I'm puzzled: Amplifiers gain vs power, vintage vs modern

[alexk0il]

My 40 yo Hitachi SR-604 has died, can't find a replacement unit on ebay, so I'm evaluating modern amps. Regardless of how the sound quality of the modern gear compares to vintage, I found one confusing difference.

The vintage amp was playing at the same volumes with ~10db less amplifications at my dac. Quiet tracks - casually playing at -20db, loud tracks - about -30db. Speakers sensitivity 93Db (specs), 88Db (measured). Also have a proper sub.

Some of the new gear I tested (Class-D, measured 50-100Wt per channel), plays at ~-10db at my dac for normal listening. If I want to shake the house per my teenager's request I have to max out amplification at source and at amps, then my son starts complaining about vocal distortions. He also claims nothing plays music as good as the old one, but that's a different story.

Anyway, I'm confused, why did the industry decided to stop building high sensitivity/high gain amps and went for low sensitivity/high power amps instead? It is almost impossible to compensate the loss of 10db of sensitivity with power, even going from 100Wt to 200Wt only adds 3db, still a far cry from the 10-15db difference.

What do I miss?

[Macca]

Most amps have the same amount of gain they always did, around 26dB is standard.

If you can share specifics - what amplifiers and DAC you have been using, we might be able to work out what is occurring.

The vocal distortion does sound like the amp is being run into clipping.

[alexk0il]

Pretty much all Class D amps from Douk, Fosi, Ampapa. Highly praised Ampapa D1, for example.

The SQ is quite impressive actually. I never thought a class d could be so gentle and polite, yet resolving and dynamic.

But input sensitivity of 2V is a dhow-stopper for me. You are right about clipping, I saw a YouTube review showing the clipping before it reaches it's max.

Also, just realised I made an error, 150mV vs 2V is 22.5 db difference, not 10.

Anyway, please do ignore the Vintage vs Class D subtletylies, I am aware and just exercising an idea of having a tube pre-amp with clean class d power amp. Not as easy as I would have thought, bit it's a different story.

[Macca]

Most modern sources - DACs, CD players, streamers with DACs output 2 volts or close to that (single ended ofc). The CD redbook standard is 2 volts output.

So it does make sense for modern amps to have 2 volt input sensitivity (or 4 volts for balanced input). If the output of the source is lower, maximum possible output from the amp is reduced.

If input sensitivity is lower you run the risk of clipping the input with a 2 volt source which sounds bad as many early adopters of CD discovered when they plugged them into their ancient amplifiers.

(I'm guessing you already understand this but worth explaining for others who may be reading).

So I suspect you're encountering two problems - '

1) Distortion from the loudspeakers - they can't manage the SPL levels your son is after
2) Amp is running into clipping.

Due to 1) there's really no point getting a more powerful amp. You need a speaker with much better power handling.

Also your Klipsch speakers have elevated treble response, like most Klipsch. That probably isn't helping.

Only puzzle is why these same issues were not evident with the old Hitachi amp. Only thing I can think of is that many class D have harmonic distortion that rises with frequency. Usually it is still so low it is not an audible issue, but perhaps there is some sort of 'perfect storm' occurring with the high SPLs demanded and the elevated HF response of the speakers?

I'm speculating wildly of course. :)

[alexk0il]

I have to respectfully disagree.

1. I've tested a coupe of Marantz amps already, all have input sensitivity below 200mV. Some of the top amp models still have the old school sensitivity, easy to check: Nait XS3: 130mV, CA Edge A: 380mV.
2. The tests that show clipping were done with a simple 8Ohm resistance, not with real speakers. Clipped by D1 around 80% of volume. Takes away 5-10db of it's dynamic range, I guess.
3. Additionally, D1 has 2V sensitivity, ie 2V for full power, it is effectively 22.5 db of less gain compared to 150mV from an old school amp.

So here is the math, old Hitachi amp - input sensitivity 150mv. Means I only need 150mv to get to full 35+35Wt (G class) power. Source at -30 to -20dB gave me plenty of headroom for comfortable running with amp always at max volume. No clipping by the analog amp, I suspect I never the 35W of pure A class.

With D1 -30 db now becomes -30 + 22.5 db = ~-8Db. And this is where the clipping by D1 starts.

Anyway, I'd rather return D1 as not fit for my purpose, and go for a different amp to gain my 22db back. Might upgrade the speakers later, but my thought process is currently: good source -> good pre -> good amp -> matching speakers.

