Would You Rather Own A Good SET Amp, Or A Great Push Pull Amp?

All are good,it depends on you.You have to listen to your purchase and in your place...you have to Live with the choice you made.Best of Luck.

@atmasphere Amp in question is special order Tektron, Italian company, amp was Robin at Robyatt Audio personal amp.

Thanks to everyone who contributed their thoughts and input. After deeply and carefully exploring my own thoughts, and pondering everything commented on here, in particular your insights, Ralph, I made the decision, yesterday to purchase the PP dream amp (monoblocks). 

I had been looking for this particular 2a3 PP amplifier for over eight years, and had a daily checkpoint, over that entire time, to see if Hi-Fi Shark showed any coming up for sale in North America. Over that period, I had become enamored with SETS, possibly due to not having any other great options. But this amplifier finally appearing got me back on track. 

The story, unfortunately, has a rather sad conclusion. When I called to purchase it, they were literally on another line with someone who had beaten me to the punch by perhaps two minutes, at most. This is almost certainly the end of that dream. I’m retired, who knows if I will even be alive when another one appears, IF it even does, or for this kind of price. Not to mention, my hearing is still fine, but I have any number of friends who have been impacted with age related hearing loss. And once that happens, you only choice is to sell off your quality audio gear and try to learn to cope with listening to music reproduced in much lesser quality, since, at that point, it won’t matter anyway. 

It appears that my existing "good" SET is what I will have to live with. 

I thought I had a bit of time to purchase the amp because it was not generally even known to be on sale. And then, either Monday night, or early yesterday, some hours before I called to purchase it, all of that changed, which I was unaware of. It’s my unforgiveable mistake to live with. And regrets for likely as long as I can listen to music. Oh well, it could be worse, I could be dead or imprisoned.

@nightfall - so sorry to hear this 😔 - it will take awhile, but the gutted feeling does go away. 
 

@sns - I have this thought about music playback in relation to fidelity - that while most of us seek the most true-to-life presentation of what we hear, the highest fidelity does not happen when the performer/artists of each track we hear is brought into our listening space with utter realism in homogenising timbre and tone, but rather when we are transported from our listening rooms to the specific venue of resonant air where the recording took place, in the process of differentiation. High fidelity is about sorting the differences of realism from track to track rather than homogenising similarity of realistic outcome - the idea that preference for a more forward presentation does not represent high fidelity, because not all performances were recorded with emphasis on a forward presentation. I wondered what your thoughts are about this.

In friendship - kevin

@nightfall - Selling your gear because of age-related hearing loss may be YOUR only option, but it's certainly not the only option for people like me who wear hearing aids, and I've worn them for over 20 years. Just bought a fine new tube preamp yesterday, in fact. 

@kevn There is still plenty of differentation and or uniqueness amongst recordings, this includes sound staging. Still you can't get past inherent or a static generalized pattern of sound staging with any particular system, therefore, my system will always project images further forward than a system that has a more distant sound stage. A more forward sound stage also inherent to horn speakers in general, this why horn shape so critical to get right. Formerly had exponential mid horn in my Khorns, went to Volti tractrix horn, loads the room far more naturally than the exponential horn. Back to sound staging, my extensively modded Khorns (includes tweeter on custom baffle for proper physical time allignement)  have plenty of center image depth, so performers aren't right on top of you, thing is with necessary placement of khorns in corners and horn relative directionality hard panned info is necessarily projected out more into room and sticks to speaker more than other designs. I previously used Merlin VSM, you probably couldn't find a speaker that images, sound stages more different than khorns, Merlins well out into room away from all walls, narrow baffle, speakers disappear, more distant sound stage even with SET, with push pulls a bit more distant. And I've had any number of open baffle speakers over the years which sound stage and image in their own unique ways.  Bottom line is I've always managed to achieve a believable sound stage with every setup and system. Room treatments absolutely critical for this so have wide variety of treatments on hand for use whenever needed. 

So, what I'm trying to say is, there are so many presentations available to audiophiles, I've heard entirely engaging systems that span the entire spectrum of audio, so everything from high power SS with very inefficient, highly complex crossover speakers to flea power SET with highly efficient speakers, and everything in between. I'm actually totally agnostic when it comes to topology, if it sounds good it sounds good, how it gets there is for the designers/engineers to figure out. If I had the money and the room I'd have a dozen systems, each completely unique in regard to topology and sound qualities. 

@atmasphere. I'm only getting to respond to your response to my comment now due to the fact that in transferring to fibre broadband, my ISP managed to leave me with no internet access at all for several days - only resolved just now.

The original Halcro amplifier was, for me, the most vivid demonstration of how an ultra low distortion amp with multiple levels of NF can sound utterly unmusical.

yoyoyaya,

I too was shocked at how lifeless, flat, and unmusical was the sound of Halcro amps.  Clearly, low measured distortion is far from the only thing that matters.  

@yoyoyaya I remember hearing something like that too. 

