Occasionally we have discussed the kinds of distortion that most
please us, or that we find least objectionable. The subject is of
interest because it is often claimed to be an important part of the
difference between valve and SS amplifiers.

Another aspect of that difference is said to be how the distortion
is distributed with respect to power output.

What is the ideal distribution? Let me give two simple examples. Amp
1 produces zero distortion at 1mW, rising steadily to 1% at 10W. Amp
2 starts at zero and rises exponentially to 5% at 10W. Assuming
that, for a particular piece of music at a particular volume, the
total distortion for the whole piece is the same, and the
distribution of distortion components is the same, how will the two
amps differ in their contribution to the sound, and which might most
please, or least offend? What is the best shape?

Personal opinions or objective criteria welcome.

cheers, Ian

Ian Iveson wrote:

Occasionally we have discussed the kinds of distortion that most
please us, or that we find least objectionable. The subject is of
interest because it is often claimed to be an important part of the
difference between valve and SS amplifiers.

It is indeed an old claim, and I remain vexed as to why
an SET amp with a pile of thd blows away some
SS amps with 100 times less.

The Halcro SS amp is said to make 0.0001% thd at 200 watts,
declining to unmeasurable levels below that high power.
Yet already I have heard ppl say that SE tube amps still are better.


Another aspect of that difference is said to be how the distortion
is distributed with respect to power output.

Well indeed, and no two amps measure the same.



What is the ideal distribution? Let me give two simple examples. Amp
1 produces zero distortion at 1mW, rising steadily to 1% at 10W.

Do you mean rising linearly with power,
or voltage at the output?
Consider that its with voltage output.
1% of 10 watts into 8 ohms is 8.9vrms, so Dn = 89 mV
1 watt into 8 ohms is 2,82vrms, so
we'd expect Dn = 2.82/8.9 x 1% = 0.316%


Amp 2 starts at zero and rises exponentially to 5% at 10W. Assuming
that, for a particular piece of music at a particular volume, the
total distortion for the whole piece is the same, and the
distribution of distortion components is the same, how will the two
amps differ in their contribution to the sound, and which might most
please, or least offend? What is the best shape?

What exactly is the curve which would describe "exponential"?

Suppose the amp under consideration is a 100 watt amp.

Then the curve in the line which usually describes the
exponentiality between 0 and 10 watts will be different to r
that of a 10 watt amp, bacause at clipping in a 10 watt amp,
the rise in Dn is very severe after clipping.


Personal opinions or objective criteria welcome.

Consider the graph for output voltage, horizontal, vs thd, vertical,
with 0/0 at the bottom LHS.

Gnerally speaking, class A amps have a rising
thd level which is like a rope hung from the 1% at the RHS,
down to the 0.02 %, at around 0.5 watts
and then there is a sudden plummet to 0.0% into the noise.
The spectra is usually a mix of 2H and 3H,
and other higher order stuff at lower and hard to see levels, even at a
watt.

Some AB tube amps have a distortion rise and a peak at 2 volts out of
perhaps 0.1%,
then a fall to say 0.05% at 10 v, then as the AB action starts,
it rises rapidly.
Such amps usually have cross over bias problems, or they
have iron cored ISTs in the signal path.

Its very easy to build an amp with such an initially humped thd profile,

especially if its an SS type.

In the real world, a 300B will have 5% at 7 watts,
which is 7.5 vrms into 8 ohms.
at 3.75 v out, perhaps it will be 1.5%,
and at 1.875v it might be 0.5%.

Up to 5 vrms its nealy all 2H.
ppl rave about such amps as being accurate, and wonderful
conveyors of sound to speakers.

Well with 93 dB sensitive speakers, 2.83 vrms
makes 1 watt, and 93 dB SPL.
If we'd drawn a graph using the typical above figures for a SE triode
amp,
then we may expect about 0.7% of thd at a watt.
This means the thd voltage is 20 mV.
Now if we could have somebody play a flute tone,
and then have someone also play the same tone from the flute,
but twice the F, and at a 20 mV level, would we not hear
when it was added?
I would say yes. 20mV played separately from a separtae speaker
would be audible, and it would blend into
what music was being played.
But then we have the potpourie of IMD products,
and they are ugly, and not so nice to hear on their own.

So we come back to a 10 watt amp, if we only ever use a maximum of
2 watts then we should prefer the "exponential rise" of thd
and we'd have no care about the suddenly higher thd at over 2 watts.

Measurementally, the 10 watt class AB amp whilst it runs in class A
should
have less thd than any SE amp at first, but as soon as it moves into
class AB at say 3 watts,
a 10 watt SE amp may measure and sound better.


cheers, Ian

There are so many profiles one can design for,
and I suggest a suck it and see approach is the way to go.

When I test the 300 watt amps of mine, they have
0.015% thd at 3 v output into either 4 or 8 ohms,
then the thd remains below 0.1% to the end of the class A portion, and
then
the thd rapidly rises to 1.2% at 290 watts into 4 ohms just before
clipping.
Into 8 ohms, there is only 0.24% at 0.24% at 210 watts, mainly all class
A.
These amps use 12 x EH6550.
Now with just 2 x EH6550, one gets a very similar thd profile up
to the same 3 v into 4/8 ohms.
So just because you parallel a pile of tubes, it don't mean you get
much reduction of thd at a given low power.
perhaps its the effects of the iron in the OPT.
But anyone worried about 0.015% at a couple of watts?
Not I.

The thd at such low power in a high powered amp is mainly combination of

2H and 3H, and then the 2H dominates, because I use an SE triode
at the front end of the amp, and there isn't much global FB,
so the 2H of the driver tube rapidly over takes the 3H
as the main source of thd until the amp starts to go class AB, when
the 3H and higher harmonics become predominant.

In fact on my 300 watters, the input 6CG7
has to make around 8vrms to drive the following
stages to full power.
Its thd is only reduced about 6 dB by the 8 dB of global FB.
So we have this FB thing to worry about.
A srpp V1 set up would remove the 2H, and lower the
thd figures, but then it may not sound as warm.

I have grown fond of acoustical output stages,
and the 300 watt design of mine now incorporates it, with
the CFB amount = 20%, and the fixed andf regulated screen voltage is at
380V,
with B+ at +515v, and cathode bias.
The design uses only 40 mA of idle current per tube, so they
are running at less than 1/2 their rated Pd,
and what effect if any this has on the sonics compared to any other
method
I am not sure, but it sure sounds OK to me.

he, he, If all this don't put ppl to sleep, what will?

Patrick Turner.

Patrick Turner wrote in
:



Ian Iveson wrote:

Occasionally we have discussed the kinds of distortion that most
please us, or that we find least objectionable. The subject is of
interest because it is often claimed to be an important part of the
difference between valve and SS amplifiers.

It is indeed an old claim, and I remain vexed as to why
an SET amp with a pile of thd blows away some
SS amps with 100 times less.

THD is not a good measure of how it will sound.


The Halcro SS amp is said to make 0.0001% thd at 200 watts,
declining to unmeasurable levels below that high power.
Yet already I have heard ppl say that SE tube amps still are better.


Quite a coincidence, but have a look at this thesis (not mine) that I just
posted about http://web.mit.edu/cheever/www/cheever_thesis.pdf. It
addresses these issues


Another aspect of that difference is said to be how the distortion
is distributed with respect to power output.

Indeed. It is very important.

