[Trevor Wilson]

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

**It's not misleading. It's fact.

Nothing is a fact until all sides agree it is, and you have to proove what
you
say.

**Just look at the curves.






Unfortunately for your arguments this is not the case with regard to
voltage
gain.
Transistors need only a small range of turn on voltage applied between
their
base and emitter,
and their gm varies enormously between being on the threshold of turn
on
of
collector-emitter current and saturation.

**Transistors are controlled by CURRENT, not Voltage. The CURRENT gain of
modern transistors is extremely linear, over a wide range of Collector
currents and Voltages.

Current relationships between base and collector currents are
substantially
linear.

**EXTREMELY linear.

The voltage relationship is not.

**Irrelevant.

You don't have the concepts correctly in your mind.

**Sure I do. Modern BJTs are extremely linear amplification devices.





Only enormous amounts of externally applied NFB corrects the inherent
voltage
non-linearity
of all BJTs where there is substantial current change as in the case of
all
power transistors.

**Transistors (BJTs) are CURRENT amplification devices.

Sure, we all know that, but their voltage linearity is intrinsically poor.

**Er, that's because they're current amplifiers. And in that sense, they are
exceptionally linear.






Try setting up any bjt in common emitter mode without any emitter
resistor
and
apply
a voltage from a very low impedance to the base. The bjt is
disgustingly
non
linear
compared to a triode using no external applied NFB.

**Transistors (BJTs) are CURRENT amplification devices. Apply a current
to
the Base and measure the change in Collector current. Get back to me when
you've done this.

The current applied to the base is propelled by the applied voltage.
No voltage = no current.

**Duh. No current flow = no Voltage. What's your point?


Set up a small bjt in bread board as a an SE device in class A with zero
NFB and
record the
voltage linearity.

**Points:
* No amplifier on the planet uses zero NFB. Why would you want to attempt to
build one?
* Se amplifiers are dumb. Push pull makes more sense.

Its terrible.
Lots of NFB is needed to linearize the bjt, which is easy, because they
have a
lot of gain...


Even most pentodes which do not have internal electrostatic NFB like a
triode
can easily
out perform a bjt.

**Not even close. Pentodes are vastly less linear than even average BJTs.
Triodes, at least, are reasonably linear.

Your ignorance of tube behaviour is breathtaking.

Try reading RDH4 about the comparison of triode connected pentodes
and the same tube running as a pentode.
At low levels the pentode was more linear than the triode.

**Are we discussing POWER amps, or something else?





I am not here to discuss whether the bjt amp can sound well, or measure
well.
Both are known to be possible so it is believed by most folks.

**Me either. I'm here to correct Jute's lies.


Without external loops of NFB including the emitter follower connection
which is a high amount of applied series voltage NFB, the BJT is a
hopeless
voltage amp

**Transistors (BJTs) are CURRENT amplification devices.

No need to emulate a trained galah.

**It seems you've forgotten this point.




unless we try to restrict the current change in the device, and then
the
gain is

enormous and subject to the perils of the high collector resistance and
circuit slowness due to capacitances, and the non linear input
resistances
of
other
devices in the signal path.

A class A single bjt voltage amp stage with a single MJE340 with
another
for a
constant current source
collector load can have a gain of maybe 1,000, or like a pentode with a
CCS load
in its anode circuit.
If external loops of shunt NFB are applied around either device, then
the
linearity can be quite good for
either device, because for a gain reduction of 100, from 1,000 to 10,
the
thd is
also reduced about
100 fold.

The linearity is dependant on the NFB applied, and not on the basic
VOLTAGE
linearity
of the device without NFB, which for a bjt is relatively poor, and
poorest
when
considerable
current change occurs in the device.

**Transistors (BJTs) are CURRENT amplification devices.

Gee, I think I heard that parrot again.

**Until you figure that out, I need to repeat it.


Maybe you could confirm what i am saying if you
went right back to basics with a breadboard and a CRO
and examined the basic workings of the basic devices that mankind has
invented over the last 100 years.

**Sure. Right after you recall that BJTs are current amplification devices
and cannot be used in identical topologies to tubes. They operate in a
fundamentally different way and need to be treated accordingly.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[robert casey]

How does negative feedback work?

Negative feedback is simply a negative voltage fed back from the output
to the input amplifying device to offset part of the harmonic
distortion which is present as a positive voltage. It costs nothing
except a loss of gain and a few side effects such as phase shift and
possible instability which are well known in the mathematical
literature and more or less easily guarded against depending on the
level of NFB.