[Macca]

Not sure what you're disagreeing with except my speculative bit at the end which I did say was speculative.

Yes you can still find amps with lower input sensitivity. If that suits your use case better, that's what you need.

Clipping tests are always done with a dummy load as speaker impedance varies wildly so it's the only way to test which will be comparative between amps, although I agree it doesn't really represent a 'real world' load.

[alexk0il]

I guess I disagree with the assertion that increasing the sensitivity all the way to 2V to protect speakers is justified considering how the DR is lost. After all, there is still plenty of modern amps that still use high sensitivity gains. Apologies if this is not what you meant.

[Macca]

I guess I disagree with the assertion that increasing the sensitivity all the way to 2V to protect speakers is justified considering how the DR is lost. After all, there is still plenty of modern amps that still use high sensitivity gains. Apologies if this is not what you meant.

No it wasn't!

I think primarily the reason is to avoid input clipping with sources with higher output voltages. Not uncommon to find amps that, with a 2 volt input, have very little travel on the volume control before it's too loud. Naim, for example, making it very hard to set the exact level you want.

Also, would it not make it easier to get a better SNR? Not sure about that.

[Barry]

Your Hitachi SR-604 produces 35W into 8Ohm. That is a maximum voltage of 16.73V at the speaker terminals. This is for an input voltage of 150mV, so the gain is 20.log(16.73/0.15) = 40.9dB.

In general you find you need an input level some 20 to 30dB lower for comfortable listening. This makes sense as the difference between 2V and 150mV is 22.5dB

The digital amplifiers you list are all designed to accept the 2V output from a digital source.
In particular the Ampapa D1 digital amplifier produces 102W into 8Ohm or 28.57V at the speaker terminals. The gain is thus 20.log(28.57/2) = 23.1dB, or 17.8dB less gain than the Hitachi amplifier. This is similar figure to the 22.5dB of attenuation you need to use with the Hitachi amplifier.

So you need 5 to 10dB less attenuation with digital amplifiers, compared to the 20 to 30dB you need with the Hitachi.

[Lawrence001]

I don't trust the wild claims made about a lot of these chip amps, I suspect the power output is similar to the PMPO measure they used to give for Amstrad and Alba hifi you could buy in Rumbelows. Which would explain why your son complained about distortion, trying to maintain a continuous power level it can only really do in bursts won't sound good and won't do the amp or speakers much good either.

If you want to shake the house you're better off buying a PA system for him and keeping a nice sounding hifi for yourself.

[Macca]

I don't trust the wild claims made about a lot of these chip amps, I suspect the power output is similar to the PMPO measure they used to give for Amstrad and Alba hifi you could buy in Rumbelows. Which would explain why your son complained about distortion, trying to maintain a continuous power level it can only really do in bursts won't sound good and won't do the amp or speakers much good either.

If you want to shake the house you're better off buying a PA system for him and keeping a nice sounding hifi for yourself.

They tend to quote power into 4 ohm load at 1 percent THD when it should really be 8 ohm load at 0.1 percent. So it is inflated if doing a like by like comparison to old amps.

But quite a few have been tested on ASR and most do meet or come close to the claimed power. Ability to maintain voltage into lower impedances is a different matter and that will make the amp clip if it falls short. Some have optional higher voltage power packs to offset this.

[Pigmy Pony]

They tend to quote power into 4 ohm load at 1 percent THD when it should really be 8 ohm load at 0.1 percent. So it is inflated if doing a like by like comparison to old amps.

But quite a few have been tested on ASR and most do meet or come close to the claimed power. Ability to maintain voltage into lower impedances is a different matter and that will make the amp clip if it falls short. Some have optional higher voltage power packs to offset this.

Like these Temple Audio monobloks with Supercharger power supplies, which did a superb job driving my speakers at the last NEBO.

https://i.postimg.cc/t4NjS8sT/20251004-183620-1.jpg

[alexk0il]

Your Hitachi SR-604 produces 35W into 8Ohm. That is a maximum voltage of 16.73V at the speaker terminals. This is for an input voltage of 150mV, so the gain is 20.log(16.73/0.15) = 40.9dB.

It's G class, it adds 35dB when the pure G class mode kicks in, so ~44dB in total.




