SE distortion

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Because no distortions born equal. If you in order to remove distortions that are measurable but less audible add distortions that are more audible you actually add more distortions, no matter what your measurements suggest. Also, when distortions go down when sounds decay it is natural and much less audible than when distortions go up when sounds decay. Strictly speaking, you want cleaner sound, not cleaner picture on the screen, because this pictures not so well corellate.

Why it so? Because everything in the Nature that conducts, reflects, generate sounds, is non-linear. But sounds that you hear in the nature do not sound distorted. But electronic additions different from what your ears and brain expect to hear draw attention and sound as distortions, while distortions of SE amp are almost ignored by ears and brain.
 
It is not like black and white. Any distortions are distortions.

I am just saying that some distortions sound more as distortions than other distortions. Like, 1 grain of salt and 1 grain of cyanide in a glass of water tastes differently, even if 1 grain of salt is easier measured by sensitive ohmmeter. Or, a glass cleaned by acetone looks cleaner than cleaned by water, but in which glass a water tastes better? The same with electronics: means to remove "visible" distortions often add distortions that are more audible, even though they are less visible. Second tube in push-pull adds audible distortions, but decreases visible distortions.

Push pull is very good approach when you want more output power loosing less for heating the air. Like, I can easily get 100W of clean power from a pair of GU-50 tubes. To get such power from tubes in SE will be quite problematic.
 
Last edited:
Administrator
Joined 2004
Paid Member
Is it mostly 2nd order?
It doesn't have to be. A good design will have a harmonic spectrum that falls off evenly. 2nd, 3rd, 4th, 5th, 6th, etc.

So what you are saying is its distortion that can be measured you just don't hear it as distortion?
Indeed. There is research going back to the 1930s (at least) that shows the masking properties of harmonics. If they fall off in a certain way, at certain levels, you don't hear them. I.E., you can't tell them from a pure sine.
 
Because we can!

Why is single ended distortion so high? In looking at tube data it seems SE distortion specs for tubes is much higher than for push pull. If that is the case why would anyone build an SE amp in the first place? Is it mostly 2nd order?

Why would any one build any kind of tube amplifier at all? Go buy 500 watts of Solid State amplifier for less than a 5 watt 300B if that is your thing. (I must confess)

Back in the early days, think 45, 2A3 and 300B, gain and watts were precious. Two or 3 watts at 5% was heaven, to some it still is. Pentodes, Push Pull and Ultra-Linier feedback came along came along and added greater efficiency and less total distortion. This was heaven and to some it still is.

Now some build hybrid tube / solid-state riaa preamplifiers with vanishing distortion. Why do this if you can get less distortion with $10.00 dollars of op-amps?

We do this human dance with technology, because we can and we like it.

It is all just for fun!

DT
 
It doesn't have to be. A good design will have a harmonic spectrum that falls off evenly. 2nd, 3rd, 4th, 5th, 6th, etc.


Indeed. There is research going back to the 1930s (at least) that shows the masking properties of harmonics. If they fall off in a certain way, at certain levels, you don't hear them. I.E., you can't tell them from a pure sine.

It's your brain that is configured to ignore them. Your ears and your brain cavity create a type of distortion, your brain is tuned to ignore these distortions and if your amp creates these similar distortions they will be ignored by your brain as well. SE amps tend to create distortions that are similar to the natural distortions your ear creates which your brain likes because the brain will ignore/filter them out. My interpretation of a lot of reading regarding NFB and why it all works when on paper it should not. my .02

Cheers,
Bob
 
Hey DualTriode,
I have an SP3A that I rebuilt with Curcio's mods. A pair of Dynaco MarkVI's that I completely rebuilt that with the recent addition of some Mundorf Silver in Oil coupling caps are really starting to sound good. (12Au7 front end KT88 outputs.) I built an 8b from scratch last year. A pair of Mark III's from scratch. I'm almost done with an SCA35 rebuild. And on deck a set of Sansui/Tango EL84 transformers, A 6V6 Magnavox, a pair of Acrosound TO 300's, I just sold a point to point 6922 Aikido. And I'm waiting for Antek to get a AN-0209 so I can start Artur Loesch's 6922 preamp.

I LOVE tubes. Sunday I bought an old Zenith EL84 SE amp to see what they're all about. I don't even like using voltage regulators!
 
Administrator
Joined 2004
Paid Member
It doesn't have to be. A good design will have a harmonic spectrum that falls off evenly. 2nd, 3rd, 4th, 5th, 6th, etc.


<snip>

My general preference is for amps with predominantly 2nd and 3rd order distortions and very little above that. (with 3rd being at least 10dB lower than 2nd) I also generally aim for the lowest levels I can get with a given topology. I don't use global feedback in my current designs.
 
There was no mention another type of distortion in valve amps and its origin is electromagnetic.

