Interesting article on SET amps

[Feanor]

...
The good news is that both the designs and technology have changed in the past forty years. My '81 Stasis uses no global feedback between input and output stages. Devices have improved greatly allowing for lower amounts - or zero feedback found in models by a number of manufacturers like Ayre, Pass Labs, etc.

Nelson Pass has always been a big advocate for zero feedback -- or as close to it a possible, and you can't argue with his results. Boy, oh, boy, I'd love to try his latest, over-the-top amps, the Xs series monoblocks --each "monoblock" is actually two blocks: see pic.

I think this Xs 300 is $85k, (don't recall whether that's each or per pair) ...



On the other hand I recently read Rod Elliott's treatise on the subject of negative feedback, (here). I didn't understand everything he said, but I do believe he made the following assertions:

• = Negative feedback is better that the distortion that would otherwise be present; (he was referring principally to s/s amps).
• = Global feedback is more effective than, and preferable to, local feedback, (not to be confused with "degenerations" which I don't entirely understand). The result is lower distortion overall.
• = Contrary to audiophile myth, feedback is an instantaneous phenomenon, not as series of iterations. The propagation time through the feedback loop is a few microseconds at worst and too short to allow iterations in the audio bandwidth.
• = Contrary to audiophile myth, feedback doesn't cause higher order distortion, it them to be more easily measured.
• = Square waves are not relevant to measuring audio distortion in general since the relevant high-order harmonics are far outside the audible range.

[E-Stat]

Nelson Pass has always been a big advocate for zero feedback -- or as close to it a possible, and you can't argue with his results.

Agreed. As opposed to guys with disproven theories like...

On the other hand I recently read Rod Elliott's treatise on the subject of negative feedback...

Contrary to audiophile myth, feedback is an instantaneous phenomenon, not as series of iterations. The propagation time through the feedback loop is a few microseconds at worst and too short to allow iterations in the audio bandwidth.

Not in the audio bandwidth? Have you read his cable comparison methodology? Connect two sets of interconnects via Y adapters between a source and preamp and switch between the two inputs using the tape monitor. You have now effectively combined the LCR characteristics of both cables together. and end up comparing - both to both. (edit: since I have a capacitance meter, I just hooked up a Y adapter to an IC and measured the value at the Y. Then added another cable. The value increased by the amount of the second cable.) How can you take someone like that seriously?

One one of my trips to Sea Cliff a few years back, I noticed a pair of the big Halcro monoblocks sitting aside. How do they sound? HP never reviewed them because he felt they were unnaturally sterile and clinical sounding. They didn't stay in the system for long. If you recall, they were high feedback designs. The latest company touting high amounts of feedback is Soulution. I haven't heard them, but I suspect they would strike me the same way. The error correction ends up throwing some of the baby out with the wash.

[Feanor]

Agreed. As opposed to guys with disproven theories like ... Rod Elliott: Distortion & Feedback
....

In fairness to Elliott, his article isn't just a series of assertions. It's fairly long 28 pages wherein he argues from theory, simulation modelling, and constructed test examples. I would be interested to hear his arguments addressed with some degree of rigour.

...
Not in the audio bandwidth? Have you read his cable comparison methodology? Connect two sets of interconnects via Y adapters between a source and preamp and switch between the two inputs using the tape monitor. You have now effectively combined the LCR characteristics of both cables together. and end up comparing - both to both. (edit: since I have a capacitance meter, I just hooked up a Y adapter to an IC and measured the value at the Y. Then added another cable. The value increased by the amount of the second cable.) How can you take someone like that seriously...

I'm not sure I follow you; are you talking about the capacitance in the cables? If so, there would be a difference of course.

This is what Elliott actually said; (did I misunderstand?) ...

