Why Tube Amps Sound Different (and better) Than SS Amps

[mep]

I must say mep, I don't get your fascination with this electron beam vs semi-conductor sourced electrons. The electrons move due to voltage differences they don't know where they come from. You think an electron that was sourced from a cloud in a vacuum rattles down the wire toward the speaker differently than one sourced from a semiconductor filling a hole in a crystal lattice? Maybe the one from the lattice of a semiconductor is more accurately herded versus one in a cloud bouncing around willy nilly if we must anthropomorphize the electron.

I must tell you that I never said I have a fascination with the electron beam vs. a semiconductor. I simply stated what many other people have said over many years when trying to speculate why tubes sound different than SS. I will have more to say to you later vis a vis your little experiment.

 

[Whatmore]

Surely all those things are measurable?

The measurements of the SF amps and the McIntosh MC-275 have been well documented in Stereophile. I trust JA when he makes measurements. As for the sound differences (if any) between a vacuum tube where the electrons are flowing in a stream/beam in a vacuum vice moving through a semiconductor, all bets are off. I don't know if that has ever been attempted and how it would possibly be measured.

I guess my question is: the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

 

[mep]

I guess my question is: the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

Because people invariably do just that?

 

[microstrip]

I guess my question is: the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

Whatmore,

I do not think that we can say the things that cause valves and SS to sound different seem to be very well understood and measurable. This is true for most older designs, where some easily perceived coloration was associated to technical limitations, but more recent tube equipment (and recent transistor equipment that measure excellently and are said to sound tube like) does not allow anymore the comfort of such statements.

The problem is similar to the one we have looking at the commonly shown measurements of two solid state amplifiers - most of the time we do not know how they sound just looking at the measurements.

But Mark has a good point remembering that electrons going through semiconductor lattices face very different torments from those traveling in vacuum subjected to electrical fields created by grids and electrodes . The problem is expressing these properties in electrical parameters that can be associated to sound quality.

 

[microstrip]

There is no problem Micro, except folks trying to muddy the waters with techno blabo.

The speaker has no idea that there was a tube stirring up electrons vs a solid state device. charges are a tangled bunch fo things that we take the aggregate of them when we use them in standard electronics.

All differences between tubes and solid state are measurable Whatmore, despite wild claims to the otherwise.

Audio amplification is SIMPLE electronics.

Tom,

You fail to understand the different between raw data and analyzed data, and correlating data with sound quality.

And yes, audio electronics is SIMPLE. I even build a galena radio when I was eight years old!

 

[esldude]

I guess my question is: the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

Because people like to believe in magic.

 

[microstrip]

I guess my question is: the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

Because audio experts some of us trust, like Prof. Keith Johnson, of Spectral, and Gérard Perrot, of Lavardin, and several others who designed excellent sounding amplifiers and wrote articles saying so. And yes, things are not so simple as many people want them to be.

 

[microstrip]

Teach me, and the rest of us what you mean by this statement! Perhaps its your failure not mine...explain!

Tom,

Sorry, we went through this some months ago in a debate about measurements - I am not wanting to rewind and start it all again.

 

[microstrip]

Because people like to believe in magic.

Surely. All magic has a price, and audiophiles love spending!

 

[Ethan Winer]

I have been surprised by people who learn of this method and yet aren't convinced at all by it. Usually the reaction is to double down on faith.

Yes, this surprises me too because it is proof positive. But faith is strong, and it's the right word. It took several hundred years after Galileo's telescopes proved the earth is round before the church "forgave" him for his blasphemy.

--Ethan

 

[Ethan Winer]

the things that cause valves and SS to sound different seem to be very well understood and measurable. So why do we need to resort to conjecture about sound differences caused by the medium in which electrons flow?

Because people like to believe in magic.

Yes, and Yes again. This is why TV shows about alien abductions and Bigfoot sightings are so popular.

