Built a single mono block amplifier of my own design, with a phase invertor, 1 ECC99 driver, 2 300Bs, and 2 SE transformers. It sounded wonderful.
The speaker was a JBL D130 Woofer in a horn enclosure, and a JBL 075 Tweeter.
We heard things that we had never heard before.
Just a sample of one.
The speaker was a JBL D130 Woofer in a horn enclosure, and a JBL 075 Tweeter.
We heard things that we had never heard before.
Just a sample of one.
The Cohen article is using current drive of the OTs rather than the usual voltage drive mode. That causes the kinks to appear from magnetizing current. Really no surprise there.
pg 45, right column:
"The output valves had high-value cathode resistors with feedback from their cathodes back to the input stage. The result was a linear current drive to the output transformer, with no (V) feedback from the transformer itself. "
( I don't see it specified in the article whether the P-P test case is still using the SE OTs in circuit or not for B+, no B+ is shown on the P-P OT CT in the diagram, which would make for huge magnetizing currents from the SE OTs acting like para-phase inductors. )
For normal voltage drive, the magnetizing current effects are well suppressed by the low drive impedance. (no doubt explaining the popularity of local shunt "Schade" Fdbk) Seems that for the usual voltage drive mode, all this doubled up SE stuff would just be a hard way of doing it.
However, the acoustic combination idea is interesting (two speakers per amp, each SE) That would give an interesting 2nd harmonic spacial effect due to opposite 2nd harmonic phase from each speaker.
pg 45, right column:
"The output valves had high-value cathode resistors with feedback from their cathodes back to the input stage. The result was a linear current drive to the output transformer, with no (V) feedback from the transformer itself. "
( I don't see it specified in the article whether the P-P test case is still using the SE OTs in circuit or not for B+, no B+ is shown on the P-P OT CT in the diagram, which would make for huge magnetizing currents from the SE OTs acting like para-phase inductors. )
For normal voltage drive, the magnetizing current effects are well suppressed by the low drive impedance. (no doubt explaining the popularity of local shunt "Schade" Fdbk) Seems that for the usual voltage drive mode, all this doubled up SE stuff would just be a hard way of doing it.
However, the acoustic combination idea is interesting (two speakers per amp, each SE) That would give an interesting 2nd harmonic spacial effect due to opposite 2nd harmonic phase from each speaker.
Last edited:
Put a low distortion 100Hz tone into a mono SE amplifier, connect a full range speaker (or a woofer) to the amp, + to +, - to -. Turn the signal up to a fairly moderate level.
Listen to the tone.
Now reverse the phase connection from the amp to the speaker, + to -, and - to +.
Listen to the tone.
If you hear a difference, it is the addition of the amp and speaker 2nd harmonic distortion, versus the partial cancellation of the 2nd harmonic distortion resulting from the other phase connection. Which is more, and which is less depends on the particular direction of the 2nd harmonic distortion of the amp, and the particular direction of the 2nd harmonic distortion of the speaker.
This is facilitated by using a DPDT switch between the amp and speaker, and having one person flip the switch back and forth (and the listener not seeing what position the switch is in; single blindfold test).
Now, take 2 woofers in a cabinet, with one facing into the baffle board, and one facing out of the baffle board. Connect them to the amp, but in opposite phases of one driver to the other driver. The speaker 2nd harmonic will now be quite effectively cancelled (acoustically).
Listen to the tone.
Now reverse the phase connection from the amp to the speaker, + to -, and - to +.
Listen to the tone.
If you hear a difference, it is the addition of the amp and speaker 2nd harmonic distortion, versus the partial cancellation of the 2nd harmonic distortion resulting from the other phase connection. Which is more, and which is less depends on the particular direction of the 2nd harmonic distortion of the amp, and the particular direction of the 2nd harmonic distortion of the speaker.
This is facilitated by using a DPDT switch between the amp and speaker, and having one person flip the switch back and forth (and the listener not seeing what position the switch is in; single blindfold test).