6.0 - Amplification Circuit Delay:
It is obvious that nothing in life is instantaneous. When a signal is applied to the input of an amplifier, there is a delay before the amplifier can react to the change, and this is determined by the speed of the devices used. Logic circuits typically have nanosecond delays from input to output, and this is also the order of delay one can expect before an amplifier as shown in Figure 9 will react to a change of input. According to the simulator, it takes about 5ns for the amp to respond to the fact that the input has changed - this is still using the very fast squarewave as an input. The output then swings in the appropriate direction at its maximum slew rate until the voltage at the inverting input again equals that at the non-inverting input. Once the voltages are equal, it takes about 220ns for the output to stabilise, settling so that the two input voltages are exactly the same. These times are very short - it takes the output 1.3us to change from +11V to -11V, so the 'reaction' time is close to negligible. It would be pointless to try to reproduce all the waveforms, so I suggest that you download the simulations. The files are in SIMetrix format, and are ready to run.

Note that any delay has nothing to do with electrons 'slowing down' - there is typically nothing in an amplifier circuit that does any such thing. The delays are simply the result of the devices taking a finite time to conduct (or switch off) after a signal has been applied or removed, an issue that affects all amplifying devices. While painstaking engineering is needed to minimise these delays (especially for very high speed switching), it is generally not needed for audio - not because audio is slow (although it is very slow compared to the logic in a fast micro-processor), but because analogue amplifiers are not switching, so are normally inherently fast. We actually have to slow them down deliberately with a capacitor (the Miller or dominant pole cap) to prevent oscillation.

However, the above test was all done with a signal that is much faster than the amplifier can handle (and much faster than any signal it is expected to handle for music reproduction), and it is more useful to examine what happens when the input slew rate is limited to something sensible. By adding a filter to the squarewave signal, the rise time can be limited to a somewhat more realistic value. A 32kHz, 24dB/octave filter was used, and this limits the output signal from the amplifier to 1.85V/us - well within its range, but still a great deal faster than any real music signal will create. Everything is now within the linear capability of the amplifier. The output is delayed by 46ns compared to the input, but this is inconsequential. Of more importance is how the amplifier reacts to the combined sine and square wave signal. It is not immediately apparent from the output, but in fact the sinewave is almost completely unaffected - the portion that would otherwise be cut off due to slew rate limiting now simply 'rides' the slope of the squarewave - if compared (after correcting for the level difference), the input and output are virtually identical - there is no evidence whatsoever of anything that could be classified as transient distortion - even with a 100kHz signal.

[E-Stat]

In fairness to Elliott, his article isn't just a series of assertions. It's fairly long 28 pages wherein he argues from theory, simulation modelling, and constructed test examples.

Mostly theory and computer simulations evaluating numbers. Not listening to real products reproducing real music. Which is where I find that real designers have a decided edge. Let's look at another one of Elliott's "assertions":

"Contrary to audiophile myth, feedback doesn't cause higher order distortion, it them to be more easily measured."

Pass would just smile since he has proven otherwise. While I don't believe for a minute he's the only guy out there doing good stuff, I think he is the best communicator of the concepts. Read his document on distortion and in particular the story of a real world amplifier he built using variable amounts of NFB which empirically counters Elliott's speculations. As did my two minute test measuring the results of connecting two cables via a Y adapter.

Pass on the complex world of distortion

I'm not sure I follow you; are you talking about the capacitance in the cables? If so, there would be a difference of course.

He is asserting that his test can determine whether or not you can hear the difference between two different cables possessing different characteristics.

The test fails to work miserably, however, because the methodology ends up blending the metrics of the two cables such that you are not comparing one cable by itself to another by itself. You hear the combined results in input A and you hear the same combined results in input B. In reality, you're not comparing anything. You're hearing the same result in either input. It is not at all like connecting one set of cables - then disconnecting them and connecting the other set of cables. He doesn't even bother to validate his theory with the simple test I performed. No wonder he thinks all "well designed cables" sound the same. He's not comparing anything!

ABX testing using a box fails for the same reasons. In order to avoid horrible switching transients, you must necessary share the grounds. So you are blending cable characteristics or even feedback loops between separate amplifiers. It was Frank Van Alstine that first made me aware of the fallacy behind the use of ABX boxes.

