Interesting that the description does not tie in with the schematic. Each stage clearly has local NFB and there is global NFB from the output to the input stage.
I have said it many times before and I make no apologies for repeating it, you cannot describe NFB without specifying both how it is derived and how it is applied. The local NFB in both stages in current derived, voltage applied and the global NFB is voltage applied, voltage applied.
Neither bears any relation to Schade NFB which is voltage derived, current applied.
Cheer
Ian
Cheers, Shoog.
I think you had the common drain examples from the F5 manual in mind, they show
drain to gate feedback:
Both those schematics are source followers with the drain connected directly to the power supply (just like cathode followers). There is no feedback from drain to gate but there is feedback from source to to gate.
Cheers
Ian
Yes, The global NFB in the F5, in your terminology, is shunt derived , series applied.
Cheers
Ian
Sorry, yes you are right. This time I have mixed up the schematics.
This one is correct (10k drain to gate, single stage amplifier):
Attachments
At normal listening levels - which is all that is important.
Would you describe the 5% 2nd harmonic typical of SE at clipping as representative of the SE sound.
Two members have produced their typical results which illustrate the point - so at this stage I can only think you are looking for an argument.
Shoog
Last edited:
You've got it wrong again. At this stage, I'm looking to learn more so I'm asking questions. You sure post a lot on this subject so I thought you knew something.
What does that mean for the numbers at the amplification stage?
Last edited:
My PP amps put out between 7 and 20Watts of Class A. My speakers run loud at around 1w, so speaking from personal experience - i cannot tell you how they sound at heavy clipping. This is what zigzag audio also stated to be his situation. The reality is that no one should be making the gross mistake of listening to their PP or SET amps at gross distortion levels, and those who have built them are generally careful to follow this sensible rule.
The profiles of distortion shown in this thread have shown that a triode Class A PP amp with zgnfb played well within its power capability has a generally similar distortion profile to a SET played within its power range. Of course its a lot more difficult to achieve this with a SET - but it does nothing to change the basic fact.
Is that informative enough. The fact that very few people have personal experience of a Triode Class A PP amp with ZgNFB means that very few people have anything meaningful to say about the comparison been discussed here.
If there is a major difference in the sound it seems to me that it resides within the differences between the types of OT rather than the valves themselves. People have even gone to the trouble of applying unbalanced DC to their PP OT in order to emulate the behaviour of the SE OT.
The simple point I have been attempting to make is that the claimed sound been misrepresented as a PP amp is nothing more than the sound of the typical Williamson style PP amplifiers most people have listened to.
Can I suggest you go and build a Triode Class A PP ZgNFB (using the same output valve as your prefered SET amplifier) and report back on what you find. I am certain we would all be more than interested in your findings.
Shoog
Last edited:
I hesitate to get involved again, as it seems the thread has devolved into unfriendly banter, but I will repeat my $0.02 one last time.
To add to shoog's statements, I would note that not only does Class A PP measure quite similarly to SE in general pattern, but at significantly lower levels of individual harmonic distortion when comparing output watt to output watt.
One can go PSE if they like to claim a fair comparison, but I think watt for watt is a good playing field either way. Some data has been provided, largely ignored, as there seems to be a trend of personal taste. That's where I drop out of the discussion, because you will never get anywhere when discussing personal preferences.
For me, mushrooms taste like dirt, and raw oysters are perfect for breakfast, lunch and dinner. I don't expect the next person to share my opinions, but it is possible to deconstruct the two foods and discuss which one has more nutrition. Not saying oysters win, just saying there are discussions that can be had apart from personal preference.
To add to shoog's statements, I would note that not only does Class A PP measure quite similarly to SE in general pattern, but at significantly lower levels of individual harmonic distortion when comparing output watt to output watt.
One can go PSE if they like to claim a fair comparison, but I think watt for watt is a good playing field either way. Some data has been provided, largely ignored, as there seems to be a trend of personal taste. That's where I drop out of the discussion, because you will never get anywhere when discussing personal preferences.