--Ethan

 

[microstrip]

(...) But faith is strong, and it's the right word. It took several hundred years after Galileo's telescopes proved the earth is round before the church "forgave" him for his blasphemy.

--Ethan

A common misconception. Galileo was never accused of blasphemy, but considered "suspect of heresy". If he had been considered heretic, he would have been convicted to death and burnt. In fact, the debate was not a scientific or religious trial, but mainly political. Most educated clergyman already had accepted the heliocentric model by this time, what was being considered was the effect of admitting that the Holy Scripture told otherwise. In order to solve the problem the heliocentric model should however be considered as a mathematical model, good for calculations but not describing reality.

It is why the sentence was very soft by Inquisition standards - he was put in house arrest at the luxury house of the Archbishop of Sienna, who protected him, and later in his own house in Florence, where we went on working until his death, publishing the well known book "Two New Sciences".

At this period of history, power and religion were very interconnected and it is not easy to separate them. IMHO Galileo is a genius, and can not be mixed in any way with our entertaining audio convictions and debates.

 

[esldude]

At this period of history, power and religion were very interconnected and it is not easy to separate them. IMHO Galileo is a genius, and can not be mixed in any way with our entertaining audio convictions and debates.

I agree with your last two sentences at least.

 

[Al M.]

A common misconception. Galileo was never accused of blasphemy, but considered "suspect of heresy". If he had been considered heretic, he would have been convicted to death and burnt. In fact, the debate was not a scientific or religious trial, but mainly political. Most educated clergyman already had accepted the heliocentric model by this time, what was being considered was the effect of admitting that the Holy Scripture told otherwise. In order to solve the problem the heliocentric model should however be considered as a mathematical model, good for calculations but not describing reality.

It is why the sentence was very soft by Inquisition standards - he was put in house arrest at the luxury house of the Archbishop of Sienna, who protected him, and later in his own house in Florence, where we went on working until his death, publishing the well known book "Two New Sciences".

At this period of history, power and religion were very interconnected and it is not easy to separate them. IMHO Galileo is a genius, and can not be mixed in any way with our entertaining audio convictions and debates.

Nice to see you reply to the commonly spread misconceptions. It was indeed mainly a political affair. Also, Galileo didn't prove anything:

http://www.catholic.com/tracts/the-galileo-controversy

The real scientific evidence came later.

 

[MylesBAstor]

Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.

Albert Einstein

 

[mep]

Esldude-I think the story you told merits discussion because maybe it sort of gets at the core point I was really trying to make and that was tube amps sound better (to me) from 40Hz up. With regards to the comparison between your VTL amps and your Spectral amp and how much better your VTL amps sounded, your original comment was:

It had more space, 3d effects, smoother, more dynamic, more nuanced, more musical. The Spectral was a good sounding SS amp, but not equal to the VTL. The Spectral was subjectively about 2/3 as good as the VTL.

The above comments you made are very powerful if that is what you truly believed before you ran your little test. Some people are spending over 5 figures in order to chase subjective single digit percentage points of improvement in the sound of their system (the law of diminishing returns). Using the subjective WAG system, you calculated that your VTL amps represented a 33 1/3% improvement over the sound of the Spectral. That’s an incredible improvement in sound quality. That’s huge. Monstrous even. All of the audiophile buzzwords that audiophile music lovers crave when they try to improve their systems are what you described for the differences between your VTL amps and your Spectral. More space, 3D effects, smoother, more dynamic, more nuanced, more musical. People are paying serious money trying to create those types of improvements in their system.

And then one day something told you that you to run an experiment. Your conclusion from the experiment was that the VTL was telling you sweet lies and the Spectral was a truth teller that only sounded 2/3 as good as the lying VTL amps. Guess what? If I had an amp (or amps) that sounded 33 1/3 better than my reference amp in the ways that you described and I could afford it, my reference amp would be up on Audiogon waiting for a new home with someone who liked specs better than sound quality.