Do you remember Skeptic/Soundmind? He had a similar "shunt test" whereby he contends you can determine whether or not a cable is absolutely perfect by switching it in or out of a tape monitor loop. He has no comprehension of the logical fallacies involved with his set of assumptions.

Too many guys with good intentions arrive at false conclusions due to the set of unproven assumptions they make with their test methodology. Which is why seasoned ears prefer simpler circuits like SETs to complex switching designs that measure far better.

[Feanor]

Mostly theory and computer simulations evaluating numbers. Not listening to real products reproducing real music. Which is where I find that real designers have a decided edge. ...

Elliott claims to have done both. He is hands-on experimenter like Pass. Personally I can't judge the validity of the work of either since I have no electrical engineering background.

...
Let's look at another one of Elliott's "assertions":

"Contrary to audiophile myth, feedback doesn't cause higher order distortion, it them to be more easily measured." ...

That quote was my paraphrase, though it is my understanding of what he said.

Pass would just smile since he has proven otherwise. While I don't believe for a minute he's the only guy out there doing good stuff, I think he is the best communicator of the concepts. Read his document on distortion and in particular the story of a real world amplifier he built using variable amounts of NFB which empirically counters Elliott's speculations. As did my two minute test measuring the results of connecting two cables via a Y adapter.

Pass on the complex world of distortion
...

I believe I have read the article before, but will do so again -- not tonight, though, because my eyes are shot.

[E-Stat]

Elliott claims to have done both.

The other guys, by contrast, have graphically demonstrated their point. What is Elliott's amplifier reference? Eval system? Given his complete misunderstanding of cable interaction, I suspect he would fall into one of Skeptic/Soundmind's favorite fallacies. The following is how he evaluates whether or not a line stage is audibly perfect:

1. Start with a mediocre preamp in today's world such as a H-K Citation 11. I owned and enjoyed one from 1974 to 1976 until the Frank Van Alstine modified FET-5 outperformed that classic. Not bad, just not particularly spectacular as compared with what is available today.

2. Pipe another linestage's output through one of the unity gain (usually buffered) tape monitor loops through two more set of high capacitance red and white patch cords.

3. Play music and switch between tape monitor in and out. If you cannot tell any difference, then the DUT is audibly perfect. Understandably, he found quite a few preamps which met that criteria.

He is hands-on experimenter like Pass. Personally I can't judge the validity of the work of either since I have no electrical engineering background.

Wouldn't it be fun to compare a line stage using Eliott's $18 OPA2134 based preamp board vs a Pass Labs XP-30? Or his MOSFET amplifier board based amp to an XA-200.5.?

I believe I have read the article before, but will do so again -- not tonight, though, because my eyes are shot.

By all means wait since it is a really good read.

[Feanor]

The other guys, by contrast, have graphically demonstrated their point. What is Elliott's amplifier reference? Eval system? Given his complete misunderstanding of cable interaction, I suspect he would fall into one of Skeptic/Soundmind's favorite fallacies...

On the issue of cables, I still don't understand: which Elliott article are you referring to?

I'm certainly not arguing that preamps sound the same, and neither would Elliott.

Wouldn't it be fun to compare a line stage using Eliott's $18 OPA2134 based preamp board vs a Pass Labs XP-30? Or his MOSFET amplifier board based amp to an XA-200.5.? ...

It would certainly be interesting to compare. Will you supply the XP-? I'll supply the $18 board.

...
By all means wait since it is a really good read.

Which I did, so it might seem that the basic disagreement is whether feedback causes high-order HD -- but note that Elliott doesn't deny it exists.

Pass refers to purported experimental evidence that feedback does cause HOHD. He show a graph that shows the corresponding relationship between amount of feedback and amount of distortion -- note that the relation ship is inverse: beyond a point, more feedback does reduce distortion. The point of maximum HOHD is about 15dB, a very small amount of negative feedback in practice.