The thing about distortion is we don't put a weight on the harmonics generated. The higher ordered harmonics are the ones that really mess with the ear- the 2nd and 3rd, hardly at all. 

But most amps have what looks like low higher ordered harmonic distortion on paper. If those harmonics are masked by lower orders (the 2nd and 3rd) then they aren't heard. SETs have the most higher ordered harmonics of any kind of amp made, but because the 2nd and 3rd harmonics they also make are prodigious enough, the amp sounds nice and smooth. 

Where I'm going with this is that the distortion signature of the amp is more important than the THD. The amp might be quite low in distortion but if the higher orders are not masked it will not sound musical. 

@sns  - thank you so much for your thorough and detailed reply : ) - you have raised important issues there to consider. If I may ask further - have you encountered any particular typology or signal chain to speaker that is not biased to either a forward or recessed presentation but is neutral as a general characteristic?

In friendship - kevin

SETs need high efficiency speakers and ideally subwoofers to support them. Compared to PP, which offer balanced operation, in my experience SETs major benefit is a better rendition of attack and decay.They subjectively have a faster attack (particularly noticeable on the reeds in oboes and clarinets and more organic reverb (best heard on good piano recordings).While @atmasphere is firmly advocating balanced connection I have so far been unable to replicate this SET advantage in a balanced setup. Even with highest efficiency horn speakers, headbangers listening to amplified instruments should stick to PP since in most instances the above SET benefits have been obviated by the PP amplifiers used when recording.

Having been around this crazy hobby for 40 years+ I have yet to find a better amp for listening to classical and jazz than Nobu Shishido’s Wavac EC300B with Takatsuki tubes.

Please continue to comment as a number of you have raised some very interesting points about the strengths of types of speakers. 

Meanwhile, I am still heartbroken about losing the amp, and by minutes. Something almost certainly never recoverable from. Eight years of effort, and hope, lost all at once. 

But I am learning things from the alternate discussion, which is never a bad thing. 

It is regrettable that the one you really wanted got away.  But, there are a lot of fish in the sea, and perhaps you will find something tat you will really like if you continue to do what you are doing which is to search for alternatives.  Absent more information, it is hard to make specific recommendations, but, I have some general observations.  First, I tend to like low-powered tube amps, whether SET or pushpull.   SET amps tend to be much simpler in design, which leads a lot of people to assume that they are cheaper to build.  Indeed, there are a lot of low-cost Chinese-built SET amps on the market.  But, SET amps require the air-gapped transformer that Atmasphere described above and good ones tend to be quite expensive, so a good SET amp is usually quite expensive.

I particularly tend to not like the sound of high-powered pushpull amps that use a lot of output tubes of the like of KT88, KT120 and KT150 tubes.  The sound is, to me anyway, "brittle" (hard and edgy).  I much prefer lower-powered pentode and tetrode tubes in pushpull amps utilizing only two tubes per channel.  My favorite types are 6L6, KT66 and EL84.  The EL84 is apparently a very easy to work with tube because there are MANY examples of low cost, good sounding amps utilizing that tube.  Low-powered triode tubes are also very good in pushpull configuration although they still should only be used with very efficient speakers.  I had a wonderful sounding 45 pushpull amp that put out about 4 watts per channel.  My current 349 pushpull amp puts out about 5 watts.

Meanwhile, I am still heartbroken about losing the amp, and by minutes. Something almost certainly never recoverable from. Eight years of effort, and hope, lost all at once. 

Well, if it’s any consolation, I usually find that when I’ve placed a certain product up on the proverbial pedestal for a long time, I typically end up disappointed when/if I finally experience said product. I bet there was a strong chance you would’ve found fault with those amps in short order. 
 

I’ve still yet to encounter an audiophile system that sounded “perfect,” to my ears, regardless of price. This leads me to suspect there is no real “perfect” component out there either. Tube amps are especially imperfect IME. All of them have some rather significant shortcoming, whether it be noise, distortion level, lack of headroom, dynamics, detail etc… Same can be said for good solid state, but to a lesser extent IME. 

You might be better off with a class A solid state amp like the Pass XA25 or one of the wide heatsink Codas (5.5, 12, 16 etc). These amps do nearly everything better than much of what’s out there in tube form. 
 

@kevn I've experienced any number of sound stages over the decades, this both from my own systems and others. It's been a long term migration towards SET and horns, simply my preference at this time, could change somewhat as I continue to experiment with different amps. 

One should always be mindful that preamps, sources, rooms also greatly affect presentations. My DHT, fully balanced,  transformer volume control pre changed things to a large degree vs 6sn7 and 12 series tube based pre's I previously used. More recently had a Pass XP-22 pre, this made for yet another presentation. And then we have footers, racks, etc, we can alter presentation to some extent here. And then we have subs, adding a pair couple years ago, opened up sound stage to an amazing degree.