In general you find you need an input level some 20 to 30dB lower for comfortable listening. This makes sense as the difference between 2V and 150mV is 22.5dB

The digital amplifiers you list are all designed to accept the 2V output from a digital source.
In particular the Ampapa D1 digital amplifier produces 102W into 8Ohm or 28.57V at the speaker terminals. The gain is thus 20.log(28.57/2) = 23.1dB, or 17.8dB less gain than the Hitachi amplifier. This is similar figure to the 22.5dB of attenuation you need to use with the Hitachi amplifier.

I definitely need this 22.5dB of attenuation with my Hitachi for the headroom to accommodate the differences between quiet and loud recording, plus shaking the house while watching actions movies vs quiet background music for lazy breakfasts.

Someone on youtube actually measured it to be around 80Wt into 8 Ohm, similar to Hitahi's G class, roughly -1.2dB to be accurate.
[/QUOTE]

So all together, your math could be corrected by extra 2+1.2 = 4.2 dB of extra for Hitachi case.

[Lawrence001]

Am I right in thinking your speakers are standmounts? As the idea of trying to "shake the house" with those would be way too much to expect. You'd want a pair of beefy floorstanders or 15" woofers or something similar for that.

[Barry]

I only went by the manufacturer’s specification for power into 8 Ohm and for the input sensitivity. How your Hitachi amp creates 35W into 8 Ohm for an input voltage level of 150mV, is all that matters, and implies an amplifier gain of 41dB.

If someone has measured the maximum power output of the Ampapa D1 to be 80W into 8 Ohm for an input of 2V, then I agree that reduces the gain; so you need less attenuation to prevent the amp exceeding the 35W maximum of your Hitachi.

Either way I would recommend you use a 10dB attenuator with the Ampapa D1, and say 25 - 30dB with your Hitachi. This will keep your speakers safe and give you more control of the volume level.

[Pigmy Pony]

Am I right in thinking your speakers are standmounts? As the idea of trying to "shake the house" with those would be way too much to expect. You'd want a pair of beefy floorstanders or 15" woofers or something similar for that.

Or a powered sub or two, that might give your speakers an easier life when 'room shaking' is called for.

[Lawrence001]

Or a powered sub or two, that might give your speakers an easier life when 'room shaking' is called for.Good idea, especially if the subs have a high pass speaker out that cuts the bass off to the standmounts.

[alexk0il]

I only went by the manufacturer’s specification for power into 8 Ohm and for the input sensitivity. How your Hitachi amp creates 35W into 8 Ohm for an input voltage level of 150mV, is all that matters, and implies an amplifier gain of 41dB.
.

The extra 35Wt are for those peaks when you need it. I am sure I never used it.

https://www.petervis.com/gallery/Vintage%20Advertisements/hitachi-hi-fi-1979/hitachi-hi-fi-1979/hitachi-sr-604.webp

[alexk0il]

Am I right in thinking your speakers are standmounts? As the idea of trying to "shake the house" with those would be way too much to expect. You'd want a pair of beefy floorstanders or 15" woofers or something similar for that.

The shaking is done by the sub, the speakers are only matching sub's energies in mids and highs.

[Barry]

The extra 35Wt are for those peaks when you need it. I am sure I never used it.

https://www.petervis.com/gallery/Vintage%20Advertisements/hitachi-hi-fi-1979/hitachi-hi-fi-1979/hitachi-sr-604.webp

Typical misleading specifications: there is no such thing as RMS power, it is mean power and is the only quoted power specification that is honest, and independent of signal waveform shape.

There are other manufacturers who quote spurious and undefined power specifications such as "Peak effective power", or similar.

[alexk0il]

The specs are aligned with principles of G class operation. A class provides 35W, G class adds incrementally additional 35W when needed. Either for transients or for longer use. Two separate power supplies, dedicated circuitry. 75wt are real.

https://www.arcam.co.uk/ugc/tor/AVR30/Class%20G%20Explained/Class_G_the_ultimate_amplifier_technology_150714_A .pdf

[Barry]

Thanks for the link Alex.

Before you mentioned your Hitachi amplifier, I had not heard of Class G amplification. I had heard of Class A, AB, C and D (which I thought covered all digital amps), but not Class G (was there ever a Class E and Class F?).

You, and the article explain it well in that Class G amplifiers have two power supplies and two voltage rails, the second, higher voltage supply, switching in when extra power is needed. The fact that it can do this without any audible artifacts: clicks or a minute gap of silence is impressive. Which brings me to the question of gain. Assuming that a full 70 W is attainable with an input of 0.15 V (that is a gain of 43.9dB ), does that mean you only need 0.106 V for an output of 35 W, as the gain is the same? What happens when the input signal voltage varies around 0.106 V - does the higher voltage power supply switch in and out at the same rate, and if so is it inaudible? I have since learned there are Class H amplifiers which have three power supplies.