" wrote:

Patrick Turner wrote in
:



Ian Iveson wrote:

Occasionally we have discussed the kinds of distortion that most
please us, or that we find least objectionable. The subject is of
interest because it is often claimed to be an important part of the
difference between valve and SS amplifiers.

It is indeed an old claim, and I remain vexed as to why
an SET amp with a pile of thd blows away some
SS amps with 100 times less.

THD is not a good measure of how it will sound.


The Halcro SS amp is said to make 0.0001% thd at 200 watts,
declining to unmeasurable levels below that high power.
Yet already I have heard ppl say that SE tube amps still are better.

Quite a coincidence, but have a look at this thesis (not mine) that I just
posted about http://web.mit.edu/cheever/www/cheever_thesis.pdf. It
addresses these issues

Interesting, and it would take some time to study.
I did not see any tube amp recipes listed in the contents to bring anyone
closer to better sound, whatever that may mean.

Thanks for the reference.

Patrick Turner.




Another aspect of that difference is said to be how the distortion
is distributed with respect to power output.

Indeed. It is very important.

Patrick Turner wrote in
:


Quite a coincidence, but have a look at this thesis (not mine) that I
just posted about http://web.mit.edu/cheever/www/cheever_thesis.pdf.
It addresses these issues

Interesting, and it would take some time to study.
I did not see any tube amp recipes listed in the contents to bring
anyone closer to better sound, whatever that may mean.


He tests a 45 DHT based amp. It, of course, gets a much better T.A.D.,
hence is a good example of subjective 'quality' being indicated by his new
rating method. Is that what you meant?


Thanks for the reference.


NP

Patrick Turner.

On Fri, 07 Nov 2003 03:50:08 +0000, the highly esteemed Ian Iveson
enlightened us with these pearls of wisdom:

Occasionally we have discussed the kinds of distortion that most please
us, or that we find least objectionable. The subject is of interest
because it is often claimed to be an important part of the difference
between valve and SS amplifiers.

Another aspect of that difference is said to be how the distortion is
distributed with respect to power output.

What is the ideal distribution? Let me give two simple examples. Amp 1
produces zero distortion at 1mW, rising steadily to 1% at 10W. Amp 2
starts at zero and rises exponentially to 5% at 10W. Assuming that, for a
particular piece of music at a particular volume, the total distortion for
the whole piece is the same, and the distribution of distortion components
is the same, how will the two amps differ in their contribution to the
sound, and which might most please, or least offend? What is the best
shape?

Personal opinions or objective criteria welcome.

cheers, Ian

I don't believe that distortion, per se, is what makes the amplifier sound
good or bad. As Patrick pointed out in his reply, the Halcro amp has a
miniscule amount of distortion at full power, and virtually undetectable
amounts at lower levels. Yet, many people (outside of $tereophile) don't
find this otherwise superb piece of engineering to be the pinnacle of
musicality, and still prefer tube gear or other SS amps. I have no doubt
that the Halcro is as neutral as a Swiss passport and would be an ideal
reference amplifier for the "accuracy freak", even though it is debatable
as to whether this really constitutes true accuracy. The reproduced sound
is still not going to be identical to the actual recorded session - heck,
even the recording isn't an actual representation of the session.

I believe what is more important is how the amplifier interacts with the
complex loudspeaker load. For that matter, the load presented to each
active device alters its behavior, as we all know. While dyed-in-the-wool
engineers, $tereophile reviewers, and the "gurus" on RAO would have you
believe that a zero-output-impedance amp (i.e. a "perfect" voltage source)
is most desirable, I disagree. I don't believe that simply raising the
output impedance is tha answer (i.e. by inserting resistance in the
speaker lead). Rather, it is how that complex load affects the amplifiers
behavior, and vice versa. It is my opinion that this interaction, not
simple distortion or other individual characteristics, is what causes an
amplifier to "sound" the way it does. It is why the technically poor
performing SE amps sound so good, despite all their inherent "flaws".
It isn't "euphonic distortion", which is what the pundits dismiss it as.
It is more of an interaction response.

Whatever it is, it sounds good and I LIKE IT! So there! :-)

--
Greg

--The software said it requires Win2000 or better, so I installed Linux.

" wrote:

Patrick Turner wrote in
:


Quite a coincidence, but have a look at this thesis (not mine) that I
just posted about http://web.mit.edu/cheever/www/cheever_thesis.pdf.
It addresses these issues

Interesting, and it would take some time to study.
I did not see any tube amp recipes listed in the contents to bring
anyone closer to better sound, whatever that may mean.


He tests a 45 DHT based amp. It, of course, gets a much better T.A.D.,
hence is a good example of subjective 'quality' being indicated by his new
rating method. Is that what you meant?

I will have to read the thesis further to find out more and about what you
mean
about T.A.D.

The subjective experiences of audio are what is most important,
and to get the best subjective experiences, the measurements only
need to be good enough, and any betterment in measure,
may/may not do very much.
I really don't know why it is that a single ended tube design
can make music so more live, and enjoyable, when
some touted great measuring SS designs have you thinking
the sound is dry, clinical, and lifeless.
Sure, its accurate, but where is the life? is not the life a part of accuracy?

Some might say any recording has had the sound go via
piles of crap SS opamps in mixing desks and the like,
yet oh what a difference tubes make at the end!

Some things are there just to puzzle us.

Patrick Turner.


Thanks for the reference.


NP

Patrick Turner.

Patrick Turner wrote in
:


He tests a 45 DHT based amp. It, of course, gets a much better
T.A.D., hence is a good example of subjective 'quality' being
indicated by his new rating method. Is that what you meant?

I will have to read the thesis further to find out more and about what
you mean
about T.A.D.

T.A.D. is this guys figure of merit. The idea is to use it, rather than
T.H.D. and other specs in evaluating how good an amp will sound.


The subjective experiences of audio are what is most important,
and to get the best subjective experiences, the measurements only
need to be good enough, and any betterment in measure,
may/may not do very much.

As you'll see in the article, he establishes that good distortion figures,
for example, may sound worse, if they are the wrong types, such as high
order harmonics. A good T.H.D. may result, but it will sound worse than an
amp, that might have poorer T.H.D., but with the distortion is places where
it is masked. He looks at the natural distortion occurring in the ear
itself, and how real distortion in the amp can be masked if it coincides
with where the brain expects the ear generated distortion to be. If I read
him correctly, it also matters that the distortion is in harmonics in the
correct ratios. I think it happens that distortion in a more natyrally
linear device (like a DHT) is closer to this, and further, if no global NFB
is applied, the harmonics are not pushed higher - where, even though they
are lower in magnitude, they are more irritating and perceivable to the
brain.

I really don't know why it is that a single ended tube design
can make music so more live, and enjoyable, when
some touted great measuring SS designs have you thinking
the sound is dry, clinical, and lifeless.
Sure, its accurate, but where is the life? is not the life a part of
accuracy?

Some might say any recording has had the sound go via
piles of crap SS opamps in mixing desks and the like,
yet oh what a difference tubes make at the end!

Some things are there just to puzzle us.

Ain't it true!

"Ian Iveson" wrote in message
...
Occasionally we have discussed the kinds of distortion that most
please us, or that we find least objectionable. The subject is of
interest because it is often claimed to be an important part of the
difference between valve and SS amplifiers.

Another aspect of that difference is said to be how the distortion
is distributed with respect to power output.