Poll Time

1 - Listen to an Audio System or Audio Device added to an Audio System,  making an evaluation based on Experiences had of the Produced End Sound

2 - Assess Measured Data Belonging to a Device/Devices followed by Listening to an Audio System or Device added to an Audio System, making an evaluation based on understanding of Measured Data and Experiences had of  the Produced End Sound 

Add a Category if felt further is required.

Poll Time

1 - Listen to an Audio System or Audio Device added to an Audio System,  making an evaluation based on Experiences had of the Produced End Sound

pindac - Top of the List - I trust my experiences had - I trust some not all with the Math, but don't typically require any data to support what is being produced as a Sound.

I use data especially when matching speaker and amps. For instance, at first glance it would seem any flea powered SET would be fine for my 103db efficient Khorns, but upon closer examination there are two impedance dips down to around 3ohms and 43 degree  phase angles, perhaps not so good for current limited amp. But then we have differences in power supply reserves in amps, any two amps within a tube family may not be created equal.

OP

maybe i am biased but i owned SET. my audiophile friend has a great push pull amp well matched with his speakers. sounds very very good, but i still like mine (call it biased) but there is a something i hear on my gear that simply was missing on his.

So, as all people have posted above, SET has limitations. IF one is going to choose on the amplifier alone, then i would say a great Push Pull.

However, i like to point out, if your build your system  based on synergy (example, MY 845 tube amp, driving Falcon gold badge and supplemented with subwoofers) then the ball game changes. all limitations of the SET has been addressed. and the glory of the the SET is now on center stage. 

before closing, just a few days ago, a non audiophile came over, and he was interested to listen to my gear, so i did a showoff between my SS versus Tube gear, using my Falcon speakers, supplemented by subs. To him, he just wanted the sound of the 845 tube because what he hears it's involving and palpable. more than my SS. I hear it, but for me its a toss up. but for my non audiophile friend, its more than that. he is by profession a music director of a local singer/band who does shows and produces records

They subjectively have a faster attack (particularly noticeable on the reeds in oboes and clarinets and more organic reverb (best heard on good piano recordings).While @atmasphere is firmly advocating balanced connection I have so far been unable to replicate this SET advantage in a balanced setup.

@antigrunge2 If you want to hear speed in a tube amp OTLs are the king of that. 

But here’s an interesting tidbit FWIW: SETs are the slowest amps made in terms of rise time (or slew rate, whatever you want to call it). They also have the least bandwidth. 

So ’speed’ isn’t what you were hearing- its something else, and that ’else’ is distortion. If the speaker used lacked the efficiency really needed to show off an SET, the result is you will use power beyond about 20-25% of full power, at which point the onset of higher ordered harmonics will cause the amp to sound ’dynamic’. Once you understand that for what it is, the result may ruin the experience for you. If that happens, my apologies. 

If you really want to hear what PP is about, the thing to do is find a PP amp that is fully balanced and differential from input to output. The reason to do so has to do with the resulting distortion signature.

SETs have a quadratic non-linearity, which results in a prominent 2nd harmonic, which is the source of their ’magic’ and smooth sound.

An amp that has a fully balanced, differential circuit from input to output has a cubic non-linearity. This results in a dominant 3rd harmonic. If the amp is zero feedback, that third will be at slightly less amplitude than you will see in an SET of the same power at full power, IOW considerably less distortion. 

The ear treats the 2nd and 3rd in much the same way in that they are both innocuous and can mask higher ordered harmonics, while imbuing a quality audiophiles call ’warmth’.

The thing about the cubic non-linearity is succeeding harmonics fall off in amplitude at a faster rate as the order of the harmonic is decreased than seen in a circuit with a quadratic non-linearity, as distortion is compounded less from stage to stage within the amp. Its inherently lower distortion. 

Now if you mix single-ended and PP operation in the same amp (like in a Dynaco ST70) you get both non-linearities, which results in a prominent 5th harmonic (which will make the amp sound harder) thru algebraic summing. I suspect the limited set of PP amps that SET lovers tend to use to compare SETs to PP are of this type, as such designs are quite common.

When you use a limited sample size, the resulting conclusion is based on a logical fallacy- IOW its easy to fool yourself, since the results you got may have been very real but don’t describe the entirety of the pantheon. 

When you compare such an amp (which requires some digging) to an SET you find that SETs really don’t have any sonic advantage.  

@atmasphere,

As a woodwind player myself, I look for typical breathing, wind and reed rasping sounds at the very beginning of a note. Having had amps from Accuphase, Chord and a Graaf GM20 before the Wavac I can safely say that none of these as well as numerous amps in other peoples’ systems have been able to reproduce a lifelike resemblance equal to the Wavac. If you call that distortion, then be it. The amp certainly wasn’t driven too hard.

