• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

If "Tube Sound" Is a Myth, Why Tubes?

MakeMineVinyl

Addicted to Fun and Learning
Manufacturer
Joined
Jun 5, 2020
Messages
862
Likes
1,056
Location
Santa Fe, NM
Back in the 1950s and early 60s Stephens Electronics made an output transformerless power amplifier but it was meant to be used with their 500 ohm speakers.
 

dfuller

Senior Member
Joined
Apr 26, 2020
Messages
414
Likes
390
OTL tube amps do exist... for some reason... But they're obnoxiously huge, use hilarious numbers of tubes (the Atma-Sphere MA-3 uses no less than forty-two 6AS7Gs per channel!)... and barely approach the distortion performance of a decent Class AB solid state amp.
 

MakeMineVinyl

Addicted to Fun and Learning
Manufacturer
Joined
Jun 5, 2020
Messages
862
Likes
1,056
Location
Santa Fe, NM
OTL tube amps do exist... for some reason... But they're obnoxiously huge, use hilarious numbers of tubes (the Atma-Sphere MA-3 uses no less than forty-two 6AS7Gs per channel!)... and barely approach the distortion performance of a decent Class AB solid state amp.
But hey they were able to eliminate the output transformer - not that it matters with all the kludge circuitry they had to use to go about it. It makes me wonder how many output tubes would be required to make a real arc welder. :eek:
 

Robin L

Major Contributor
Joined
Sep 2, 2019
Messages
1,764
Likes
1,873
Location
1 mile east of Sleater Kinney Rd
OTL tube amps do exist... for some reason... But they're obnoxiously huge, use hilarious numbers of tubes (the Atma-Sphere MA-3 uses no less than forty-two 6AS7Gs per channel!)... and barely approach the distortion performance of a decent Class AB solid state amp.
Also not so huge headphone amps for electrostatic headphones.

Stax SRM-T1 Review (kenrockwell.com)

D3S_5315-front-0640.jpg
 
Joined
Jan 9, 2021
Messages
10
Likes
3
@mhardy6647 maybe I wasn't clear (again I'm here to learn). I meant this specific part:

"
Comparison
If an imaginary speaker had no internal Reactance and behaved as perfect 8R resistor across the frequency spectrum, there would be little auditory difference between a valve and solid-state amp. All cone speakers are highly Reactive and the Impedance varies over the frequency spectrum (approx 400%). Therefore, because Valve and solid-state amps operate in opposite ways to how power is delivered into a changing Impedance, the resultant audible difference is very noticeable. With most cone speakers, solid-state amps appear to sound flat and lifeless, the bass and high frequencies appear dead, as if a blanket is put over the speaker.


  • Valve amp power is directly proportional to speaker Impedance
    therefore power increases as the speaker Impedance rises.

  • Solid-state power is inversely proportional to speaker Impedance
    therefore power decreases as the speaker Impedance rises.
When listening to a valve amp, the bass and hi-frequencies appear to spring to life with clarity and detail, or may appear to sound exaggerated."
 
Joined
Jan 9, 2021
Messages
10
Likes
3
And this:

Ultra-linear. The Screen acts as a second Grid modulating the gain of the valve in opposite phase to the Anode. This return path acts as an automatic Negative feedback that removes most if not all residual distortion created by the output valves and output transformer.

Ultra-linear
This local feedback from the 50% primary taps to the Screen results in the speaker connected to the secondary being included in the UL feedback structure. The output Impedance from the secondary winding tends to reflect the speaker Impedance. A remarkable result is that this reflected Impedance tends to naturally maintain constant power to the speaker, over the frequency spectrum, providing the ideal damping that enables the speaker to give the most pleasing musical sound.
 

dfuller

Senior Member
Joined
Apr 26, 2020
Messages
414
Likes
390
If an imaginary speaker had no internal Reactance and behaved as perfect 8R resistor across the frequency spectrum, there would be little auditory difference between a valve and solid-state amp.
Accurate.
All cone speakers are highly Reactive and the Impedance varies over the frequency spectrum (approx 400%).
Not just cones - any speaker that uses a traditional voice coil does this.

Therefore, because Valve and solid-state amps operate in opposite ways to how power is delivered into a changing Impedance, the resultant audible difference is very noticeable.
If by this one means "solid state amps are far less affected by impedance dips because of their lower output impedance than tube amps" I would agree. Noticeable, that entirely depends on the design.

With most cone speakers, solid-state amps appear to sound flat and lifeless, the bass and high frequencies appear dead, as if a blanket is put over the speaker.
Absolute audiophool nonsense.

  • Valve amp power is directly proportional to speaker Impedance
    therefore power increases as the speaker Impedance rises.

