Tubes vs. Solid State Amps

[adrianywu]

I addressed some of the issues you raised in my article here:

To Test or Not to Test, That is the Question (Part One) – PS Audio

www.psaudio.com

Psychoacoustics is more complicated than just odd and even order harmonics, because each harmonic has its own character. The sensitivity to distortion is also frequency dependent. In general, the sensitivity is much higher for high frequency signals than for frequencies from the mid-bass down. Some odd order harmonics sound objectionable because they are dissonant, and the brain will notice a very low level of such harmonics. What Jean Hiraga found in his experiments more than 30 years ago is that generally, each harmonic is masked by the harmonic lower down. Therefore, if the 4th is at a higher level than the 5th, for example, the high one will have less effect. SET amplifiers do not generate euphonic distortion per se, but the high levels of second and third harmonics serve to mask the higher ones. Second and third order harmonics are generally well tolerated, because musical instruments also generate a lot of these harmonics. Adding these harmonics to the fundamental causes what listeners describe as a " muddling effect", but if you add a dissonant harmonic, such as the 7th, even at a very low level, listeners will find it highly unpleasant.
Early transistor amplifiers did add mainly odd order harmonics, and this might be a function of high negative feedback with inadequate open loop gain. Intermodulation distortion is another subject, and this distortion is all dissonant.
Is it possible to design solid state amplifiers that have a "pleasant" harmonic spectrum while avoiding IMD ? Absolutely. You just need to know what you are doing. Equally, it is possible to design tube amplifiers that have objectionable harmonic spectra, although you would need to try hard.

 

[DonH50]

I was looking for an answer from someone who understands and builds quality amplifiers based upon the principles of short signal paths, 99.9999 pure silver and the use of the best quality parts, but limiting parts as much as possible as each additional part affects the sound adversely. I am asking for people like Masaki Ashizawa of Kondo Japan, Ken Shindo of Shindo electronics and Don Garber from fi who build exceptional sound reproduction systems based upon these design concepts explain them in a way that even audio engineers like DonH50 will understand and appreciate.

I am done here. I do not meet your qualifications for competence. I'll peek in now and then to see if one of your preferred designers chimes in to cure my stupidity, but for now have other places to be. For the record, my day job is not audio circuit design, though I have done some in the past, but fundamental engineering principles and physics do not change (actually, it gets harder as you reach into the GHz range).

 

[morricab]

Actually, the fact that nearly all transistor amps are push/pull is the main reason they are primarily generating odd order harmonics. Push/pull cancels even orders by design. Push/pull tube amps do so as well but there is more residual even order harmonics in most designs (tube matching issues??). Negative feedback tends to suppress all harmonics but it suppresses the low orders more and has the side effect of generating new, higher order harmonics. This can lead to a signal modulated "noise" floor that looks like noise but is really a myriad of distortion products. Being signal modulated and correlated it is not something one can "hear below" like one can with true noise.

The correct weighting of harmonics has been the subject of a number of papers over the decades and at least one thesis. One of the first serious attempts was a BBC engineer named D.E.L Shorter who though the weighting would be the square of the harmonic order but Cheever found that this didn't actually go far enough and that there is also an SPL dependency that needs to be accounted for.

 

[Rensselaer]

Thank you adrianywu for that informative article.

So, if I read this correctly, Harold Leak says that he never listened to his amplifiers, instead designing solely by looking at measurements.

It also suggested that some manufacturers traditionally used specific measurements to sell products, measurements which might not make a difference. I suppose this is referring to when they printed THD levels in large letters in adds and on packaging, the lowest levels attained by applying the most global feedback, (which in large amounts can adversely affect sound too), in order to sell products by measurements (rather than sound).

But most applicable to this thread, Hiraga’s study into harmonics. He found that the removal of even order harmonics so that only odd order remain, sounds horrible. Also, that a large amount of a lower order harmonic distortion can pleasantly mask unpleasant higher order harmonic distortions. This observation suggests to me that the reason why many DHSET’s, which allegedly put out much more 2nd order harmonic distortion than SS, sound great to many people.

DonH50, I am confused by how you interpreted my comments so negatively as I never intended such, however please accept my apologies for making you feel I had.

 

[morricab]

Thank you adrianywu for that informative article.

So, if I read this correctly, Harold Leak says that he never listened to his amplifiers, instead designing solely by looking at measurements.