What is the ideal distribution? Let me give two simple examples. Amp
1 produces zero distortion at 1mW, rising steadily to 1% at 10W. Amp
2 starts at zero and rises exponentially to 5% at 10W. Assuming
that, for a particular piece of music at a particular volume, the
total distortion for the whole piece is the same, and the
distribution of distortion components is the same, how will the two
amps differ in their contribution to the sound, and which might most
please, or least offend? What is the best shape?

Personal opinions or objective criteria welcome.

cheers, Ian



Hi, Ian,
Valve amps almost certainly sound different/more pleasureable because of the
distortion they introduce. Where global NFB from the OPT secondary is used,
this effect can easily be seen by putting a pot in the NFB line, and
adjusting the volume control to compensate for the increased gain
This is all subjective and influenced by the listening room and the speakers
you use but it is my impression that with NFB, the sound is more tightly
controlled and is obviously generated by the speakers. Without NFB, the
sound steps out from the speakers, ie. the sound stage widens and comes
forward, and the sound becomes warmer and fuller.
You could almost use it as a 'goodness' control.

Whether you can produce a 'best sound' distortion generator is unlikely.
Even harmonics are pleasureable to the ear, odd ones are not, so it is not
just the % total distortion which matters, it is the breakdown of what this
'distortion' actually is...
I'm pretty dim, as you know, and measure distortion with a sharp stick and
six cans of lager, ....
That's not a joke, really, it's a measure of what is pleasureable to me...
I actually have a very clever 'scope which stacks up input/output and can
add/subtract etc.. etc..

I also appreciate that my speakers will introduce 99.99% distortion in
converting electrical energy into sound and that if my wife rearranges the
furniture, I could have to rebuild the power amp.

After 35 years of ****ing about, I think I may have got a handle on this
stuff...... Transistors are unnatural, American, things.......... Valves
are warm, friendly and British..... Distortion was invented in 1953 and we
have worried about it ever since..... A third harmonic is irrelevant if you
can apply a sixth harmonic half as fast with an RC network..... Beam
tetrodes generate mainly even harmonics and pentodes generate mainly odd
ones..

I am currently listening to a 2 x 6L6GC x 2 stereo pair, something like
Williamson, and the big Zenith disco speakers

It actually sounds much better with no NFB.......

kind regards
jim

Greg Pierce wrote:

On Fri, 07 Nov 2003 03:50:08 +0000, the highly esteemed Ian Iveson
enlightened us with these pearls of wisdom:

Occasionally we have discussed the kinds of distortion that most please
us, or that we find least objectionable. The subject is of interest
because it is often claimed to be an important part of the difference
between valve and SS amplifiers.

Another aspect of that difference is said to be how the distortion is
distributed with respect to power output.

What is the ideal distribution? Let me give two simple examples. Amp 1
produces zero distortion at 1mW, rising steadily to 1% at 10W. Amp 2
starts at zero and rises exponentially to 5% at 10W. Assuming that, for a
particular piece of music at a particular volume, the total distortion for
the whole piece is the same, and the distribution of distortion components
is the same, how will the two amps differ in their contribution to the
sound, and which might most please, or least offend? What is the best
shape?

Personal opinions or objective criteria welcome.

cheers, Ian

I don't believe that distortion, per se, is what makes the amplifier sound
good or bad. As Patrick pointed out in his reply, the Halcro amp has a
miniscule amount of distortion at full power, and virtually undetectable
amounts at lower levels. Yet, many people (outside of $tereophile) don't
find this otherwise superb piece of engineering to be the pinnacle of
musicality, and still prefer tube gear or other SS amps. I have no doubt
that the Halcro is as neutral as a Swiss passport and would be an ideal
reference amplifier for the "accuracy freak", even though it is debatable
as to whether this really constitutes true accuracy. The reproduced sound
is still not going to be identical to the actual recorded session - heck,
even the recording isn't an actual representation of the session.

I believe what is more important is how the amplifier interacts with the
complex loudspeaker load. For that matter, the load presented to each
active device alters its behavior, as we all know. While dyed-in-the-wool
engineers, $tereophile reviewers, and the "gurus" on RAO would have you
believe that a zero-output-impedance amp (i.e. a "perfect" voltage source)
is most desirable, I disagree. I don't believe that simply raising the
output impedance is tha answer (i.e. by inserting resistance in the
speaker lead). Rather, it is how that complex load affects the amplifiers
behavior, and vice versa. It is my opinion that this interaction, not
simple distortion or other individual characteristics, is what causes an
amplifier to "sound" the way it does. It is why the technically poor
performing SE amps sound so good, despite all their inherent "flaws".
It isn't "euphonic distortion", which is what the pundits dismiss it as.
It is more of an interaction response.

Whatever it is, it sounds good and I LIKE IT! So there! :-)

We don't have to understand the stars we see from so far away
to feel their majesty.

The interaction between speakers and amps, and the dynamic
electron flow, and the dynamically varying cathode emission
are perhaps all not expressed in terms of thd, or IMD,
using just two steady sine waves.
The best test is something akin to music,
and I like Mozartian and Bethovenistic waves to have some idea of
clarity, punch, detail, sweetness, warmth, done
by listening forawhile with a few friends gathered round.
Sure, I got a few meters, and I measure fervently,
and the numbers add up, but at the end of the day, is it any better
than what I made last week?

Those technically poor measuring SE amps are not so bad measuring
if you get the power required from them down to low levels per
unit of vacuum.
The 8 watt 300B SET amp is seen as a wimp of a thing by many,
especially those troglodytes who cannot bear to be without
200 watts of muscle power offered by some SS behemouth.

I have a colleague in Sydney who makes more amps than I do,
and he built a couple of 80 watt amps for a cane farmer using 10 x 300B all in
parallel.
On a good day, maybe they expected to make 10 watts.
But average levels of music would be gained with just 2 watts,
like most ppl with normal modern speakers,
so the amps were working so only 0.2 watts were made by each 300B.

If we draw a curve of the typical thd of an SET,
we may find we have 5% at 8 watts, which is 8vrms.
at 4vo, ie, 2 watts, thd might be 1%,
and at 1.25 vo, ie, 0.2 watts, it might be 0.3%
Its depressingly high, compared to a PP amp.

So our friend up north with a cane farm still
gets 0.3% at 2 watts, even with 10 x 300B on his side,
( for each channel )

Methinks that if the RL was changed so the load for such an amp was
raised to halve the maximum power output,
then he'd get 40 watts, but maybe 1/3 the thd at 40 watts,
mainly because the load line was more horizontal
and for triodes, the more horizontal the load line is, the less thd you get.
So perhaps at 2 watts, the farmer might hear 0.15%.
I'd have to calculate it out to verify, maube somebody would like to.

Another alternative is to use horn speakers, to reduce our power needs,
and further reduce thd, because we are away from the high thd region.

The third alternative is to use a shirtload of NFB.
I might wait until those who don't like NFB to finish vomiting
up over their PCs, before moving right along.

The fourth alternative is to use voltage canceling
of the distortions. This isn't a feedback method of thd minimisation.
Of all the electronic freebies, 2H distortion current cancelling
seems to have become pride of place very early on in the history
of amplifiers.
This is achieved in the everyday use of PP circuits where the 2H currents
are juxteposed to cancel each other, and you get only 3H.
PP 300B amps are remarkably linear.