I don’t know how you have your ears/brain reconcile/rationalize the fact that you left 33 1/3 of the improvements in sound quality you had on the table but somehow now you are happier and enjoying your music more because your experiment told you the Spectral was more accurate. That sounds crazy to me.

And here’s a little clue for you all besides the fact that the Walrus was Paul: electronics that are judged to be “accurate” or “neutral” based on their fantastic specifications whether they are printed or measured can also sound quite boring. And based on the descriptors that you used to describe the differences between the sound of your VTLs and the sound of your Spectral, your Spectral sounds as boring as waiting for customers at a hamburger stand in the middle of a Hindu festival. I’ll take the amps that you judged to be far superior to the Spectral any day and forget the little resistor test.

 

[Atmasphere]

From what I know, it all comes down to the harmonic signature of the 2 formats
Tubes = even order harmonics
SS= odd order

Some are more a tuned to the harmonics than others..

A lot depends on topology. If the circuit is fully balanced and differential, even orders will be absent, even if tube. You can tailor the distortion created by the circuit by mixing single-ended and balanced circuits. This fact should be included in the above statement; otherwise it is oversimplified and not entirely accurate.

What I believe the clear sound difference can be attributed to is based on the following;

1) A "hollow state" device can switch high voltages 'faster' than a tube can switch high currents. I know this was true in the '90s as I worked on a missile that had a transmitter (active homing) and we originally had a solid state radar frequency amplifier - this tech did not work and we went with tubes. Yes. This missile is called an AMRAAM. Go look it up. Apparently transistors (of some stripe I am not sure if bipolar of FET) have been sufficiently advanced that modern radars use transistors exclusively - but then again most modern military radars are phase-arrays so the power requirements for each transmission element is rather low - or at least much lower than in single aperture radars.

So the increased switching speed available with a tube device may be a contributor to the different sound. Quick - not unlike live music. All that current sloshing around within a bipolar transistor tends to make the device slow (which maybe why our genius solid state guys - like Nelson Pass - do NOT use bipolar silicon but instead work with FETs (much higher voltage across the device).

2) The biggest difference is that huge chunk of iron sitting on the output of your tube amplifier - that transformer which translates those high voltages into high currents for use by the speaker motors. Transformers, like it or not, ring like a bell. That is to say when a transformer is hit with a change in applied voltage it actually works to induce a voltage of the opposite polarity - di/dt - fact of life folks. Transformers resist, in a very active and known sense AC changes.

Not only does the transformer create a very non-linear response to changing voltages a transformer is NOT linear in its time response to a changing voltage. To tune a transformer "up" (increase voltage) proceeds much more quickly than to tune a transformer "down" (lower the voltage). I would estimate by a factor of 3 dB (double the time).

This non-linear response of a transformer, coupled with the massive EMF "kickback" associated with large speaker motor means that most tube amps produce a very pleasant distortion that closely resembles the sound of "slap echo" that is heard in any large concert hall (particularly in my local concert hall - the Meyerson in Dallas - very pronounced slap echo, i.e. you can hear the echo from the walls of the hall - slight, and very close in time to the incident signal but nevertheless very real as it is outside the Haas effect window). The EMF kickback - as it crosses the coils of the transformer AND the non-linear response of the output transformer to a changing voltage means that the sound tends to be more than a bit "ringy" - which is EXACTLY the sound of a real un-amplified instrument played in a real concert size hall.

Hence the "tube heads" think they are getting "better" sound - but what they are actually getting is a "ring" that is nothing more than a distortion. Is this ringy sound pleasant? Of course, especially if you are playing symphonic music through some high quality speakers. More so if your speakers are somewhat "lean and thin" sounding - particularly like a Martin-Logan electrostat (or ANY electrostat save a big SoundLab). The tubes "put some flesh" on the sound - but what those precious tubes are doing is ringing like a church bell at a drunken wedding.