As you can see there is a lot I don't know about these high efficiency class amplifiers.

Which brings me on to the power rating adjectives. I was somewhat arrogant and dogmatic when I said that there was no such thing as 'RMS power'. Another Member here pointed out to me that the phrase 'rms power' is used a lot in audio, especially in connection with the power handling of loudspeakers. So I Googled 'rms power' to find out. The AI reply was that 'rms power' is the continuous or average power an audio device such as a speaker or amplifier can sustain. It then goes on to say that rms power is a measure of continuous heating effect (equivalent to DC power) of the AC signal in a resistor, whereas 'mean power' refers to the arithmetic average over time (which can be zero for pure AC (???) ), but often means 'average power' calculated from rms voltage and rms current

Now I regard the second clause of the last sentence as obvious nonsense. The AC mains that comes into your house is close to being a pure AC voltage at a frequency of 50Hz, yet it certainly dissipates power in the light bulbs and other appliances (i.e. your audio system). If it did not it would be of any use. The average energy dissipated in a resistor or a loudspeaker voice coil is the same for both the positive and the negative going parts of the AC signal.

So how do we describe the power rating of your Hitachi amplifier? As I have said, I think the term 'rms power' is confusing, if not incorrect. Both 35W and 70W to my mind are mean powers, so perhaps we should say "lower" and "upper" mean power? That could also be confusing to some, so it is probably best left 'as is'.

So thanks for the link - I certainly have learnt a lot today, and my apologies if I have come across as dogmatic, pedantic and arrogant. The last three undesirable traits are something I will have to guard against, especially as I get older!

[Macca]

Thanks for the link Alex.

Before you mentioned your Hitachi amplifier, I had not heard of Class G amplification. I had heard of Class A, AB, C and D (which I thought covered all digital amps), but not Class G (was there ever a Class E and Class F?).


You had a listen to a class G amplifier at the meet we had in Stoke, Justin brought one along and we hooked it up to the QUAD ESLs.

Class E and F are I think switching amplifiers for high power RF transmission, not audio amplifiers.

[Barry]

I don't remember that Martin. It might have been demonstrated whilst we we were having lunch away from the hotel.

[Macca]

I don't remember that Martin. It might have been demonstrated whilst we we were having lunch away from the hotel.

Possibly, yes. It was a long time ago now. I know Arcam make some class G amps but I think Justin's was from a pro brand.

[User211]

Possibly, yes. It was a long time ago now. I know Arcam make some class G amps but I think Justin's was from a pro brand.That's right, Martin. I just had to look behind the lounge telly cos I couldn't remember the name of it. It was/is an Apart Champ One. I still use it to drive those white Mark&Daniel Sapphire loudspeakers I also brought to that bake off.

It was amusing watching it drive Alan's Quads with the clipping lights coming in when it is supposed to deliver 600 odd Watts. It doesn't do that driving my Apogees. For the £200 I paid for it it sounds remarkably decent I reckon.

That bake off was good fun. Ali's OBs were great I seem to recall.

[Barry]

Agreed about Ali's OB speakers and of his whole system he was demonstrating at the get-together.

It was a long time ago but it was great fun; especially meeting and putting a face to some of our Members here. Perhaps we should try to organise another one.

[Macca]

That bake off was good fun. Ali's OBs were great I seem to recall.

They were, we also had a couple of sets of vintage JBLs that went down well. People brought along some really good kit.

Would have done another but that hotel closed down a couple of months later and has never re-opened.

[Pigmy Pony]

Agreed about Ali's OB speakers and of his whole system he was demonstrating at the get-together.

It was a long time ago but it was great fun; especially meeting and putting a face to some of our Members here. Perhaps we should try to organise another one.

NEBO is a great day out, but we could also do with something more 'darn sarf', more central.

[Barry]

Sadly NEBO is just too far for me. Somewhere in the Midlands would be better.

[Pigmy Pony]

Yes that's what I was thinking, though we're probably too lazy to organise it. I know I am. And the Maverick show (formerly 'Wam show) is too far north for many.

[Macca]

I tried to book a conference room at the university. It's ideal as there's loads of parking, and there's food, drink and accommodation on site - Marriot hotel.