Pass the proceeds with a graph the shows amount of HOHD at various levels of feedback, the highest being 15dB, the highest point of HOHD distortion according to the above. WTF? What would be the result with significantly higher feedback? Pass doesn't provide that info because it doesn't suit his argument.



From what I read, Pass' only real argument against a lot of feedback is that it might require more stages of amplification to provide the high non-feedback, ("open loop"), gain to accommodate the high feedback.

Elliott, on the other hand, would consider only 15dB feedback bad design. As I mentioned earlier, he also strongly believes in global, i.e. multi-stage spanning, feedback because it corrects distortion at all stages. Of course if you accept Pass' argument that any feedback is bad, then you are constrained to his approach of minimizing the number of amplification stages.

Apparently there isn't universal agreement among designers that high-feedback is bad, e.g. Halcro, Boulder, and others. Some reviews like these amps a lot, though others like Harry Pearson, (the hi-fi God??), do not.

[E-Stat]

On the issue of cables, I still don't understand: which Elliott article are you referring to?

Cable articles

Anytime you find an article about cables that starts with "The Truth"... you know that's from a labcoat who doesn't understand simple concepts. I just have to smile about this comment:

"On the other hand, many subjectivists claim that anything other than a listening test is invalid, and commonly and even vigorously eschew ABX testing ..This is very confronting, and to have one's beliefs shattered is not a pleasant experience."

Yeah, I'd love to see his face when someone points out the obvious fault with his pathetically inept "test"! Do you understand how incredibly stupid that is?

I'm certainly not arguing that preamps sound the same, and neither would Elliott.

What he claims is that the byproducts of NFB lie outside the audible band. Right! His op amp based preamp uses copious amounts of NFB. I'm convinced he *believes* that it would sound as good as a superbly engineered discrete design with little feedback.

Which I did, so it might seem that the basic disagreement is whether feedback causes high-order HD -- but note that Elliott doesn't deny it exists.

He remains clueless.

maximum[/I] HOHD is about 15dB, a very small amount of negative feedback in practice.

15 db is actually a relatively modest amount. That's what is used with my VTL amps. The newest Pass amps use about 10 db. 70s era amps used 40-50 db. Op amps require even more. Note that 40 db is a thousand times more than 10 db.

What would be the result with significantly higher feedback? Pass doesn't provide that info because it doesn't suit his argument.

Sure it does! Don't you see the huge higher order spikes that are non-existent at lower levels? Read the commentary!

"Paradoxically, you can visualize instances of feedback pyramid schemes, in which more gain stages are added to generate more feedback to partially correct for the distortions generated by the additional gain stage...

The resulting complexity creates distortion which is unlike the simple harmonics associated with musical instruments, and we see that these complex waves can gather to create the occasional tsunami of distortion, peaking at values far above those imagined by the distortion specifications."

From what I read, Pass' only real argument against a lot of feedback is that it might require more stages of amplification to provide the high non-feedback, ("open loop"), gain to accommodate the high feedback.

Then you missed reading critical parts of the article along with the conclusion.

Elliott, on the other hand, would consider only 15dB feedback bad design. As I mentioned earlier, he also strongly believes in global, i.e. multi-stage spanning, feedback because it corrects distortion at all stages.

Obviously. That's what poor designers do. Finding and working with the most linear discrete devices requires more effort and ability.

Of course if you accept Pass' argument that any feedback is bad, then you are constrained to his approach of minimizing the number of amplification stages.

And? Yes! That's certainly true. My VTL amps are two stage. Unless of course, you prefer the harshness of a Crown or the sterility of a Halcro.

Apparently there isn't universal agreement among designers that high-feedback is bad, e.g. Halcro, Boulder, and others. Some reviews like these amps a lot, though others like Harry Pearson, (the hi-fi God??), do not.

I don't base my opinions on how many reviewers like that approach vs those who do not. I prefer to listen to the results myself.

[Feanor]

Cable articles...

Thank you, I'll read it sometime. Meanwhile I've no comment.

