sser2,
Is the proper curve diagram of an SE transformer one that looks something like a distorted Ellipse in a single quadrant?
If the signal is strong enough to cause saturation, then . . .
Saturation is at the peak that is furthest from the origin.
Is the proper curve diagram of a PP transformer one that looks something like a distorted Ellipse in one quadrant, and another distorted Ellipse in the opposite quadrant?
If the signal is strong enough to cause saturation, then . . .
A Saturation is at the peak that is furthest away from the origin in one quadrant.
And . . . Saturation in the other direction is at the peak that is furthest away from the origin in the opposite quadrant.
At 78 years old, I can not remember all that I have seen.
Anybody, please post those curve diagrams!
Is the proper curve diagram of an SE transformer one that looks something like a distorted Ellipse in a single quadrant?
If the signal is strong enough to cause saturation, then . . .
Saturation is at the peak that is furthest from the origin.
Is the proper curve diagram of a PP transformer one that looks something like a distorted Ellipse in one quadrant, and another distorted Ellipse in the opposite quadrant?
If the signal is strong enough to cause saturation, then . . .
A Saturation is at the peak that is furthest away from the origin in one quadrant.
And . . . Saturation in the other direction is at the peak that is furthest away from the origin in the opposite quadrant.
At 78 years old, I can not remember all that I have seen.
Anybody, please post those curve diagrams!
it is like this, it is not driven by the screen.
2 examples from
https://www.ne.jp/asahi/evo/amp/guide.htm
.
more about the concept
http://www.gem.hi-ho.ne.jp/katsu-san/audio/what_STC1_english.html
Last edited:
I saw that exact circuit with other names: https://www-ne-jp.translate.goog/as..._sl=ja&_x_tr_tl=en&_x_tr_hl=it&_x_tr_pto=wapp
We discussed about it in the pentode retaliation thread:
https://www.diyaudio.com/community/threads/single-ended-the-pentode-retaliation.373257/post-7049242
This is what I’ve always seen called as Super Triode:
https://www.diyaudio.com/community/threads/super-triode-single-ended-amp.393097/
We discussed about it in the pentode retaliation thread:
https://www.diyaudio.com/community/threads/single-ended-the-pentode-retaliation.373257/post-7049242
This is what I’ve always seen called as Super Triode:
https://www.diyaudio.com/community/threads/super-triode-single-ended-amp.393097/
See Landee, Davis and Albrecht Electronic Designers Handbook 1957 p.14-3 Fig 14.3 which shows the normal magnetization curve derivation as the locus of the tips of the family of magnetization curves of various levels, on a B/H grid. It shows how the slope of the B/H curve decreases from its maximum when approaching zero crossing, and then decreases again when approaching saturation. The slope of the curve is permeability, and inductance varies directly with permeability (almost by definition).
All good fortune,
Chris
@6A3sUMMER
I don’t want to abuse your availability, but do you know where I can find typical harmonics vs power plots for the 300B (or other famous triodes)?
I would like to have a better idea of how well sounding SET amps should behave.
I don’t want to abuse your availability, but do you know where I can find typical harmonics vs power plots for the 300B (or other famous triodes)?
I would like to have a better idea of how well sounding SET amps should behave.
zintolo,
sser2,
Thanks for the Graph, zintolo.
I was wrong, sser2.
The Hysteresis curve for Push Pull is not a 2 Quadrant curve.
It is a 4 Quadrant curve.
Funny what tricks my 78 year old brain played on me.
Gee, that dotted line with the jog in the middle, looks like a Zero Crossing Error (it looks that way to me, make your own conclusions).
But it does not cause me to stop designing, building, and listening to both balanced amplifiers, and push pull amplifiers.
Now you know why I built the dual SE amplifier, drove them out of phase, and paralleled the outputs by reversing the + and common connections on one of them before paralleling them.
As I previously said, the French beat me to that topology.
Note: there is no "jog in the middle" when you combine 2 SE amplifiers that way (Just my theory, yet to be proven, or to be disproven, by those who are more knowledgeable about those things than me).
