What are the advantages/disadvantages of different amp topologies?

[bonzo75]

not bad showing my 1 million dollar system if someone think I just google and I have no experience:

View attachment 124576

are these multi-way horns? please explain how a million dollar system of highly inefficient loud speakers brought you to the conclusion that only the multi way approach followed by vice Olympian and Romy is correct.

 

[Rensselaer]

What size and design do you consider the output transformer to be good?
Do you pay attention to the technical characteristics of these devices?

I don’t know … I am seeking truth.

 

[SVS]

I don’t know … I am seeking truth.

I will tell you my truth about output transformers, maybe it will be useful to you.
I make output transformers in my amplifiers with a lower operating frequency of at least 7-8 Hz at the level of - 3 dB at half the maximum output power. Such a transformer provides a much more natural sound than, for example, a transformer with a frequency of 20 Hz. Of course, both transformers have no problems at high frequencies.
The weight of such a device for a SET amp with a power of 20-25 W is 8-8.5 kg.

 

[Atmasphere]

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?

To your last question the simple answer is 'yes'. That's already happened, and a long time ago at that. To your first comment, there's no agreement on that at all!

SETs were created by engineers. They can be engineered poorly by people that don't know what they are doing or they can be done properly. You really can't talk about this topic without some engineering speak. Its essential. Otherwise, don't use electricity, cars, the Internet and so on... SETs are engineered products. I really shouldn't have to say that, but there it is.

The reason to not use feedback is due to inadequacy in the design- it lacks the Gain Bandwidth Product to support the feedback and the designer is ignorant of how to apply it without causing trouble anyway so its not done. For example, our OTLs (which easily take on the amps you mentioned above as long as both are compared on a loudspeaker with which both amps are happy) use little or no feedback because we recognized that tube amps simply can't be made with enough Gain Bandwidth Product to support the feedback you need to really make them neutral.

Technology has been marching on for a long time since SETs were created; in order to understand how SETs have been eclipsed, you have to drop the seemingly religious made-up stories around SETs. As an engineer and also because audio is one of my hobbies, I've been designing and building amps of all kinds now for over 45 years, which includes SETs. I'm working on one now that will use either UX250s or graphite plate 300bs, hopefully interchangeably. My last SET project was a 45 based amp; 2A3 prior to that. I've also built PP amps using those same power tubes and one thing that comes up about that is when SET guys try to compare amps, they seem to never place them on a level playing field. Invariably it seems the SET is compared to some kind of PP amp that on paper shares a lot in common with a Dynaco ST70.

If you really want to know how bad SETs really are, you have to start by doing 2 things: build or find a PP amp using the same power tubes as the SET and build(t) it to the same build quality. Minimize variables such as B+ Voltage, feedback, class of operation, grounding and so on. What you find out is that the PP amp will not take a back seat to the SET in any way. The second thing to do is build a PP amp of the same power as the SET under comparison, and simply give that PP amp a reasonable chance with good quality design and components. You get the same result. The issue here with amps that have output transformers, in case its not obvious, is the smaller tube amps usually sound better because the output transformer has wider bandwidth and the circuit is usually simpler. An ST70 makes 35 Watts which is quite a lot more than most SETs.

That is why I mentioned that the topology is important because of the kind of distortion that results. A PP amp like an ST70 combines both quadratic non-linearities (there's that engineering speak again) along with cubic non-linearities, resulting in a prominant 5th harmonic. Norman Crowhurst wrote about this about 65 years ago so this shouldn't be news. What this means is why you compare the ST70 to an SET you compare apples and oranges- the ST70 in addition to the 5th harmonic problem has feedback, is class AB and uses pentode power tubes! How is that a valid comparison if what you want to know is whether SET is good or not??

The technology that has marched along since the 1920s has now produced class D which was first invented in the 1950s. It did poorly at first with tubes but as semiconductors improved for switching purposes class D got better. In the last 20 years switching devices have appeared that are so good that class D has been able to make enourmous strides such that it easily challenges the prior art. Everyone here uses it, because phones and computers use it now rather than the older class AB designs. The only real question is does it challenge high end audio amps of any kind (including SETs) and the answer is 'yes, it does'. Look earlier on this thread and you'll see how this is so. I mentioned earlier, not all class D amps are identical; they vary in sound quite a lot so its unwise (if you really want to know the truth of the matter) to think that because you've heard a few over the years that somehow you have heard them all and don't have to think about this anymore.