So we are back at the fact that the attraction of a tube amplifier is about its woeful non-linearity and the complete inability of a coupling transformer to respond in a linear matter to changing AC voltages.

Sorry guys - if you like tubes, their heat, their lack of power, their utter lack of reliability - then I say knock yourself out. But don't come around here and try to tell me that "tubes" are better.

If you want a pleasant distortion - go get one of those QoL thingies.

Proceed to nuke me.

Having been building OTLs now for over 38 years, I can say without reservation that your statements above are not correct- they are an impression. First, tubes can switch as fast or faster than transistors and this fact is well-known and has been for a very long time. It is why Eimac is still in business, and why the military still works with cold-cathode topologies. Switching is of course a function of bandwidth; we've built OTLs that could function easily as RF power amplifiers at 30MHz, and yet entirely stable doing so. You simply cannot do that with a transistor amp in this day and age.

The second notion is of course obviated by the fact of OTLs. OTLs do it without a transformer and so do lack the non-linearities you mention. Yet they can still come off smoother sounding than transistor amps; imagine one that can do that with a risetime of 600V/usec...

The sound difference IOW is neither of these things. One really important reason has a lot to do with human hearing. If you don't understand human hearing rules, you will not be able to build a decent sounding amp. You will only be able to build one that looks good on paper, which, it turns out, is not that important. Humans use higher ordered and particularly odd ordered harmonics to sort out how loud sounds are. Because of this, trace amounts of distortion that solid state amps have that tube amps don't cause transistor amps to sound brighter and harsher than the actual signal really is. This distortion issue is why two amps can measure the same on the bench, but one will be bright and the other will not. Its *all* about distortion, and tipping points in the brain.

Myles; hey - tubes are not as reliable as transistors. Think about a chunk of ARC gear where the tubes are "ganged" in pairs and when they fail - they take out both themselves and a local resistor. Tubes are mostly junk. While you may very well have a Phd - you do NOT possess a PhD in electronics. Wishful thinking does not a fact make.

In regards 'stats; sorry dude - to these ears they have always sounded "lean and thin" hence the need for some ringing tubes to flesh them out. With respect to OTL - they don't have the long sustains and generally euphoric presentation of transformer coupled tubes. Though, in all honesty, they do have that tube sound to a lesser degree - as to why? You have me on that one.

There is confusion in the above statements about execution which causes them to be false. Not all tubes fail. Not all amps fail when a tube fails. Stuff like that.

Again I ask:

it's up to you to explain specifically how and why one I times E ratio "sounds different" than another.



Of course, and no matter what impedance is stated, it varies with frequency. I assumed we all know that by "8 ohms" I meant "an 8 ohm loudspeaker."



Since you're the one putting forth a theory that defies well known properties of electricity, the burden is on you to explain why it won't sound the same. Or that the differences defy what is already understood such as frequency response and noise.

--Ethan

You are right in that if 100 watts is going into 8 ohms, the current and voltage will be the same regardless of the amp that made the power. But that is where it ends. One thing that is not being discussed here is the impact of output impedance. In short its poorly understood. For example, its easy to overdamp a speaker and then it won't sound right. For another example, **no speaker made** needs more than 20:1 in terms of damping factor. Many need quite a bit less, some are overdamped if the damping is 1:1.

You might want to look at this link http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

I suppose I brought up colorations. At least for my post you know the speakers, and the equipment by model number. Specific enough?

You have the audiophile idea of transparency that a device can pass on the signal without altering or coloring it. The Spectral transparently passed on the sound of the tube amp. The tube amp couldn't pass on the sound of the Spectral because it wasn't transparent. Simple enough even a PhD might understand, unless he doesn't wish to do so. Again, good SS amps are more accurate nothing wrong with saying you prefer tube sound, just don't present it as superior rather than your preference.

The VTL is one of hundreds of amps; IMO not by any means the most transparent. There is a baby with the bath problem with your argument.