For me, mushrooms taste like dirt, and raw oysters are perfect for breakfast, lunch and dinner. I don't expect the next person to share my opinions, but it is possible to deconstruct the two foods and discuss which one has more nutrition. Not saying oysters win, just saying there are discussions that can be had apart from personal preference.
More......
"What may be deduced about a system based on Harmonic distortion ?
Read D.E.L. Shorter and now more recently Dr. Earl Geddes, The GedLee
Metric. Will shed some light on this issue.
Geddes showed that there was little or no perceptible difference
between amps of WIDELY varying "THD", but that there were differences in amps of different spectra of harmonics."
"What may be deduced about a system based on Harmonic distortion ?
Read D.E.L. Shorter and now more recently Dr. Earl Geddes, The GedLee
Metric. Will shed some light on this issue.
Geddes showed that there was little or no perceptible difference
between amps of WIDELY varying "THD", but that there were differences in amps of different spectra of harmonics."
Hiraga on distortion spectrum psychoacoustics
The following is an excerpt from an article posted to a private
mailing list. I am not the author! My opinion of this information is
that it is interesting speculation (as opposed to senseless
speculation). If the conjecture is supported by valid psychoacoustic
data, then it may be of use in explaining certain subjective
characteristics of amplifiers.
Remember, I didn't write this. Apologies to the author for
reposting without his permission, but he did make the article public
in the first place...
-----
In the 70's Jean Hiraga (french native but his fathers were
japanese) spend 9 years in Japan and an important part of his work
(for Musen To Jikken = MJ and others) was dedicated to the
understanding of distortion behaviour of amplifiers.
You may find in his book:
"Initiation aux amplis a tubes", Jean Hiraga, 1990,
2nd edition, 152 p., (ISSN 0753-7409)
some results he obtained by spectral analysis on famous tubes and
transistors amplifiers (a work published in 1975 in La Nouvelle Revue
du Son).
During this period he did a lot of work on the WE300B and most of the
curves still published about 300B published in MJ are from Jean
Hiraga.
When he came back in France to work at "L'Audiophile" and "La Nouvelle
Revue du Son" him and his coworkers enriched that work, especially
with the results of psychoacoustical studies and numerous listening
sessions..
It results from that huge work what I accept (may be I am wrong about
that point...) as a good answer to the common question: "why, even
with high level of THD few amps seems to have no audible distortion".
What I wrote here is not totally published and result from personnal
discussions some of them were with Jean Hiraga or few ones of his
coworkers. I hope I don't betray the meaning of what they say to me:
Our ear and especially our internal ear is non linear. When a pure
sinus come to the ear it is distorted all along its path and specially
in the internal ear. (All of this is classical ) When it comes inside
the internal ear the distorted sine is enriched with even and odd
harmonics the level of which disminished linearly with frequency
(don't ask me why I don't know, this was said to me by one of the Jean
Hiraga's coworker).
If you submitted to the external ear such a predistorted sinus
enriched with even and odd harmonics with linearly disminishing levels
versus frequency, it will be recognized as a non distorted sound, even
it will be like our ears was larger and if the sound source was
subjectively nearer . (Probably a reason while listening to very good
triode SE amps you feel like if you are "inside" the music.)
On the basis of well done listening tests (see intro.) Hiraga and
others have point out that the best sounding amps were those
possessing such regular slope of the distortion spectrum . If you
take their distortion spectrum (level in dB versus log(frequency) you
are able to draw a straigth line though the top of the peaks for the
whole set of harmonics (this should be true at every output level
until clipping). For the best amplifiers such a line possess a slope
between -18 and -24 dB by octave. But don't believe that those with
the higher slope are better because of the higher harmonics have a
lower level, this is not true! Between those amps, the best ones are
few monotriodes with tubes like 300B (by example the Tanaka or the
Audiophile Legend), RE604, PX25 and few others ones (yes I listened to
them) and they possess only a slope around -18 dB/octave.
Amplifiers with zero distortion doesn't exit so the goal is not to
design the amp with the lowest level of THD (or even 3rd order HD /
2nd order HD) no, the goal is to design an amplifier the distortion of
which is not recognized by our ears as distortion. From what I learn
from those I accept as more experienced and from the 30 and more well
prepared listening sessions I attended to, I assume that those amps
has to possessed a distortion containing harmonics up to the 10th
order (and more but the high order ones are imbedded in the noise) the
level of which regularly decreases in level with frequency at a rate
around -18 dB/octave.
