Eh, no. There are measurements for various power levels and 4/8 R loads, noise measurements, as well IMD measurements in that and one another related thread. There are precise DF measurements over frequency in another thread. There is more.THD_1k
1W(rms) into static 8 Ohm
I think Ed would like to go a little further
Bye,
HBt.
Oh yes,
I took a look at the first posts - everything is fine.
But you mustn't take this personally, but I myself am not interested in a complementary MOSFET output stage as a power extension to the 828 operational amplifier - that is not meant in a judgmental way.
Despite the personal lack of interest - a sincere compliment.
I find the corresponding thread title somewhat misleading, because a very specific expectation is raised (but not quite fulfilled).
Nevertheless,
a good job
I took a look at the first posts - everything is fine.
But you mustn't take this personally, but I myself am not interested in a complementary MOSFET output stage as a power extension to the 828 operational amplifier - that is not meant in a judgmental way.
Despite the personal lack of interest - a sincere compliment.
I find the corresponding thread title somewhat misleading, because a very specific expectation is raised (but not quite fulfilled).
Nevertheless,
a good job
I recall that Stereophile uses a low-pass filter only on class D amplifiers. Using it on class AB would make the numbers equally dishonest.
tombo56 and knutn - Class A achieves truly low distortion without the switching that can goof up class AB. I am more inclined to believe simulated THD <0.01% on class A than class AB.
Ed
Before we speakers go round in circles, I would like to politely apologize - and say goodbye to this thread.
I assume that everything has been said as far as possible.
#
I would like to add that even a commercial mass product of the 1980s was able to achieve a flat THD curve up to 20kHz of less than 0.002% at 50W(rms) - and is still good for surprises today.
For example, the SU-V50 Matsushita / Technics brand hi-fi amplifier.
Measurements were made with the HP3045 at the time.
kind regards,
HBt.
I assume that everything has been said as far as possible.
#
I would like to add that even a commercial mass product of the 1980s was able to achieve a flat THD curve up to 20kHz of less than 0.002% at 50W(rms) - and is still good for surprises today.
For example, the SU-V50 Matsushita / Technics brand hi-fi amplifier.
Measurements were made with the HP3045 at the time.
kind regards,
HBt.
Because it was back in the 80’s when audio had finally matured. Things were understood well enough for distortion to be dominated by the things you can’t control, and to get it down to the point where improvements generally don’t matter. Semiconductors capable of doing the job were readily available. There will always be those that want performance an order of magnitude or two better for whatever reason, and they are free to spend as much to get it as they wish. But most people buying or building an amplifier simply don’t have to. Personally, I’ve achieved the sound I wanted from an amp I built in 1990. Every iteration since has either been improvements in the physical construction, scaling it up to higher power, or making practical daily use versions. A few years ago I picked some of this up after being dormant for a decade. I applied what I learned along the way to rebuilding old equipment and to a couple builds based on some classic retro designs (you, know, back when things had output caps and only NPN’s). The sound becomes indistinguishable from the big ones, when played below clip. And retain the same tolerance to overdrive as I have come to expect from the iron pig PA gear.
Unfortunately I don't have that fantastic Panasonic analyser nor manual, it was measured in Thailand.
It was measured at 200W on 20kHz 0.00216% an simulation result was 0.000534% or 4 times better.
BR Damir
True. 80’s class D sucked. Much of it still sucks today, but it doesn’t have to. Just like in the 80’s. Much of the class AB sucked, but it didn’t have to. We were capable of getting distortion close to the limits of perceptibility, but we still had the LM380 and TDA2002. Lots of them. And plenty of discrete amps that sounded like *** too.
I design for my enjoyment and my own use. My goal is to have the output match the input under all conditions (a criterion that is constantly growing and expanding). Others design for an output sound. Both are fine by me. Different goals. different measures of success.
As for THD+N as a useful metric, it is useful for me. Extremely low measured THD+noise does tell me that considerable attention has been given to PCB layout, to chassis layout and power supply design as well as other aspects of design. It is only one of very many measurements used in development.
This leads to simulation. First simulate what you make and make what you simulate. The theory works, your/my application of the theory may not. If my build is way off from the simulation, I usually find that I did not make what I simulated.
Again simulate what you make and make what you simulate.
As for THD+N as a useful metric, it is useful for me. Extremely low measured THD+noise does tell me that considerable attention has been given to PCB layout, to chassis layout and power supply design as well as other aspects of design. It is only one of very many measurements used in development.
This leads to simulation. First simulate what you make and make what you simulate. The theory works, your/my application of the theory may not. If my build is way off from the simulation, I usually find that I did not make what I simulated.
Again simulate what you make and make what you simulate.
You can't just state there's little point to low distortion capable amplification just because of poorly produced or inherently distorted recordings which don't require it.
Distortion adds up along the whole component chain, some of which can complement the perceived audio quality (2nd order HD) and some which makes it sound worse. Its not a matter of a blanket statement being an absolute suitable definition.
I don't listen to poorly recorded music and appreciate excellent recordings listened to on capable gear which delivers little if any perceived distortion. Thats really a critical trait of higher end audio reproduction. I can't vouch for the other people who can't hear a difference between a chain saw or violin on their horrible air pods or plastic TV sound bar. While many people consume music this way, I don't and never will. My priorities are simply different.
One of the things which really gets me ticked off is when people generalize saying most people don't care or can't tell the difference, then lump everyone else based on that criteria into one group as the defining majority. That's a very socialist way of thinking which forces those who care to be alienated.
Its like those people who say a well handling, high HP automobile is pointless to own as you can't "legally" take advantage of the performance and speed its capable of. Some of us really enjoy higher end, well made and performing audio equipment. I'm thankful there's a choice and that matters alot. I won't dumb myself down to the level of the masses just because someone says its the sensible or morally correct thing to do, let alone claim I won't notice a difference. I don't need people speaking for me that way, making my mind up for me or the rest of us who care.