...
What he claims is that the byproducts of NFB lie outside the audible band. Right! His op amp based preamp uses copious amounts of NFB. I'm convinced he *believes* that it would sound as good as a superbly engineered discrete design with little feedback. ...

I haven't heard Rod's preamp: have you?

BTW, Rod Elliott allows that opamps don't all sound the same but he does like the OPA2134's. I recently install a pair of these (replacing 5532's) plus an OPA2604 (replacing as I recall an OP275) in my latest DAC and the sound is extremely transparent and grain free.

In any case the internal feedback propagation of an opamp is measured in nano or even picoseconds which is effectively instantaneous being an order of magnitude above the audio spectrum.

...
15 db is actually a relatively modest amount. That's what is used with my VTL amps. The newest Pass amps use about 10 db. 70s era amps used 40-50 db. Op amps require even more. Note that 40 db is a thousand times more than 10 db. ...

Or maybe you're missing the point. Yes, 15dB is a modest amount of feedback ... but note, according to the graph Pass provides, it produces the highest level of HOHD -- even more feedback produces less HD of all orders.

...
"Paradoxically, you can visualize instances of feedback pyramid schemes, in which more gain stages are added to generate more feedback to partially correct for the distortions generated by the additional gain stage...

The resulting complexity creates distortion which is unlike the simple harmonics associated with musical instruments, and we see that these complex waves can gather to create the occasional tsunami of distortion, peaking at values far above those imagined by the distortion specifications."
...

Sure, each stage of amplification adds its own distortion. Elliott argues that sufficient global feedback will minimize all of them. Again, global feedback propagation (according to Rod) is <2 uS which corresponds to a frequency of 500kHz and is hence effectively instantaneous in the audio spectrum.

In a perfect world amplification devices would be inherently totally linear. Who would argue with that?

...
Then you missed reading critical parts of the article along with the conclusion. ...

Or maybe I'm just questioning them.

...
And? Yes! That's certainly true. My VTL amps are two stage. Unless of course, you prefer the harshness of a Crown or the sterility of a Halcro. ...

The very worst amp I've heard in the last decade was my old Phase Linear 400, so I'm not arguing the high feedback is an panacea. How I lived with that for 20 years I'll never understand (unless it degraded a lot towards the end). I haven't heard the Halcro. My current class D amp is the best amp I've heard on the last decade; it's the most transparent and grain-free that I personally have ever heard, (at least in my own system). I would not describe it as "warm", "tonally rich", "organic", or whatever the opposite of "sterile" is, but it reproduces instrument sounds more accurately than the low-feedback Monarchys that I was using 'till recently.

In the end I feel two points are to be made. (1) There are multiple amp design philosophies; either might work depending on the design & implementation. (2) There are different tastes in sound reproduction that will bias a person towards one or another actual product.

[E-Stat]

I haven't heard Rod's preamp: have you?

Effectively, yes. The output stage of my GamuT CD-1 uses an OPA2134 followed by an AD712. Nice, but better exists. I've had that player for ten years now.

In any case the internal feedback propagation of an opamp is measured in nano or even picoseconds which is effectively instantaneous being an order of magnitude above the audio spectrum.

Back to theory and assuming that is the only factor which causes audible differences. Like his pathetic cable test.

Or maybe you're missing the point. ]even more[/I] feedback produces less HD of all orders.

While the overall level reduces, the complexity of the distortion continues to increase. That is the problem.

Sure, each stage of amplification adds its own distortion. Elliott argues that sufficient global feedback will minimize all of them.

He can speculate all he pleases. Real designers test their hypotheses in the real world. The overall level drops a bit, but the complexity increases.

AIn a perfect world amplification devices would be inherently totally linear. Who would argue with that?

Which is why the very best gear I've heard, both SS and tube alike uses simple circuits with the most inherently linear devices that don't require severe crutches.

(2) There are different tastes in sound reproduction that will bias a person towards one or another actual product.

Fair enough. The clueless guys like Elliott can speculate all they please while the high fidelity industry quietly ignores them.