The class D amp I have in my system has beaten all comers: our class A triode OTLs, all the SETs I been able to find (the best was the Viva so far) all the PP tube amps and I've not heard a class A or AB solid state amp as good. I run horns and its nice to have the low noise as well. So this idea that there is a consensus about how SETs are the best really doen't hold water.

 

[DasguteOhr]

In the search for realness in sound amplification, what make/type filtering capacitor do you use at the (star) grounding point? What make of transformer do you prefer (Peerless, Tango, Magnaquest) or do you wind your own? Is it better to use C-core (potted or not) or toroidal or depends? Does 99.999 pure silver hookup wire sound more real than 99.999 pure copper or even tinned copper? Oil and silk insulated , PVC or teflon?

I am hopeful of learning from you all, the best out there, which design topology (and specific component details) will give me the greatest musical enjoyment if what I want is to replace my Ayon with an amplifier that, in combination with my Altec speakers, will get me closest to live that I can get.

I'm a big fan of mundorf evos aluminium foil in oil at this postion not too expensive good sound.
For wiring the amps I always use cables from this shop. Either teflon or kapton wire shielded (signal) and unshielded for the other cabling.

Sicherheitsmaßnahme

www.ebay.de

OPT

Welcome to MONOLITH MAGNETICS | MONOLITH MAGNETICS

www.monolithmagnetics.com

Really good

Audioröhrentechnik Ausgangstrafo Übertrager Bandfilter Spulen Röhren Audio Fassungen 516201 516228 516236 516260 516686 NF-Übertrager Drossel Trafo Hifi Highend

Bauteile für die Röhrentechnik,Ausgangsübertrager, Ausgangstrafo,HiFi,HighEnd,Drossel,Netztrafo,Transformator,Röhren,Röhrenfassungen,Übertrager,Miniatur-Übertrager,516201,516228,516236,516260,516686,NF-Übertrager

www.roehrentechnik.de

Reinhöfer has been building Transformers and OPTS for decades. also as a custom-made product whatt you want.

Ausgangsübertrager / Transformatoren / Röhren-Verstaerker-Upgrade / Steuerung / Automation

Audio Interstage & Output Transformers, Exklusive Tubes Audio Amplifiers, Steuerung, Automation, Labortechnik

www.tubesoundelectronics.de

Phantastic opt for set amp with all important measurements, which is not always a given.

The important factor of a transformer is the transverse inductance, which determines the lower limit frequency fu (Hz). To lower this, transformers are nested. Be careful not to overdo this as this will increase leakage inductance and lower the upper transmissible frequency. The output winding should have the lowest possible impedance. This increases the damping factor and thus the control over the speaker.

 

[SVS]

The class D amp I have in my system has beaten all comers: our class A triode OTLs, all the SETs I been able to find (the best was the Viva so far) all the PP tube amps and I've not heard a class A or AB solid state amp as good. I run horns and its nice to have the low noise as well. So this idea that there is a consensus about how SETs are the best really doen't hold water.

If you added that this is your opinion and not the truth, then it would be correct. Very reminiscent of the legend of the Procrustean bed.

 

[SVS]

I'm a big fan of mundorf evos aluminium foil in oil at this postion not too expensive good sound.
For wiring the amps I always use cables from this shop. Either teflon or kapton wire shielded (signal) and unshielded for the other cabling.

I believe that the shielded wire kills the sound. An exclusive place for its use is the phono cable. I don't use it in amps, not even phono preamps. At the same time, there is not the slightest hum.

 

[DasguteOhr]

I believe that the shielded wire kills the sound. An exclusive place for its use is the phono cable. I don't use it in amps, not even phono preamps. At the same time, there is not the slightest hum.