The fifth alternative is to retain the idea of SE circuitry,
but use either SEUL, or preferably acoustical,
and use voltage canceling of thd.
Paralleling many tubes is possible, which favours simplicity
of transformer design, since although the OPT has to be large,
it can have fewer turns of thicker wire than if winding
something small and fiddly for one tube.
The use of multigrid tetrodes and pentodes can be entertained,
except that we have a distortion problem because
with plain old tetrode or pentode, you get a lot more
nasty thd/imd products besides the 2H, which changes level and phase
with the RL you choose.
But if the multi grids are arranged in SEUL, there is a magical change to the
harmonic spectra, from lots of harmonics, down to
mainly only 2H, like a triode.
The trouble is that the SEUL's 2H is no better than a pentode's 2H,
although it is less load sensitive.
The Acoustical connection, like Quad II, but done SE,
also has the same effect on the harmonic spread,
but because the local NFB acts in not only the G2 circuit, but also around the
G1
circuit, the effect on Ra and thd/imd is far more impressive than plain old UL.
Also we can tailor G2 supply voltage to any point we want.

The result of using acoustical with say 16% CFB means you have maybe only
2% of thd at 9 watts from an EL34, all mainly 2H,
and the Ro is lower than a triode, and the drive requirements are barely worse
than a triode.

I first woke up about SE acoustical many years ago by adding turns to a speaker
secondary on a small OPT, and using this longer winding
as a FB winding to the cathode of the EL84, as well as
using 1/2 the winding as a speaker winding.
For many years, it was my delightful workshop radio amp,
which made the FM and AM sound sweet as.

With larger power octals, the result is that if we needed say 35vrms voltage
input to this SE Acoustical stage, then the driver thd, if it is a humble 12AU7,
or 6CG7, etc,
will also make about 2% at 35 vrms even if its set up optimally
with a nice high value RL.
And not only that, the rate of increase of thd is similar
to the rate of increase in thd of the output stage, at least for the rated load.

The result is that the 2H VOLTAGE of the driver cancels most of the
2H voltage produced in the output stage.
Its not a perfect scenario, but without trying too hard,
its possible to get a 12 dB reduction in 2H right across the
power width. Nothing is perfect, and one gets some 3H
at the "end of the voltage swings" but that's not a worry at
the lower power level, where we want the beauty of the tubes
to shine through the murk of the distortion.
Some 4H is made in this process, but its at a very low level.

If some additional global NFB is applied, it don't have to be such a high
amount,
and we end up with an SE amp which measures better than many PP varieties,
throughout the sweetzone, ie, the first few watts.

At present, I am building a pair of 45 watt amps using this fifth SEA idea
with 4 x EH 6CA7.
It is possible to plug in 3 x KT88/6550, or 4 x 6L6/5881EL34/KT66.
and I have NO clue as to whether they will really shine
as well as some of the other things I have tried.
But unless something is tried, nothing is known,
and no amount of hot air on the subject here at RAT would ever predict
an outcome.
That will be decided in the coming months,
with listening sessions where some other amps are used as references.

A sixth alternative is to use the acoustical and have a decent
LTP triode phase inverter, which is loaded with a high impedance DC supply
element, such as a choke with a CT, and then RC coupled to the outputs.
Thus the driver thd is minimised, and the CFB in the output stage is also
minimised,
and that leaves only the input SE triode to worry about, and
in my latest efforts on all this, the thd is predominantly 2H from the
input triode, not the 3H one sees from a totally balanced amp.
Its usually under 0.1% up to 1/3 full power, since I use some 8 dB of global FB,

in addition to the 8 dB CFB in the output stage.
I found this sounds better than plain UL, with all the FB being
16 dB of global.

If I were to use a SRPP triode input circuit, the thd could probaly be reduced
at least
8 dB in the first 1/3 of the power spread.

A seventh alternative is to use solid state, but that's a whole other
series of conundrums.

So yeah, I guess there are quite a few ways to design distortion, that's for
sure.
You can either put up with heaps of it,
or reduce it by designing it out of the picture.

If a lot of class A is used, either in SE, or PP,
there's a real chance of real fidelity.


Patrick Turner.



--
Greg

--The software said it requires Win2000 or better, so I installed Linux.

" wrote:

Patrick Turner wrote in
:


He tests a 45 DHT based amp. It, of course, gets a much better
T.A.D., hence is a good example of subjective 'quality' being
indicated by his new rating method. Is that what you meant?

I will have to read the thesis further to find out more and about what
you mean
about T.A.D.

T.A.D. is this guys figure of merit.

I thought it might mean Total Audiophile Discernment, or such like.

The idea is to use it, rather than
T.H.D. and other specs in evaluating how good an amp will sound.

I have to read a bit.....

The subjective experiences of audio are what is most important,
and to get the best subjective experiences, the measurements only
need to be good enough, and any betterment in measure,
may/may not do very much.

As you'll see in the article, he establishes that good distortion figures,
for example, may sound worse, if they are the wrong types, such as high
order harmonics. A good T.H.D. may result, but it will sound worse than an
amp, that might have poorer T.H.D., but with the distortion is places where
it is masked.

Yes, with a high amount of 2H and a low amount of applied NFB,
and a poor open loop bandwidth, although you reduce the 2H,
the 3H gets higher, and other H that were not there at all begin to appear,
and thay appear where open loop gain is low, so they are not reduced much.

He looks at the natural distortion occurring in the ear
itself, and how real distortion in the amp can be masked if it coincides
with where the brain expects the ear generated distortion to be.

And when you get older, the ears have reduced dynamic range,
and changes occur which ca make you rather critical of the sound your'e
listening to, whilst when you are young, almost anything goes;
the mind is on other things, such as the shiela across the room who is grinning
at you.....

If I read
him correctly, it also matters that the distortion is in harmonics in the
correct ratios. I think it happens that distortion in a more natyrally
linear device (like a DHT) is closer to this, and further, if no global NFB
is applied, the harmonics are not pushed higher - where, even though they
are lower in magnitude, they are more irritating and perceivable to the
brain.

There is a weighting formula for thd, the effective % equivalent to
an amount of 2H is = Nsquared / 4, where N is the harmonic.
So if you have say 0.1% of 5th H, then its equal to 25 / 4 times 0.1%
of 2H, or equl to 0.625% of 2H.
Things get crook when you have 0.1% of 13th H, its equal to
4.2 % of 2H.
This is terribly over simplified, and a bit silly, since when are you ever
going to get the 13th H at 0.1% in a class A tube amp at normal listening?
Some dispute the weighting and say it should be N cubed / 4.

PPL rave about leaving cathodes unbypassed in small signal triode amps,
and typically, 3 db of current NFB is applied.
This generates some 3 H, but where the innitial 2H is low,
the 3H produced isn't high.

Patrick Turner.



I really don't know why it is that a single ended tube design
can make music so more live, and enjoyable, when
some touted great measuring SS designs have you thinking
the sound is dry, clinical, and lifeless.
Sure, its accurate, but where is the life? is not the life a part of
accuracy?

Some might say any recording has had the sound go via
piles of crap SS opamps in mixing desks and the like,
yet oh what a difference tubes make at the end!

Some things are there just to puzzle us.

Ain't it true!

"jim" wrote in
:


Whether you can produce a 'best sound' distortion generator is
unlikely. Even harmonics are pleasureable to the ear, odd ones are
not, so it is not just the % total distortion which matters, it is the
breakdown of what this 'distortion' actually is...


Yes and it's not just even vs odd apparently. The best mix may well be
that that tracks the internal distortion 'envelope' of the ear in terms of
which harmonics, and what levels. The brain may be happier if it matches
this, because it has learned that that type of distortion is 'normal' or
natural, and cal successfully null it out.

wrote in message
...
"jim" wrote in
:


Whether you can produce a 'best sound' distortion generator is
unlikely. Even harmonics are pleasureable to the ear, odd ones are
not, so it is not just the % total distortion which matters, it is the
breakdown of what this 'distortion' actually is...