  • Solid-state power is inversely proportional to speaker Impedance
    therefore power decreases as the speaker Impedance rises.
Tubes and transistors aren't honestly very different in a linear region amplifier (i.e., not class D) - transistors are lower voltage, lower impedance, higher current. This doesn't make any sense at all.

Ultra-linear. The Screen acts as a second Grid modulating the gain of the valve in opposite phase to the Anode. This return path acts as an automatic Negative feedback that removes most if not all residual distortion created by the output valves and output transformer.
UL uses the screen as really intense local NFB by taking advantage of some interesting phenomena related to screen and plate voltage swing, that's correct.

Ultra-linear
This local feedback from the 50% primary taps to the Screen results in the speaker connected to the secondary being included in the UL feedback structure.
Yes, that's because the impedance of the speaker is reflected across the output transformer. Nothing crazy here. By the way, the most common UL tap is about 43%.

A remarkable result is that this reflected Impedance tends to naturally maintain constant power to the speaker, over the frequency spectrum, providing the ideal damping that enables the speaker to give the most pleasing musical sound.
This is literal gibberish and can be disregarded.
 

dfuller

Senior Member
Joined
Apr 26, 2020
Messages
414
Likes
390
Also not so huge headphone amps for electrostatic headphones.
Electrostatics are an entirely different can of worms. They're super high impedance and as such don't need an impedance matching transformer.
 

Robin L

Major Contributor
Joined
Sep 2, 2019
Messages
1,764
Likes
1,873
Location
1 mile east of Sleater Kinney Rd
Electrostatics are an entirely different can of worms. They're super high impedance and as such don't need an impedance matching transformer.
So? It's still OTL. I own that particular model. As far as I can tell, the transformers are a greater part of the classic sound signature of vintage amplifiers than the tubes themselves. The Stax headphone amp is the best demonstration of that. Compared to Stax's TOPL solid state amp on offer at the time [1988], the Stax SRM T-1 was grainless, smooth but detailed on top. Much like the Topping L30 in that regard. Using the Stax energizer sans amplification, I hooked up the Marantz 8b. Smooth sounding top to bottom, but increasingly out of focus at the frequency extremes.
 

mhardy6647

Major Contributor
Joined
Dec 12, 2019
Messages
1,974
Likes
3,356
Back in the 1950s and early 60s Stephens Electronics made an output transformerless power amplifier but it was meant to be used with their 500 ohm speakers.
Yes, indeedy, thanks for throwing that into this goulash!
(in all seriousnesss) I was gonna go there next! :)
I got involved with some household chores -- and filling the birdfeeders.


1610223849008.png

source: http://www.hifilit.com/Stephens/Stephens.htm

Yes, that is the other way of doing OTL without quite as much messiness on the amplifier side of the equation.
Philips, also, made high-impedance speakers and an OTL to go with them (or, perhaps, vice versa).
 
Last edited:
Joined
Jan 9, 2021
Messages
10
Likes
3
So can the "magic of tubes" be attributed to amps having output transformers?

Also "If by this one means "solid state amps are far less affected by impedance dips because of their lower output impedance than tube amps" I would agree. Noticeable, that entirely depends on the design."

I think he meant quite the opposite. Valve amps being less sensitive to speaker's changing impedance.
 

mhardy6647

Major Contributor
Joined
Dec 12, 2019
Messages
1,974
Likes
3,356
Valve and solid-state amps operate in opposite ways to how power is delivered into a changing Impedance
Umm... no. No they don't.
The analogies between a class AB vacuum tube & solid state amplifier are pretty apparent by looking at two schematics of two relatively simple (power) amplifiers using the two flavors of active devices (tubes and transistors). The biggest difference is in coupling strategies, because an output tube is a high impedance device, and a transistor is a low impedance device. In practice tubes could be said to be "better at amplifying voltage" and transistors "better at amplifying current", but we in fact need power amplifiers and voltage, current, power and resistance (or, for AC circuits, such as those amplifying audio signals, impedance) all enter into the equation (literally).
 

mhardy6647

Major Contributor
Joined
Dec 12, 2019
Messages
1,974
Likes
3,356
Not just cones - any speaker that uses a traditional voice coil does this.
Electrostatics are an entirely different can of worms. They're super high impedance and as such don't need an impedance matching transformer.
And electrostatics look like big ol' (physically speaking) capacitors -- as does their characteristic impedance behavior, e.g., the classic Quad "ESL-57".
(the transformer is built into the loudspeaker -- at least in cases like the Quad)