But they seemed to be booked up in perpetuity.

Hard to find somewhere that ticks all the boxes and doesn't cost a fortune.

[Barry]

Is it booked up in the student holidays? They might be hosting Open University meetings. Pity, as we would only need it for one day, and I'm sure we could all share the hire costs.

[Macca]

Is it booked up in the student holidays? They might be hosting Open University meetings. Pity, as we would only need it for one day, and I'm sure we could all share the hire costs.

I tried booking in the Easter holidays, but no joy. That was last year though.

[alexk0il]

You had a listen to a class G amplifier at the meet we had in Stoke, Justin brought one along and we hooked it up to the QUAD ESLs.


Please tell me you loved it. Even though my Hitachi is dead, I am still sentimentally attached to that 50 yo all G-class tech. except for modern Rotel's :) :)

Anyway, my journey with modern audio is only started, I will share on this forum my experience when I have time. Hint, no more recapping of vintage masterpieces for me, the world has changed.

[Lawrence001]

Thanks for the link Alex.

Before you mentioned your Hitachi amplifier, I had not heard of Class G amplification. I had heard of Class A, AB, C and D (which I thought covered all digital amps), but not Class G (was there ever a Class E and Class F?).

You, and the article explain it well in that Class G amplifiers have two power supplies and two voltage rails, the second, higher voltage supply, switching in when extra power is needed. The fact that it can do this without any audible artifacts: clicks or a minute gap of silence is impressive. Which brings me to the question of gain. Assuming that a full 70 W is attainable with an input of 0.15 V (that is a gain of 43.9dB ), does that mean you only need 0.106 V for an output of 35 W, as the gain is the same? What happens when the input signal voltage varies around 0.106 V - does the higher voltage power supply switch in and out at the same rate, and if so is it inaudible? I have since learned there are Class H amplifiers which have three power supplies.

As you can see there is a lot I don't know about these high efficiency class amplifiers.

Which brings me on to the power rating adjectives. I was somewhat arrogant and dogmatic when I said that there was no such thing as 'RMS power'. Another Member here pointed out to me that the phrase 'rms power' is used a lot in audio, especially in connection with the power handling of loudspeakers. So I Googled 'rms power' to find out. The AI reply was that 'rms power' is the continuous or average power an audio device such as a speaker or amplifier can sustain. It then goes on to say that rms power is a measure of continuous heating effect (equivalent to DC power) of the AC signal in a resistor, whereas 'mean power' refers to the arithmetic average over time (which can be zero for pure AC (???) ), but often means 'average power' calculated from rms voltage and rms current

Now I regard the second clause of the last sentence as obvious nonsense. The AC mains that comes into your house is close to being a pure AC voltage at a frequency of 50Hz, yet it certainly dissipates power in the light bulbs and other appliances (i.e. your audio system). If it did not it would be of any use. The average energy dissipated in a resistor or a loudspeaker voice coil is the same for both the positive and the negative going parts of the AC signal.

So how do we describe the power rating of your Hitachi amplifier? As I have said, I think the term 'rms power' is confusing, if not incorrect. Both 35W and 70W to my mind are mean powers, so perhaps we should say "lower" and "upper" mean power? That could also be confusing to some, so it is probably best left 'as is'.

So thanks for the link - I certainly have learnt a lot today, and my apologies if I have come across as dogmatic, pedantic and arrogant. The last three undesirable traits are something I will have to guard against, especially as I get older!
You had a listen to a class G amplifier at the meet we had in Stoke, Justin brought one along and we hooked it up to the QUAD ESLs.

Class E and F are I think switching amplifiers for high power RF transmission, not audio amplifiers.Don't manufacturers just make it up as they go along, like "New Class A" and "Class AA", does Class G add anything more to other attempts to get close to class A with the efficiency approaching class AB?

[Macca]

Don't manufacturers just make it up as they go along, like "New Class A" and "Class AA", does Class G add anything more to other attempts to get close to class A with the efficiency approaching class AB?

amplifier classes are just the order in which they were developed with class A being the first and the simplest. They don't have any relevance to sound quality, which depends on the implementation.

[Lawrence001]

What I mean is who decides if a design is different enough to justify being classified as a new class or just a variation on an old one?

[Macca]

What I mean is who decides if a design is different enough to justify being classified as a new class or just a variation on an old one?

I don't know and it seems the internet doesn't have the answer either.