'Wow!' those meeting NFB for the first time will now say, 'Something
for free! I'll grab some of that for my amp.' Hey, I said it, and I am
a professional intellectual, by definition an infinite skeptic. NFB is
a thing of beauty that will draw you in. It is like an electronic
Marxism which admits of no contrary arguments because it has subsumed
them all into The Holy Measurements. To question The Measurements is to
commit heresy. You need to be of strong mind to resist the
blandishments of such a universal panacea and of strong stomach to
withstand the hysterical assaults of the lesser engineers defending
their holy grail. (And when you do get hold of a superior engineer to
explain NFB to you, you need to be high-domed indeed because suddenly
NFB can turn very intricate.)

Unfortunately NFB doesn't come without a price. It levies a cruel
charge on the perceived quality of the sound. Negative feedback is what
gives all those 'blameless' transistor and big PP tube amps their
chillingly unnatural sound.

Then how did NFB come to be such a panacea in amplifier design?

Your guess is as good as mine. Hi-fi design is not prestige work...
The exceptions to this rule are
normally audio enthusiasts in charge of their own small audio
manufactories with niche markets; those who grow larger from this base
follow the mainstream mantra of "mo' NFB give lowa' THD" because the
marketing channels demand it from them if they wish to grow. At this
point they usually cease to offer anything different, only the
exclusivity of a very high price. (I know, because a sub-board I
designed for a supplier to the trade turns up in so many very expensive
amps with so many different big names neatly silkscreened on it... it
strikes me as the sort of detail a real designer, as distinct from a
marketer, would take under his own control.) Those very few makers who
will sell you an ultrafi amp without any NFB operate even tinier shops,
usually one man and a cat, just hanging on.

The mechanism by which NFB wrecks your sound



2. In real life, as distinct from simplified theory, NFB adds artifacts
of its own. Remember, it is a loop. The signal starts at the input and
is amplified by devices until it reaches the output. From the output a
part of the signal called the negative feedback is fed back to the
input. Here a loop is completed and the combination, less distorted,
reaches the output again, a part of the combination is fed back,
endlessly. The artifacts we want to consider here are created by the
fed-back residue of harmonic distortions adding to both the fundamental
and the distortions already created by the amplifier, then some portion
of the sum of the original and the feedback distortion is fed back
again and added on, until the ooh-ah bird flies up its own fundament.
It looks marginally less disgusting as a recursive mathematical formula
with lots of nested parenthetical parcels of noise being loaded onto
your music. But it is a monkey on the back of your sound, with a
smaller monkey on the back of the first monkey, a still smaller monkey
on the back of the second monkey, and so on ad infinitum. These
additive artifacts are all higher harmonics and the more dominant ones
are all odd. Suppose, for the sake of simplicity, a superbly designed
ultrafidelista amp with some second harmonic and zero odd harmonics
before NFB. Add NFB and the second harmonic will be lowered but the
recombinant new loop now contains newly added intermodulation effects
between the fundamental and the residual second harmonic, and that is
third harmonic. In the next cycle a small but nasty dose of fifth
harmonic that wasn't there before is added by interaction between the
still residual second harmonic and reduced newly added third harmonic.
In short, the artifacts NFB adds to the distortion mix are all of the
most harmful kind. But, say the proponents of NFB, so what? Every time
the loop cycles the added artifacts are smaller, even if there are more
of them... The whole affair starts to smell of trying to argue with a
Marxist who simply declares any inconvenient truth 'an anomaly'. (If
this sounds like a mess from which you should run a mile, you have come
to the right conclusion. Start running now. It gets worse.)

I've had similar thoughts on NFB. That the fed back correction signal
suffers the same distortion in the amp as it tries to cancel the
original distortion. Thus more higher harmonics, and more intermod.

3. We thus arrive at a situation where distortion has been lowered
by NFB but where the most disturbing odd harmonic distortions are still
present to some measure, with the added disadvantage that new and
extremely disturbing artifacts of higher harmonic distortions have been
created by the very process of using negative feedback to lower
distortion. Regardless of the absolute level of THD, or the volume
setting, the mix of harmonics has been adversely affected and now
includes a higher proportion of third and higher harmonics than before
NFB. Let me say that again: after NFB, third and higher harmonics will
make up a greater part of the distortion than before.