It also suggested that some manufacturers traditionally used specific measurements to sell products, measurements which might not make a difference. I suppose this is referring to when they printed THD levels in large letters in adds and on packaging, the lowest levels attained by applying the most global feedback, (which in large amounts can adversely affect sound too), in order to sell products by measurements (rather than sound).

But most applicable to this thread, Hiraga’s study into harmonics. He found that the removal of even order harmonics so that only odd order remain, sounds horrible. Also, that a large amount of a lower order harmonic distortion can pleasantly mask unpleasant higher order harmonic distortions. This observation suggests to me that the reason why many DHSET’s, which allegedly put out much more 2nd order harmonic distortion than SS, sound great to many people.

DonH50, I am confused by how you interpreted my comments so negatively as I never intended such, however please accept my apologies for making you feel I had.

Studies have shown that up to at least 2% 2nd order harmonic is inaudible...that audibility increases exponentially with increasing harmonic order...

 

[PeterA]

Studies have shown that up to at least 2% 2nd order harmonic is inaudible...that audibility increases exponentially with increasing harmonic order...

Here is a graph of how my left channel ML2 measures from the manual. I figure I use a fraction of a watt with my 105 dB/16 ohm speakers at listening levels of around 70-80 dB with rare peaks into the 90s, so much less than 1% THD plus noise. I don't know about 2nd harmonic or if this graph is relevant to the discussion. Even at 5 watts (which I do not think I ever hit even when blasting at 100+dB), it is around 1% if I am reading this correctly.

 

[wil]

Here is a graph of how my left channel ML2 measures from the manual. I figure I use a fraction of a watt with my 105 dB/16 ohm speakers at listening levels of around 70-80 dB with rare peaks into the 90s, so much less than 1% THD plus noise. I don't know about 2nd harmonic or if this graph is relevant to the discussion. Even at 5 watts (which I do not think I ever hit even when blasting at 100+dB), it is around 1% if I am reading this correctly.

View attachment 102808 T

Quick side note question @PeterA ? Are you using a phone spl meter app, C weighted, and if so which one? I’ve found they tend to vary a lot.

 

[PeterA]

Quick side note question @PeterA ? Are you using a phone spl meter app, C weighted, and if so which one? I’ve found they tend to vary a lot.

No, I am not using the telephone app. I have a handheld meter that member Al M gave me and I confirm it with my radio shack SPL. Some more dynamic classical music is in the 50-60 range with peaks much higher.

 

[morricab]

Here is a graph of how my left channel ML2 measures from the manual. I figure I use a fraction of a watt with my 105 dB/16 ohm speakers at listening levels of around 70-80 dB with rare peaks into the 90s, so much less than 1% THD plus noise. I don't know about 2nd harmonic or if this graph is relevant to the discussion. Even at 5 watts (which I do not think I ever hit even when blasting at 100+dB), it is around 1% if I am reading this correctly.

View attachment 102808

This shows distortion vs. frequency. WHat you will often see is that distortion rises with frequency (and with it the character of that distortion is changing as well). The fact that the ML2 doesn't do this is a good thing. What it doesn't tell you though is the nature of that distortion. You will only see that with a FFT plot. It probably means though that the highs will sound clean and natural if the basic distortion is following the "right" pattern.

I think there are some stereophile measurements of the ML2.1 or ML2.2 that do show that data.

Lamm ML2.2 monoblock power amplifier Measurements | Stereophile.com

Note that the ML2.2 has the same flat distortion vs. frequency plot (perhaps even better).

Plot 12 shows the FFT distortion at 50Hz and shows the dominant low order harmonics. what is important about the distortion vs. frequency plot is that it means the high frequency distortion pattern will look pretty much the same as the 50hz spectrum. You cannot say the same for amps with rising distortion vs. frequency.

The IMD is also pretty good for a SET or otherwise.


Lamm ML2.1 monoblock power amplifier Measurements | Stereophile.com

figures 13 and 14 confirm that the pattern remains the same for both low and higher frequencies. What is interesting is the ML2.1 has a somewhat different distortion vs. power and distortion vs. frequency from the ML2.2...not a huge difference but a difference.

These are quite good measurements for a SET.

One of my amps, the Amplifon SET 42 SE, which uses the GM70 (like the ML3) has very good measurements for a SET as well and sounds very transparent but with great tonal expression. I can imagine the Lamm to sound a lot like that.