. . . Let's hear from you, if you have the proof or the disproof, please.
sser2,
Thanks for the Graph, zintolo.
I was wrong, sser2.
The Hysteresis curve for Push Pull is not a 2 Quadrant curve.
It is a 4 Quadrant curve.
Funny what tricks my 78 year old brain played on me.
Gee, that dotted line with the jog in the middle, looks like a Zero Crossing Error (it looks that way to me, make your own conclusions).
But it does not cause me to stop designing, building, and listening to both balanced amplifiers, and push pull amplifiers.
Now you know why I built the dual SE amplifier, drove them out of phase, and paralleled the outputs by reversing the + and common connections on one of them before paralleling them.
As I previously said, the French beat me to that topology.
Note: there is no "jog in the middle" when you combine 2 SE amplifiers that way (Just my theory, yet to be proven, or to be disproven, by those who are more knowledgeable about those things than me).
. . . Let's hear from you, if you have the proof or the disproof, please.
Last edited:
zintolo,
Thanks for the explanations of Super Triode!
It reminds me of a 2 stage amplifier that I built. Instead of Schade negative feedback from output plate to driver plate,
I connected the output plate to the cathode of the driver, by using a series Rf and Cf.
The negative feedback amount was dependent on the driver cathode impedance, cathode self bias resistor, and the feedback resistor, Rf.
But, with a series RC, the negative feedback reduced to at very low frequencies (to zero at DC for sure), according to Rf and Cf.
Thanks for the explanations of Super Triode!
It reminds me of a 2 stage amplifier that I built. Instead of Schade negative feedback from output plate to driver plate,
I connected the output plate to the cathode of the driver, by using a series Rf and Cf.
The negative feedback amount was dependent on the driver cathode impedance, cathode self bias resistor, and the feedback resistor, Rf.
But, with a series RC, the negative feedback reduced to at very low frequencies (to zero at DC for sure), according to Rf and Cf.
zintolo,
I can not remember the title of the book, but a Bernie _ somebody __ wrote a book about WE tubes, it had the graphs you are wanting.
There were curves of power out, 2nd harmonic, 3rd harmonic, versus the Load impedance on the plate.
It had the WE 300B.
(*) But the most memorable set of curves, was of the WE 275 Triode. It had a super sharp, super deep null of the 3rd harmonic versus plate load impedance. Wow!
But do not get too excited, using that tube to drive an output transformer and a real speaker would seldom be at the 3rd harmonic null;
instead it would be going below the RL null impedance, and above the RL null impedance, depending on the speaker impedance at various frequencies.
Also, that null was for Real Resistive Impedance Load, and not for a real world complex loudspeaker impedance that had R&C, R&L, or RLC.
Real Resistive load impedance of loudspeakers are only at the Peaks of impedance, and at the nulls of impedance;
everywhere else is an elliptical load, sorry.
And there is the fly-in-the-ointment.
But, our playback systems can sound excellent, in spite of all the above (*).
Have Fun Everybody!
I can not remember the title of the book, but a Bernie _ somebody __ wrote a book about WE tubes, it had the graphs you are wanting.
There were curves of power out, 2nd harmonic, 3rd harmonic, versus the Load impedance on the plate.
It had the WE 300B.
(*) But the most memorable set of curves, was of the WE 275 Triode. It had a super sharp, super deep null of the 3rd harmonic versus plate load impedance. Wow!
But do not get too excited, using that tube to drive an output transformer and a real speaker would seldom be at the 3rd harmonic null;
instead it would be going below the RL null impedance, and above the RL null impedance, depending on the speaker impedance at various frequencies.
Also, that null was for Real Resistive Impedance Load, and not for a real world complex loudspeaker impedance that had R&C, R&L, or RLC.
Real Resistive load impedance of loudspeakers are only at the Peaks of impedance, and at the nulls of impedance;
everywhere else is an elliptical load, sorry.
And there is the fly-in-the-ointment.
But, our playback systems can sound excellent, in spite of all the above (*).
Have Fun Everybody!
