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

[Gregadd]

A manufacturer's personal preference tends to find its way into his products. Thus it is more relevant than that of individual
I will avoid the temptation to name the impressive list of solid state amps present in the high end. But for tube glow they have similar characteristics to tube amps. If you want glow you get it from meters and l.e.d.'s

 

[esldude]

A manufacturer's personal preference tends to find its way into his products. Thus it is more relevant than that of individual
I will avoid the temptation to name the impressive list of solid state amps present in the high end. But for tube glow they have similar characteristics to tube amps. If you want glow you get it from meters and l.e.d.'s

One preference of manufacturers is to sell more units, and make more money. As for a list of impressive ss amps in high end, what of it? It has no bearing on what I posted.

 

[mep]

'

Well see if we get past personality and preference whether someone is a tube guy or not is of no consequence. One is of superior fidelity to the input or they both equally are. We know they generally as manufactured aren't of equal fidelity.

Tim deParavicini has said he can get identical sound from tubes or SS. When asked why he used tubes he said that is where the market is. People will pay more for and pay more attention to tubes than an identical sounding SS unit. I suspect Bob Carver has found the same. The visual impression that you have something special in his Super Silver Seven is beyond that of a solid state unit. The heft, weight, all the glowing tubes is not something that seems as substantial in transistor form.

BTW, Mr. Carver worked with Mr. deParavicini on his current design. Mr. deParavicini desinged the input and driver stages, they collaborated on how do to the feedback while Mr. Carver handled the output stage and power supply.

First I heard of this and I find it very interesting.

 

[MylesBAstor]

First I heard of this and I find it very interesting.

I think it was written up in a Stereophile show report a while back.

 

[mep]

Those are two towering intellects/designers in both worlds of SS and tube audio design. Does anyone know who is winding the output transformers for Carver?

 

[treitz3]

The last I talked to Bob, he is doing them himself.

Tom

 

[andromedaaudio]

Ive tested it during the time I had the CAT preamp and krell evo 400s and later the Boulder 1060 with CAT , the CAT( Still have it, ultimate mk 2 ) being the tube pre with exxcelent bass control
During that time I also owned the integrated Zanden 300 B amp (I still have it ).
Although SS certainly has its attributes , high quality tubes gives you more air around the speakers , you get more of a certain live soundbubble .
This test is done on the same speaker which is a good test imo
On electronic music/and some movies i still miss the boulder , it has tremendous drive , much more as krell ,eventhough the evos had more watts/larger powersupply go figure

 

[andromedaaudio]

On my zanden no need for bias adjustments , and 300 B tubes last a very very long time , thus far 1,5 year and it being a 5 year old second hand product when I bought it ,I never had any troubles.
Only downside ;putting your ear next to the amp you will hear a very slight transformer hum , besides that its dead quit

 

[Mark (Basspig) Weiss]

Isn't it interesting that today Bob Carver makes tube gear? If he could produce the same sound with SS, why doesn't he now? It would be a lot cheaper. So one way or the other he's a charlatan? Can't have it both ways.

I'm going to go out on a limb here and, based on my tests on two of his VT amplifiers, state that he is catering to a market. It's not that his tube amps are 'better', but it's what he perceives the public wants. The problem is that the collective expertise of UTC, Thordarson, Stancor, etc, are interred with the engineers who designed the best audio output transformers in the mid 1950s to mid 1960s. Bob's attempts and the Chinese attempts at recreating audio output transformers have been abyssmal. But since most buyers of tube amps listen to acoustic music with no bass, the deficiencies are less likely to be noticed. The reality is, the amplifiers perform terribly.

I've only tested two very excellent performing tube amps. The McIntosh MC series (30 and 60 watt and 2 x 25 watt), and the Altec 1568A. Virtually every modern-day VT amp brought to my bench yielded lackluster performance. The Carver amps were especially bad. I've done a full workup on it at Amplifier Experts Facebook page, for anyone that's interested in the nuts and volts of it.