Yes and it's not just even vs odd apparently. The best mix may well be
that that tracks the internal distortion 'envelope' of the ear in terms of
which harmonics, and what levels. The brain may be happier if it matches
this, because it has learned that that type of distortion is 'normal' or
natural, and cal successfully null it out.

This stuff is not quantifiable... Chasing a minimum distortion figure, in
electrical terms, is not related to sound...... Sound and electronics are
only very loosely related.. Sound is about speakers and rooms.... Rooms are
external resonant chambers which the speakers work into... You pack a
speaker cabinet with sound absorbent material to modify or nullify its
resonant frequency...... move the furniture round and you can achieve the
same effect. You want to hear what effect the room has ...... set up the
speakers in the garden and see how they sound.. Reflected sound within the
listening room is as important as the sound directly radiated by the
speakers...
Then we move on to how my ears/brain perceive things relative to yours,
given that anything over 12kHz is unknown territory, now, to me, at 53, but
you may hear out to 16Khz and tell me the sound is thin and lacking in bass
??
My brain hurts
regards
jim

" wrote:

"jim" wrote in
:


Whether you can produce a 'best sound' distortion generator is
unlikely. Even harmonics are pleasureable to the ear, odd ones are
not, so it is not just the % total distortion which matters, it is the
breakdown of what this 'distortion' actually is...


Yes and it's not just even vs odd apparently. The best mix may well be
that that tracks the internal distortion 'envelope' of the ear in terms of
which harmonics, and what levels. The brain may be happier if it matches
this, because it has learned that that type of distortion is 'normal' or
natural, and cal successfully null it out.

Gee, what does your audiologist say about this.

I guess there is some 2,3,4,5,6,7,8,9H produced by
the mechanisms in the ear, and one wonders which exact profile they'd have,
but perhaps every person is different, so designing
to null out, or cancel, or compliment the human Dn might be very difficult
indeed.

I tend to think all distortion is bad, and some types
are worse than others.

Patrick Turner.

"jim" wrote in
:

Yes and it's not just even vs odd apparently. The best mix may well
be that that tracks the internal distortion 'envelope' of the ear in
terms of which harmonics, and what levels. The brain may be happier
if it matches this, because it has learned that that type of
distortion is 'normal' or natural, and cal successfully null it out.

This stuff is not quantifiable...

No?


Chasing a minimum distortion
figure, in electrical terms, is not related to sound......

I thought the sound was what it was all about.

Sound and
electronics are only very loosely related.. Sound is about speakers
and rooms....

All part of it yes.

Rooms are external resonant chambers which the speakers
work into... You pack a speaker cabinet with sound absorbent material
to modify or nullify its resonant frequency...... move the furniture
round and you can achieve the same effect. You want to hear what
effect the room has ...... set up the speakers in the garden and see
how they sound.. Reflected sound within the listening room is as
important as the sound directly radiated by the speakers...

er, yes, however, still one amp can sound better than another, all else
being the same.

Then we move on to how my ears/brain perceive things relative to
yours, given that anything over 12kHz is unknown territory, now, to
me, at 53, but you may hear out to 16Khz and tell me the sound is thin
and lacking in bass ??

and still, we may both decide one amp is better than another, and it may
not be the amp with lowest distortion and/or other 'specs'.


My brain hurts
regards
jim

Patrick Turner wrote in
:

Yes and it's not just even vs odd apparently. The best mix may well
be that that tracks the internal distortion 'envelope' of the ear in
terms of which harmonics, and what levels. The brain may be happier
if it matches this, because it has learned that that type of
distortion is 'normal' or natural, and cal successfully null it out.

Gee, what does your audiologist say about this.

That's where I got the idea :-)

I guess there is some 2,3,4,5,6,7,8,9H produced by
the mechanisms in the ear, and one wonders which exact profile they'd
have,

studies have apparently been done

but perhaps every person is different, so designing
to null out, or cancel, or compliment the human Dn might be very
difficult indeed.

Designing to match it might be difficult, but I think the point is that
amps that we subjectively say sound better, are the ones that are more
naturally like it. The further from it, the less we tolerate the
aberrations.


I tend to think all distortion is bad, and some types
are worse than others.


Exactly. The weighting against higher harmonics, that you mentioned in
another post, is an example. But there may be a better weighting than the
simple mathematical formula you mentioned. It may be that the weighting
should be the inverse of that which the brain can automatically discount,
or stated differently, that which is masked by the naturally occurring
distortions of the ear.

There is a sort of plausibility to this explanation - which is better put
in the thesis I posted about, that appeals.

There is a sort of plausibility to this explanation - which is better put
in the thesis I posted about, that appeals.



** Are you still talking about hi-fi - or is it Scientology ???




........... Phil

" wrote:

"jim" wrote in
:

Yes and it's not just even vs odd apparently. The best mix may well
be that that tracks the internal distortion 'envelope' of the ear in
terms of which harmonics, and what levels. The brain may be happier
if it matches this, because it has learned that that type of
distortion is 'normal' or natural, and cal successfully null it out.

This stuff is not quantifiable...

No?

Chasing a minimum distortion
figure, in electrical terms, is not related to sound......

I thought the sound was what it was all about.

Well when using electronics, some attention has to be paid to N&D.

Its like going on a boat cruise, sure, the views are wonderful,
the babes are swanning around in bikinis,
and the martinies slide down the hatch nicely,
but if some dude leaves the muffler disconnected off the motor,
the noise kinda gets to ya.

Patrick Turner

" wrote:

Patrick Turner wrote in
:

Yes and it's not just even vs odd apparently. The best mix may well
be that that tracks the internal distortion 'envelope' of the ear in
terms of which harmonics, and what levels. The brain may be happier
if it matches this, because it has learned that that type of
distortion is 'normal' or natural, and cal successfully null it out.

Gee, what does your audiologist say about this.

That's where I got the idea :-)

I guess there is some 2,3,4,5,6,7,8,9H produced by
the mechanisms in the ear, and one wonders which exact profile they'd
have,

studies have apparently been done

but perhaps every person is different, so designing
to null out, or cancel, or compliment the human Dn might be very
difficult indeed.

Designing to match it might be difficult, but I think the point is that
amps that we subjectively say sound better, are the ones that are more
naturally like it. The further from it, the less we tolerate the
aberrations.


I tend to think all distortion is bad, and some types
are worse than others.

Exactly. The weighting against higher harmonics, that you mentioned in
another post, is an example. But there may be a better weighting than the
simple mathematical formula you mentioned. It may be that the weighting
should be the inverse of that which the brain can automatically discount,
or stated differently, that which is masked by the naturally occurring
distortions of the ear.

There is a sort of plausibility to this explanation - which is better put
in the thesis I posted about, that appeals.

Tryin to convertus plausibilitieum, and decipherum thesisiticus argumentus,
with explanatious complicateae thrown in for good measure
is beyond my triode addled brain at this late hour.

Patrick Turner.

wrote in message
...
"jim" wrote in
:

Yes and it's not just even vs odd apparently. The best mix may well
be that that tracks the internal distortion 'envelope' of the ear in
terms of which harmonics, and what levels. The brain may be happier
if it matches this, because it has learned that that type of
distortion is 'normal' or natural, and cal successfully null it out.

This stuff is not quantifiable...

No?



NO IT ISN'T !

