What, measurement wise, is "tube sound?"

[MakeMineVinyl]

One difference in single ended triode amps with no feedback (as most of them are) is that it is very hard if not impossible to drive them into hard clipping. The waveform just gets rounder. When driven hard with less than extremely efficient speakers, this mostly even order distortion would be well audible and could be pleasant depending on the listener.

 

[Wes]

I'm not sure whether I've ever heard anyone say exactly that. Perhaps, "On average, men are taller than women." More to the point, your example is concerned with a particular and exact attribute, height. Perhaps it would make sense to speak of tube sound if there were a fuzzy blending with overlap, with respect to some very specific sonic attribute, but only if the sonic attribute is very specific and very well defined. And only if it has been properly established that this very particular sonic attribute is in fact the attribute by which people distinguish the sound of tube amplifiers from the sound of solid state amplifiers.

No exact attribute is required. That is why I specifically mentioned principal components analysis (which is used to ordinate multi-variate factors).

 

[SIY]

Just anecdotally, I've thought the sound of push pull tube amps is in the transformers mostly. I've had a few tube pre's back when people had the idea you could get much of tube sound by using a tube pre into a ss power amp. I didn't find it to be the case. Good tube preamps didn't really have a sound. Most put out more voltage than necessary, and sometimes adjusting gain was a good step, but they were not noticeably different than good clean SS or FET preamps.

Now I've done some careful comparisons of tube power amps, but all sighted. Discount that as you see fit, but I found Ultralinear tube amps to have a similar and non-transparent sound, triode push-pulls had a sound (which was my favorite) and SET's have a sound (though these varied more than the the others). It was most apparent when I did series tests where I put a power amp with loaded output between the source and the power amps driving the speakers. I could then switch the power amp under test in and out of the circuit. I found there were FR differences, but also they seemed to create a 3d space, good ones seemed smoother, and usually seemed to punch up subjective dynamics. The FR differences were real and easily measured. The creation of space, smoothness and dynamics not so easily measured. Other gear with transformers often seems to have a smoothness, and ability to add some spaciousness to sound. Not always does it add the dynamics, so maybe more than just the transformers did that or maybe at power amp levels where saturation can be an issue it creates false dynamics. I've thought a simple push-pull triode preamp using 12ax7s or 12at7's which fed a small output transformer might be a good thing to do. Basically a mini push pull triode amp to pre-color the signal fed onto your power amps. Tim's E.A.R. preamps used transformer coupling at the output.

What I found is that good SS amps can fully and faithfully re-create all that spacious, dynamic and smooth sound of a triode power amp. Because if you put a triode amp in front of the SS amp you get the full triode sound.

But none of this answers definitively the OP's question.

Interestingly, the one person here who actually tried a blind level-matched comparison of tube vs solid state used Dynaco, which had an Ultralinear output stage topology.

 

[Blumlein 88]

Interestingly, the one person here who actually tried a blind level-matched comparison of tube vs solid state used Dynaco, which had an Ultralinear output stage topology.

It is my opinion that the Ultralinear was the least colored in my listening.

 

[Newman]

I seem to recall an interview where Tim said he could get the sound he wanted with SS or tube.

I remember that too. I think it was in HFNRR.

 

[Jim Shaw]

Just anecdotally, I've thought the sound of push pull tube amps is in the transformers mostly. I've had a few tube pre's back when people had the idea you could get much of tube sound by using a tube pre into a ss power amp. I didn't find it to be the case. Good tube preamps didn't really have a sound. Most put out more voltage than necessary, and sometimes adjusting gain was a good step, but they were not noticeably different than good clean SS or FET preamps.

Now I've done some careful comparisons of tube power amps, but all sighted. Discount that as you see fit, but I found Ultralinear tube amps to have a similar and non-transparent sound, triode push-pulls had a sound (which was my favorite) and SET's have a sound (though these varied more than the the others). It was most apparent when I did series tests where I put a power amp with loaded output between the source and the power amps driving the speakers. I could then switch the power amp under test in and out of the circuit. I found there were FR differences, but also they seemed to create a 3d space, good ones seemed smoother, and usually seemed to punch up subjective dynamics. The FR differences were real and easily measured. The creation of space, smoothness and dynamics not so easily measured. Other gear with transformers often seems to have a smoothness, and ability to add some spaciousness to sound. Not always does it add the dynamics, so maybe more than just the transformers did that or maybe at power amp levels where saturation can be an issue it creates false dynamics. I've thought a simple push-pull triode preamp using 12ax7s or 12at7's which fed a small output transformer might be a good thing to do. Basically a mini push pull triode amp to pre-color the signal fed onto your power amps. Tim's E.A.R. preamps used transformer coupling at the output.