Now, take 2 woofers in a cabinet, with one facing into the baffle board, and one facing out of the baffle board. Connect them to the amp, but in opposite phases of one driver to the other driver. The speaker 2nd harmonic will now be quite effectively cancelled (acoustically).
Sorry, I do not have reading links for you. Here is what I have:
2nd harmonic is caused by asymmetry, one direction of a sine wave (alternation) is shorter and fatter than the other direction of a sine wave (opposite alternation).
The output stage of an SE amp has this asymmetry.
The output stage of an SE amp tends to have more 2nd harmonic than 3rd harmonic.
What the whole amp has for the relative levels of 3rd and 2nd harmonic also depends on what the earlier stages in the amp do, not just the output stage.
Push Pull uses the opposite directions of the 2 output tubes to partially cancel this output asymmetry (2nd harmonic is partially cancelled).
A single ended driver tube has 2nd harmonic distortion in one direction, and if it drives an SE output stage (it also has 2nd harmonic distortion, but is of the opposite phase) will have a partial cancellation of the 2nd harmonic (the total 2nd harmonic distortion of the whole amp).
Now to speakers, some principles from Physics:
1. Consider Hooke's Law for springs. Use a long linear, symmetrical spring. There are no non-symmetrical changes in the spring versus motion.
Now consider a half roll outer suspension on a woofer; the angle and shape of the half roll changes, but non symmetrically . . . resulting in 2nd order harmonic.
2. Consider a magnetic gap and voice coil that are the same length (depth); but the speaker voice coil is not centered in the gap.
Or consider a voice coil that is longer than the magnetic gap, but again what if the voice coil is not exactly centered in the gap, and there are large excursions?
Or consider a voice coil that is shorter than the magnetic gap, and the voice coil is not centered in the gap, but again what if the excursions are large?
In all these cases, 2nd harmonic will be produced due to the asymmetry of the motion.
Does that explain it?
Back to amps:
The 3rd harmonic of a sine wave is a shortening and fattening of both the positive alternation and the negative alternation.
The output stage of a Push Pull amp tends to have more 3rd harmonic than 2nd harmonic.
What the whole amp has for the relative levels of 3rd and 2nd harmonic also depends on what the earlier stages in the amp do, not just the output stage.
2nd harmonic is caused by asymmetry, one direction of a sine wave (alternation) is shorter and fatter than the other direction of a sine wave (opposite alternation).
The output stage of an SE amp has this asymmetry.
The output stage of an SE amp tends to have more 2nd harmonic than 3rd harmonic.
What the whole amp has for the relative levels of 3rd and 2nd harmonic also depends on what the earlier stages in the amp do, not just the output stage.
Push Pull uses the opposite directions of the 2 output tubes to partially cancel this output asymmetry (2nd harmonic is partially cancelled).
A single ended driver tube has 2nd harmonic distortion in one direction, and if it drives an SE output stage (it also has 2nd harmonic distortion, but is of the opposite phase) will have a partial cancellation of the 2nd harmonic (the total 2nd harmonic distortion of the whole amp).
Now to speakers, some principles from Physics:
1. Consider Hooke's Law for springs. Use a long linear, symmetrical spring. There are no non-symmetrical changes in the spring versus motion.
Now consider a half roll outer suspension on a woofer; the angle and shape of the half roll changes, but non symmetrically . . . resulting in 2nd order harmonic.
2. Consider a magnetic gap and voice coil that are the same length (depth); but the speaker voice coil is not centered in the gap.
Or consider a voice coil that is longer than the magnetic gap, but again what if the voice coil is not exactly centered in the gap, and there are large excursions?
Or consider a voice coil that is shorter than the magnetic gap, and the voice coil is not centered in the gap, but again what if the excursions are large?
In all these cases, 2nd harmonic will be produced due to the asymmetry of the motion.
Does that explain it?
Back to amps:
The 3rd harmonic of a sine wave is a shortening and fattening of both the positive alternation and the negative alternation.