Chasing a minimum distortion
figure, in electrical terms, is not related to sound......

I thought the sound was what it was all about.

Sound and
electronics are only very loosely related.. Sound is about speakers
and rooms....

All part of it yes.


A very big part of it ......

You pack a speaker cabinet with sound absorbent material
to modify or nullify its resonant frequency...... move the furniture
round and you can achieve the same effect. You want to hear what
effect the room has ...... set up the speakers in the garden and see
how they sound.. Reflected sound within the listening room is as
important as the sound directly radiated by the speakers...

er, yes, however, still one amp can sound better than another, all else
being the same.

Of course..... but you are talking sound and not distortion figures, and are
we in my room or yours ???

Then we move on to how my ears/brain perceive things relative to
yours, given that anything over 12kHz is unknown territory, now, to
me, at 53, but you may hear out to 16Khz and tell me the sound is thin
and lacking in bass ??

and still, we may both decide one amp is better than another, and it may
not be the amp with lowest distortion and/or other 'specs'.

But are we in the kitchen or the dining room with the supersoft Dunlopillo
three peice suite with my speakers, or yours ? Have you got a carpet with
12mm underlay or a hardwood floor ??
I keep saying it and somebody, one day, will take some notice........ sound
and electronics are only slightly related .......


Kind regards
jim

"Patrick Turner" wrote in message
...


" wrote:

"jim" wrote in
:


Whether you can produce a 'best sound' distortion generator is
unlikely. Even harmonics are pleasureable to the ear, odd ones are
not, so it is not just the % total distortion which matters, it is the
breakdown of what this 'distortion' actually is...


Yes and it's not just even vs odd apparently. The best mix may well be
that that tracks the internal distortion 'envelope' of the ear in terms
of
which harmonics, and what levels. The brain may be happier if it
matches
this, because it has learned that that type of distortion is 'normal' or
natural, and cal successfully null it out.

Gee, what does your audiologist say about this.

I guess there is some 2,3,4,5,6,7,8,9H produced by
the mechanisms in the ear, and one wonders which exact profile they'd
have,
but perhaps every person is different, so designing
to null out, or cancel, or compliment the human Dn might be very
difficult
indeed.

I tend to think all distortion is bad, and some types
are worse than others.

Patrick Turner.


C'mon, PT, stop taking the ****..... the profile of my ear is no different
to yours. After 50 it hears **** all above 12 kHz and our perception of
what is good or bad is irrelevant....Yer daft old *******....
Janice will **** about with the cushions in the front room and I might have
a resonance at 16kHz. Fortunately, I won't be able to hear it....... My
children might ...
All distortion isn't bad .. at our age, even harmonics are pleasureable
and odd ones are inaudible....
With a magnet and a paper cone which translates electricity into sound, in a
wooden box, with less than 1% efficiency and 99.9% distortion, or less, or
more, in my room, or yours, ... should I just snap open another can of
lager and look at the blue glow from the 6L6GCs ??
kind regards
jim

"jim" wrote in message
...
shortenin snip
: I tend to think all distortion is bad, and some types
: are worse than others.
:
: Patrick Turner.
:
:
: C'mon, PT, stop taking the ****..... the profile of my ear is no
different
: to yours. After 50 it hears **** all above 12 kHz and our perception of
: what is good or bad is irrelevant....Yer daft old *******....
: Janice will **** about with the cushions in the front room and I might
have
: a resonance at 16kHz. Fortunately, I won't be able to hear it....... My
: children might ...

Hm, well, things are not as clearcut there as you presume.
For one thing, perception isn't anything like what most people seem to
assume, some receptor(s) converting a stream of data from one domain
to another (electrical/chemical). It is an active process.
And what constitutes hearing ? You assume it's the air pressure modulation,
setting the mechanics of your ear into motion, eventually 'rustling' the
nerves. However, it has been established that acoustic information is
also conveyed to the area in the brain doing the processing by transmission
through the skull's bonestructure. This is the reason that people who
don't 'hear a damn thing' above 10 KHz, say, still have no problem
differentiating
between a sound A - unfiltered and a sound B - same, but LPF at 15 kHz.
And this is also a good argument against the boringly repeated 20 KHz
upper limit of reproduction being fine - it is not.

Just adding some info,
Rudy

: All distortion isn't bad .. at our age, even harmonics are pleasureable
: and odd ones are inaudible....
: With a magnet and a paper cone which translates electricity into sound, in
a
: wooden box, with less than 1% efficiency and 99.9% distortion, or less, or
: more, in my room, or yours, ... should I just snap open another can of
: lager and look at the blue glow from the 6L6GCs ??
: kind regards
: jim
:
:
:
:
:

"jim" wrote in message


I keep saying it and somebody, one day, will take some notice........
sound
and electronics are only slightly related .......


** But much more closely related to each other than anything a pommy
cretin called "jim" has ever said is related to the truth.




.............. Phil

"Ruud Broens" .


However, it has been established that acoustic information is
also conveyed to the area in the brain doing the processing by
transmission
through the skull's bonestructure.


** What ? Direct conversion in the grey matter ??

That has never been established by anyone.



This is the reason that people who don't 'hear a damn thing' above 10 KHz,
say, still have no problem
differentiating between a sound A - unfiltered and a sound B - same, but
LPF at 15 kHz.


** There are lies, dam lies, statistics - and then there is all the
****ing drivel posted by lunatics on usenet.



And this is also a good argument against the boringly repeated 20 KHz
upper limit of reproduction being fine - it is not.



** He says with all the science in the world contradicting him.




............... Phil

"Phil Allison" wrote in
:


There is a sort of plausibility to this explanation - which is better
put in the thesis I posted about, that appeals.



** Are you still talking about hi-fi - or is it Scientology
???





I have no problem when a well constructed scientific argument is
plausible.Of course, the whole idea of a new theory is that it should fit
the observed phenomena better than previous attempts. This one at least has
some correlation between subjective results and measured results.
Something that's been sorely lacking for some time now.

Why not have a read of the paper?

wrote in message
...
"Phil Allison" wrote in
:


There is a sort of plausibility to this explanation - which is better
put in the thesis I posted about, that appeals.

** Are you still talking about hi-fi - or is it Scientology
???


I have no problem when a well constructed scientific argument is
plausible.


** A well constructed *scientific* argument needs to be a damn site more
than simply "plausible" to some fool.


Of course, the whole idea of a new theory is that it should fit
the observed phenomena better than previous attempts.


** New theories ( hypotheses really) need to explain existing facts and
predict new ones - this desperate load of cobblers does neither.


This one at least has some correlation between subjective results and
measured results.


** Which is far inferior to the track record of established theory.


Something that's been sorely lacking for some time now.

** Bull****.



Why not have a read of the paper?


** Wallowing through that pile of lies, half truths and plain gobbledegook
from a demented audiophool gives any sane person a headache. That thesis is
now 14 years old and has been studiously ignored by everyone.

That means it is both worthless in terms of information AND worthless as
a money spinning idea.




.............. Phil

"Phil Allison" wrote in
:




Why not have a read of the paper?


** Wallowing through that pile of lies, half truths and plain
gobbledegook
from a demented audiophool gives any sane person a headache. That
thesis is now 14 years old and has been studiously ignored by
everyone.


Dated 2001, you have mistakenly read the date when he attained his BSEE.


That means it is both worthless in terms of information AND
worthless as
a money spinning idea.




............. Phil

"Jon Yaeger"

"Ruud Broens" .