The different of distortion in class A and class AB is the harmonic profile. Usually class A have monotonic harmonic profile. But THD don't care about how harmonic profile look alike, human ears can detect the different as long as the distortion is not too small.I recall that Stereophile uses a low-pass filter only on class D amplifiers. Using it on class AB would make the numbers equally dishonest.
tombo56 and knutn - Class A achieves truly low distortion without the switching that can goof up class AB. I am more inclined to believe simulated THD <0.01% on class A than class AB.
Ed
bimo, I'm glad you mentioned harmonic profile. In my experience those amplifiers with graded harmonics tend to sound better.
My main gripe is why is everybody chasing low THD? THD is the least of our worries when designing amplifiers. Intermodulation, TID and plenty of other Non-Harmonic distortions are the ones that are way more audible than THD. There are plenty of tests showing that THD below 1% is very difficult to identify and there are plenty of highly respected and great sounding tube/valve amplifier designs which do not have vanishingly low THD that go towards proving the case.
Historically this all started with Harold Leak's showmanship with his Point One amplifier promotions in the '50s which meant that, suddenly, every amplifier designer started chasing ridiculously low THD and that specification, almost alone, became an arbiter for amplifier quality. Then, with faster semiconductor devices, designers started applying more and more feedback, often with deleterious results for other aspects of sonic performance.
In my view it's all got out of hand, a bit like 0-60mph acceleration times for cars which, apart from being unobtainable by the user, don't tell you anything about what they're like to drive. We ought, instead of THD, to be looking at all the non-harmonic and dynamic distortions that amplifiers introduce in order to really understand why amps sound different. But THD is easy to measure and has become the defacto standard, so that's all people concentrate on.
Discuss!
My main gripe is why is everybody chasing low THD? THD is the least of our worries when designing amplifiers. Intermodulation, TID and plenty of other Non-Harmonic distortions are the ones that are way more audible than THD. There are plenty of tests showing that THD below 1% is very difficult to identify and there are plenty of highly respected and great sounding tube/valve amplifier designs which do not have vanishingly low THD that go towards proving the case.
Historically this all started with Harold Leak's showmanship with his Point One amplifier promotions in the '50s which meant that, suddenly, every amplifier designer started chasing ridiculously low THD and that specification, almost alone, became an arbiter for amplifier quality. Then, with faster semiconductor devices, designers started applying more and more feedback, often with deleterious results for other aspects of sonic performance.
In my view it's all got out of hand, a bit like 0-60mph acceleration times for cars which, apart from being unobtainable by the user, don't tell you anything about what they're like to drive. We ought, instead of THD, to be looking at all the non-harmonic and dynamic distortions that amplifiers introduce in order to really understand why amps sound different. But THD is easy to measure and has become the defacto standard, so that's all people concentrate on.
Discuss!
Those Dual CV series were decent and quite good sounding overall, only problem were those weird switches with poor reliability.
You are much too focussed on subtle non-linear effects. The few people on this forum who did double-blind amplifier tests found that there are only two things that matter, besides accurate gain matching:
Frequency response with an actual loudspeaker load
Clipping behaviour, if the amplifier is driven into clipping.
I don't want this thread to turn into a debate about the audibility of distortion.
THD is a measure of linearity. The techniques are well-understood for designing amplifiers so linear that all distortion is reduced to below the threshold of audibility. There is no downside to doing this.
IMO, one area that should be getting more attention is S/N. A system can disappear only if it has nearly-perfect linearity, is time-invariant, and has a dead silent background. From reading this board, one can see that hum and noise are much more problematic in practice.
Ed
THD is a measure of linearity. The techniques are well-understood for designing amplifiers so linear that all distortion is reduced to below the threshold of audibility. There is no downside to doing this.
IMO, one area that should be getting more attention is S/N. A system can disappear only if it has nearly-perfect linearity, is time-invariant, and has a dead silent background. From reading this board, one can see that hum and noise are much more problematic in practice.
Ed
Class D has its own problems. None that can’t be solved, but not with 80’s technology. And all these cheap one-chip solutions don’t necessarily go about things in the best possible way.
Personally I like the G and H class solutions. They are efficient ENOUGH for most uses, even for very high power. Waste heat gracefully increases as you go to very high power or low impedance loads. Class D can fall off a cliff quickly when switching losses run away and efficiency goes from 95% to 80. Then you end up in a situation where the heat sinking is insufficient. Class G would have been designed to deal with that 20% heat, not the typical 95% efficiency class D cases.
The commutation noise is never a problem in PA amp service. It is always masked by the music and effectively buried. In hi fi use it is usually true as well, but you might run into a situation where it can be audible. The non-switching variety, using linear amp transistors in the upper bank instead of a hard switch, along with Schottky steering diodes will get rid of the VCE glitch. And regardless of which type, I only use topologies with a f***-ton of NFB in these classes.
Personally I like the G and H class solutions. They are efficient ENOUGH for most uses, even for very high power. Waste heat gracefully increases as you go to very high power or low impedance loads. Class D can fall off a cliff quickly when switching losses run away and efficiency goes from 95% to 80. Then you end up in a situation where the heat sinking is insufficient. Class G would have been designed to deal with that 20% heat, not the typical 95% efficiency class D cases.
The commutation noise is never a problem in PA amp service. It is always masked by the music and effectively buried. In hi fi use it is usually true as well, but you might run into a situation where it can be audible. The non-switching variety, using linear amp transistors in the upper bank instead of a hard switch, along with Schottky steering diodes will get rid of the VCE glitch. And regardless of which type, I only use topologies with a f***-ton of NFB in these classes.