What I found is that good SS amps can fully and faithfully re-create all that spacious, dynamic and smooth sound of a triode power amp. Because if you put a triode amp in front of the SS amp you get the full triode sound.

But none of this answers definitively the OP's question.

Thanks for your reply. Another factor of transformer output coupling is expected to be hysteresis of the iron/steel. This should be more apparent at low signal levels. -JS

 

[b4nt]

Tube sound or transformer sound? I've read this last days:

The F6 is a Class-A MOSFET power amplifier, with a JFET driver stage. It utilizes a small-signal transformer for phase inversion, so both the output MOSFET are N-channel. This transformer greatly contributes to the overall sonic character of the amp which is very highly regarded and considered to be similar and having all the benefits and positive characteristics of a tube amp... making some observers speculate that much of the positive 'tube amp' sound is the transformers, not the tubes. The F6 is a regular kind of amp - a voltage source amplifier, requiring nothing out of the ordinary from the preamp or the speakers. https://diyaudiostore.com/products/f6-board-transformer-kit

 

[MakeMineVinyl]

Tube sound or transformer sound? I've read this last days:

The F6 is a Class-A MOSFET power amplifier, with a JFET driver stage. It utilizes a small-signal transformer for phase inversion, so both the output MOSFET are N-channel. This transformer greatly contributes to the overall sonic character of the amp which is very highly regarded and considered to be similar and having all the benefits and positive characteristics of a tube amp... making some observers speculate that much of the positive 'tube amp' sound is the transformers, not the tubes. The F6 is a regular kind of amp - a voltage source amplifier, requiring nothing out of the ordinary from the preamp or the speakers. https://diyaudiostore.com/products/f6-board-transformer-kit

A lower level interstage transformer is a whole different beast than an output transformer. The transformer used in this phase inversion is going to typically have lower distortion than an output transformer, and it doesn't have to deal with the varying impedance of a loudspeaker load.

 

[b4nt]

A lower level interstage transformer is a whole different beast than an output transformer. The transformer used in this phase inversion is going to typically have lower distortion than an output transformer, and it doesn't have to deal with the varying impedance of a loudspeaker load.

There a more notes and comments on the internet about that amp. When it was created and compared, it was classified close to tube amps for its sound, not with the solid state group.

That F6 is a particular amp, Nelson Pass is designing low distortion A class amps.

 

[Wes]

When Bob Carver aced that "challenge" - was it against a tube amp? I forgit, but f so then he obviously figured out what tube sound meant.

 

[DonH56]

When Bob Carver aced that "challenge" - was it against a tube amp? I forgit, but f so then he obviously figured out what tube sound meant.

Yes, a CJ (Conrad Johnson) amp. He said later the biggest factor was the high output impedance, though there were other factors, and added a switch to a line of amplifiers that essentially added a resistor in series with the output to create "tube sound".

 

[levimax]

Yes, a CJ (Conrad Johnson) amp. He said later the biggest factor was the high output impedance, though there were other factors, and added a switch to a line of amplifiers that essentially added a resistor in series with the output to create "tube sound".

I did a blind level matched test between a Neurochrome Mod86 and Dynaco ST 70 and could tell no difference. While I had everything set up and level matched I did a frequency sweep of the two amps and then added in a 0.9 Ohm resistor to speaker cable of the SS Neurochrome amp to simulate the output impedance of the Dynaco tube amp and it did indeed change the FR to almost perfectly match the Dynaco. Even without the resistor I could not hear the very subtle FR differences between the amps although maybe someone could. If using a SET tube amp with no feedback and 5 ohms of output impedance the FR differences would be much bigger but for most tube amps designed to be accurate and low distortion (including 60 year old ones) they really don't sound different than a modern SS amp.

 

[Wes]

Yes, a CJ (Conrad Johnson) amp. He said later the biggest factor was the high output impedance, though there were other factors, and added a switch to a line of amplifiers that essentially added a resistor in series with the output to create "tube sound".

