Bruno Putzeys blickar tillbaka över utvecklingen av klass D

[pLudio]

https://hometheaterhifi.com/30th-annive ... mplifiers/

Relaterat; https://www.stereophile.com/content/john-ulrick - Spectron (tyvärr med Jack Bybee-uppgraderingar o dyl, har inte sett några mätningar)

[guldfisk]

Började precis läsa och verkar mycket lovande - tack!

[Kraniet]

Ja det va en intressant och underhållande text.

Nu vill jag ha ett gensvar från Morello. Tydligen är Brunos klass-D slutdestinationen för sann spitzenklasse-audio

[Maarten]

Tackar, intressant från en ledande snubbe inom området. Några citat:

Class D amplifiers were first described in Burnice Bedford’s patent (US1874159), filed in 1930.
...
it is now a proven fact and not an opinion that self-oscillating amplifiers are inherently more stable, and inherently allow more loop gain, than clocked amplifiers.
...
Carsten Nielsen, former CEO of ICEpower, enjoyed explaining how his Ph.D. student Niels Anderskouv one day forgot to turn on the triangle wave on an amplifier he was working on. As if by magic, performance improved.
…
What followed was merely consolidation and refinement. That it looks otherwise to many audiophiles is because in practice, these “refinements” broke one performance barrier after the other. Every generation posted 10 times lower distortion than the one before. In my case, the iterations were called UcD (2000), Ncore (2008), and Eigentakt (2016). I ran out of inspiration, so the next one is still called Eigentakt.
…
Many audiophiles still think that anti-measurement and anti-feedback sentiments are somehow countercultural and refreshing. They’re not. They’ve been the established culture for over 50 years, perpetuated by self-proclaimed engineers with more charisma than math chops who’ve been taking you for a ride.
….
But how then do we verify that an amplifier can even handle signals above 10kHz? A single sine wave won’t cut it because all harmonics are outside the measuring range (and inaudible too). The amplifier could be grossly clipping, and we wouldn’t know.

The solution is to blast the amplifier to near-clipping with two sine waves right at the end of the audio band and to inspect the resulting spectrum. You’ll agree that this is just about the worst possible test signal that still technically qualifies as “audio”. Nothing like it ever occurs in real music, so it’s a proper stress test.
The two tall poles 18.5 kHz and 19.5 kHz are the test tones
…
Technologically, class D is hard. Just the PCB layout is a nightmare to figure out, let alone the circuit. All truly high-performance designs are made by a handful of people who’ve made a career out of it. And they’re not exactly publishing how-to guides

Och så detta:
'

A typical design cycle in esoteric high-end circles consists of taking a product one really likes the sound of and calling that the “reference”. Then, more or less randomly, bits and pieces of a new prototype are tweaked until it sounds “better” than the reference. The prototype goes into production and becomes the new reference.

This Brownian motion of sonic ideals and outcomes explains how some “Hi-Fi” components can sound so utterly unalike. This practice of “designing by ear” attempts to short-circuit the most powerful design resource we have at our disposal, which is the bit that sits between those ears. Sound quality is not a mystery. It’s a tractable problem if only we care to think it though.

[sprudel]

Ja det va en intressant och underhållande text.

Nu vill jag ha ett gensvar från Morello. Tydligen är Brunos klass-D slutdestinationen för sann spitzenklasse-audio

Ja, enligt Bruno.

[avr7000]

Jag äger Ucd, Ncore, Nilai, Eigentakt, nån billig kines och har ägt Ice500 och Ice1000....

Jag skulle säga att Nilai500 sticker ut och är tydligt bättre än de andra...