1610224326531.png


http://quadrevisionspot.blogspot.com/2009/06/impedance-of-quad-esls-and-monoblocks.html

quadsnvinyl0609 (1)
by Mark Hardy, on Flickr
 

dfuller

Senior Member
Joined
Apr 26, 2020
Messages
414
Likes
390
I think he meant quite the opposite. Valve amps being less sensitive to speaker's changing impedance.
Then he's got it absolutely ass backwards. Nominal load impedance should ideally be at least 10x that of the source impedance for linear reproduction. Solid state amps have much lower Zout (generally in the ballpark of 0.1 ohms if not lower) than tube amps (generally around 1 ohm if not slightly less with feedback, quite a bit more without).
Yeah, I think so, too.
Which is part and parcel of what makes that article (or white paper or whatever) pretty darned hard to treat as anything but noise. :(
Pretty much.
 

Robin L

Major Contributor
Joined
Sep 2, 2019
Messages
1,764
Likes
1,873
Location
1 mile east of Sleater Kinney Rd
Yes, that's the point -- as long as the load impedance is (quite) high, a vacuum tube OTL amplifier becomes less of a Rube Goldberg contraption (viz. the aforementioned 500 ohm Stephens coax loudspeaker and its companion OTL "direct drive" amp).
Thus tubes & horn speakers.
 
Joined
Oct 23, 2020
Messages
61
Likes
46
Location
Melbourne, FL, USA
After they invited me to the front of the class a couple times for algebra and physics lessons, everyone in the class pretty much agreed their explanations were better suited to the audience because they were accessible, though admittedly fantastic.
”Do you want a correct answer or a simple answer?” :)
 

mhardy6647

Major Contributor
Joined
Dec 12, 2019
Messages
1,974
Likes
3,356
Frankly, @Rzezniq, a few of us have invested a fair amount of time, effort, keystrokes ;) and forum bandwidth in explaning why nonsense is nonsense -- not a good use of any form of resource, human or silicon.

Leave that "paper" (or whatever it is) behind and read some classic literature on audio reproduction then work your way back towards modernity (or not, as you wish). The "best" practicioners of the art will understand all of it well enough 1) to make intelligent and rational choices and 2) to explain them rationally, accurately, and clearly to an audience at any level of technical proficiency. (EDIT: The audience does have to be interested, though!)

As I like to tell my students when I teach 'em "old stuff" so that they have some insight into "new stuff" -- Picasso had to learn to "do" art (e.g., painting) "the right way" before he could detonate the whole field!

1610225799439.png

Then he's got it absolutely ass backwards. Nominal load impedance should ideally be at least 10x that of the source impedance for linear reproduction. Solid state amps have much lower Zout (generally in the ballpark of 0.1 ohms if not lower) than tube amps (generally around 1 ohm if not slightly less with feedback, quite a bit more without).
Indeed.
This is very fundamental and has about as much to do with the differences in (typical) vacuum tube vs. solid state audio amplification as anything.

Just to be really clear -- I think I understand this stuff reasonably well for someone whose interest in the subject is purely recreational. That said, I choose to do most of my listening via archaic amplification technology using single-ended (class "A") triode power amplification, transformer-coupled*, with no global (nor any externally applied local) negative feedback. The damping factor of the amp I use is about 8 (i.e., 1 ohm output impedance) and the power output is about 3.5 watts (per channel) at clipping. Is it hifi? Well, perhaps not, and it may be that what I like about it is euphonic distortion. But I like it, and I choose to listen to it.

DSC_0124
by Mark Hardy, on Flickr

(and it looks nice in the dark :) )


___________
* and the requirements of single-ended vs. push-pull output transformers is another issue entirely! ;) Suffice it to say that, in "normal" single-ended (SE) amplifier designs, the OPT has to deal with the DC plate voltage applied to the amplifier tube; thus the SE transformers have an "air gap" and have to be very large & heavy to avoid saturation by the DC voltage. These are weeds not worth getting into in this particular discussion, but it's of profound importance if one wants to build a decent single-ended vacuum tube power amplifier.
 
Last edited:

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,434
Likes
8,451
Equally - slightly off-topic but germane to our type of discussions - there were mixers and masterers who literally didn't know and didn't care what equipment they were using. I knew a multiple Grammy-winning engineer who couldn't have told you what monitors he worked on. He was staring at them 10 or 12 hours a day and the brand simply didn't register. That's how it was for non-superstar engineers, freelancers especially - it was like being a truck driver, showing up at the depot, getting handed a set of keys. You got what you got. The pickiness audiophiles show would have been regarded as weird.
That's fascinating! My experience with mixing and mastering engineers has generally been the opposite - many (but not all) I know have audiophile-like tendencies, espousing technically incoherent fantasies about how audio equipment, accoustics and psychoacoustics work.
 
Top Bottom