In a push-pull amp, the DC current flows through the two halves of the OPT primary , then the magnetic field is canceled.
Fantastic, do not need an air gap in the OPT core, can also be smaller.
But the hysteresis loop is a closed curve, not a straight line, regardless of the quality of the core material.
This means distortion, not only for the non-linearity, but also for the zero crossing of hysteresis loop, the core must be magnetized cycle by cycle.

In a SE amp the DC current flows all the time through OPT primary, to avoid saturation of the core, we must add an air gap, the core must be larger.
But all is not lost, the air gap reduces the area of the hysteresis curve, and also the core is magnetized by the DC component all the time.
As a result we have a more linear hysteresis curve, and no zero crossing distortion.
This is a part of "SE Magic"
 
It's pretty revealing to take a PC-based signal generator an dial in say 1000Hz on one channel and then listen to how it sounds when you mix in a 2000Hz signal at 5% level on the other channel. The difference is barely audible and not very displeasing.
Then proceed with 3, 4, 5,.....kHz. You will notice that 3rd harmonic is more audible but not extremely disturbing whereas higher harmonics tend to sound very dissonant.

/Olof
 
My SE experiment, going to use Sovtek EL84m's. In the RH84 circuit.
 

Attachments

  • DSCN0293.JPG
    DSCN0293.JPG
    735.2 KB · Views: 641
There was no mention another type of distortion in valve amps and its origin is electromagnetic.

In a push-pull amp, the DC current flows through the two halves of the OPT primary , then the magnetic field is canceled.
Fantastic, do not need an air gap in the OPT core, can also be smaller.
But the hysteresis loop is a closed curve, not a straight line, regardless of the quality of the core material.
This means distortion, not only for the non-linearity, but also for the zero crossing of hysteresis loop, the core must be magnetized cycle by cycle.

In a SE amp the DC current flows all the time through OPT primary, to avoid saturation of the core, we must add an air gap, the core must be larger.
But all is not lost, the air gap reduces the area of the hysteresis curve, and also the core is magnetized by the DC component all the time.
As a result we have a more linear hysteresis curve, and no zero crossing distortion.
This is a part of "SE Magic"

Fascinating!

Can you please point me to a reference that shows a measurement of this phenomenon? I would like to see the nonlinearity of the magnetization curve without DC bias.

Thanks!
 
Fascinating!

Can you please point me to a reference that shows a measurement of this phenomenon? I would like to see the nonlinearity of the magnetization curve without DC bias.

Thanks!

Thanks, but I do not deserve.
You can see Maxwell equations here

http://www.diyaudio.com/forums/tubes-valves/181072-designing-transformers-j-c-maxwell.html

From the constitutive relationship

B = u H

For ferromagnetic materials u is far from constant, then B and H are related by a curve called magnetic hysteresis loop, magnetization is related but not the same (it can be proved from Maxwell's equations) for now see the attachments.


But not so far away, from high school I know that in an ideal transformer (hysteresis in a straight line) a sinusoidal voltage corresponds to a sinusoidal current, in the real world transformers that does not happen, due to hysteresis.
In the book "Basic Electrical Theory and Practice" by Wolfgang Müller-Schwartz, Siemens Aktiengesellschaft, page 182 you can see nice pictures that best illustrate what I.
Also you can measure for yourself, you just need a transformer and an oscilloscope.

Best regards
Johann
 

Attachments

  • Hysteresis.gif
    Hysteresis.gif
    14.6 KB · Views: 626
  • Magnetization.png
    Magnetization.png
    12 KB · Views: 593
Last edited:
When somebody listen some or determinate live music, IE, a Philharmonic orchestra you listen the fundamental of each timbre in each instrument. Ergo" "You listen live music".
Live music is impossible to rise with any electronic reproduction system. The problem is the distortion. The harmonics, even or odd, are harmonic. And those are distortion. Some of them are "euphonics", probable the even one. But if you listen a philharmonic orchestra with even forttisimos you eat the fundamentals more a lot of music (euphonic) that nobody generate with a particular intention, is collateral music, is a damage to the original program.
SE have pairs or even harmonics, the music that them produce are beautifully but because you listen a voice with a flute or a guitar for that reason SE amps are good for chamber or jazz. However when you try to listen Bruckner in a stupid SE you listen a barbecue with chorizzo and longaniza music with some reminiscence to the original one.
There are a sad story, sad for the mortal common listeners and HiFi followers about one of the big orchestra director of all the times, Celibidache avoid to record music because he considered the poor quality of the reproduction of the systems in those times and sadly for us because his entire most important performances we never couldn't today listen.
When somebody put a hand in a Piano instrument. If put the C and C, is euphonic, if you put a C and e bemol is disphonic or odd. Is an harmonico with a determinate character (major or minor) happy or sad, imagine with the music in a orchestra listen all the C of each instrument in a big fortíssimo plus the C of upper pairs are C but they never generate them. All if you not at this moment sum the odd harmonics.
Talk more about is talk about the sex of angels.
At least I recognize by myself i will continue to follow this kind of conversation:(
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.