power showed distortion dropping steadily as power increased until the amp was at its rated maximum. It looked as if the distortion was a constant small value which became a progressively diminishing factor as the power output increased. From a headroom perspective this has to be good news.


Nice…looks like a wonderful gift for sure.

1 Like

I was gifted an entire system. Out of the box it sounded so much better than anything I have ever heard before. And it became a starting point for a foolishly expensive system that is tickling my ears at this moment.

My amps are Hypex Ncore based. They strike me as being extremely clean. When I first got them I was thinking oh no, class D. But after a bit of study and a lot of listening I got over that…