Here is the Audio Beatnik on this subject: https://www.dagogo.com/beatnik-pet-peeves-part-one/

Finally had a good listen to the Tektron  Bendix 6094 mono blocks pp. So we finally find a pp tube I can really appreciate, tonally balanced, refined, resolving, transparent, dynamic. In comparison to my 845, two 300B mono blocks this a contender. Just a few hours in so final conclusions not reached, but this just a bit more resolving than my custom build 300B, Audio Note Quest, Coincident 845, just a bit more open on top, sparkly yet refined, mids really challenge my WE 300B, natural and sophisticated harmonic development/timbre, slightly more recessed and  layered sound stage, a bit more width which really shows itself on mono recordings,  just a bit more macro dynamics.  Now here's the thing, at this point prefer the custom build 300B micro dynamics. Up to now I haven't spoken on the one thing that really draws me to SET, and that is the micro dynamics, I've yet to hear any push pull that quite equals a really nice SET. I've been through a few pp such as Conrad Johnson, Cary, highly modified Prima Luna, and I've heard really top flight pp at shows, audio dealers, private homes. My custom build 300B provide this really special 'breath of life' that makes performers come alive in my room, this close, not quite there. May have some mods to do, Duelund CAST caps and nude Vishay (Texas Components TX2575) resistors, Duelund Silver bypass caps on electrolytic caps in power supply really 'woke' up the custom 300B's.

And I do like OTL, Joule with the Merlins always one of my favs at shows.

I own both types of amps and listen to both types constantly. I have several amplifiers along with speakers In my listening room. When I first began in this hobby, I wished one day I would own a really nice sounding audiophile system. Well, 35 yrs later I’ve surpassed my initial goal mostly because I was fortunate to do much better financially than what I expected. One one of my setups, I switch back and forth between a 100lb 100-watt EL34 push pull amp and a 100lb 45-watt single-end amp that gets its power using a 805 tube with a 300b tube per channel. Both amplifiers are very high level and the best quality within its category. It’s the reason I selected them. I use the exact cables, preamps, front-end so nothing is different other than the amplifiers when I do the swap. What I hear between mine, neither one is better, it’s personal preference of what someone prefers. In my setup, my single-end amplifier give me a slightly more sharper sound. Images are tighter and the presentation is more forward. I’m not suggesting in a bad way, just more forward than my push-pull EL34 amplifier. The single-end amp projects the music closer to me. So when I listen to my system with the single-end amp, I feel like the performance is here with me.  When I listen to my Push-pull EL34 amplifier, I get a slightly less focused imaging and the soundstage is more recessed. The soundstage also appears to be larger. With both amps I’m getting perfect detail and resolution and both are very open sounding but I feel the soundstage is larger with my push-pull amp. The presentation with the EL34 amplifier is that the music doesn’t sound like it’s here, it sounds like it’s there somewhere between my speakers and behind them. I feel if someone prefers to here vocals and instruments intimately and wants to feel a really close connection with that, then single-end amps should be better. If someone values a large soundstage and like to see the performance laid out in front of them, push-pull should be good for that. I’ve tweaked both my amps with various tubes and fuses so that I would be satisfied with the sound each one is giving me. I feel I really like both amps. I’m so connected to my single-end amps that uses 1940’s GE 805 tubes. What a special amp I have. But when I switch to my EL34 push-pull amp, I many times feel I prefer it over the single-end amps because it is so relaxing and laidback with so a nice laid and layered big soundstage.  Both designs are excellent and unless you can do what I’ve done, you’ll need to decide what’s your preference. Actually I shouldn’t tell you mine but I will, I’m sold so much to heart on my single-end amplifier and I love the combo of the GE 805 tubes with 300b sound from this amp. I like the sound so much that I’ve purchased 10 backup GE 805 NOS tubes.  But the crazy thing is that when I listen to my push-pull EL34 amp, I think I slightly prefer it overall over my single-end amp. I’m never gonna sell either amp and I rotate between them on a monthly basis and I can live happily with either but if I had to choose solely on which I’m most satisfied with as an audiophile in sound quality, it would be the 100lb 100-watt Push-pull EL34 amp. 

Both amps are very transparent, both have great bass and excellent highs. Both are very open. Both have wonderful sounding midrange and very palpable and natural. But I find the single-end amp is more focused in the midrange. The push-pull sounds wider in its soundstage. 

Speaking in broad generalizations, SET (or single-ended amps in general) tend to be more lifelike, transparent and have fantastic mid-range. Push pull amps, on the other hand, tend to be more dynamic, have better high/low extension and way more versatile due to their higher wattage. A person's taste in music will also be a big factor. I tend to be a mid-range guy, so I prefer single ended amps, but I can get down with a good push pull.

@sns - thank you so much for your thorough replies : ) - kevin 

A class A Amplifier fits in the middle Pathos inpole has tubes  in its unique patented  circuit. MK 2

If you prefer the SET sound, but want to use less sensitive speakers, such as Magnepan, SET amps are so simple to design and build that you do not have to spend tens or even hundreds of thousands of dollars to add an 833A triode gain stage. I couple the globe 45 output through a Lundahl 2765 transformer to the grid of the 833A. I use a Hammond 1642 output transformer which is more robust than far more expensive transformers that are built the same way. I put 1000 Volts on the plate of the 833A' and bias the grid at zero Volts. I also use polypropylene capacitors in the power supplies and run the pi-filter chokes on the ground negative side for safety. This gives me a sound that closely matches what I hear in live performances in the opera house, which has good acoustics.  