The shielding increases the capacity somewhat. I always use it from the input socket to the first tube. sometimes you just have longer distances. So far I haven't heard anything negative.
You can also simply twist wires, but it has the same effect

 

[Atmasphere]

The output winding should have the lowest possible impedance. This increases the damping factor and thus the control over the speaker.

I think you will find that without feedback, the amplifier behaves more as a power source than a voltage source, which means the change in damping is nearly meaningless. What is far more important is making sure that the speaker design is compatible with an amplifier that has a higher output impedance to begin with. Such speakers usually don't make big damping demands of the amp driving them.

For more information see Paradigms in Amplifier Design

The shielding increases the capacity somewhat. I always use it from the input socket to the first tube. sometimes you just have longer distances. So far I haven't heard anything negative.
You can also simply twist wires, but it has the same effect

If shielding is handled correctly it works fine. I've heard of people saying shielding messes things up, but so far every time I've investigated such claims, I've found ignorance of how to use it. If it didn't work, recordings would not exist...

 

[acousticsguru]

No the LV speaker, which I love the sound of, does not use DSP…it is also not time aligned to the best of my knowledge. The horns in that design are small enough to allow reasonable placement so the disparity in propagation isn’t so bad. I think the midbass in that design is also backloaded so the mids are easier to align with the upper mid horn. Add to that Kevin Scott spent about 30 years working on that design, which started as the Air Partner

AFAIK the LV VO is time-aligned - at least Kevin Scott said so.

Greetings from Switzerland, David.

 

[SVS]

The shielding increases the capacity somewhat. I always use it from the input socket to the first tube. sometimes you just have longer distances. So far I haven't heard anything negative.
You can also simply twist wires, but it has the same effect

Try a cable with twisted wires on the MM input of the preamp, then you will see the same effect or a different one.

 

[Argonaut]

If you added that this is your opinion and not the truth, then it would be correct. Very reminiscent of the legend of the Procrustean bed.

We seem to have ‘Dropped Out For A Commercial Break ‘ normal service on this topic should resume shortly.

 

[DasguteOhr]

Try a cable with twisted wires on the MM input of the preamp, then you will see the same effect or a different one.

Some mm cartridge require a lot of capacitance 150-400pf in order to have a linear frequency response. 30 cm more cables have no effect. This can be annoying when used with sut. We can argue about this for a long time now, let's agree that everyone has their own way to reach their goal.

 

[DasguteOhr]

I think you will find that without feedback, the amplifier behaves more as a power source than a voltage source, which means the change in damping is nearly meaningless. What is far more important is making sure that the speaker design is compatible with an amplifier that has a higher output impedance to begin with. Such speakers usually don't make big damping demands of the amp driving them.

For more information see Paradigms in Amplifier Design


If shielding is handled correctly it works fine. I've heard of people saying shielding messes things up, but so far every time I've investigated such claims, I've found ignorance of how to use it. If it didn't work, recordings would not exist...

I see it a little differently, I try to get as much control over the speaker as possible. This means a low-resistance output winding and a speaker cable with a larger wire cross-section. I don't have any measurements, just what I hear.

 

[morricab]

I see it a little differently, I try to get as much control over the speaker as possible. This means a low-resistance output winding and a speaker cable with a larger wire cross-section. I don't have any measurements, just what I hear.

While that is likely necessary for the typical vented box, a horn loaded bass, open baffle or we damped sealed box doesn’t require a high damping factor and in many cases this would be detrimental to good bass.

 

[DasguteOhr]

While that is likely necessary for the typical vented box, a horn loaded bass, open baffle or we damped sealed box doesn’t require a high damping factor and in many cases this would be detrimental to good bass.

do a pulse measurement with 50hz tone with thick cables and thin cables. This indicates whether the damping factor is audible or not. I don't do it, I hear a difference. OK, I don't have any xover in front of my bass+ midrange, maybe a xover swallow up this advantage.
Inductance and resistance play a big role in speaker cables my opion.

 

[SVS]

Some mm cartridge require a lot of capacitance 150-400pf in order to have a linear frequency response. 30 cm more cables have no effect. This can be annoying when used with sut. We can argue about this for a long time now, let's agree that everyone has their own way to reach their goal.