[Feanor]

...
While the overall level reduces, the complexity of the distortion continues to increase. That is the problem.
...
He can speculate all he pleases. Real designers test their hypotheses in the real world. The overall level drops a bit, but the complexity increases.
...

The complexity is irrelevant if it's inaudible -- but I won't convince you of that.

I think we've beaten this subject to death for the time being.

[E-Stat]

I think we've beaten this subject to death for the time being.

Sorry. My long term experience is unaffected by his speculation.

I still can't get over his cable test. Too funny!

[RGA]

Atma-Sphere on TAS and SETs

Atma-Sphere music systems, inc.» Blog Archive » Thoughts about the new TAS Buyer’s Guide….

[Feanor]

Atma-Sphere on TAS and SETs

Atma-Sphere music systems, inc.» Blog Archive » Thoughts about the new TAS Buyer’s Guide….

Same old, same old.

[RGA]

Feanor

Maybe you or someone can explain to me why so many people spend thousands of hours researching articles - reading white papers, and thousands more posting on forums and in virtually every single case none of them have ever spent 1 hour auditioning the amplifiers they are blasting away at.

I mean it can't be hard to make a list of known high end SET amplifiers and say right - I will give the technology a fair shake by listening to four of them. Just four to say "right I have sampled the technology fairly. It doesn't require a lot of time or money or effort.

Make a list -

Cary, Sugden's A21a, Audio Note, Shindo, Line Magnetic, Antique Sound Labs, Wavac, Almarro. Wyetech Labs, Fi 2a3, Yamamoto, Melody, Sound Master, Grant Fidelity.

Sampling different tube outputs and across brands and price points. Why is this so hard? Sure maybe a few are not located in reasonable driving distance but unless you're completely in the sticks somebody is selling SET somewhere in every major city.

Line Magnetic and Shindo are in Oregon, Audio Note has several dealers in California, San Francisco, New York, Florida, MA, PA, Virginia, Michigan, Colorado. 4 more in Canada.

Or better yet CES in Vegas. The Rooms that sounded jaw dropping brilliant with SE amplifiers - gee they rarely whine about rooms and treatments using that as a crutch every single year year in and year out no matter what the room.

Why do people distrust their ears so much? I'm sick and tired of talking about this garbage with a bunch of wankers who read some graph or some report from some engineer (they pick and choose whichever engineer they agree with). That would be fine if they actually heard the products fairly - and decide to trust the engineers their ears agree with. I can at least see the rationale.

[Poultrygeist]

FWIW, I live in the largest city in S.C. yet there is no opportunity to audition a SET here. I also do not know of any audio stores in Charlotte, N.C. that carry SETs. I bought two of mine on ebay and one on C.L.

I believe the reason many high end audio stores are reluctant to carry SETs is it forces them to also carry HE speakers ( often single driver FR ) which have less market appeal and can not command the higher price tags.

I read so many articles extolling the virtues of single ended triodes, buying one was only a slight leap of faith.

[Feanor]

The Atma-Sphere blurb is the standard anti-solid state cant. E-Stat and I were just finished grinding through all that stuff, which is why I might have sounded dismissive.

Why don't people don't people just believe and get baptised in the SET/HE faith? Why are there so many religious people? Beats me.

Well, except you say "believe your ears". My ears are telling me I'm getting great sound right now with my ripped CDs, s/s DAC, and class D amp. Could there be better sound? Likely, but I just don't have any incentive to change.

BTW, I'm not complacent: I'll switch when I feel the need. I changed from my low feedback, "tonally rich", more tube-like Monarchy amps because I felt I was missing maximal resolution, transparency, and most accurate instrument timbres. My current amp is better -- I'm believing my ears.

[E-Stat]

Why don't people don't people just believe and get baptised in the SET/HE faith?

You're mixing messages. Ralph Karsten doesn't make single ended triode amps. Triode, yes. Just like my VTLs which are switchable.

Only SETs limit your choice of speakers.

I changed from my low feedback, "tonally rich", more tube-like Monarchy amps because...