EL34 certainly has the natural mids, highs can sound a bit rolled but I found various methods to solve this. Bass was the one area I found EL34 a bit wanting, sometimes a bit soft and warm, never could get impact AND tuneful, articulate bass. Sure I could have attained better with more work and finding sympathetic partnering equipment, attempts at this ended with discovery of 845 SET. And then we come to present Frankenstein Khorns, 845 is excess, its like a brute compared to the subtlety and refinement of 300B. The EL84 PP is right up there in refinement, subtlety, natural presentation, also has extended highs, more articulate, forceful bass than EL34. Bass very similar to 300B in my setup, virtually exact same settings on volume/gain, no change in sub settings. And this with REL subs, so high level settings connect to amp binding post and chassis ground, supposedly REL's take on personality of amps to some extent. 300B sometimes gets bad rap on bass, I hear no issues and I'm a stickler for articulate, tuneful bass.

Bottom line, I can very much relate to lowtubes in regard to comparison between SET and PP, in my case this being 300B and EL84. Difference with micro dynamics is present issue, but this early days.

This question is wildly subjective but that's ok.

I think the first question is what kind of speakers you are using. I've been the proud owner of a pair of Klipschorns for the past 48 years, so I can use any amp I want and I have owned many over the years.

The best set up for my system is bi-amping. I use a pair of Klein-Hummel 80 WPC channel transistor amps for the bass and I keep in rotation 10 different amps of various power and circuit configurations for the mids and highs in both tube and transistor output.

I have experienced stunning results with my 30 WPC Sony V-fet and Accuphase 30 WPC Pure class A transistor amps. My push-pull 8 WPC Eastern Electric 6BM8 tube amp offers phenomenal performance. I have realized similar results with my restored 15 WPC 1958 built Fisher 30-A tube mono blocs and my Dennis Had built tube amp that can use 7 different output tubes: the 5881, 6550, and KT88 are my favorites, again running 5 to 12 WPC.

Bottom line for me? SET at the end of the day. With the K-horns SET is simply magical!!

@hwr stated " I've been the proud owner of a pair of Klipschorns for the past 48 years, " 

Hats Off to You , being dedicated to a particular End Sound and the knowing of how it is capable of impressing is a very good place to be.

My ESL 57's both as Single Height and Stacked Array share similarities in their age with you own Speakers and a Pair are with me since the 90's.

They have been used on a variety of Valve Amp's and I have also heard the same model on a extended range of Amp's in PP - SET - OTL - SS.

I am always left impressed.

Are Vintage ESL's the best Speaker, certainly not, I have other Speakers that are parity in their attraction and leave very good impression as an End Sound.

Vintage ESL's and Valves have been part of my audio interest for more than half of my adult life, they have a place that is permanent as long as I am with health and hearing.  

I am a converted SET user, I failed to Wed to them at another time in life and very recently have found it difficult to consider anything else to listen to, but the 300b SET recently brought into the home system.

Helping out a Widower with my becoming a buyer of a few of the deceased Spouses audio devices has certainly been a game changer in my little world of audio, in both Amp' Design and Speaker Design. 

Hello, this is an interesting discussion that is relevant to solid state designs too, of course. Speaking of that, and  bringing this discussion down to the level of physics and transmission of sound waves through air, what about this assertion from Nelson Pass:

https://www.passlabs.com/technical_article/single-ended-class-a/

Virtually all the amplifiers on the market are based on a push-pull symmetry model. The push-pull symmetry topology has no particular basis in nature. Is it valid to use air’s characteristic as a model for designing an amplifier? If you accept that all processing leaves its signature on the music, the answer is yes.

Enjoyed reading this thread.  Educational, many different points of view, civil discussion so lacking in other threads these days.  Well done!

The Pass read very interesting, never heard the analogy to air argument previously. .

My earliest references for audiophile sound quality was Audio Research with Maggies and Dahlquist DQ10, this and the Dynaco ST35 with Dynaco A25 at the other end of things. Being a complete novice, I'd listen and half believe the crap Stereo Review mag threw out, as did much of the audio audience. Problem was going through all this great measuring class a/b amps I never came close to achieving sound quality I heard with my references. Over time I migrated to my first class a ss amp, N.E.W. DCA66, this battery power, resembled Pass designs. Finally getting closer to my references, still no cigar. From here I finally decided to take the plunge back into tubes, Conrad Johnson MV55, sweet, golden glow tube sound, loved that romantic presentation for a few years, what vocals! And then we go on and on over the years with a variety of push pulls looking for something between the romance and the analytical. Fairly successful, and then I get the horns and SET, took some years to get it all exactly to my liking, never thought I'd go back to push pull, and then this new Bendix 6094 mono blocks, now reconsidering push pull although this will never wholly replace SET, will always maintain a 300B amongst the collection. Now considering adding a Type 50 tube SET to collection. A First Watt always a lure as well. I'm always mindful every watt counts and should be highest quality, far more difficult to produce vast number of quality watts vs a single or just a few. 12wpc may be my max.