 

[Mark (Basspig) Weiss]

Mark, if you have ever had the chance to hear one of the smaller tube amps he recently designed.....you may be right. The one I heard below bested a VTL tube amplifier and not by a small margin either.

Tom

This was the first Carver tube amp that I tested. What I found when I pulled the cover off the transformer was like discovering that the Emperor was naked. Inside was a $17 transformer from Edcore, with a power rating of 10W and a frequency response of 70-15,000Hz, which matched up with the power bandwidth I measured. A far cry from "8-43KHz" that he claims. Measured power was only 12W, not the 20 he claims.

Needless to say, I was disheartened and disappointed, as I had always had the highest respect for Bob. His earlier work on SS products was nothing short of top end genius. What happened here, to cause him to produce such a poor performing amplifier, well, I can only assume rough financial times. But I think it's a major miscalculation on his part, as it will backfire and ruin his formerly good name.

I also tested his VT 180, as a favor to a fellow audiophile within driving distance, who'd read my test results on the VT20S. This is essentially a VOX AC15, in terms of transfer function and bandwidth. It's harmonic profile matches the British guitar amp nicely.

Oh Bob, how could you stoop so far?

 

[esldude]

I have been hunting the net for info on Bob's dc-restorer circuit but there are snips about, and he has come up with a way to idle the tubes at a lower level (thus less heat on the tube) and yet maintain full power level at that idle setting of the tubes...and thus the key is the tubes dont have to idle at such a high standing current to get nice low distortion from super small to super big input signals. I am not sure how he translates that into a 20 year tube life, but surely it helps "some" to idle at a lower current, but meanwhile, the cathode is still losing "electrons" all the time its on....but maybe not as many are generated since not as many are being attracted to the plate at low idle current(pure guess work here)....chemistry and metal mixtures on tube cathodes is not my strongpoint for sure.

To do what he did, and apply it to audio, shows his reasoning skills extraordinaire...my jaw dropped........

although dc-restoration is used in TV circuits, he figured something out another way, and that is genius. Brilliant job Bob!

Oh Bob is the genius I think. The article I read, some time ago, sorry couldn't find it again was from Dick Olsher.

Now Bob since doing the Stereophile challenge has always insisted the tube sound was mostly the high output impedance. In the article by Olsher he sort of spilled the rest of the beans. In 2012. He speaks of speakers being microphones. Even when playing they respond with back EMF from sound in the room. With high damping factor it gets swamped out. But with high output impedance and moderate or low feedback, that voltage EMF gets cycled back to the amp input by the feedback, and enhances overtones etc. ESL panels are especially good with tubes as they are better, cleaner microphones for this effect. He describes how he and deParavicini added an extra nested loop of feedback in his Super Silver Sevens. Not just global, but also to the driver stage. How he ended up thinking an output impedance of 1.7 ohms was optimum or close to it. I do believe Atmasphere has 1.75 ohms for what it is worth.

Reminds me of a friend who was digitizing his prized albums. Did so without speakers playing. He wasn't quite satisfied with the results. I told him his phono cartridge and tubed phono stage were getting acoustic feedback and he needed to record with speakers blaring. He doubted me. I had him place his stylus on a record with turntable not on. Record the output from his tube phono digitally while he played a CD. I then played the recording, and turned up though amongst some noise there was the music. He believed me and recorded LP's with speakers on after that point.

 

[Mark (Basspig) Weiss]

Mark,

Thanks. I am astonished by this result - it suggests that the engineers who designed the power supply did a poor job and the power supply was not well matched to the amplifier - or the tube rectifiers you were using had low cathode emission. Replacing tube rectifiers with diodes increases the voltage of the B+power supply under load, and as sometimes the drivers and input take their supplies from the same source this effect must be considered - typical tube rectifiers have an equivalent dynamic resistance around 50 - 100 ohms. Most probably it was a low feedback design, and the increase in distortion due to power supply modulation could not be compensated.