You will never understand until you have a really good tube amp in your system for a while. There are many which do well with Magneplanars.

[Feanor]

You're mixing messages. Ralph Karsten doesn't make single ended triode amps. Triode, yes. Just like my VTLs which are switchable.

Only SETs limit your choice of speakers. ...

Yep, I just don't see how 2-8 watts/ch is going to work for me. HE speakers are a whole other matter, and I'm not going to restrict myself to horn or single driver just so I can check out SET.

...
You will never understand until you have a really good tube amp in your system for a while. There are many which do well with Magneplanars.

Some fine day I will try a tube amp. The problem is it won't be the likes of your VTL. Even so, I will try some sort of push-pull that will do 50+ wpc. 30 wpc would probably work well enough for my chamber music listening but I think I'd be making big compromises in case of orchestral. Possibly I could bi-amp with an s/s tube combo (?).

Meanwhile I'm remarkably satisfied with my current setup

[Poultrygeist]

If you can not reach a listening level with that first watt you miss all of what's happening there. The first watt is where the magic lies. It contains 10 times the dynamic range of the next watt. SET/HE owners are fortunate enough to hear what they're getting in that all important first watt. Megga watt amps driving 83 db multi-driver speakers just can't go there.

No amp can resolve detail at low levels like a single ended triode driving HE speakers and that's probably the main reason folks who eventually land there seldom find anything they like better.

I was remarkably satisfied with my Classe and Aerials until I heard a SET driving HE FR speakers.

[E-Stat]

If you can not reach a listening level with that first watt you miss all of what's happening there.

I don't know of any systems where that is not possible. A 256 watt amp has 24 db more output than a one watt system. Many recordings achieve better than 60 db of usable dynamic range.

edit: I am currently listening to the double New Advents in the garage at approximately the two watt level (-30db). Which for the Stasis means that the class A voltage amp is totally driving the speakers.

The first watt is where the magic lies. It contains 10 times the dynamic range of the next watt.

I concur that the treasure lies at the micro dynamic level, but don't agree with your math. Doubling power represent only a 3db change. A 10 db change requires ten times the power. And dynamic range is relative to the noise level of the system.

No amp can resolve detail at low levels like a single ended triode driving HE speakers.

So, what is your point of reference for the best system in your experience that is not SET/horn based - on which your opinion is based? I'd be curious as to what you are comparing.

[Poultrygeist]

A pair of 96 db speakers playing with 1 watt against the average noise floor in one's listening room has 40 dbs of dynamic range but adding a second watt only increases the dynamic range by 3 dbs. For every additional 3 dbs you would need to double the power. It's all about the resolution in hearing the inner detail of that first watt. High efficiency full range drivers have very low mass and are faster and more accurate than inefficient drivers using heavier parts. The ultra thin banana paper cones of my Fostex and Tang Band drivers are like butterfly wings.

I recently auditioned a pair of $32k Aerial 10T powered by a high end solid state CJ amp. In no way did I hear the low level inner detail I'm accustomed to with my homespun SET/FR OB's. I found myself turning up the volume to hear more but I never heard more. With my home system I find myself turning the volume down to hear more.

A few years ago I purchased my first SET, a $240 Miniwatt. I was so bowled over I sold my Classe CAP-150 on ebay within a matter of days. The liquid warmth of the Miniwatt and great detail far surpassed the sterile sounding $3000 Classe. The $300 FR Tektons I bought at the same time replaced my $2500 Aerial Model 5's. It was an epiphany for me.

My son-in-law has some $30k powered Meridians ( once belonged to Emmitt Smith ) but they can't produce the holographic imaging and stunning clarity I get from my Bottlehead 2a3/OBs. The bloom and decay with sonic naturalness is what I get with SETs and I just don't hear this with the better solid state amps.

Since the single ended triode is the simplest of all amp circuits and uses the least number of parts perhaps there's just less to get in the way of the music.

[RGA]

Feanor

You don't strike me as a headbanger.