However, it has been established that acoustic information is
also conveyed to the area in the brain doing the processing by
transmission through the skull's bonestructure.


** What ? Direct conversion in the grey matter ??

That has never been established by anyone.



Unfortunately you're wrong, Phil.


** No way I am wrong !!!!!

Go learn to read - you pathetic cretin.



The concept has been established for quite some time.

Quote:

" Bone conduction refers to the response of the bones of the skull to audio
and higher frequency vibrations, and to the transmission of such vibrations
to, and the reception by, the auditory organ. Direct stimulation of the
cochlea by a vibrator through bone conduction is routinely employed in
audiological evaluations ..... "


** I see no mention of direct conversion in the *brain* !!!!

Quite the opposite - the normal hearing organs are the receptors as
usual.





.............. Phil

"Ruud Broens"

:
OK, let me expand on that: it's the 'shaking' of the cochlean fluid
by this transmission that's causing electrical signals, obviously.
So we don't *perceive* this as hearing, but live tomography
imaging has shown a response to ultrasonic stimuli from the brain
without a shadow of a doubt.



** So this para of yours was really a pack of damn lies:


" However, it has been established that acoustic information is
also conveyed to the area in the brain doing the processing by transmission
through the skull's bonestructure. This is the reason that people who
don't 'hear a damn thing' above 10 KHz, say, still have no problem
differentiating
between a sound A - unfiltered and a sound B - same, but LPF at 15 kHz.
And this is also a good argument against the boringly repeated 20 KHz
upper limit of reproduction being fine - it is not. "





........... Phil

From: "Phil Allison"
Newsgroups: rec.audio.tubes
Date: Sat, 8 Nov 2003 09:30:23 +1100
Subject: Designing distortion


"Ruud Broens" .


However, it has been established that acoustic information is
also conveyed to the area in the brain doing the processing by
transmission
through the skull's bonestructure.


** What ? Direct conversion in the grey matter ??

That has never been established by anyone.


Unfortunately you're wrong, Phil. The concept has been established for
quite some time.

For example, see

http://216.239.37.104/search?q=cache...aquiet.com/Pag
es/Human%2520Skull%2520Response.PDF+sound+transmissio n+cranium&hl=en&ie=UTF-
8


- Jon

jim wrote:

"Patrick Turner" wrote in message
...


" wrote:

"jim" wrote in
:


Whether you can produce a 'best sound' distortion generator is
unlikely. Even harmonics are pleasureable to the ear, odd ones are
not, so it is not just the % total distortion which matters, it is the
breakdown of what this 'distortion' actually is...


Yes and it's not just even vs odd apparently. The best mix may well be
that that tracks the internal distortion 'envelope' of the ear in terms
of
which harmonics, and what levels. The brain may be happier if it
matches
this, because it has learned that that type of distortion is 'normal' or
natural, and cal successfully null it out.

Gee, what does your audiologist say about this.

I guess there is some 2,3,4,5,6,7,8,9H produced by
the mechanisms in the ear, and one wonders which exact profile they'd
have,
but perhaps every person is different, so designing
to null out, or cancel, or compliment the human Dn might be very
difficult
indeed.

I tend to think all distortion is bad, and some types
are worse than others.

Patrick Turner.


C'mon, PT, stop taking the ****..... the profile of my ear is no different
to yours.

I am sure some audiolgists would disagree strongly,
since they test ppl routinely for F response, and there are big differences,
not just between people, but between L and R ears.

After 50 it hears **** all above 12 kHz and our perception of
what is good or bad is irrelevant....Yer daft old *******....

Yeah, but heck, although we live in a decentructionalist age,
where dudes have their own realities, not all determined by the Catholic Church,

I still reckon I need to say ********!, or Balderdash!,
or What a load of Crap!, during solemn and serious meetings where
things are discussed earnestly.


Janice will **** about with the cushions in the front room and I might have
a resonance at 16kHz. Fortunately, I won't be able to hear it....... My
children might ...
All distortion isn't bad .. at our age, even harmonics are pleasureable
and odd ones are inaudible....

I actually find that the less distortion the better,
and it lets the tubes get on with what they are good at,
and that's handling music.


With a magnet and a paper cone which translates electricity into sound, in a
wooden box, with less than 1% efficiency and 99.9% distortion, or less, or
more, in my room, or yours, ... should I just snap open another can of
lager and look at the blue glow from the 6L6GCs ??
kind regards
jim

Rooms and lager are important, Mr Jim, I know, even blue glows are pretty,
but I like it all without much thd/imd.

Those who like to add it in are welcome to do so.

Each unto their own.

Patrick Turner.

Patrick Turner wrote in
:


Those who like to add it in are welcome to do so.


Just so long as you don't think I was saying you need to add to it. I
think that you have to be careful where it is though, and not just reah
conclusions on the basis of a low thd/imd figure alone.

If an amp has low values for these figures, and the distortions consist of
the more benign harmonics, you are a long way towards having a truly great
amp. If, however, they are nasty harmonics, then you might be better off
with another amp, maybe with and despite less impressive thd/imd.

wrote in message
...
Patrick Turner wrote in
:


Those who like to add it in are welcome to do so.


Just so long as you don't think I was saying you need to add to it. I
think that you have to be careful where it is though, and not just reah
conclusions on the basis of a low thd/imd figure alone.

If an amp has low values for these figures, and the distortions consist of
the more benign harmonics, you are a long way towards having a truly great
amp. If, however, they are nasty harmonics, then you might be better off
with another amp, maybe with and despite less impressive thd/imd.


** Has the concept behind the term "inaudible" ever entered your
calculations ??

Have you ever considered the real world "time domain" view of amplifier
performance instead of the artificial "frequency domain" one ?? (
Cheever ( Cheater?) avoided it like a zombie avoids garlic )

In the time domain there is a continuous error signal - defined as the
difference voltage at any instant between what the output should* be
compared to what it is. ( * the input signal times some fixed number.)

If this error signal were to be zero with music programme and driving a
real speaker - then the amp would be *absolutely perfect*.

If this error signal could be maintained at a level during reproduction
of musical programme that was always below the threshold of human hearing -
then the amp would be *audibly perfect*.

Due to the masking effect of one sound on another smaller one, even the
above criterion is much too stringent for no audible imperfection to be the
result.

Amplifiers that meet the criterion of "audible perfection" are on sale
everywhere.


** The reasons folk often do not like the reproduced sound from
commercial recordings lie elsewhere than the amplification.

Adding any deliberate non linearity is about as desperate as someone
re-arranging the deck chairs on the Titanic as she goes down.





............ Phil

" wrote:

Patrick Turner wrote in
:


Those who like to add it in are welcome to do so.


Just so long as you don't think I was saying you need to add to it. I
think that you have to be careful where it is though, and not just reah
conclusions on the basis of a low thd/imd figure alone.

And I like to get it low BEFORE adding some NFB to reduce it futher.
With class A designs, there is usually very little harmonics
other than the 2nd or 3rd in the class A part of the working range
used for listening.
I prefer that the thd up to the first 10 watts is mainly all 2H, and less than

1%, before nfb is added, and when it is, you just need enough
to reduce the Ro of ther amp to suit the speakers.
This usually means a total of about 20 dB for pentode amps, 15 db for UL,
and 12 dB for triode amps, but feel free to experiment.
As I have said before, a couple of SE 300B paralleled
can be mated with a 5k to 5 ohm OPT and each run far below max ratings,
and the Ro is 0.4 ohms, with no FB, but you may only get 6 watts,
but the thd will be very low.

