What are the advantages/disadvantages of different amp topologies?

[morricab]

“single ended“, I am assuming, meaning the signal is simply amplified without adding complicated (and sound worsening) distortion reduction or power increasing circuitry as found in some OTL and Push Pull amps (respectively).

The distortion, allegedly caused by transformer saturation, I read is also reduced/eliminated by building the transformers of the correct materials and size, carefully. High power triodes are few and far between, operating voltages lethal, so perhaps we should concentrate our attention on perfecting the higher efficiency speaker so that we can get by with low parts count flea-powered SET amplifiers (like what used to be norm).

Single ended means that a single active element or multiple elements in parallel are amplifying the whole waveform. OTLs are also Push pull as are most SS and most tube amps. Push pull amps have at least one active element for each half of the signal. One for the positive going signal and one for the negative going signal.

Hi power triodes are not so uncommon. Your Ayon uses a high power variant of the 300b. It is a single ended amp as it has only 1 output tube per channel.

 

[DasguteOhr]

There are several factors that should be taken into account when choosing a set amp. Yes, the OPT is a big factor, but the power supply has just as much influence. E.g. the last capacitor (CLC) of the filtering has an extreme influence on the sound. By the way, this capacitor should be the main grounding point (stargrounding) of the amp circuit and it has proven successful for me. Please use high quality there, you will be rewarded as promised. the OPT should be generously dimensioned an example. Good power transformer, choke filter ,oil caps and a good OPT are a good path to success.
300b/2a3 opt

 

[bonzo75]

Set 100 is a hybrid amp

the model I was referring to is Aelius

 

[Cellcbern]

Single ended means that a single active element or multiple elements in parallel are amplifying the whole waveform. OTLs are also Push pull as are most SS and most tube amps. Push pull amps have at least one active element for each half of the signal. One for the positive going signal and one for the negative going signal.

Hi power triodes are not so uncommon. Your Ayon uses a high power variant of the 300b. It is a single ended amp as it has only 1 output tube per channel.

FYI: Best sounding "high power" SET amp I've heard:

833 Single Ended Power Amplifier | alexusaudio

 

[Cellcbern]

FYI: Best sounding "high power" SET amp I've heard:

833 Single Ended Power Amplifier | alexusaudio

FYI: These two Alexus SET amps are also outstanding:

33SE Tube Power Amplifier | alexusaudio

845SE Single-Ended Class A Power Amplifier | alexusaudio

Alexus Audio 845SE Single-Ended mono amplifiers

After moving on from my long-time reference Eminent Technology LFT-8b’s loudspeakers (inefficient but wonderful) ) to the Tekton Double Impact’s (SUPER efficient and wonderful), I received word that Alex Chorine of Alexus Audio had a new pair of mono blocks for me to try out. As much as I love...

www.stereotimes.com

 

[Amir]

- Imagine the best topology is zero feedback SET DHT
- Imagine the best is using just one tube not paralleling two or more tubes
- Imagine the best tubes are low voltages (low power)

what you find is a low power amplifier than can not drive most loudspeakers so you need to go for horns .

if you want good sound from horns you should have 5way or 6way horns not two way or three way.

finally you will find the Romy the Cat audio rules are the best rules in Audio.

Romy The Cat's Audio Rules

GoodSoundClub - Romy the Cat's personal High-End Audio site

goodsoundclub.com

 

[Amir]

About hybrids:

I would like to ask you two questions about hybrids :
1- do you think all hybrids (input stage tubes / outputstage SS) have coherent harmonic structure?
2- do you think using tubes in line level (dac, pre , input stage of power) is transparent enough in a reference transparent system?

I ask these questions because I never had good experience with hybrids

 

[bonzo75]

if you want good sound from horns you should have 5way or 6way horns not two way or three way

please state some listening experience for this ridiculous statement rather than simply quoting Romy

 

[Amir]

please state some listening experience for this ridiculous statement rather than simply quoting Romy

 

[bonzo75]

View attachment 124529

Bill and I are close friends chat daily and well aligned on our horn strategies. After that post he bought a fibre glass horn, single driver, also my favourite, and now is building a dual FLH, also my favourite. Our top two carts and TT are also same, and we spend time exchanging notes on records and music and videos

what are you doing apart from googling

 

[Atmasphere]

So how do we get that “right” sound, as close to unamplified acoustic instruments as possible?