According to your measurements we can think that the amplifier was optimized for an higher power supply voltage than what it really used!

The engineering was typical of 1956 era. The rectifier was good. But the fact remains that plate resistance close to 100 ohms vs. the single digit resistance of solid state rectifiers has a very big influence on source impedance of the amplifier. The power supply is the first link in that chain. The output tubes next, and finally the transformer. (We'll exclude external things like power lines and interconnects to stay focused on the difference between power supply topologies). If any one of these links is weak, the overall performance will be affected. A low source impedance depends on a power supply with a low source impedance. Fifty ohms doesn't cut it. I would not have believed it myself, had I not been using distortion analyzers and spectrum analyzer to monitor the results of my work. It was unexpected. My intent was to increase the power output a bit. The 10:1 drop in THD was an unexpected bonus, as was the increase in damping factor. This is another reason why McIntosh tube amps sound so clean and tight. It's the solid state power supply.

I'm really excited about completing my OTL amplifier this spring, as all the remaining components arrive this week. I'm using 30A Schottky rectifiers and 12,000uF of capacitance in the power supplies. Each tube is capable of 1400mA plate current and there are a dozen of them that form the output stages. If the baseline design works well, my power supply topology allows me to double the B+ rails and I may get crazy and implement a Class H design using MOSFETs to modulate the B+ rails based on signal input demands, to get quadruple the continuous power output, which my design goal is 125WPC. A stiff power supply is absolutely essential to driving loudspeakers with OTL VT.

 

[MylesBAstor]

I'm going to go out on a limb here and, based on my tests on two of his VT amplifiers, state that he is catering to a market. It's not that his tube amps are 'better', but it's what he perceives the public wants. The problem is that the collective expertise of UTC, Thordarson, Stancor, etc, are interred with the engineers who designed the best audio output transformers in the mid 1950s to mid 1960s. Bob's attempts and the Chinese attempts at recreating audio output transformers have been abyssmal. But since most buyers of tube amps listen to acoustic music with no bass, the deficiencies are less likely to be noticed. The reality is, the amplifiers perform terribly.

I've only tested two very excellent performing tube amps. The McIntosh MC series (30 and 60 watt and 2 x 25 watt), and the Altec 1568A. Virtually every modern-day VT amp brought to my bench yielded lackluster performance. The Carver amps were especially bad. I've done a full workup on it at Amplifier Experts Facebook page, for anyone that's interested in the nuts and volts of it.

Have you measured the output transformers from *current* Jadis, EAR, cj, CAT or ARC tube amplifiers? A good friend measured the Jadis and said they are the best measuring OTs he's ever seen.

Also, I'd hate to say that Tim de Paravicini is slouch when it comes to designing and winding transformers.

 

[Tom B.]

I'm really excited about completing my OTL amplifier this spring, as all the remaining components arrive this week. I'm using 30A Schottky rectifiers and 12,000uF of capacitance in the power supplies. Each tube is capable of 1400mA plate current and there are a dozen of them that form the output stages. If the baseline design works well, my power supply topology allows me to double the B+ rails and I may get crazy and implement a Class H design using MOSFETs to modulate the B+ rails based on signal input demands, to get quadruple the continuous power output, which my design goal is 125WPC. A stiff power supply is absolutely essential to driving loudspeakers with OTL VT.

1400mA per tube? And you're only shooting for 125 watts/channel......

750mA for each output stage bank gets most OTL's close to 125 watts.

Tom

 

[microstrip]

I'm really excited about completing my OTL amplifier this spring, as all the remaining components arrive this week. I'm using 30A Schottky rectifiers and 12,000uF of capacitance in the power supplies. Each tube is capable of 1400mA plate current and there are a dozen of them that form the output stages. If the baseline design works well, my power supply topology allows me to double the B+ rails and I may get crazy and implement a Class H design using MOSFETs to modulate the B+ rails based on signal input demands, to get quadruple the continuous power output, which my design goal is 125WPC. A stiff power supply is absolutely essential to driving loudspeakers with OTL VT.