I do love to romanticize about a sweet 300B, 2A3, or 845 amp, and I keep saying I'll end up with one.....but we'll see. I love my Aric Audio "Push Pull" EL34 amp, and will die with it.

For years I've have a Bob Latino Tubes4hifi ST-120 amp that can run in "triode" or "ultralinear" mode. It's a $2,000 amp including tubes, and I just don't give it enough credit; it's a little brilliant amp. It responds really really well to tube rolling too.

Virtually all the amplifiers on the market are based on a push-pull symmetry model. The push-pull symmetry topology has no particular basis in nature.

Vacuum tubes and transistors don't either. This is a red herring.

I don't normally disagree with Nelson, but in this case I do. 

If physics is a part of nature, then we have an access to see that indeed, push pull does have a basis in nature. Two men with a large saw can cut down trees a lot faster than a single person can. You can see that analogies can be used in ways to make any argument you want 

We've been selling OTLs for the last 49 years. They have one stage of gain so are simpler than an SET which might have 3 gain stages including the output. 

So you can see there are a lot of ways to twist the simplicity argument around too. 

The bottom line is regardless of the amplifier class of operation or topology, the result must be musical, so that amplifier will have a very good first Watt as well as the succeeding Watts. 

But, with proper speakers and a quality build, SET amps deliver a very "pure" sound that can be quite seductive.  While bass tends to be not as tight and punchy as that of a good pushpull amp, the bass has its own positive attributes--it has natural and subtle differences of tone while pushpull bass is much tighter and punchy at the cost of sounding a touch "mechanical" (sameness).  I have no disagreement with someone finding the balance of attributes favoring SET amps.

I can't describe better my experience with SET and push-pull amps. I personally prefer a natural bass of a SET than a push pull punchy bass. My set amplifier has a very high distortion (THD) but it sounds clearer than my push-pull and solid state amplifiers with much less THD. The comment made by Ralph somehow explains this phenomenon. 

So if you listen mainly listen to (an amplified) rock or electronic music, push-pull would be fine. But if you want to fully enjoy an acoustic music, you will do that better with a set amplifier.

It surprises me that the THD is still a major parameter for measuring the sound quality. It is quite irrelevant to the quality of the sound as we perceive it. Even the formula by which it is calculated doesn't look relevant. I guess the engineers were unable to develop  a more adequate parameter for measuring the sound quality so far. I suggest this is the only reason why THD still remains a standard. 

So if you listen mainly listen to (an amplified) rock or electronic music, push-pull would be fine. But if you want to fully enjoy an acoustic music, you will do that better with a set amplifier.

@niodari This statement is misleading at best. Electronics really don’t care what kind of music you play. An orchestra is acoustic and it can play as loud as a rock band and can have just as much bass energy.

I don’t regard THD as having much to do with sound quality. It is a metric that is only useful after the distortion spectrum has been analyzed. 

For example the harmonic spectrum of a typical SET is pretty good; imagine what that would be like if it was the same spectrum but a 10th as much at full power. Distortion obscures detail, so an amp of the latter distortion character would be more revealing at any power level and more relaxed. A good PP amp can do exactly that. 

When doing analysis of any amplifier, IME/IMO its a good idea to see what harmonics are showing up at low power, at 6dB below full power and then see how the amp behaves when overdriven. 

Its also a good idea to look at distortion (and its spectrum) not just at 1KHz, but at 20Hz and 5KHz also. Tube amplifiers often fall apart at lower frequencies (SETs in particular); you put in a sine wave but the output can look pretty dreadful. So 20Hz response really can tell a lot about an amp. 

At what frequency does the output fall off? If there is no feedback, phase shift will affect tonality and sound stage presentation to 1/10th or 10x the cutoff frequency, depending on which end of the audio band we’re talking about. 

From those measurements you can get a good idea of how musical the amp can be. 

The thing that isn’t being discussed here is the vast difference in power between the SET vs the PP amp used. If you really want to hear what that is about the SET should be the same power as the PP amp. For example if you have a stock Dynaco ST70 compared to a 35 Watt/channel SET, you’ll find out really quickly how the two actually compare. 

Since high power SETs are not easy to find and really expensive if you do find one, its a bit easier to try to find a PP amp of the same power as the SET. If you can do that again you find the SET has no musical merit over the PP amp, if both are built to good quality construction standards (often PP amps of low power are not given the same quality of parts as a 35 Watt amp might have). 