I was using Quad ESL63 speakers, amps were VTL 75/75 with some upgraded parts connected in triode, and a Spectral DMA50. Has been a few years, but I believe I was using a Meridian DAC at the time. Only one of the resistors to do the voltage division were in the direct signal path. As I recall it was a pairing of Vishay and Holco to get unity gain from resistors I had on hand. Interconnect was some homemade silver and teflon.

You can read post #13. To recap, series connected a VTL loaded with a power resistor and the output divided to achieve unity gain feeding a DMA50. The only thing extra in the direct signal path was a Vishay resistor, and silver interconnect. Sound was that of a VTL which at the time I thought superior to and a better higher fidelity amp of the two. Reversed positions and heard no change. Could insert or remove the Spectral and not hear any indication it was in circuit or out. Sounds like a definition of straight wire with gain or complete audible transparency to me.

Would be nice if some audio publication such as yours did an article about such things. I have suggested before that someone with a love for SET's but power hungry speakers could use the procedure to feed an SET which fed a good SS amp to have their cake and eat it too. Do you think the audiophile public would be interested in such a thing?

How did you load each amp? Did you do it to idealize either one? You make no mention of the load resistance but its pretty important- otherwise you may not have done this in a scientific manner.

 

[DonH50]

FWIWFM, there is a thread in the technical subforum showing the response of simulated amp and speaker loads. It includes various output impedances to mimic SS and tube, and includes rising impedance with frequency as is typical of both amps but not transformer nonlinearity. I added higher tube Zout and nastier speaker loads in a post on the second page.

For the record my earlier post specifically excluded OTL's, largely because of their fundamentally different output stage. When I win the lottery I'd love to try a pair of Atma-Sphere monoblocks... It has been a long, long time since I listened to Futterman or my DIY OTL designs.

Perhaps the best sound I ever had in my system, though perhaps not the most technically accurate, was using a tube preamp driving an active crossover to bi-amp my Maggies with a tube mid/upper amp and hybrid (Counterpoint, tube input/MOSFET output) bass amp. The Magnepans are a pretty benign load and I always thought they worked well with tube amps, though bass was better with SS at least IMO.

 

[puroagave]

Given that I've hear both tubes and solid state sound both harsh and smooth, flat and spacious, sloppy and controlled, I'm inclined to believe that it is the combination of these two things:

1. design and execution

2. how a particular amplifier will perform matched with any specific speaker

what jack said. it so happens I have a spectral amp in the system driving an ESL and i'm staying put for now. I never thought id have SS amp driving the CLSs let alone spectral but the synergy is undeniable and I have new found admiration for spectral as a whole (great customer service, btw). I'm still interested in trying more tube amps with the CLSs but im in no rush:b

 

[mep]

what jack said. it so happens I have a spectral amp in the system driving an ESL and i'm staying put for now. I never thought id have SS amp driving the CLSs let alone spectral but the synergy is undeniable and I have new found admiration for spectral as a whole (great customer service, btw). I'm still interested in trying more tube amps with the CLSs but im in no rush:b

The ironic thing is that I started this thread and the only tubes that I currently own that are in my system is my pair of Ampex 350s. I swore off tubes and sold all of my tube gear except for my Ampex 350s. While my Krell KSA-250 was making round trips back to the factory, I did buy an ARC VS115 tube amp to tide me over and I know exactly how long that has been since it was in my system. And that is due to the fact that when I shipped the ARC to its new owner after my KSA-250 came back, I ended up rupturing the bicep tendon in my right arm. That happened in September of 2013. Two weeks ago, a pair of tube amps and a tube preamp showed up on my door for review. So basically this was the first time I have had tube amps back in my system for 16 months. I'm not letting the cat out of the bag yet with regards to what amps I'm listening to, but I would be lying if I said they didn't have me rethinking owning tube amps again.