 

[morricab]

This shows distortion vs. frequency. WHat you will often see is that distortion rises with frequency (and with it the character of that distortion is changing as well). The fact that the ML2 doesn't do this is a good thing. What it doesn't tell you though is the nature of that distortion. You will only see that with a FFT plot. It probably means though that the highs will sound clean and natural if the basic distortion is following the "right" pattern.

I think there are some stereophile measurements of the ML2.1 or ML2.2 that do show that data.

Lamm ML2.2 monoblock power amplifier Measurements | Stereophile.com

Note that the ML2.2 has the same flat distortion vs. frequency plot (perhaps even better).

Plot 12 shows the FFT distortion at 50Hz and shows the dominant low order harmonics. what is important about the distortion vs. frequency plot is that it means the high frequency distortion pattern will look pretty much the same as the 50hz spectrum. You cannot say the same for amps with rising distortion vs. frequency.

The IMD is also pretty good for a SET or otherwise.


Lamm ML2.1 monoblock power amplifier Measurements | Stereophile.com

figures 13 and 14 confirm that the pattern remains the same for both low and higher frequencies. What is interesting is the ML2.1 has a somewhat different distortion vs. power and distortion vs. frequency from the ML2.2...not a huge difference but a difference.

These are quite good measurements for a SET.

One of my amps, the Amplifon SET 42 SE, which uses the GM70 (like the ML3) has very good measurements for a SET as well and sounds very transparent but with great tonal expression. I can imagine the Lamm to sound a lot like that.

SoundStage! Measurements - Einstein Absolute Tune Integrated Amplifier (10/2007) (soundstagenetwork.com)

Here is an example of an amp that I though SHOULD sound good...so I bought one many years ago. I had been favorably impressed with the Einstein OTLs and preamp so I though this hybrid integrated would deliver a good dose of the big boys. I also found these measurements that have a flat distortion vs. frequency, a mild rise with power (not quite like a SET) and a distortion pattern that shows a steep drop off after the 5th harmonic, which is arguably too high but but is quite low in absolute terms.

How did it sound? Frankly, rather flat and boring and more "SS" character than I was expecting. This was one of my great disappointments in amp purchases. Now, it didn't sound bad just didn't make you want to listen through it.

SoundStage! Measurements - Wyetech Labs Topaz 572B Stereo Amplifier (6/2003) (soundstagenetwork.com)

This one looks relatively poor but this amp sounds very good with the right speaker and was heavily lauded in it's day.

SoundStage! Measurements - Monarchy Audio SE-160 Mono Amplifiers (12/2003) (soundstagenetwork.com)

Here is a hybrid that measures very much like a SET...and it sounds really good!! Superb sounding amp for very little money. If you didn't know it was a hybrid you would swear it is a SET from the test data. Does it sound just like a SET? Not quite but if I had a tight budget and/or heat etc. was an issue or had harder to drive speakers this would be a top choice.

General
The Monarchy Audio SE-160 is an interesting attempt to duplicate some of the characteristics of a single-ended (SE) tube amplifier in a hybrid solid-state design. What is very SE-tube-like is the amount and way the distortion rises with power output, with the second harmonic being dominant. This distortion characteristic, no doubt, is generated in the vacuum-tube front end of this design. What is different from most tube SE amplifiers is the wide bandwidth and low out impedance of the SE-160.
Chart 1 shows the frequency response of the amp with varying loads. As can be seen in the chart, the high-frequency bandwidth is about 100kHz and is nicely controlled in shape as a function of loading. In the case of the NHT dummy load, the variation is about a harmless +/-0.25dB. Chart 2 illustrates how total harmonic distortion plus noise versus power varies for 1kHz and SMPTE IM test signals and amplifier output load. As can be seen, attainable power is greater for the 4-ohm load, as is usual for most power amplifiers. Note the SE-tube-like smooth increase in distortion over the whole power range. Also note that the distortion is less for a 4-ohm load. Total harmonic distortion plus noise as a function of frequency at several different power levels is plotted in Chart 3. Admirable is the low increase in distortion at the higher frequencies. Damping factor versus frequency is shown in Chart 4. A spectrum of the harmonic distortion and noise residue is plotted in chart 5 for an 8-ohm load. The AC-line harmonic spectrum is composed of odd harmonics, and there are some modulation effects of the line frequency around the second harmonic of the signal test frequency of 1kHz. The signal frequency harmonic components fall off in a nice manner with the second harmonic most dominant. This is said to have desirable sonic consequences.