[Lawrence001]

So I suspect they just make it up as they go along!

[Macca]

I'm wondering if it has something to do with the patent applications.

[Pigmy Pony]

Would class A amplifiers have been called class A before class B came along, because that would have expected obsolescence written all over it :scratch:

[Macca]

I'd guess not.

Like a mark one Cortina wasn't called the mark one until the mark two came out.

[Pigmy Pony]

And similarly the SME Series II Improved would not have been preceded by the SME Series II Could Be Better.

[Lawrence001]

How many manufacturers have released something they say could never be improved?

[Pigmy Pony]

I think CDs used to be marketed as "perfect sound forever".

[Macca]

There's a few products with 'Ultimate' in the name.

I think NVA did an amp called 'The Definitive Statement'

[Barry]

The answer to all these questions is not clear cut or straightforward. One has to look into the history of amplifying devices first, then consider developments (especially with the advent of solid state electronics).

A very good article to read is "What are the different types of audio amplifier classes?" at https://www.audioholics.com. Or there is the redoubtable Wikipedia. Distilling down the information from the audioholics site, I will try to provide a synopsis.

Class A
The defining principle of Class A operation is that all of an amplifier's output devices must be conducting through the full 360 degree cycle of a waveform. There is a further division between single-ended and push-pull amplifiers.
Push-pull amplifiers diverge from the basic explanation above in that a pair of output devices is used. Whilst both devices are conducting through the full 360 degree cycle, one device will 'shoulder' more of the load during the first half of the cycle, with the second device doing the same for the second half. The primary advantage of this arrangement is reduced distortion relative to single ended designs; as even order harmonics are cancelled out.
The drawback to Class A amplification is one of efficiency. The requirement of Class A designs to have all output devices to be conducting at all times results in a significant amount of wasted power, which is converted into heat. This is further exacerbated by the high level of quiescent current, the amount of current flowing through the ouput devices when the amplifier is producing zero output. This means the theoretical maximum efficiency is only 50%, and in practice will only be 15 - 35%.

Class B
Whilst all of the output devices in a Class A amplifier are conducting all of the time, Class B amplifiers use a push-pull arrangement so that only half of the output devices are conducting at any one time. As a consequence Class B amplifiers more efficient, with a theoretical maximum of 78.5% (= pi/4). However a serious disadvantage of pure Class B is that of crossover distortion; a problem with the delay in handing over one half of the waveform from one device to its partner to handle the second half of the waveform.

Class AB
This class combines the best of Class A and Class B to create an amplifier without the drawbacks of either. Whereas Class B uses a push-pull arrangement with each half of the output stage conducting for 180 degree, Class AB increases that to 181 - 200 degree. In so doing so, there is far less of a 'gap' in the cycle to occur, and consequently the crossover distortion is diminished to acceptable levels. Depending on how the output devices are biased, that is when in Class A mode at low powers to avoid distortion, and moving towards Class B operation at higher powers, efficiency rates of 50 - 70% can be achieved.

Class C
Here the device(s) conduct for significantly less than half the input cycle. They are highly efficient but produce heavy distortion, so is used in non-audio applications such as RF transmitters, where a tuned circuit can be used to reconstruct the signal.

Class D
This is a highly efficient system that uses pulse width modulation. Efficiency rates are > 90%.


Class G and H
Neither of these classes are officially recognised, Instead they are considered as variations on Class AB, using voltage rail switching and voltage rail modulation respectively.


As to manufacturers developing their own output circuit topologies and having to give them names, this is done largely to impress the buying public, with tenuous relation to electronic theory. An example is of Technics using the term 'New Class A' and later 'Class AA'. The two are quite different, but confusion arose as the general public were led to believe they were synonymous.
According to contributors to Pink Fish Media, and in particular to Jez Arkless (one time Member of this parish), Class AA was a non-switching design by Dr Sandman, which uses a low bias Class B amplifier to do the 'heavy lifting' and drive the load so that a small, more linear amp can 'fill in' because it doesn't 'see' a low load impedance. There was a legal dispute between Dr Sandman and Technics (Panasonic and hence Matsushita) over the implementation, and one supposes that Matsushita could afford to employ better or more skilled patent lawyers, to their advantage.

There were several design using a low power Class A amplifier and 'feed forward' to devices which would provide the necessary current when high power was required. Quad very sensibly simply called this 'Current Dumping' rather than "Class Q" ( :lol: ), although there are still many circuit analysts who are still unclear as to how Quad's idea actually works. All they can agree on is that to have perfect balance, the components used in the bridge circuit need to have very close tolerance on their values.