A good start - but each of the Atma-sphere MA2's has over 100,000 uF total capacitance. Can we know what output tubes are you using?

 

[Atmasphere]

Got any references for this?

Plenty, being that it is one of the fundamental reasons we have a SS/tubes debate.

Several years is a lifetime in that technology. I have a second generation POD. It's probably 60% of the real thing. FWIW, I don't buy your response by a long shot either. I was figuring we just had a slightly different experience until you said "don't have the guts." Then I knew we were talking about completely different products.

How did you record that modeling amp? Which amp was it? What kind of mic did you use?

Tim

We record with a variety of mics- for guitar amps could be a lowly SM-57 but could be a Neumann U-67 too. Seems to me in the case I remember we were using a condenser mic but not the U-67s. The amp itself was not memorable, but I know you can buy it at Guitar Center. Apparently, from your post above its an evolving art, with a get-what-you-pay for thing going on; in that regard really not a lot different from the rest of the guitar amp market.
.

I was extremely clear in my explanation. All you have is "you're wrong" which is not compelling. Please read what I wrote again, and maybe it will make sense to you. If it doesn't, please cite my specific points individually, and hopefully we'll come to an understanding about what is audible and what is not and why. Here it is again for your convenience:



Next you wrote:



No we can't, and this shows a fundamental misunderstanding of how we hear and what can be measured. Now, I'm always glad to change my mind, but you'll have to prove it by posting an audio example file or similar evidence. I'm also curious where you got that from, and why you believe it's true. Can you cite a legitimate scientific reference?

--Ethan

Ethan, you did not offer an explanation, just a statement. I did not just say you were wrong, I also stated why.

General Electric did most of the work on this in the 1960s. In their study, they found that people did not object to even ordered harmonic distortion so much, and would tolerate up to 30% without objection. This was not true with odd orders, where 0.1% was found to be quite intolerable. I have not found the results of this study online so far. However it is pretty obvious that the industry has for the most part chosen to ignore it.

I don't need to cite a study to show that the ear uses odd orders to process how loud a sound is- here is an easy test that anyone with simple test gear can show that this is the case. All you need is a VU meter, an amplifier, a speaker and signal generator, capable of sine and square waves. It works best if you have a low distortion sine oscillator rather than a generator.

Run a sine wave into the amp and speaker. Set it at 0VU on the meter. Now cover up the meter and run up the square wave from nothing until you have it at the same apparent volume as the sine wave. Uncover the meter. You will find that it is somewhere between -30 and -20 db depending on how clean your sine wave source is. Square waves are composed of odd ordered harmonics. When a signal that is -30 or -25 db can sound to our ears as if it has the same energy as a sound that is over 100x louder when its not, that should tell you something.

 

[esldude]

Plenty, being that it is one of the fundamental reasons we have a SS/tubes debate.


..

Okay, then how about giving us a couple references. Just for reference.

Now you mention tests showing odd order distortion audible around .1% which is a couple zeroes shy of your previous remark. .1% is about the threshold for distortion in testing done. Some of what I have seen shows our threshold is only that low around 2-5 khz where hearing is most sensitive. No quibble with that. Add some zeroes behind the decimal however, and that changes things.

 

[esldude]

I don't need to cite a study to show that the ear uses odd orders to process how loud a sound is- here is an easy test that anyone with simple test gear can show that this is the case. All you need is a VU meter, an amplifier, a speaker and signal generator, capable of sine and square waves. It works best if you have a low distortion sine oscillator rather than a generator.