Alternatively you could compare an SET against a PP amp which uses the same tube complement, in particular the power tubes.

If one amp has feedback and the other does not that too is a tremendous variable!

It is on the account of these three variables that most comparisons between SETs and PP amps don’t hold water; they are instead examples of a logical fallacy known as a limited sample size. 

@atmasphere (Ralph), ​My judgment comes not from a particular knowledge of electronics but from the practice, from my (humble) experience of using and auditioning audio equipment. Hence, it can be subjective. Since your judgment is clearly  based on your extended knowledge of electronics, it should be objective. But, as you have also mentioned in a number of different threads, measurement and theory do not necessarily correctly reflect how the sound is actually reproduced. Much depends on the models the measurement parameters are derived from (they can be misleading). A good example of that is the formula for the THD, as you likely agreed. In the light of your comments in this and the other threats, a more truthful parameter would take into account how an amplifier performs on high harmonics and how low harmonics mask the distortion in high harmonics, and not only at 1khz. It is not impossible to construct such a formula and create a new parameter, call it an Overall Distortion Measure (ODM) or whatever, amean for all tested frequencies or for each frequency range separately. Electronic engineers like you, could surely contribute to this.  

My comment that you call "at least misleading" comes from the earlier cited observations by @larryi  , which completely coincide with my experience using the PP and SET amps, though of different power (I comment just a bit later about my perception regarding amplifiers of different power). Based on your comments, these comparative observations may not be correct because of different power capacities that SET and PP amps typically have. Suggesting that the above observations are correct (see my note below), a simple implication is that a SET amp should manage better with the reproduction of acoustic music (that includes classical music, most of jazz and unplugged rock performances) since it does that in the "purest possible" way at a reasonable volume level. At the same time, for instance, there is no "natural" base in an amplified rock music and a tight and punchy bass can be more pleasing, just as an example. Regarding your note, although the bass energy in a symphony can be about the same as in rock, the base is still natural (acoustic) in any kind of classical music. 

Regarding the power differences in SET and PP amps that you addressed, I personally, very rarely raise the volume of my 5.5W SET amp more than 1/4 and my normal listening level is below that on any kind of music including rock or symphonic or music (and my Thiel CS 6 speakers are not sensible at all, they are at 87). In general, on all amplifiers I have and had in the past, I try to keep a similar volume level, at least below 1/2 even if for rock, independently of the speakers I use. To me, the distortion raises too much above a reasonable volume level,  independently  from how powerful the amp is; that is to say, on e.g., 1/4th volume level, I do not perceive too much louder the music played on my 600 watts SS amp than on the 5.5 watts SET amp on the same volume level. To my perception, independently of the power an amp has, there is more or less the same threshold level above which the distortion is notable, a reason why I don't care too much about the power in watts. 

But, as you have also mentioned in a number of different threads, measurement and theory do not necessarily correctly reflect how the sound is actually reproduced.

@niodari Actually I've said quite the opposite, although decades ago we could hear things we couldn't measure. Those days are gone as measurement technology has vastly improved. Apparently the education of what the measurements mean is still lacking.  

a simple implication is that a SET amp should manage better with the reproduction of acoustic music (that includes classical music, most of jazz and unplugged rock performances) since it does that in the "purest possible" way at a reasonable volume level. At the same time, for instance, there is no "natural" base in an amplified rock music and a tight and punchy bass can be more pleasing, just as an example.

Not sure what "purest possible" way might mean, but if you mean that an SET does that we'll simply have to disagree. 

At any rate there's no genre of music that does not have full spectrum although there are examples in those genres that do not. One has to be careful to avoid committing a logical fallacy due to a limited sample size. I've found that if you really want to get the system to sound right, one important aspect is to set things up so your amps are not working hard. This is very difficult to do with an SET because they only make about 20-25% usable power, above which, as we see in this comment:

on e.g., 1/4th volume level, I do not perceive too much louder the music played on my 600 watts SS amp than on the 5.5 watts SET amp on the same volume level. To my perception, independently of the power an amp has, there is more or less the same threshold level above which the distortion is notable, a reason why I don't care too much about the power in watts. 

-the higher ordered harmonics show up in greater amounts, imparting a sense of loudness due to the fact they are used by the ear to sort out how loud sounds are. So if your amp is generating more of them it will sound 'loud'. A sound pressure meter will show the truth of the matter. 

I've found it the mark of the best systems that they don't sound loud; instead are relaxed and detailed while maintaining authority at any volume. If you read a lot of comments by SET owners, this 'loud' issue is quite common, expressed in comments like '7 Watts is all I need as it plays plenty loud'. If you have clean power you will find yourself naturally and easily playing the system at higher volume levels because the artificial 'loudness cues' generated by the electronics are vastly reduced. 

If you are really intent on using an SET, you would do yourself quite a favor by using a speaker of greater efficiency!