It's not about the capacity. If you just put your hand to the RIAA input tubes of the open design preamp, you will hear a hum. The reaction of unshielded input cables in these devices is similar.

 

[DasguteOhr]

It's not about the capacity. If you just put your hand to the RIAA input tubes of the open design preamp, you will hear a hum. The reaction of unshielded input cables in these devices is similar.

But that has nothing to do with the cable but with the high-impedance input of mm. They are more sensitive to interference from outside. low impendance input react little to interference like mc input except grounding phono shematics are bad.

 

[Rensselaer]

To your last question the simple answer is 'yes'. That's already happened, and a long time ago at that. To your first comment, there's no agreement on that at all!

SETs were created by engineers. They can be engineered poorly by people that don't know what they are doing or they can be done properly. You really can't talk about this topic without some engineering speak. Its essential. Otherwise, don't use electricity, cars, the Internet and so on... SETs are engineered products. I really shouldn't have to say that, but there it is.

The reason to not use feedback is due to inadequacy in the design- it lacks the Gain Bandwidth Product to support the feedback and the designer is ignorant of how to apply it without causing trouble anyway so its not done. For example, our OTLs (which easily take on the amps you mentioned above as long as both are compared on a loudspeaker with which both amps are happy) use little or no feedback because we recognized that tube amps simply can't be made with enough Gain Bandwidth Product to support the feedback you need to really make them neutral.

Technology has been marching on for a long time since SETs were created; in order to understand how SETs have been eclipsed, you have to drop the seemingly religious made-up stories around SETs. As an engineer and also because audio is one of my hobbies, I've been designing and building amps of all kinds now for over 45 years, which includes SETs. I'm working on one now that will use either UX250s or graphite plate 300bs, hopefully interchangeably. My last SET project was a 45 based amp; 2A3 prior to that. I've also built PP amps using those same power tubes and one thing that comes up about that is when SET guys try to compare amps, they seem to never place them on a level playing field. Invariably it seems the SET is compared to some kind of PP amp that on paper shares a lot in common with a Dynaco ST70.

If you really want to know how bad SETs really are, you have to start by doing 2 things: build or find a PP amp using the same power tubes as the SET and build(t) it to the same build quality. Minimize variables such as B+ Voltage, feedback, class of operation, grounding and so on. What you find out is that the PP amp will not take a back seat to the SET in any way. The second thing to do is build a PP amp of the same power as the SET under comparison, and simply give that PP amp a reasonable chance with good quality design and components. You get the same result. The issue here with amps that have output transformers, in case its not obvious, is the smaller tube amps usually sound better because the output transformer has wider bandwidth and the circuit is usually simpler. An ST70 makes 35 Watts which is quite a lot more than most SETs.

That is why I mentioned that the topology is important because of the kind of distortion that results. A PP amp like an ST70 combines both quadratic non-linearities (there's that engineering speak again) along with cubic non-linearities, resulting in a prominant 5th harmonic. Norman Crowhurst wrote about this about 65 years ago so this shouldn't be news. What this means is why you compare the ST70 to an SET you compare apples and oranges- the ST70 in addition to the 5th harmonic problem has feedback, is class AB and uses pentode power tubes! How is that a valid comparison if what you want to know is whether SET is good or not??

The technology that has marched along since the 1920s has now produced class D which was first invented in the 1950s. It did poorly at first with tubes but as semiconductors improved for switching purposes class D got better. In the last 20 years switching devices have appeared that are so good that class D has been able to make enourmous strides such that it easily challenges the prior art. Everyone here uses it, because phones and computers use it now rather than the older class AB designs. The only real question is does it challenge high end audio amps of any kind (including SETs) and the answer is 'yes, it does'. Look earlier on this thread and you'll see how this is so. I mentioned earlier, not all class D amps are identical; they vary in sound quite a lot so its unwise (if you really want to know the truth of the matter) to think that because you've heard a few over the years that somehow you have heard them all and don't have to think about this anymore.