[Firebottle]

The description of the various classes are by no means recent. A snippet from the web:

Formal classification (1930s-1940s): The terms Class A, Class B, and Class AB were formally introduced during this period.

[Barry]

The three classes were almost introduced from day one with the Lee Forest's invention of the triode valve.

[Macca]

The description of the various classes are by no means recent. A snippet from the web:

Formal classification (1930s-1940s): The terms Class A, Class B, and Class AB were formally introduced during this period.

That doesn't answer the question of who decides on whether a circuit is a new 'class' or not. QUADs current dumping is still class AB but rail switching gets its own class (or possibly two, I think class H is also rail switching?).

[Barry]

I would imagine it is the relevant professional institutions who decide, such as the IEE in the UK and the IEEE in America, and whoever it is for the EU.

[Barry]

If you type: "Who decides if the design of an audio amplifier should be designated by a new Class?" into your smartphone, you get this AI overview response:



There is no single governing body that officially designates new amplifier classes. Instead, new classes are established through a combination of engineering practice, academic research, and industry consensus, often driven by manufacturers developing proprietary designs.

The fundamental classifications (Class A, B, AB, C and D} are defined by the operating point and the conduction angle of the amplifying device. More recent or specialised classes (e.g. Class G,H,I,T,AD) typically build upon these basic topologies with proprietary improvements and are named by the engineers and companies that develop them.

The process generally follow these steps:

Engineering Innovation: An engineer or design team develops a new amplifier topology to meet specific performance goals, such as improved efficiency, reduced distortion, or smaller size.

Documentation and Prototyping: The designers may publish their findings in technical papers, apply for a patents, or present their innovations at industry conferences (e.g. those hosted by the Audio Engineering Society (AES)).

Industry Adoption and Marketing: Manufacturers then adopt the design, often using the new "class" designation in their marketing materials to highlight the unique aspects of their product.

Consensus: The new class enters general usage if it offers a clear, repeatable advantage and is widely adopted by other manufacturers, engineers, and technical reviewers in the audio community.


Ultimately, the market and the broader community decide if a new class designation has merit and becomes a recognised standard or remains a specific proprietary term.




This would appear to be the definative answer, but if it isn't then I can't be bothered to do any more research (though I'm still curious about Class E and F, if they ever existed.)

[Lawrence001]

The answer to all these questions is not clear cut or straightforward. One has to look into the history of amplifying devices first, then consider developments (especially with the advent of solid state electronics).

A very good article to read is "What are the different types of audio amplifier classes?" at https://www.audioholics.com. Or there is the redoubtable Wikipedia. Distilling down the information from the audioholics site, I will try to provide a synopsis.

Class A
The defining principle of Class A operation is that all of an amplifier's output devices must be conducting through the full 360 degree cycle of a waveform. There is a further division between single-ended and push-pull amplifiers.
Push-pull amplifiers diverge from the basic explanation above in that a pair of output devices is used. Whilst both devices are conducting through the full 360 degree cycle, one device will 'shoulder' more of the load during the first half of the cycle, with the second device doing the same for the second half. The primary advantage of this arrangement is reduced distortion relative to single ended designs; as even order harmonics are cancelled out.
The drawback to Class A amplification is one of efficiency. The requirement of Class A designs to have all output devices to be conducting at all times results in a significant amount of wasted power, which is converted into heat. This is further exacerbated by the high level of quiescent current, the amount of current flowing through the ouput devices when the amplifier is producing zero output. This means the theoretical maximum efficiency is only 50%, and in practice will only be 15 - 35%.

Class B
Whilst all of the output devices in a Class A amplifier are conducting all of the time, Class B amplifiers use a push-pull arrangement so that only half of the output devices are conducting at any one time. As a consequence Class B amplifiers more efficient, with a theoretical maximum of 78.5% (= pi/4). However a serious disadvantage of pure Class B is that of crossover distortion; a problem with the delay in handing over one half of the waveform from one device to its partner to handle the second half of the waveform.

Class AB
This class combines the best of Class A and Class B to create an amplifier without the drawbacks of either. Whereas Class B uses a push-pull arrangement with each half of the output stage conducting for 180 degree, Class AB increases that to 181 - 200 degree. In so doing so, there is far less of a 'gap' in the cycle to occur, and consequently the crossover distortion is diminished to acceptable levels. Depending on how the output devices are biased, that is when in Class A mode at low powers to avoid distortion, and moving towards Class B operation at higher powers, efficiency rates of 50 - 70% can be achieved.