The triode has its own internal electrostaic loop
of FB built in.

It was put there by the God of Triodes, and it isn't
something done with external loops of NFB, which is mankind's
trickery, to correct errors.
Trickeratory behaviour was put into man's craniums by the General God,
who has tricked us about a lot of things...........

But a world full of amps with no feedback and only triodes
would be a real set back for the world, when you think about it,
away from the passions of a few audiophiles, because the mainstream
world seems quite content with opamps, transitor electronics,
all loaded to the hilt with FB paths.

Many of the greatest recordings were made using studio gear with
pentode amps galore, all with loops of FB..


If an amp has low values for these figures, and the distortions consist of
the more benign harmonics, you are a long way towards having a truly great
amp. If, however, they are nasty harmonics, then you might be better off
with another amp, maybe with and despite less impressive thd/imd.

Its ok to have a high thd final amp for audio;
one gets away with it.
But where the audio has to go through 50 amps,
like a telephone signal or communications signal might,
or maybe a studio signal might,
then N&D should be very low in each device, hence
NFB has its use

If you had 50 x 300B amps all cascaded,
I suggest the sound at the output would be the purest mud.

Patrick Turner.

"Phil Allison" wrote in
:


Just so long as you don't think I was saying you need to add to it.
I think that you have to be careful where it is though, and not just
reah conclusions on the basis of a low thd/imd figure alone.

If an amp has low values for these figures, and the distortions
consist of the more benign harmonics, you are a long way towards
having a truly great amp. If, however, they are nasty harmonics,
then you might be better off with another amp, maybe with and despite
less impressive thd/imd.


** Has the concept behind the term "inaudible" ever entered your
calculations ??


My calculations?


Have you ever considered the real world "time domain" view of
amplifier

Yes

performance instead of the artificial "frequency domain" one ??

It simply a representation of reality. I have no problem with reality.


(
Cheever ( Cheater?) avoided it like a zombie avoids garlic )

In the time domain there is a continuous error signal - defined
as the
difference voltage at any instant between what the output should* be
compared to what it is. ( * the input signal times some fixed
number.)

yes



If this error signal were to be zero with music programme and
driving a
real speaker - then the amp would be *absolutely perfect*.

If this error signal could be maintained at a level during
reproduction
of musical programme that was always below the threshold of human
hearing - then the amp would be *audibly perfect*.

Fine except, it isn't error signal we listen to (but I know what you meant)



Due to the masking effect of one sound on another smaller one,
even the
above criterion is much too stringent for no audible imperfection to
be the result.

That is another effect that can be observed, yes.


Amplifiers that meet the criterion of "audible perfection" are on
sale
everywhere.


No. If that were true, then no-one would be able to observe a difference
between them.




** The reasons folk often do not like the reproduced sound from
commercial recordings lie elsewhere than the amplification.

All faults in a system contribute.



Adding any deliberate non linearity is about as desperate as
someone
re-arranging the deck chairs on the Titanic as she goes down.



I would never deliberately add non-linearity. Why would you want to?







........... Phil

Patrick Turner wrote in
:

nd I like to get it low BEFORE adding some NFB to reduce it futher.
With class A designs, there is usually very little harmonics
other than the 2nd or 3rd in the class A part of the working range
used for listening.
I prefer that the thd up to the first 10 watts is mainly all 2H, and
less than

1%, before nfb is added, and when it is, you just need enough
to reduce the Ro of ther amp to suit the speakers.
This usually means a total of about 20 dB for pentode amps, 15 db for
UL, and 12 dB for triode amps, but feel free to experiment.

But now the 'mainly 2H' is distributed (albeit at lower total levels) in
higher harmonics. Which one sounds better, that is what is important?

.... and would both sound better or worse than a super high open loop gain
monster with high global NFB? Critical listeners seem to be saying , yes.

wrote in message
...
"Phil Allison" wrote in
:


Just so long as you don't think I was saying you need to add to it.
I think that you have to be careful where it is though, and not just
reah conclusions on the basis of a low thd/imd figure alone.

If an amp has low values for these figures, and the distortions
consist of the more benign harmonics, you are a long way towards
having a truly great amp. If, however, they are nasty harmonics,
then you might be better off with another amp, maybe with and despite
less impressive thd/imd.


** Has the concept behind the term "inaudible" ever entered your
calculations ??


My calculations?

** Its was an allegorical remark.




Have you ever considered the real world "time domain" view of
amplifier

Yes

performance instead of the artificial "frequency domain" one ??

It simply a representation of reality.


** The frequency domain does not exist - it is a mathematical fiction.


I have no problem with reality.

** Says a man who believes in things that do not really exist.



(
Cheever ( Cheater?) avoided it like a zombie avoids garlic )

In the time domain there is a continuous error signal - defined
as the
difference voltage at any instant between what the output should* be
compared to what it is. ( * the input signal times some fixed
number.)

yes



If this error signal were to be zero with music programme and
driving a
real speaker - then the amp would be *absolutely perfect*.

If this error signal could be maintained at a level during
reproduction
of musical programme that was always below the threshold of human
hearing - then the amp would be *audibly perfect*.


Fine except, it isn't error signal we listen to (but I know what you
meant)


** It is this error signal that YOU allege *is* audible and spoils the
reproduced sound.




Due to the masking effect of one sound on another smaller one,
even the above criterion is much too stringent for no audible
imperfection to
be the result.


That is another effect that can be observed, yes.



Amplifiers that meet the criterion of "audible perfection" are on
sale everywhere.


No.

** Yes they are - you ignoramus.



If that were true, then no-one would be able to observe a difference
between them.


** That is true.

There is no audible difference do to non-linear error.

Most so called "differences" are in the imagination of the listener.

See this for a simple method to prove that:
http://sound.westhost.com/absw.htm




** The reasons folk often do not like the reproduced sound from
commercial recordings lie elsewhere than the amplification.

All faults in a system contribute.


** Absurd cant.


Adding any deliberate non linearity is about as desperate as
someone re-arranging the deck chairs on the Titanic as she goes
down.

I would never deliberately add non-linearity. Why would you want to?



** Your whole line of bull**** here ASSUMES that audible non-linearity
exists in all amps - or possibly needs adding to improve the sound. That
is a crazy fallacy.



.......... Phil

"Phil Allison" wrote in
:


It simply a representation of reality.


** The frequency domain does not exist - it is a mathematical fiction.


The two statements, yours and mine, and not contradictory.

wrote in message
...
"Phil Allison" wrote in
:


It simply a representation of reality.


** The frequency domain does not exist - it is a mathematical
fiction.


The two statements, yours and mine, and not contradictory.


** The actual words you posted we

"It simply a representation of reality. I have no problem with
reality."


** The contradiction is clear.


Now Mr Frog, if you can do no better that to continually snip virtually
my entire, carefully worded post out and merely post a dopey nit pick -
then kindly got get rooted.






................. Phil

"Phil Allison" wrote in
:


wrote in message
...
"Phil Allison" wrote in
:


It simply a representation of reality.


** The frequency domain does not exist - it is a mathematical
fiction.


The two statements, yours and mine, and not contradictory.


** The actual words you posted we

"It simply a representation of reality. I have no problem with
reality."


** The contradiction is clear.



There is no contradiction in my statements either.


Now Mr Frog, if you can do no better that to continually snip
virtually
my entire, carefully worded post out and merely post a dopey nit pick
- then kindly got get rooted.

There was nothing else to which I wished to reply.