In a modest size room say 13 X 18 X 9 my 4 watt (undistorted) Audio Note OTO with the notoriously wimpy EL 84 output tubes had no trouble playing the AN K louder than I am comfortable listening.

The AN K is 90db (perhaps less given how AN is deemed to over spec their sensitivity ratings.

This is hardly a super efficient speaker - it's a little above average if their spec is correct. So you don't need a horn and you don't need a single driver.

I am currently reviewing the Class T amps from Trends - I've tried them on three loudspeakers including their own 88dB speakers. It also has not trouble driving any of the speakers to loud levels without me cranking the volume - again far louder than I would ever listen - (including classical).

The T amps are rated 15 watts but looking at the numbers more carefully it's only 6 watts into 8ohms with acceptable distortion - 10 watts into 8 ohms with 10% THD. 15 watts into 4ohms with 10% THD.

As the watts drop the distortion drops dramatically - similar to SET.

Again Monarchy isn't a SET. It's another SS amp trying to sound like a SET. ie; veiled - I only heard them once and maybe it was the speakers - but I'm not surprised you found better.

[E-Stat]

... but adding a second watt only increases the dynamic range by 3 dbs.

I'm glad we got that issue solved.

For every additional 3 dbs you would need to double the power.

I'm delighted you agree with me about that.

It's all about the resolution in hearing the inner detail of that first watt.

You have created a religion around the first watt. Admittedly, Nelson Pass would smile at your concept.

High efficiency full range drivers have very low mass and are faster and more accurate than inefficient drivers using heavier parts.

On that topic we are in complete agreement. Since a full range driver possess one coil of wire in addition to the cone itself, the mass is relatively low as compared with other cone speakers. The mass of thirty square feet of the U-1's diaphragm, however, is less than the air around it. I really do understand.

I recently auditioned a pair of $32k Aerial 10T powered by a high end solid state CJ amp. In no way did I hear the low level inner detail I'm accustomed to with my homespun SET/FR OB's. I found myself turning up the volume to hear more but I never heard more. With my home system I find myself turning the volume down to hear more.

Would you care to complete the description of the system?

A few years ago I purchased my first SET, a $240 Miniwatt. I was so bowled over I sold my Classe CAP-150 on ebay within a matter of days. The liquid warmth of the Miniwatt and great detail far surpassed the sterile sounding $3000 Classe. The $300 FR Tektons I bought at the same time replaced my $2500 Aerial Model 5's. It was an epiphany for me.

I looked to see what I could find about ithe Classe and located a schematic. It uses ten op amps in the front end and has but a 34 joule power supply. Your description of its sound makes sense to me.

My son-in-law has some $30k powered Meridians ( once belonged to Emmitt Smith ) but they can't produce the holographic imaging and stunning clarity I get from my Bottlehead 2a3/OBs. The bloom and decay with sonic naturalness is what I get with SETs and I just don't hear this with the better solid state amps.

I confess that I've never really favored powered speakers either. I had some active Braun speakers back in the mid 70s.

Since the single ended triode is the simplest of all amp circuits and uses the least number of parts perhaps there's just less to get in the way of the music.

Yes, the power amp is but one of about eight factors in the system.

[Feanor]

Feanor

You don't strike me as a headbanger. ...

Right: I'm not. But if I'm going to spring bucks (of which I have too few), I don't want to waste it on an under powered amp. I know 50 watts would do for the Maggies but I'm not so sure about fewer.

...
The T amps are rated 15 watts but looking at the numbers more carefully it's only 6 watts into 8ohms with acceptable distortion - 10 watts into 8 ohms with 10% THD. 15 watts into 4ohms with 10% THD.

As the watts drop the distortion drops dramatically - similar to SET. ...

Humm ... well class D/T amps are significantly different than AB amps.

...
Again Monarchy isn't a SET. It's another SS amp trying to sound like a SET. ie; veiled - I only heard them once and maybe it was the speakers - but I'm not surprised you found better.

It's a somewhat of an understatement to say the Monarchys are "veiled", but there better out there -- and not only SET.