The class D amp I have in my system has beaten all comers: our class A triode OTLs, all the SETs I been able to find (the best was the Viva so far) all the PP tube amps and I've not heard a class A or AB solid state amp as good. I run horns and its nice to have the low noise as well. So this idea that there is a consensus about how SETs are the best really doen't hold water.

Unless someone has pertinent peer-reviewed double-blinded controlled listening tests in front of them, their beliefs will be based upon experience and reading (some reading being of anecdotal experience). In this manner I came to the following beliefs (which may be incorrect).

In the early days of “talkies”, major companies designed and built sound systems for movie theatres consisting of low-powered SET amplifiers ( Loftin-White, etc.) and very large horn loaded speakers (Altec VOT).

When home hi-fi started, the biggest complaint was the size of the speakers. Smaller speakers followed but, being smaller, they required more power to push the same amount of air. The PP amp followed, doubling power but at the cost of sound quality (I don’t know the engineering term) at the crossover point.

During this drive for ever more power to drive ever less sensitive speakers, solid state arrived and engineers found they could attain equivalent power to valves but at a fraction of the cost, and still charge 90% of what equivalent output valve amplifiers were selling for. What’s more, they learned that the lower the impedance of the speaker, the more power the solid state amplifier puts out, exact opposite to what happens to OTL valve amps (power output increases with increasing speaker impedance). In fact, if someone with modern low-impedance speakers buys OTL amps they often need to purchase a transformer (Zero Auto Former) to place between the OTL amp and speakers to increase impedance. I believe SETs put out the same power regardless of the speaker impedance.

There are other issues with OTL amplifiers too (some the fault of one maker (Flutterman) whose amps periodically blew up). If you don’t want to replace every valve (and there’s a lot of them) you’ll need a valve tester (replacing valves is expensive). Heat, and adjustments, and bass (Ralph Karsten at Atma-Sphere allegedly recommends Audio-Kinesis sub-woofers to customers?

As to class D amplifiers, I don’t have any experience here to judge sound quality, but will avoid all amplifiers that put out more than 30 watts (and am taking a chance with 30 watts). My speakers are Altec A7’s. The high-frequency compression drivers (Great Plains Audio 802-8G Alnico) have edge-wound aluminium-ribbon voice-coil-diaphrams that have a sensitivity of 106db and a power rating of just 8 watts. Irregardless of the fact that I would not use 90% of the power available from your 100 watt per channel class D amps, even turned way down a single trumpet blast could end up tearing my compression driver diaphragms.

I am looking to replace my Ayon Spitfire SET with some design that will improve the realness and presence from my Altec speakers. I was communicating with a couple of contributors who were addressing my specific requirements, and they were predominantly in favour of SETs with certain design characteristics, the specifics I was hoping would be revealed. By starting over, saying each design (A, B, AB, PP, D, OTL) is capable of beating DHSET’s because of (list of SET failures), you basically (IMHO) cut my inquiry off and dismissed all opinions that differed with yours.

 

[acousticsguru]

While that is likely necessary for the typical vented box, a horn loaded bass, open baffle or we damped sealed box doesn’t require a high damping factor and in many cases this would be detrimental to good bass.

do a pulse measurement with 50hz tone with thick cables and thin cables. This indicates whether the damping factor is audible or not. I don't do it, I hear a difference. OK, I don't have any xover in front of my bass+ midrange, maybe a xover swallow up this advantage.
Inductance and resistance play a big role in speaker cables my opion.

A passive crossover (including wires and cables) may add .4 to 1.0 Ohms series resistance, depending on the order and type and quality of inductor(s). Cabinets without crossovers are most safely based on the assumption that the series resistance is not going to be zero, but about .2 Ohms, in order to avoid the “over-damping” @morricab mentions above (pet peeve: don’t make assumptions and/or calculate, but measure - there’s no such thing as zero series resistance, the only part that’s not going to make a measurable difference in pass-band series resistance is a reasonable or better quality capacitor).

(Edit: needless to say, I’ve built passive speaker crossovers with slightly lower series resistance, e.g. .3 Ohms including wires, connectors and cables, just saying, without measuring, the “subjective” over-damping one will hear is due to miscalculation, not the amp).

Greetings from Switzerland, David.