Hysteresis curves for SE are same as for PP, as Chris posted. Just shift the curves on the graph to account for DC magnetization.
sser2,
What Chris said:
" It shows how the slope of the B/H curve decreases from its maximum when approaching zero crossing"
I ask you sser2:
Suppose we have a typical 3k Ohm 300B air gapped single ended transformer that has 60mA quiescent current.
When signal is applied, first the 300B plate current peaks at 60mA + 50mA = 110mA
Then when the signal reverses direction, we get the other signal extreme, and we have 60mA - 50mA = 10mA
Please tell me what is the plate current (mA) that we are near the zero crossing of the transformer primary (I really want to know).
What zero crossing? And I do not accept that Chris is referring to the input music signal (music normally does not have quiescent DC, but it does have zero crossings from compression to rarefaction).
And yes, I realize the 300B is not perfectly linear, so we will have more additional current above the quiescent current;
Versus the lower current when it reduces to the minimum signal swing.
(We do not change exactly +50mA and -50mA).
Perhaps we get +51mA change, and -50mA change, that is about 2% 2nd harmonic distortion, Right?
In my old age, perhaps I am missing the nature of the real operation of an air gapped SE output transformer.
What Chris said:
" It shows how the slope of the B/H curve decreases from its maximum when approaching zero crossing"
I ask you sser2:
Suppose we have a typical 3k Ohm 300B air gapped single ended transformer that has 60mA quiescent current.
When signal is applied, first the 300B plate current peaks at 60mA + 50mA = 110mA
Then when the signal reverses direction, we get the other signal extreme, and we have 60mA - 50mA = 10mA
Please tell me what is the plate current (mA) that we are near the zero crossing of the transformer primary (I really want to know).
What zero crossing? And I do not accept that Chris is referring to the input music signal (music normally does not have quiescent DC, but it does have zero crossings from compression to rarefaction).
And yes, I realize the 300B is not perfectly linear, so we will have more additional current above the quiescent current;
Versus the lower current when it reduces to the minimum signal swing.
(We do not change exactly +50mA and -50mA).
Perhaps we get +51mA change, and -50mA change, that is about 2% 2nd harmonic distortion, Right?
In my old age, perhaps I am missing the nature of the real operation of an air gapped SE output transformer.
Last edited:
Permeability, the slope of the B/H curve, doesn't effect signal directly, but indirectly, by its determination of inductance, it effects the shunt inductive parasitic load to the output valve. Effect varies of course with the importance of that parasitic shunt, varying with frequency and level. Significant amounts of nickel in the core straighten the burble through B/H zero crossing somewhat, and a series'd air gap rotates clockwise and stretches the normal magnetization curve.
Single ended amplifiers, with DC bias through the OPT, don't operate in this region, but with all of their well understood liabilties, are an extreme example of putting monotonicity foremost in design. One could argue that monotonicity is the most important characteristic of an amplifier. But that isn't well accepted, except maybe by me.
All good fortune,
Chris
Single ended amplifiers, with DC bias through the OPT, don't operate in this region, but with all of their well understood liabilties, are an extreme example of putting monotonicity foremost in design. One could argue that monotonicity is the most important characteristic of an amplifier. But that isn't well accepted, except maybe by me.
All good fortune,
Chris
Chris,
I repeat:
Now you know why I built the dual SE amplifier, drove them out of phase, and paralleled the outputs by reversing the + and common connections on one of them before paralleling them.
As I previously said, the French beat me to that topology.
Note: there is no "jog in the middle" when you combine 2 SE amplifiers that way (Just my theory).
What do you think?
I repeat:
Now you know why I built the dual SE amplifier, drove them out of phase, and paralleled the outputs by reversing the + and common connections on one of them before paralleling them.
As I previously said, the French beat me to that topology.
Note: there is no "jog in the middle" when you combine 2 SE amplifiers that way (Just my theory).
What do you think?