Seems the industry needs to establish a new distortion measurement
standard, not THD but one that gives differing weights to say the 2nd,
the 3rd etc harmonics. My ears seem to accept and even like the 2nd
harmonic up to 45dB down, but the 3rd must be at least 10dB lower. Your
Ears May Vary (YEMV). But an amp may need to be better than this to
avoid intermod distortions in complex music.

4. Low volume levels perforce account for 99 per cent of audiophile
listening because we all have families or neighbours, and we would like
to keep our ears. ... At any lower level perceived
interference of this harmonics cocktail with the music will increase in
inverse proportion to the volume level. At low volume levels the
artifacts generated by NFB will by their nature as higher harmonic
distortions be disproportionately far more disturbing. At these normal
listening levels 0.75 per cent of second harmonic distortion may be
below the threshhold of perception for sophisticated listeners, whereas
tiny amounts of third and higher odd harmonic distortions grate.

Maybe I missed something, but I don't follow. Unless you're talking
about crossover distortion in class B amps.



It follows from the argument above that ultrafidelista should choose an
intrinsically linear topology and device which does not require added
negative feedback to 'linearize' the output. The intrinsically linear
device is the thermionic tube in either its triode form or as a pentode
hogtied to work as a triode, which can be a most pleasing alternative
both economically and sonically. The topology is often single-ended
operation...

The ultrafidelista, who are as keen on silent amps as anyone else,
accept this small potential difficulty because it is the lesser evil
compared to NFB. Unbelievers (largely unwashed, according to reports)
sneer that ultrafidelista like this approach because of the 'added
euphonics', which is bow-wow techie talk for the warmth a big chunk of
second harmonic lays on a zero negative feedback single-ended
amplifier. But competent design can easily reduce the level of second
harmonic to below the level of perception without the need for NFB and
its deleterious after-effects. In any event, it is your amplifier. You
paid for it. You have a right to tune it as you please. The key thing
is to get rid of NFB and to understand why you did it.

Can we prove any of this scientifically?

We have already. All of this is the technical subtext to my longtime
contention that what the ultrafidelista hear and love is not a directly
heated triode sound as is claimed by many enthusiasts but a Class A1,
ZNFB sound. (Admittedly, as we have seen above, the right sound is
virtually guaranteed with a ZNFB DHT SE amp of conservative provenance
but may have to be developed the hard way with more economical or
higher-power contenders.) In comparative ABX tests conducted over a
number of years, I found that professional musicians, certified golden
ears, choose the triode-linked Class A1 PP ZNFB EL34 whenever it is
present in the test over all other contenders including SE 300B and
'blameless' high-NFB silicon.

I can understand the preference for single ended class A ZNFB amps, but
what is special about DHT vs indirectly heated ones?



But transistor amps won't work at all without NFB!

It should be possible to build single ended transistor amps, General
Electric used to do it in table radios (not hifi of course, but the
topology should work for hifi given careful selection of the parts)
using a transistor that feeds an audio output transformer like those
used for tube radios. Take a look at the audio section of
http://www.geocities.com/wa2ise/radios/ssaf.gif
A hifi amp would of course use a proper power supply and not be "hot
chassis".
Of course this topology is not efficient in terms of power consumption,
but I'll take that hit.

[Eeyore]

Patrick Turner wrote:

Eeyore wrote:

If you take a single transistor amplifier circuit and do the same, let's say at ~
1mA Then for a typical small-signal silicon device the gm will be 40mA/V ( basic
physics ). Much greater than any toob, note !

If you restrict the signal level to say +-10% of Iq then the change in gm will be
from ~ 36 to 44 mA/V ( also +- 10% ) . Not exactly an *enormous change* !

But if the R load of the bjt was say 5k6 in a typical circuit with a 12V rail supply,

Who said anything about 12V ?


then 10% Ic change gives +/- 0.1mA, or 0.39vrms of output.
If the change in gm is 36 to 44, then the gain change over a swing of 0.39Vrms is 5%,
and I dare not calculate the 2H but it will be large.

The use of a triode with 56k load and with a +/-10% in Ia won't result in anything like
the
huge gm change you have in your bjt and the thd will be far less, even at 10 times the
output voltage.

Ok, I've done a couple of experiments in EWB.

One of the nice things about simulation is that you can do experiments using models that
would simply not be feasible in the real world and get astonishingly accurate results.

My criteria were simple. To accomadate the 12AX7 I made B+ = 300V, Ia = Ic = 2mA and Rl =
50k ohms.

This gives Va = 200V which I'm sure you will recognise is helpful to a triode compared to
using a lower voltage.