I would look for something about 10dB more efficient- that will allow you to reach the same sound pressure with 1/10th the power and I think you'll find the greater resolution and dynamic character a nice benefit. Finding speakers that are 97dB these days isn't really that hard.

The second most important thing to do (far more important than finding the 'right' tubes for the amp) if you want the most out of the SET is to prevent bass information from entering the amp (I explained earlier why this is so important on this thread). You'll find the audible improvement immediate and obvious. 

Actually I've said quite the opposite, although decades ago we could hear things we couldn't measure. Those days are gone as measurement technology has vastly improved. Apparently the education of what the measurements mean is still lacking.  

@atmasphere ,  My memory could have failed in this issue. Though you agree that THD does not reflect the sound quality, are there other important parameters based on which we can judge about the SQ without the audition? 

If you have clean power you will find yourself naturally and easily playing the system at higher volume levels because the artificial 'loudness cues' generated by the electronics are vastly reduced. 

If you are really intent on using an SET, you would do yourself quite a favor by using a speaker of greater efficiency!

The loudness can be subjective. I was just auditioning a cassette  deck on two different cassettes on my 50 watts PP amp  with EL 84 output tubes. The first cassette is not "loudly recorded", so I had to rise the volume up to almost 1/2 of the max power. This was an optimal level since rising higher the volume did not really result in a louder coherent reproduction. The other cassette is recorder louder, so that I had to drop the volume at about 1/4 to get the same loudness as on the first cassette. Rising the volume higher had the same affect as with the first cassette at the level 1/2.  A natural implication here is that the distortion did not actually come from the amplifier:  The sound itself becomes "distorted" if it is higher than an acceptable (in this case, for me) level. So, it looks like a completely acoustic (non-electronic) matter. 

I played my 5.5 W SET amp (EL34 tubes) alternatively with 91db and 87db speakers. Honestly, I noted only a slight difference, much less than i normally note between two different CDs recorded at slightly different loudness levels. And the SQ with the large (less sensible) Thiel speakers is much better. 

I would love to audition a non-SET SS amp with the same clarity as my SET amp (at this point, I am a bit skeptical).

@atmasphere  I can't totally agree with you on your opinion of loud, many live unamplified concerts can be very loud. How can an audio system be accurate if it isn't loud like some live unamplified music is?

How can an audio system be accurate if it isn't loud like some live unamplified music is?

@invalid It wouldn't be! It must reflect that of the signal (since that is as close to the music as an audio system can get). 

If the system sounds loud when it really isn't (for example, only 83dB), that should be a point of concern. 

are there other important parameters based on which we can judge about the SQ without the audition? 

@niodari  Yes- I mentioned some prior, such as understanding the harmonic spectrum created by the amp. IMO that is more important than the THD. 

The loudness can be subjective.

Exactly my point; it can be independent of the actual amplifier power due to the presence of distortion fooling our ears into the impression its louder than it really is (as seen on a sound pressure level meter). 

All analog tape machines have distortion but I think you drew an incorrect conclusion from your cassette anecdote. 

If you really want to hear what a 5.5 Watt SET does, you would need a speaker of at least 12dB more efficiency (in most rooms); the few dB difference you mentioned still means that the amp is being pushed too hard.

I know of a PP amp that has about the same power as your SET. I've seen it take on SETs of similar power with no problem. It, or something like it might be something to try.   

Not that I am an educated individual in such matters, but I do know how to learn basics. 

A Stray away from the tech' talk and back to basics what I know. 

I know through being involved in and being close to design talk and builds, that certain circuits are produced with the Trannxs selected to be used to enable a Valve to work at a Optimised state.

Creating the different circuits to suit the Amp' design has great potential to create discernible difference to an End Sound. Neither should be any less enjoyable as an experienced End Sound, which is what I am experiencing. 

Yes- I mentioned some prior, such as understanding the harmonic spectrum created by the amp. IMO that is more important than the THD. 

@atmasphere ,  it would be interesting to detail this point: which are these important parameters that more realistically reflect the real performance of an amplifier, or are they still to be developed? 

I am about to change my earlier  point of view that engineering is a purely applied science. I share the concern by @invalid . I think that before we talk about the amplification, we need to understand how the music performance in question was recorded. That is to say, two things are to be distinguished and seen  separately: Which kind of music we are talking about and how it was recorded (with a single microphone or each instrument separately etc.) , and then how the recorded music is reproduced by some amplifier. I think that the second issue should be discussed based on the analysis of the first one. So suppose for now that we have a "perfect" amplifier of your choice and see how its reproduction is affected by a particular kind of music and by how this music was recorded. 

I suggest that an inherently loud music (e.g., rock) is easier and more natural to reproduce loudly than an inherently non-loud music (e.g., jazz). If we force the amp to play an inherently non-loud music loudly, you may unavoidably get some distortion, regardless of how the amp is built.  By the way, among the two cassettes mentioned in my  previous post, the first one was jazz and the second one was rock music (such an analysis would understandably make less sense for CDs, i.e., it should be an analog source).