The output stage of a Push Pull amp tends to have more 3rd harmonic than 2nd harmonic.
What the whole amp has for the relative levels of 3rd and 2nd harmonic also depends on what the earlier stages in the amp do, not just the output stage.
I have spoken with the author and found upon question his reasoning less than sound.
Upon consultation with more than three US based winders, and one in Oz, three shut the idea down completely (cause they know it all, right?) The last was ok with providing samples but could not see the logic put forth in the article, and of course those samples would have come at a cost - so it never materialized at my end.
Your adventure might differ, and if it does, please let us know.
Hanze.
Depends on the phase relations of the Fundamental to the Harmonic. For 3H, that can just as easily be the symmetrical bend at the crossover. That is what the Dual SE is trying to avoid when compared to ordinary PP.
NTL, this design appears to be based on sound logic.
My opinion only.
There is logic to the scheme, but it is a matter of quantification.
P-P OTs have over an order of magnitude ( > 10x) less magnetizing current than SE OTs. (and it is not clear that the SE OTs were completely removed from the P-P OT test case in the article)
Much is made of the image of SE OTs not having 3rd harmonic zero crossing (magnetic) effects, however, they do. It is just hidden from obvious exposure by the large 1st order magnetizing current. The SE case still has an AC hysteresis curve formed around the new (DC biased) operating point. The DC just serves to pin down the higher permeability magnetic domains, leaving the remaining domains to handle the AC flux with reduced permeability. The air gap further reduces permeability, but keeps it near constant at least.
It is so easy to reduce the P-P OT magnetizing current effects via low impedance Voltage drive, especially with local or global N-FDBK, (and at least 10X less magnetizing current to begin with in P-P OTs). There seems little reason to fuss over this. Real Voltage drive amplifiers have the magnetizing effects suppressed below -60dB easily, usually -80dB.
My view of this doubled SE scheme is similar to removing the windshield from a car to avoid mosquito splats obscuring the view. Bugs in your teeth instead.
P-P OTs have over an order of magnitude ( > 10x) less magnetizing current than SE OTs. (and it is not clear that the SE OTs were completely removed from the P-P OT test case in the article)
Much is made of the image of SE OTs not having 3rd harmonic zero crossing (magnetic) effects, however, they do. It is just hidden from obvious exposure by the large 1st order magnetizing current. The SE case still has an AC hysteresis curve formed around the new (DC biased) operating point. The DC just serves to pin down the higher permeability magnetic domains, leaving the remaining domains to handle the AC flux with reduced permeability. The air gap further reduces permeability, but keeps it near constant at least.
It is so easy to reduce the P-P OT magnetizing current effects via low impedance Voltage drive, especially with local or global N-FDBK, (and at least 10X less magnetizing current to begin with in P-P OTs). There seems little reason to fuss over this. Real Voltage drive amplifiers have the magnetizing effects suppressed below -60dB easily, usually -80dB.
My view of this doubled SE scheme is similar to removing the windshield from a car to avoid mosquito splats obscuring the view. Bugs in your teeth instead.
Saturation of SE amps causes gradual increase of level and order of harmonics, starting from the 2'Nd one. It mimics what happens with eardrum when loudness increased, that's why "SE distortions" sound more like increased volume than like distortions. The same amount of distortions without 2'Nd harmonic ruins this effect. Yes, it measures better in terms of THD, but sounds as if dirtier. While in both cases it is a dirt, but differently perceived and interpreted by our perception.
A few Generalizations:
(All Generalizations have Exceptions)
1. SE Output Stage:
Without feedback, the Triode Output Stage of a properly designed SE amp has dominant 2nd Harmonic distortion, and has 3rd Harmonic distortion which is about 15 dB or more Lower than the 2nd harmonic distortion.
The upper order harmonic distortions fall off rapidly, and are far below the 2nd and 3rd harmonic distortion levels.