I also made an equivalent circuit using the same criteria using an MPSA42 300V small signal
silicon bjt http://www.fairchildsemi.com/ds/MP/MPSA42.pdf using a suitable bias source
and adding an emitter R to give the same voltage gain as the 12AX7.

So, the DC conditions are identical, as is the gain.

The 12AX7 gave a voltage gain of 48.2x
For an input voltage swing of +/- 0.1 V the voltage swing at the anode was +4.8046 and -
4.837 V
This translates to a linearity of 0.336%

For the MPSA42 the voltage gain was in fact 48.8x ( keeping to preferred values )
The voltage swing at the collector was +4.882 and - 4.885 V
This translates to a linearity of 0.034 % ( or 10 times better )

I also ran a test using an npn/pnp compound pair with all other parameters the same and the
linearity was ***0.002%***


***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Graham

[Eeyore]

robert casey wrote:

It should be possible to build single ended transistor amps

Of course it is.

Early Neve mixing consoles even used this topology as a 'line level driver'.

The thing is, you end up with both the transformer colouration and the expense
of the transformer.

If you use an active load instead, you can dispense with the transformer
entirely though ( at the expense of some efficiency )..

Graham

[Eeyore]

Trevor Wilson wrote:

"Tom Schlangen" wrote in message
...
Hi Trevor,

BTW: Modern transistors are more linear
than even the best Triodes.

I don't know much about recent transistors, so this
is new to me and sounds interesting.

**When I say recent, I mean post - 1975. The advent of the Ring Emitter
Transistor heralded the advent of spectacularly linear (power) transistors.
Unfortunately, the devices were relatively fragile and not capable of large
current capabilities. However, since the late 1990s, devices have come to
market which offer very impressive linearity, high current capacity and low
cost.

It seems you're referring here to the effect of the output stage loading on the
previous voltage gain stage.

I avoid the issue entirely by using several stages of low-current emitter
followers at this point to avoid loading the gain stage at all. ( like the Quad
triples if you like but I use quadruples ) It's bad practice to reflect the load
back in any case, not least for stability reasons !

Graham

[Eeyore]

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...

Maybe you could confirm what i am saying if you
went right back to basics with a breadboard and a CRO
and examined the basic workings of the basic devices that mankind has
invented over the last 100 years.

**Sure. Right after you recall that BJTs are current amplification devices
and cannot be used in identical topologies to tubes. They operate in a
fundamentally different way and need to be treated accordingly.

Trying to force fit the transistor into the same circuit as the toob produced
that shockingly misleading piece of buffoonery by the appropriately named Mr
Hamm ! that even managed to get published in the AES journal. A classic piece of
disengenuity if ever there was one.

Graham

[Prune]

Regarding voltage nonlinearity, why not use current amplifiers, with
current feedback? Many speakers work fine when driven with current
output, and for a source, well, most DAC chips already have current
output. So simply do the whole system current mode throughout, and
solid state should perform excellently. No? Why not?


Patrick Turner wrote in
:



Arny Krueger wrote:

"Patrick Turner" wrote in message
...


Arny Krueger wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

Unfortunately for your arguments this is not the case with
regard to voltage
gain.
Transistors need only a small range of turn on voltage applied
between their
base and emitter,
and their gm varies enormously between being on the threshold of
turn on
of
collector-emitter current and saturation.

It's almost like in Turner-world there are no such things as
emitter resistors.

I am forced to used Re to linearize bjts....

So what? Re also helps with bias stability.

I am not complaining about having to use NFB.
I have readily used it in the solid state amps I have designed and
built.

I am merely saying that NFB is necessary because bjts are not very
linear, and collector resistance is much higher than the loads to
which an amp is connected.

Its negative FB that makes a transistor circuit with fairly non linear
devices linear.
People have said here that SS devices are linear voltage amplifiers.
They are not.
But they do have a lot of gain and NFB can easily be used to provide
the linearity
that is not built into the device.







Only enormous amounts of externally applied NFB corrects the
inherent voltage
non-linearity of all BJTs where there is substantial current
change as in
the case of all
power transistors.

It's almost like in Turner-world there are no such things as
emitter followers.

This is a huge amount of series voltage NFB....

If NFB is such a sonic terror, why did almost every tubed audio amp
and preamp of the days of tubes have it?