Last edited:
6A3sUMMER:
I believe your approach unnecessarily complicates the matter. Plate current of the SE output stage has two components: DC bias current and AC signal current superimposed on it. Magnetic flux in SE output transformer follows the magnetizing current, so it also has the two components: constant flux caused by DC current, and alternating flux superimposed on it. Presented graphically, the B-H loop for DC+AC is the same as for AC alone, but shifted upwards for B caused by DC current. The shape of the B-H loop, which represents hysteresis, is exactly same in both cases ( of course if we stay away from saturation).
I believe your approach unnecessarily complicates the matter. Plate current of the SE output stage has two components: DC bias current and AC signal current superimposed on it. Magnetic flux in SE output transformer follows the magnetizing current, so it also has the two components: constant flux caused by DC current, and alternating flux superimposed on it. Presented graphically, the B-H loop for DC+AC is the same as for AC alone, but shifted upwards for B caused by DC current. The shape of the B-H loop, which represents hysteresis, is exactly same in both cases ( of course if we stay away from saturation).
Chris,
Monotonicity is very important in so many electronic applications.
One that few will understand is in a RBW filter of a spectrum analyzer (Resolution Bandwidth).
Non-monotonic is a No No.
Hmm, there might be something similar in speaker crossovers.
Then, there are A/Ds and D/As,
Non-monotonic is a No No.
And so many other applications.
You make me think about non-monotonic response, and then trying to correct it with negative feedback (I will not attempt that, someone else is welcome to do that).
That null of the 3rd harmonic of the WE 275 triode makes me think about the varying impedance of a woofer as it goes through its positive and negative excursions, and that drives the 3rd harmonic up, down, up.
Monotonicity is very important in so many electronic applications.
One that few will understand is in a RBW filter of a spectrum analyzer (Resolution Bandwidth).
Non-monotonic is a No No.
Hmm, there might be something similar in speaker crossovers.
Then, there are A/Ds and D/As,
Non-monotonic is a No No.
And so many other applications.
You make me think about non-monotonic response, and then trying to correct it with negative feedback (I will not attempt that, someone else is welcome to do that).
That null of the 3rd harmonic of the WE 275 triode makes me think about the varying impedance of a woofer as it goes through its positive and negative excursions, and that drives the 3rd harmonic up, down, up.
sser2,
Your curves in Post # 36 is very interesting!
Please draw in the normal magnetization dotted line that I saw on Post # 25.
Is that that the shape of the normal magnetization curve of an air gapped single ended output transformer that has quiescent DC current, plus signal current that increases and then decreases from that DC level, but that never approaches zero current, and never approaches saturation current?
I just want to know if there is any difference of that dotted curve's shape of SE versus PP.
Your curves in Post # 36 is very interesting!
Please draw in the normal magnetization dotted line that I saw on Post # 25.
Is that that the shape of the normal magnetization curve of an air gapped single ended output transformer that has quiescent DC current, plus signal current that increases and then decreases from that DC level, but that never approaches zero current, and never approaches saturation current?
I just want to know if there is any difference of that dotted curve's shape of SE versus PP.
As I understand it single ended amplifiers are at least partially current drive or are quasi current drive whereas nearly all SS amps are 100% voltage drive and this accounts for a fair bit of the sonic difference.
Lenard Audio make the claim that nearly all listeners will prefer the sound of current drive amplifiers when heard in comparison to voltage drive.
There are some current drive SS amplifiers which are reputed to sound more like single ended tube amps.SGR active speakers have an upgrade option using all current drive SS amplifiers and claim this sounds better than the standard voltage drive.Owners who have made that upgrade agree.
The fundamentals are that speaker drivers are powered by current not voltage.
http://education.lenardaudio.com/en/12_amps_8.html
Lenard Audio make the claim that nearly all listeners will prefer the sound of current drive amplifiers when heard in comparison to voltage drive.
There are some current drive SS amplifiers which are reputed to sound more like single ended tube amps.SGR active speakers have an upgrade option using all current drive SS amplifiers and claim this sounds better than the standard voltage drive.Owners who have made that upgrade agree.
The fundamentals are that speaker drivers are powered by current not voltage.
http://education.lenardaudio.com/en/12_amps_8.html