With Negative feedback, that raises the levels of the upper harmonic distortions due to what I call the “Crowhurst” effect (according the math of his ‘mostly’ well accepted writings).
2. Push Pull Output Stage:
Without feedback, a properly designed Push Pull Triode Output Stage has dominant 3rd harmonic, and 2nd harmonic that may be Lower than the 3rd harmonic by some 15 dB, 20dB, or more.
The upper order harmonic distortions fall off rapidly, and are far below the 3rd and 2nd harmonic distortion levels.
With Negative feedback, that raises the levels of the upper harmonic distortions due to what I call the “Crowhurst” effect (according the math of his ‘mostly’ well accepted writings).
3. Amplifiers with 1 or more stages preceding the Output Stage of SE and Push Pull amps,
(with or without negative feedback):
What happens to the relative levels of the 2nd, 3rd and upper harmonic distortions of the ‘Complete Amplifier’ is much more complex than case 1. and case 2. above.
Without negative feedback, the multi-stage SE amp:
Sometimes the 2nd harmonic distortion will be partially cancelled.
The 3rd harmonic distortion is far less likely to be partially cancelled (for a Triode, you might look at Child’s Law and the Taylor series as applied to harmonic distortion).
I have built some of these.
With negative feedback, the multi-stage SE amp:
The 2nd and 3rd harmonic distortion will both tend to be reduced, but the upper harmonic distortions will tend to be increased.
With negative feedback, the multi-stage Push Pull amp:
The 2nd and 3rd harmonic will both tend to be reduced, but the upper harmonic distortions will tend to be increased.
I did not comment on the multi-stage Push Pull amp Without negative feedback.
I do not have an ‘easy’ Generalization for the many different topologies of this kind of amp.
I have built some of these, and I used at least 2 different topologies.
And there are lots more topologies with even more different results.
(All Generalizations have Exceptions)
1. SE Output Stage:
Without feedback, the Triode Output Stage of a properly designed SE amp has dominant 2nd Harmonic distortion, and has 3rd Harmonic distortion which is about 15 dB or more Lower than the 2nd harmonic distortion.
The upper order harmonic distortions fall off rapidly, and are far below the 2nd and 3rd harmonic distortion levels.
With Negative feedback, that raises the levels of the upper harmonic distortions due to what I call the “Crowhurst” effect (according the math of his ‘mostly’ well accepted writings).
2. Push Pull Output Stage:
Without feedback, a properly designed Push Pull Triode Output Stage has dominant 3rd harmonic, and 2nd harmonic that may be Lower than the 3rd harmonic by some 15 dB, 20dB, or more.
The upper order harmonic distortions fall off rapidly, and are far below the 3rd and 2nd harmonic distortion levels.
With Negative feedback, that raises the levels of the upper harmonic distortions due to what I call the “Crowhurst” effect (according the math of his ‘mostly’ well accepted writings).
3. Amplifiers with 1 or more stages preceding the Output Stage of SE and Push Pull amps,
(with or without negative feedback):
What happens to the relative levels of the 2nd, 3rd and upper harmonic distortions of the ‘Complete Amplifier’ is much more complex than case 1. and case 2. above.
Without negative feedback, the multi-stage SE amp:
Sometimes the 2nd harmonic distortion will be partially cancelled.
The 3rd harmonic distortion is far less likely to be partially cancelled (for a Triode, you might look at Child’s Law and the Taylor series as applied to harmonic distortion).
I have built some of these.
With negative feedback, the multi-stage SE amp:
The 2nd and 3rd harmonic distortion will both tend to be reduced, but the upper harmonic distortions will tend to be increased.
With negative feedback, the multi-stage Push Pull amp:
The 2nd and 3rd harmonic will both tend to be reduced, but the upper harmonic distortions will tend to be increased.
I did not comment on the multi-stage Push Pull amp Without negative feedback.
I do not have an ‘easy’ Generalization for the many different topologies of this kind of amp.
I have built some of these, and I used at least 2 different topologies.
And there are lots more topologies with even more different results.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.