Class C
Here the device(s) conduct for significantly less than half the input cycle. They are highly efficient but produce heavy distortion, so is used in non-audio applications such as RF transmitters, where a tuned circuit can be used to reconstruct the signal.

Class D
This is a highly efficient system that uses pulse width modulation. Efficiency rates are > 90%.


Class G and H
Neither of these classes are officially recognised, Instead they are considered as variations on Class AB, using voltage rail switching and voltage rail modulation respectively.


As to manufacturers developing their own output circuit topologies and having to give them names, this is done largely to impress the buying public, with tenuous relation to electronic theory. An example is of Technics using the term 'New Class A' and later 'Class AA'. The two are quite different, but confusion arose as the general public were led to believe they were synonymous.
According to contributors to Pink Fish Media, and in particular to Jez Arkless (one time Member of this parish), Class AA was a non-switching design by Dr Sandman, which uses a low bias Class B amplifier to do the 'heavy lifting' and drive the load so that a small, more linear amp can 'fill in' because it doesn't 'see' a low load impedance. There was a legal dispute between Dr Sandman and Technics (Panasonic and hence Matsushita) over the implementation, and one supposes that Matsushita could afford to employ better or more skilled patent lawyers, to their advantage.

There were several design using a low power Class A amplifier and 'feed forward' to devices which would provide the necessary current when high power was required. Quad very sensibly simply called this 'Current Dumping' rather than "Class Q" ( [emoji38] ), although there are still many circuit analysts who are still unclear as to how Quad's idea actually works. All they can agree on is that to have perfect balance, the components used in the bridge circuit need to have very close tolerance on their values.This seems to agree with my point that Class G sounds like another variation on AB along the lines of New Class A and Class AA.

[helma]

Anyway, I'm confused, why did the industry decided to stop building high sensitivity/high gain amps and went for low sensitivity/high power amps instead? It is almost impossible to compensate the loss of 10db of sensitivity with power, even going from 100Wt to 200Wt only adds 3db, still a far cry from the 10-15db difference.

What do I miss?

Gain and power don't really have anything to do with each other, you could have a 1W amp with 26dB gain and a 100W amp with 26dB of gain with identical volume controls and they would be equally as loud for as long as the 1W amp can keep up. A typical volume control is just an attenuator which at full volume lets all the input signal through. So with the same source the 1W amp might start distorting beyond specs with the volume at 10 o'clock, while the 100W amp could go all the way and just keep getting louder.

A sensible amplifier will have specified input sensitivity that's in accordance with the sources it will be used with. Vintage amps had high sensitivities because that was needed with the sources of the time. A typical phono preamp or tape deck might put out something like 0.3V, so with a modern 2V source the volume control becomes deceiving. Using a CD-player with 2V output you might already be pushing the amplifiers specs with the volume control halfway up or something. In such a case it's easy to get the misconception there's something more on tap, even though the amp might already be exceeding it's distortion specs and rated power output. Some vintage amplifiers it seems to me might have done that on purpose, I've come across some integrateds where the power amp input sensitivity was so high it was practically guaranteed it would be heavily distorting at "full volume" no matter what you were using for a source. I guess it would be impressive in a showroom to have the amp really loud at 10'o clock, even though all you're really doing is limiting the useful range of the volume control. To give them the benefit of the doubt, things also didn't seem as well standardized back then, or there were many competing standards, so you couldn't very precisely assume what the source is going to put out. I guess in that case they would've made sure you can get full power even with the weakest sources.

Amplifiers don't typically fall off a cliff after their rated power, the rated power is simply a number it can do with a specified level of distortion into a specified load, so for example an amp with rated power of 75W at 0.1% THD might be capable of putting out more power but higher levels of distortion. I'm not really familiar with class D amps and what tends to happen beyond their specified power rating, so I guess it's possible they distort in a more nasty and immediately obvious way after that point, or simply shut themselves down. If that's the case, a 75W class AB amp might in practice appear more powerful than a 75W class D amp, if pushed to their limits and beyond. A bit like how valve guitar amplifiers used to "have louder watts" than their solid state counterparts, because they would distort in a more pleasing way so you could drive them well beyond their rated power, while the solid state amps would sound horrible when their power amps started distorting.