Here are the calculated relative level of the harmonics for slopes
between 18 and 24 dB/octave, the level of the 2nd harmonic taken as a
reference:
order of slope = slope = slope = slope =
harmonic 18dB/octave 20dB/octave 22dB/octave 24dB/octave
2 0 0 0 0
3 -10.5 -11.7 -12.9 -14
4 -18 -20 -22 -24
5 -23.8 -26.4 -29 -31.7
6 -28.5 -31.7 -34.9 -38
7 -32.5 -36.1 -39.8 -43.4
8 -36 -40 -44 -48
9 -39 -43.4 -47.7 -52.1
10 -41.8 -46.4 -51.1 -55.7
11 -44.3 -49.2 -54.1 -59
The relative phase of those harmonics is also to be considered but as
most of the common spectrum analyser don't deliver them (except FFT
but here you have to use the proper time window if you want to measure
corectly the low level harmonics components) and because their values
are in theirselves difficult to study the best thing is IMHO to look
at the distortion signal versus the original signal with an
oscilloscope in XY mode.
The signal should appear on your scope as sunglasses = a kind of
horizontal 8 pinched along an oblique line.
My opinion has always been to say there is no magic in the way
monotriode SE amplifiers behave, there is only work to do and things
to learn but we must accept that a lot of work have yet been done by
more experimented and maybe more clever people. That's why I refer so
often to Jean Hiraga.
The following is an excerpt from an article posted to a private
mailing list. I am not the author! My opinion of this information is
that it is interesting speculation (as opposed to senseless
speculation). If the conjecture is supported by valid psychoacoustic
data, then it may be of use in explaining certain subjective
characteristics of amplifiers.
Remember, I didn't write this. Apologies to the author for
reposting without his permission, but he did make the article public
in the first place...
-----
In the 70's Jean Hiraga (french native but his fathers were
japanese) spend 9 years in Japan and an important part of his work
(for Musen To Jikken = MJ and others) was dedicated to the
understanding of distortion behaviour of amplifiers.
You may find in his book:
"Initiation aux amplis a tubes", Jean Hiraga, 1990,
2nd edition, 152 p., (ISSN 0753-7409)
some results he obtained by spectral analysis on famous tubes and
transistors amplifiers (a work published in 1975 in La Nouvelle Revue
du Son).
During this period he did a lot of work on the WE300B and most of the
curves still published about 300B published in MJ are from Jean
Hiraga.
When he came back in France to work at "L'Audiophile" and "La Nouvelle
Revue du Son" him and his coworkers enriched that work, especially
with the results of psychoacoustical studies and numerous listening
sessions..
It results from that huge work what I accept (may be I am wrong about
that point...) as a good answer to the common question: "why, even
with high level of THD few amps seems to have no audible distortion".
What I wrote here is not totally published and result from personnal
discussions some of them were with Jean Hiraga or few ones of his
coworkers. I hope I don't betray the meaning of what they say to me:
Our ear and especially our internal ear is non linear. When a pure
sinus come to the ear it is distorted all along its path and specially
in the internal ear. (All of this is classical ) When it comes inside
the internal ear the distorted sine is enriched with even and odd
harmonics the level of which disminished linearly with frequency
(don't ask me why I don't know, this was said to me by one of the Jean
Hiraga's coworker).
If you submitted to the external ear such a predistorted sinus
enriched with even and odd harmonics with linearly disminishing levels
versus frequency, it will be recognized as a non distorted sound, even
it will be like our ears was larger and if the sound source was
subjectively nearer . (Probably a reason while listening to very good
triode SE amps you feel like if you are "inside" the music.)