​

 

[Thieliste]

Someone is tempting me to get away from solid state amps and get into tube amps - specifically Air Tight ATM-3211s.

My father use to operate tubes, he wouldn’t own a solid state two-channel system. I remember him always fiddling with them. I have some friends that own them. They sound great IMO. Great sound stage and a warmth normally heard in live natural sounding music. And I do enjoy such.

This said, what say you? And if someone does go the way of tubes would you go with Air Tight or ??? (Monos). Budget 80k list. They would be driving Magico M3s. Alon Wolf uses an Octave to drive his speakers - at least he use to, things may have changed now ...

View attachment 97002

I would stick with Vitus Monos and try VAC Master or Statement linestage or other tube linestages.

 

[Zappadaddy]

Don't throw out red herrings to attempt to invalidate my views, Mike. I have had these views long before I became a dealer of any brand and I am not comparing darTZeel to what I sell. I have heard darTZeel from it's inception, been to Herve's workshop and home (at least circe 2005...he had just come out with his preamp) and heard his reference system as well. The one time I heard the Tenors it was revelatory in just how good it sounded...no need to think about it...it was just present and alive. Never got anything remotely like that from darTZeel...just decent hifi.

how would you compare sound of Tenor and Aries Cerat Essentia or other models?

 

[JackD201]

I brought out one of my tube amps and as soon as I powered up it showed that I have 4 output tubes had gone bad. I removed the tube cage and there it was. Easy to see. Tubes ordered. I'm bummed put such is life.

Been playing with our most affordable floor stander the Marten Oscar Trio with my Quad II Eighties. I got this pair not long after Tim passed away. He was always fun to talk to at the HK shows. I remember we were cracking up because I said his 509s looked like a small box of tomatoes LOL That was actually what I wanted but the Q2s found their way to me.

The Q2s are quirky as all hell and fall short in so many ways but damn can this amp put lipstick on a pig. The Oscars and Q2 qualify for my new Audio Gateway Drug Award. LOL

 

[Argonaut]

I brought out one of my tube amps and as soon as I powered up it showed that I have 4 output tubes had gone bad. I removed the tube cage and there it was. Easy to see. Tubes ordered. I'm bummed put such is life.

Thats fairly unusual Jack , to which amplifier(s) are you referring and what tube type ?

 

[morricab]

how would you compare sound of Tenor and Aries Cerat Essentia or other models?

Been too long since I have heard Tenor to say how they compare. You mean the hybrid Tenor not the OTL Tenor amps, right?

 

[Zappadaddy]

Been too long since I have heard Tenor to say how they compare. You mean the hybrid Tenor not the OTL Tenor amps, right?

Right, since the hybrid Tenor is not what OTL used to be. If it was as great as you say it has to be deep in your head. Do you have same or similar emotinal experience with AC amps as with Tenor. That answer woudl be enough for me.

 

[morricab]

Right, since the hybrid Tenor is not what OTL used to be. If it was as great as you say it has to be deep in your head. Do you have same or similar emotinal experience with AC amps as with Tenor. That answer woudl be enough for me.

As I said, it’s been too long since I heard a Tenor to say but AC sounds wonderful.

 

[Zappadaddy]

As I said, it’s been too long since I heard a Tenor to say but AC sounds wonderful.

Thanks, I understand.

 

[JackD201]

Thats fairly unusual Jack , to which amplifier(s) are you referring and what tube type ?

842vhd. Vacuum failure. I suspect damage during transport. I should have removed all tubes first.

 

[Audiohertz2]

A correction/clarification on what I have written. I am not saying SS sounds worse than hollow state period, nor am I saying that those who enjoy SS and hybrid systems are enjoying them solely on how they measure, so please stop taking offence.

What I am trying to find out in my posting is what aspects of design are making amplifiers put out by Kondo, Shindo, fi, Morrison etc., sound so incredibly real (to me), whether or not they sacrifice frequency extremes and require horn speakers. And to put it in terms that engineer’s can appreciate but also so that lay persons like myself can understand.

I am also wondering if differences in people’s hearing and psychoacoustic perceptions are what is truly behind disagreements over solid state vs hollow state vs SET vs pure analogue tape to cutting head with or without digital, etc. etc.

The biggest difference is the imagination and spiritual connection.



Regards