If Class of amplification isn’t, is it down to limiting capacitance, simplifying design, using interstage transformers, managing voltage regulation with valves, pure silver or copper wire, hollow state over solid state, NOS valves from the classic producers over new valves from China or Russia? So far, I have not been able to discern any absolutes here.

One absolute that is quite measurable is distortion must not rise with frequency. If it does, it will be perceived as harshness and brightness.

Any zero feedback circuit can do this and that is part of why feedback has a bad rap in high end audio. But it is possible (although not until recently in audio history) to build an amp with feedback that also has this quality.

In addition, the distortion product must be benign. To that end it must be lower ordered harmonics (the 2nd and 3rd, which are treated by the ear the same way in that they add 'warmth' or 'bloom', using audiophile terms) with enough amplitude to mask any succeeding harmonics, the latter of which must fall off on an exponential curve (if the circuit is zero feedback). The exponent of the exponential curve does not seem to be important; once you exceed that of a quadratic exponent it becomes moot.

You can satisfy the above requirements with an SET, a fully differential PP tube amp, a class A or AB solid state amp (although this is very diifficult technology to work with if you are to satisfy the first requirement above) and class D.

So there is no topology that is an absolute to get to this 'right' sound. It is easiest to do with SETs and hardest to do with A or AB solid state amps. Each tech has its drawbacks. SETs for example make the most distortion of any amplifier concept and suffer bandwidth problems if making much more than about 8 Watts. So you need high efficiency speakers to work with them and even then you can only use about 20% of full power if you really want to hear what the SET is about.

If you are running zero feedback, bandwidth is important to prevent phase shift. Ideally this means 2Hz to 100KHz (since phase shift above 10KHz likely isn't audible). This bandwidth issue makes tube amps of all kinds rather tricky but in particular almost rules SETs out of the picture. Phase shift is used by the ear to assist with understanding the sound stage space so if it gets messed up you don't get the nice 3D presentation of the recording (if its done right).

If you are able to run enough feedback, you can correct phase shift and so the bandwidth becomes less important. But to do that you need a lot of feedback and that puts you at risk of distortion rising with frequency due to Gain Bandwidth Product limitations. You can compromise by having some feedback and also wide bandwidth, but you have to be careful to make sure the feedback is really doing what its supposed to do. Nearly all tube amps and most solid state amps fail at this because the feedback is applied to the wrong part of the amp, typically the cathode of the input tube, where it gets distorted prior to doing its job. That results in higher ordered harmonic generation, which does not happen if the feedback is properly applied. See Norman Crowhusrt and Peter Baxandall's writings on this matter, its not like this problem hasn't been known about for a really long time. Curiously neither provided a solution, although that problem has been easily solved for decades on now. Tradition has been the only thing preventing that solution being more commonplace.

- Imagine the best topology is zero feedback SET DHT
- Imagine the best is using just one tube not paralleling two or more tubes
- Imagine the best tubes are low voltages (low power)

what you find is a low power amplifier than can not drive most loudspeakers so you need to go for horns .

if you want good sound from horns you should have 5way or 6way horns not two way or three way.

finally you will find the Romy the Cat audio rules are the best rules in Audio.

Romy The Cat's Audio Rules

GoodSoundClub - Romy the Cat's personal High-End Audio site

goodsoundclub.com

Imagination is very different from reality. Its nice to imagine this stuff, but in practice its way easy to do better.

The 5 or 6 way horn thing is ridiculous IMO.IME. Some of the very best speakers I've heard of any kind are 3 way. I've heard some 6-way horns that cost close to a million dollars that didn't do as well.

 

[andromedaaudio]

Guys , just a observational opinion here .
Doesnt anyone in these continuous discussions has a feeling he / she is going in circles .