Run a sine wave into the amp and speaker. Set it at 0VU on the meter. Now cover up the meter and run up the square wave from nothing until you have it at the same apparent volume as the sine wave. Uncover the meter. You will find that it is somewhere between -30 and -20 db depending on how clean your sine wave source is. Square waves are composed of odd ordered harmonics. When a signal that is -30 or -25 db can sound to our ears as if it has the same energy as a sound that is over 100x louder when its not, that should tell you something.

Just did the test for myself. Using headphones to isolate the two channels. Played sine in one ear and square in the other. They sounded equally loud around -20db for me. I also constructed a waveform using only even harmonics which decreased at -6 db per octave or in this case -6db per harmonic. The even harmonics sounded equally loud to a pure sine when it was -6 db.

Interesting little test.

 

[Atmasphere]

Just did the test for myself. Using headphones to isolate the two channels. Played sine in one ear and square in the other. They sounded equally loud around -20db for me. I also constructed a waveform using only even harmonics which decreased at -6 db per octave or in this case -6db per harmonic. The even harmonics sounded equally loud to a pure sine when it was -6 db.

Interesting little test.

Its not valid done the way you did it. You want both ears to process both sounds the same way. You could run the signal mono. What you describe here is something different.

The point here is that the human ear/brain system uses the higher ordered harmonics as loudness cues. I am quite surprised that I would get any pushback on that at all. But this is why one amp can sound bright while the other does not, even though on the bench they have the same bandwidth.

Now when the bright amp measures with only 0.001% THD and yet it still sounds bright, then its pretty obvious (if you are an engineer anyway) what is going on (if you also understand how the ear works). We see this phenomena going on all the time in high end audio and it has spurred millions of lines of text.

Norman Crowhurst did expound on this issue in his writings. I'm a big fan of his- and what I find odd is that so many of the things he wrote about 50 years ago seem to have been forgotten, enough so that we rehash them in these forums quite a lot. Pete Millet has scanned a lot of his books on his website and I recommend you download them and give them a read. http://www.tubebooks.org/technical_books_online.htm

 

[Ethan Winer]

Ethan, you did not offer an explanation, just a statement.

My explanation was extremely clear for why very low levels of distortion are not a problem regardless of their makeup! Here it is yet again:

What you miss is that the magnitude of distortion matters much more than its specific makeup. Nobody will hear 0.01 percent distortion whether it's mostly odd or mostly even, and even budget amplifiers often achieve distortion that low. Further, the notion that odd-order harmonics are more objectionable is simply false. It depends on the context, and how far away (in frequency) the added harmonics are compared to the fundamental due to the masking effect. A clarinet produces mostly odd overtones, so adding a 2nd harmonic will change its timbre more noticeably than adding a 3rd harmonic which is already present anyway. Further, even or odd THD is much less important than IM distortion, which is vastly more objectionable than any THD. Since THD and IMD are always present together, IMD is the most irritating type of distortion.

If this still doesn't explain why very small amounts of distortion are not irritating (or even audible) regardless of odd versus even, please cite my specific points individually, and hopefully we'll come to an understanding about what is audible and what is not and why. As esldude explained, what you're saying now is a far cry from the 0.001 percent figure you stated originally, which is what I responded to. Are you now acknowledging that your original statement was in error?

General Electric did most of the work on this in the 1960s. In their study, they found that people did not object to even ordered harmonic distortion so much, and would tolerate up to 30% without objection. This was not true with odd orders, where 0.1% was found to be quite intolerable.

Quite intolerable? I'm certain that's not the case. You must be remembering it wrong, or maybe you read someone else's take on the study in a hi-fi magazine. As Mark Weiss explained the other day, 0.1 percent is right at the threshold of what's audible.

here is an easy test that anyone with simple test gear can show that this is the case.

I can't see why matching volume levels between a sine wave and a square wave tells anything about the audibility of odd versus even distortion. Why not just test that directly? That's why I've done. Mix a pure sine wave fundamental with a pure sine wave at three times the frequency, and see at what level the added harmonic becomes "quite intolerable." I promise you it's a lot louder than -60 dB!

--Ethan