Thx

I don't see how he could have added the harmonic distortion spectra to his amp in the alloted time... so that would mitigate against that factor as a significant contributor to 'tube sound' - at least for the CJ amp comparison.

 

[levimax]

Thx

I don't see how he could have added the harmonic distortion spectra to his amp in the alloted time... so that would mitigate against that factor as a significant contributor to 'tube sound' - at least for the CJ amp comparison.

For a well designed tube amp THD is less than -60 dB which for most people is completely inaudible when playing music. After I tried a blind level matched test for myself I have concluded Tube Sound is a myth at least for me. I would recommend people try it for themselves .... well worth the trouble I think.

 

[Wes]

I'm not comfy with the claim that -60 dB is generally inaudible for all music signals in combination.

The tests for all combinations in the factor space are simply not comprehensive.

 

[DonH56]

Thx

I don't see how he could have added the harmonic distortion spectra to his amp in the alloted time... so that would mitigate against that factor as a significant contributor to 'tube sound' - at least for the CJ amp comparison.

Pretty sure he did; it is not that hard to add controlled harmonics to a signal with a simple analog circuit.

Whenever I have created files with -60 dB (0.1%) added distortion (2nd and 3rd) to music samples few if any were able to pick them out in a listening test. Added to a single sine wave, or even a two-tone test, there were a few who could distinguish the file with the added distortion.

 

[Blumlein 88]

Thx
I don't see how he could have added the harmonic distortion spectra to his amp in the alloted time... so that would mitigate against that factor as a significant contributor to 'tube sound' - at least for the CJ amp comparison.

Don't sell Bob Carver short as he was a clever guy. I think maybe he could add distortion. I think maybe he did. I think he possibly altered some local feedback amounts in some parts of the circuit. But even he said the main factor was the output impedance. I do recall the final piece to the puzzle which caused the Stereophile people to throw in the towel and declare them identical sounding. Carver said the C-J power supply wasn't up to heavy signals below 100 hz, and he had to alter his amps power supply so it was similarly limited and this was the final thing he did. I don't know how he did that. I do remember he talked about winding coils maybe he put coils in parts to alter the current his amp could supply at high levels of low frequencies.

 

[Blumlein 88]

When Bob Carver aced that "challenge" - was it against a tube amp? I forgit, but f so then he obviously figured out what tube sound meant.

It was the Conrad-Johnson Premier Four power amp.

https://www.stereophile.com/content/carver-challenge

After the Challenge Bob Carver created his own tube power amp. The Silver Seven T.
https://www.stereophile.com/content/carver-silver-seven-t-monoblock-power-amplifier

 

[Blumlein 88]

I'm not comfy with the claim that -60 dB is generally inaudible for all music signals in combination.

The tests for all combinations in the factor space are simply not comprehensive.

I'm pretty comfortable with it. I'm, very comfortable with the idea if the error signal is -60 db it is inaudible 90% of the time with very nearly all music. I'm not so comfortable with the idea tube amps can maintain - 60 db into real world loudspeaker loads with their complex reactive impedance.

Just as a for instance most CJ amps would list frequency response as pretty flat 20 hz to 20 khz (but at low signal level). While full power specified some level of THD at full output for 30hz to 15 khz. This would have been with a resistive load. With a reactive load my guess is distortion is somewhat worse.

 

[Blumlein 88]

Stereophile didn't measure the Premier Four, but many years later did measure the Premier Twelve. Here are the spec claimed by C-J.

Sidebar 1: Specifications
Description: Vacuum-tube monoblock amplifier. Tube complement: one 5731, two 6FQ7, four 6550. Maximum power: 140W into 4 ohms (18.5dBW), 30Hz-15kHz, at no more than 1% THD or IMD. Frequency response at 10W: 20Hz-20kHz, ±1dB. Hum and noise: 98dB below rated power. Sensitivity: 900mV to rated power. Input impedance: 100k ohms.
Dimensions: 17.5" W by 7.25" H by 15.25" D. Weight: 51 lbs.
Serial numbers of units reviewed: 6101307, 6101308.
Price: $6990/pair. Approximate number of dealers: 70.

See 30-15 khz at 1% thd. That is only -40 db, not -60 db.

Now the amp did pretty well meet these specs in Stereophile's tests of it. It also had about .9 ohm output impedance which alters the response with their simulated load enough it likely would be heard.

https://www.stereophile.com/content/conrad-johnson-premier-twelve-monoblock-amplifier-measurements