These discussions have been going on since 12 years since i joined the forum without any substantial outcome

 

[christensenleif@msn.com]

if you want good sound from horns you should have 5way or 6way horns not two way or three way.

this is utter bullshit and just reveals you don´t have a clue and are hung up in oldstyle rules of law for traditional horns

 

[bonzo75]

this is utter bullshit and just reveals you don´t have a clue and are hung up in oldstyle rules of law for traditional horns

point is he has no listening experience. he bases everything on the fact that he is an astute forum reader

 

[SVS]

If you are able to run enough feedback, you can correct phase shift and so the bandwidth becomes less important. But to do that you need a lot of feedback and that puts you at risk of distortion rising with frequency due to Gain Bandwidth Product limitations. You can compromise by having some feedback and also wide bandwidth, but you have to be careful to make sure the feedback is really doing what its supposed to do. Nearly all tube amps and most solid state amps fail at this because the feedback is applied to the wrong part of the amp, typically the cathode of the input tube, where it gets distorted prior to doing its job. That results in higher ordered harmonic generation, which does not happen if the feedback is properly applied. See Norman Crowhusrt and Peter Baxandall's writings on this matter, its not like this problem hasn't been known about for a really long time. Curiously neither provided a solution, although that problem has been easily solved for decades on now. Tradition has been the only thing preventing that solution being more

The speed of propagation of the signal in the amplifier is not instantaneous, so the feedback from the output to the input, for example, to the cathode of the input tube, is the correction of the signal that arrived at this moment with a signal from the past that has already reached the input to the output and returned back to compensate for distortions. This reduces the lower harmonics, but creates a large number of higher harmonics, they are very small, but are perceived by our ears as an unnatural sound.
At the same time, local feedback, for example, a resistor in the cathode or emitter not shunted by a capacitor, has a very good effect on the sound.

 

[morricab]

- Imagine the best topology is zero feedback SET DHT
- Imagine the best is using just one tube not paralleling two or more tubes
- Imagine the best tubes are low voltages (low power)

what you find is a low power amplifier than can not drive most loudspeakers so you need to go for horns .

if you want good sound from horns you should have 5way or 6way horns not two way or three way.

finally you will find the Romy the Cat audio rules are the best rules in Audio.

Romy The Cat's Audio Rules

GoodSoundClub - Romy the Cat's personal High-End Audio site

goodsoundclub.com

That is why the single stage triode/FET and NuFet from Aries Cerat and NAT, respectively, are interesting. They are single stage designs with voltage current gain in one "hybrid" device. Haven't heard the NAT version but the AC version is stunning. Power is a more usable 20-60 watts for the AC amps and 100 watts for the NAT.

I also know of one amplifier using the Emission labs 20B that is a single tube amplifier. The 20B has enough gain that a driver stage is unnecessary...makes 5 watts...so needs 95dB + speakers typically (special circumstances might get away with lower sensitivity).

I have to say that I don't agree with Romy's requirements for a 5 or 6 way horn. It can be done well much more simply and the difficulty of getting a 5 or 6 way speaker right is absolutely horrendous and rarely ever works well in practice.

I will say that the two times I have heard the Ypsilon SET100 Ultimate single ended hybrid I was quite impressed.

 

[Atmasphere]

Guys , just a observational opinion here .
Doesnt anyone in these continuous discussions has a feeling he / she is going in circles .


These discussions have been going on since 12 years since i joined the forum without any substantial outcome

Personally, no. I've not seen all that's discussed here elsewhere. There has been some sniping at this thread a bit, but that's been pretty obvious and is easy to skip past.

 

[morricab]

Guys , just a observational opinion here .
Doesnt anyone in these continuous discussions has a feeling he / she is going in circles .


These discussions have been going on since 12 years since i joined the forum without any substantial outcome

One thing is consistent...you trying to stop the discussions...

 

[Rensselaer]

One absolute that is quite measurable is distortion must not rise with frequency. If it does, it will be perceived as harshness and brightness.

Any zero feedback circuit can do this and that is part of why feedback has a bad rap in high end audio. But it is possible (although not until recently in audio history) to build an amp with feedback that also has this quality.

In addition, the distortion product must be benign. To that end it must be lower ordered harmonics (the 2nd and 3rd, which are treated by the ear the same way in that they add 'warmth' or 'bloom', using audiophile terms) with enough amplitude to mask any succeeding harmonics, the latter of which must fall off on an exponential curve (if the circuit is zero feedback). The exponent of the exponential curve does not seem to be important; once you exceed that of a quadratic exponent it becomes moot.

You can satisfy the above requirements with an SET, a fully differential PP tube amp, a class A or AB solid state amp (although this is very diifficult technology to work with if you are to satisfy the first requirement above) and class D.