On the basis of well done listening tests (see intro.) Hiraga and
others have point out that the best sounding amps were those
possessing such regular slope of the distortion spectrum . If you
take their distortion spectrum (level in dB versus log(frequency) you
are able to draw a straigth line though the top of the peaks for the
whole set of harmonics (this should be true at every output level
until clipping). For the best amplifiers such a line possess a slope
between -18 and -24 dB by octave. But don't believe that those with
the higher slope are better because of the higher harmonics have a
lower level, this is not true! Between those amps, the best ones are
few monotriodes with tubes like 300B (by example the Tanaka or the
Audiophile Legend), RE604, PX25 and few others ones (yes I listened to
them) and they possess only a slope around -18 dB/octave.
Amplifiers with zero distortion doesn't exit so the goal is not to
design the amp with the lowest level of THD (or even 3rd order HD /
2nd order HD) no, the goal is to design an amplifier the distortion of
which is not recognized by our ears as distortion. From what I learn
from those I accept as more experienced and from the 30 and more well
prepared listening sessions I attended to, I assume that those amps
has to possessed a distortion containing harmonics up to the 10th
order (and more but the high order ones are imbedded in the noise) the
level of which regularly decreases in level with frequency at a rate
around -18 dB/octave.
Here are the calculated relative level of the harmonics for slopes
between 18 and 24 dB/octave, the level of the 2nd harmonic taken as a
reference:
order of slope = slope = slope = slope =
harmonic 18dB/octave 20dB/octave 22dB/octave 24dB/octave
2 0 0 0 0
3 -10.5 -11.7 -12.9 -14
4 -18 -20 -22 -24
5 -23.8 -26.4 -29 -31.7
6 -28.5 -31.7 -34.9 -38
7 -32.5 -36.1 -39.8 -43.4
8 -36 -40 -44 -48
9 -39 -43.4 -47.7 -52.1
10 -41.8 -46.4 -51.1 -55.7
11 -44.3 -49.2 -54.1 -59
The relative phase of those harmonics is also to be considered but as
most of the common spectrum analyser don't deliver them (except FFT
but here you have to use the proper time window if you want to measure
corectly the low level harmonics components) and because their values
are in theirselves difficult to study the best thing is IMHO to look
at the distortion signal versus the original signal with an
oscilloscope in XY mode.
The signal should appear on your scope as sunglasses = a kind of
horizontal 8 pinched along an oblique line.
My opinion has always been to say there is no magic in the way
monotriode SE amplifiers behave, there is only work to do and things
to learn but we must accept that a lot of work have yet been done by
more experimented and maybe more clever people. That's why I refer so
often to Jean Hiraga.
You are no longer talking about these? http://www.diyaudio.com/forums/tubes-valves/210925-se-distortion-14.html#post3753769 The image you linked doesn't even tell what the output level was.
One poster showed FFT measurement of 2 amps (1w, 4w of class A PP and 1w, 5w of SET). Only comparable measurement is at 1w because the other ones are not at same output. They look similar but there was noticeable difference in the dominance of second hamonic. A-PP amp has a tiny bit larger second harmonic than third. It is unlikely that such amp will do the 2nd harmonic "masking" as well as SET amp showing much stronger 2nd harmonic than 3rd.
As far as the distinctive tube sound of SET goes, those graphs say it's the other way around.
If you are talking about the audibility, no one has revealed that they did objective listening comparisons of those two amps yet. I guess that means there isn't anything meaningful to say in terms of audible aspect.
You said "If". How can you make a point about something you don't know?
Does it sound just like SET amp or typical PP amp in a matched level comparison? Oh, wait, you don't know that, right? Perhaps you should find out before going around and suggesting to people.
people talk all day and night about "ears" but they seem to forget that everything the ear hears is passed on the the brain......where the sensation is processed and manipulated if you will....
Echos one of my favorite mantras about proper description.
And to help settle any argument left, I consider the F5 as having global feedback.
Interestingly, I do have a couple of F5 variants with Schade types of
feedback, but their distortion specs are more modest.
More to point of the topic, the introduction of the P3 potentiometer between
the Jfet Sources of the F5 has given many the opportunity to "roll their own
harmonics", adjusting the relative amplitudes for 2nd vs 3rd.
While this is nicely measurable, it is not as effective sonically as some might
hope.
Oh, and the Hiraga excerpt is as good as I have seen in print.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.