The key is low-impedance drive. Whether it is achieved through local FB or inherent topology does not matter.
A regular tube PP stage without FB has high output impedance, and is maximally affected by the transformer's non-linearity, but that is precisely something tube-sound-lovers are looking for: in fact, the tube sound is mostly "iron" sound, and it bears similarities with the semi + iron sound, either in the fifties/sixties style, or with more modern approaches, like Pass's ones
That's a kind of a dilemma: how do you get good objective, numerical performances without sacrificing the magical tube sound?
It is mostly a matter of personal taste
It certainly makes a difference. How much depends on the circuit specifics
These figures are interesting, but not really significant: with a transformer having a coupling coefficient of exactly 1, primary and secondary figures would be identical, but they would include the contribution of the core and the active elements, and there is no simple way to disentangle them, especially if the OP stage has a local FB.
This means that you have no way to know in what proportion the iron and the tubes contribute to the total THD figure.
The increase seen at the secondary is just caused by the primary leakage inductance and the primary resistance.
The only way to determine reliably the transformer's contribution to the THD would be to test it in isolation, with a good test-amplifier.
Drive the secondary from the test amp with a series resistance equivalent to the 100 ohm input multiplied by 8/5000, and measure the THD for various excitation levels, at the primary or the secondary
The circuits I have in mind are push pull sweep tubes such as say ECL85, 6LR8 etc. in UL or triode mode with possibly an additional 10 or 20 dB of local feedback.
Indeed the best info would come from someone with the proper equipment to drive the OPT buy itself with various source impedances at various frequencies and output levels. I have seen Jensen data of this type on their line level OPTs but nothing for tube power OPTs. For what it is worth the distortion on the line OPTs is minuscule below saturation. My guess is that power outputs would be significantly higher.
Indeed the best info would come from someone with the proper equipment to drive the OPT buy itself with various source impedances at various frequencies and output levels. I have seen Jensen data of this type on their line level OPTs but nothing for tube power OPTs. For what it is worth the distortion on the line OPTs is minuscule below saturation. My guess is that power outputs would be significantly higher.
Member
Joined 2009
Paid Member
Later versions of this SE amp with an 826 output tube ended up with .027% distortion at 1W on the transformer secondary. I raised the amount of feedback and I estimate the driving impedance on the primary to be under 20 Ohms, just based on the measured Zout on the secondary and the winding resistances. See my Corona amplifier thread for more details.
I don't know what the irreducible distortion in a typical output transformer is, but it appears that it is quite low at midband and normal listening levels. Things are going to fall apart if you try to drive a transformer at LF with levels more than it can handle, but we know this.
To answer your basic question, I think you can make a very nice PP amplifier without any feedback around the output transformer. That's basically what my Plitron Unity-Coupled amplifier is. 50% CFB Push-pull amp, no feedback from the secondary.
I've toyed with applying my Corona SE amp feedback scheme to push-pull amps, but making the feedback kind of like the input stage of an instrumentation amplifier (see attached).
I don't know what the irreducible distortion in a typical output transformer is, but it appears that it is quite low at midband and normal listening levels. Things are going to fall apart if you try to drive a transformer at LF with levels more than it can handle, but we know this.
To answer your basic question, I think you can make a very nice PP amplifier without any feedback around the output transformer. That's basically what my Plitron Unity-Coupled amplifier is. 50% CFB Push-pull amp, no feedback from the secondary.
I've toyed with applying my Corona SE amp feedback scheme to push-pull amps, but making the feedback kind of like the input stage of an instrumentation amplifier (see attached).
Attachments
Norman Crowhurst wrote this about feedback from the primary in his Twin Coupled Amplifier:
https://www.audiofaidate.org/it/articoli/Crowth-twincoupled.pdf
https://www.audiofaidate.org/it/articoli/Crowth-twincoupled.pdf
Attachments
Hello SpreadSpectrum and all. This is an interesting and very well explained thread. Seems to me in your post above that the amp's damping factor is low in the absence of overall negative feedback. Is this OK? Would you call this consequent amp [and others like it in this thread] current or voltage source?
Thanks.
Anton
Poor Man's Definitions:
Current Source: Any change in the load resistance/impedance does Not change the amount of current in the load.
Voltage Source: Any change in the load resistance/impedance Does change the amount of current in the load.
Low Pass Filter: When the frequency going into the filter is increased, the higher the frequency is, the longer the delay is for the high frequency to get out of the filter.
The delay is so great for the very highest frequencies, so they never make the journey out of the filter.
Amplifier output classification:
All amplifier outputs are according to the following:
Current Source
Voltage Source
Resistive power source
Combination of two of those.
In the real world, resistance is present.
There are no perfect current sources
There are no perfect voltage sources
All of the above are just my opinions
(and I suspect a few other people's opinions too).
Current Source: Any change in the load resistance/impedance does Not change the amount of current in the load.
Voltage Source: Any change in the load resistance/impedance Does change the amount of current in the load.
Low Pass Filter: When the frequency going into the filter is increased, the higher the frequency is, the longer the delay is for the high frequency to get out of the filter.
The delay is so great for the very highest frequencies, so they never make the journey out of the filter.
Amplifier output classification:
All amplifier outputs are according to the following:
Current Source
Voltage Source
Resistive power source
Combination of two of those.
In the real world, resistance is present.
There are no perfect current sources
There are no perfect voltage sources
All of the above are just my opinions
(and I suspect a few other people's opinions too).
Last edited:
It is 20 Ohms driving impedance on the primary of the output transformer. That means that the amplifier Zout is almost completely dominated by transformer copper resistances. I measured 0.7 Ohms Zout on the 8ohm secondary, so a damping factor of over 10 which is unusually high for an SE amp.
Member
Joined 2009
Paid Member
Hello SpreadSpectrum. I am interested in vacuum tube power amps and their performance. My knowledge of this broad topic is embryonic. I experiment for entertainment. I am toying with mating a look-alike of a VT amp with a solid state one. This is found in the last post [and previous ones] of the thread 'Class aP amplification'/ Pass Labs Forum. Your thoughts [and others] on this experimental approach of study can only be valuable.
Best wishes
Anton.
You read my mind.
I was searching for that brilliant RCA design to answer the original question, but found it in this exactly thread.