So there is no topology that is an absolute to get to this 'right' sound. It is easiest to do with SETs and hardest to do with A or AB solid state amps. Each tech has its drawbacks. SETs for example make the most distortion of any amplifier concept and suffer bandwidth problems if making much more than about 8 Watts. So you need high efficiency speakers to work with them and even then you can only use about 20% of full power if you really want to hear what the SET is about.

If you are running zero feedback, bandwidth is important to prevent phase shift. Ideally this means 2Hz to 100KHz (since phase shift above 10KHz likely isn't audible). This bandwidth issue makes tube amps of all kinds rather tricky but in particular almost rules SETs out of the picture. Phase shift is used by the ear to assist with understanding the sound stage space so if it gets messed up you don't get the nice 3D presentation of the recording (if its done right).

If you are able to run enough feedback, you can correct phase shift and so the bandwidth becomes less important. But to do that you need a lot of feedback and that puts you at risk of distortion rising with frequency due to Gain Bandwidth Product limitations. You can compromise by having some feedback and also wide bandwidth, but you have to be careful to make sure the feedback is really doing what its supposed to do. Nearly all tube amps and most solid state amps fail at this because the feedback is applied to the wrong part of the amp, typically the cathode of the input tube, where it gets distorted prior to doing its job. That results in higher ordered harmonic generation, which does not happen if the feedback is properly applied. See Norman Crowhusrt and Peter Baxandall's writings on this matter, its not like this problem hasn't been known about for a really long time. Curiously neither provided a solution, although that problem has been easily solved for decades on now. Tradition has been the only thing preventing that solution being more commonplace.

Imagination is very different from reality. Its nice to imagine this stuff, but in practice its way easy to do better.

The 5 or 6 way horn thing is ridiculous IMO.IME. Some of the very best speakers I've heard of any kind are 3 way. I've heard some 6-way horns that cost close to a million dollars that didn't do as well.

Just when most were in agreement that a directly-heated Class A Single Ended Triode amplifier (or triode/mosfet hybrid ) without feedback, with star grounding and appropriately sized and constructed transformers, playing into efficient two-way or three-way speakers, will result in the most realistic sound reproduction of all the possible amp topologies, you confuse the thread with your engineer speak.

You repeat what was already covered, admittedly to a limited degree, that if harmonic distortion in the lower orders (2nd and 3rd, which matches that of real acoustical musical instruments so is not objectionable ) is predominant (louder), it should mask any higher order harmonic distortion. This "characteristic" of SET's might be why they sound more "real to life" despite high distortion levels.

You then argue against zero feedback, stating that it is needed to gain bandwidth in order to prevent phase shift, which, if present, apparently worsens pin-point imaging (pin-point imaging helps with presence, but not realness).

Not that I believe you intentionally wanted to turn back the direction that the thread had morphed (deciding which amplifier topology and design principles is most likely to result in the most real to life presentation of recorded music) and go back to the start deliberately (I don't think so anyway), but simply that you were trying to answer the OP's original question.

Nothing wrong with that, but could you assign values to each design consideration you mention in future? For instance, how much investment in design time, complicated circuitry and money is required to make a Push Pull amplifier sound as real as a Wavac, Kondo, Air Tight, Soundgate etc. SET? Is it worth it? Can it even be done?

 

[acousticsguru]

That is why the single stage triode/FET and NuFet from Aries Cerat and NAT, respectively, are interesting. They are single stage designs with voltage current gain in one "hybrid" device. Haven't heard the NAT version but the AC version is stunning. Power is a more usable 20-60 watts for the AC amps and 100 watts for the NAT.

I also know of one amplifier using the Emission labs 20B that is a single tube amplifier. The 20B has enough gain that a driver stage is unnecessary...makes 5 watts...so needs 95dB + speakers typically (special circumstances might get away with lower sensitivity).

I have to say that I don't agree with Romy's requirements for a 5 or 6 way horn. It can be done well much more simply and the difficulty of getting a 5 or 6 way speaker right is absolutely horrendous and rarely ever works well in practice.

I will say that the two times I have heard the Ypsilon SET100 Ultimate single ended hybrid I was quite impressed.

Off-topic, but apart from the difficulty of designing crossovers and if possible time-align multiple drivers there's the minimum listening distance that follows directly from the surface / center-to-center distance of the most distant drivers (and even that only holds true if the drivers play in time and in phase).

Greetings from Switzerland, David.