A huge % of consumer electronics avoided NFB in preamps because that
meant having gain which could be thrown away, needing extra tubes.
NFB was indeed commonly used in phono amps with two halves of a 12AX7
in countless amps.
But the tone controls and line stages rarely used NFB.
Power amps nearly always had NFB to linearize pentode/tetrode power
tubes and reduce Rout.
I didn't say NFB was a terror, I just said lots of NFB is used in SS
amps.

But triode amps can easily be built without global NFB loops and still
give low thd and
great sound.




Try setting up any bjt in common emitter mode without any
emitter resistor
and
apply a voltage from a very low impedance to the base.

Proving once again that if you do something really stupid, you
will get suboptimal results.

I am not here to discuss whether the bjt amp can sound well, or
measure well.
Both are known to be possible so it is believed by most folks.

It's almost like in Turner-world sonic transparancy is the
ultimate sin.

I get all the transparancy I want.

So do I. Only, I'm obviously harder to please than you are, Pat.

You are impossible to please.

Patrick Turner.



Why does Arny suggest that i sin with ultimate badness?

I think I just got KROOGERED!

Figure of speech.

Without external loops of NFB including the emitter follower
connection which is a high amount of applied series voltage NFB,
the BJT is a hopeless
voltage amp unless we try to restrict the current change in the
device, and then the gain is
enormous and subject to the perils of the high collector
resistance
and
circuit slowness due to capacitances, and the non linear input
resistances
of
other devices in the signal path.

Proving once again that if you do something really stupid, you
will get suboptimal results.

.....Another statement of BS, dislocated from the preceding
text....

Not BS, just accepted practice, even in the heyday of tubes, back in
the 60s.

[Prune]

Can you get some comments in on FETs here, any of you guys? There are
quite a number of audio amplifiers that use both JFETs and MOSFETs so I
think it warrants some discussion.


"Trevor Wilson" wrote in
:


"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

**It's not misleading. It's fact.

Nothing is a fact until all sides agree it is, and you have to proove
what you
say.

**Just look at the curves.






Unfortunately for your arguments this is not the case with regard
to voltage
gain.
Transistors need only a small range of turn on voltage applied
between their
base and emitter,
and their gm varies enormously between being on the threshold of
turn on
of
collector-emitter current and saturation.

**Transistors are controlled by CURRENT, not Voltage. The CURRENT
gain of modern transistors is extremely linear, over a wide range of
Collector currents and Voltages.

Current relationships between base and collector currents are
substantially
linear.

**EXTREMELY linear.

The voltage relationship is not.

**Irrelevant.

You don't have the concepts correctly in your mind.

**Sure I do. Modern BJTs are extremely linear amplification devices.





Only enormous amounts of externally applied NFB corrects the
inherent voltage
non-linearity
of all BJTs where there is substantial current change as in the
case of all
power transistors.

**Transistors (BJTs) are CURRENT amplification devices.

Sure, we all know that, but their voltage linearity is intrinsically
poor.

**Er, that's because they're current amplifiers. And in that sense,
they are exceptionally linear.






Try setting up any bjt in common emitter mode without any emitter
resistor
and
apply
a voltage from a very low impedance to the base. The bjt is
disgustingly
non
linear
compared to a triode using no external applied NFB.

**Transistors (BJTs) are CURRENT amplification devices. Apply a
current to
the Base and measure the change in Collector current. Get back to me
when you've done this.

The current applied to the base is propelled by the applied voltage.
No voltage = no current.

**Duh. No current flow = no Voltage. What's your point?


Set up a small bjt in bread board as a an SE device in class A with
zero NFB and
record the
voltage linearity.

**Points:
* No amplifier on the planet uses zero NFB. Why would you want to
attempt to build one?
* Se amplifiers are dumb. Push pull makes more sense.

Its terrible.
Lots of NFB is needed to linearize the bjt, which is easy, because
they have a
lot of gain...


Even most pentodes which do not have internal electrostatic NFB
like a triode
can easily
out perform a bjt.

**Not even close. Pentodes are vastly less linear than even average
BJTs. Triodes, at least, are reasonably linear.

Your ignorance of tube behaviour is breathtaking.

Try reading RDH4 about the comparison of triode connected pentodes
and the same tube running as a pentode.
At low levels the pentode was more linear than the triode.

**Are we discussing POWER amps, or something else?





I am not here to discuss whether the bjt amp can sound well, or
measure well.
Both are known to be possible so it is believed by most folks.

**Me either. I'm here to correct Jute's lies.


Without external loops of NFB including the emitter follower
connection which is a high amount of applied series voltage NFB,
the BJT is a hopeless
voltage amp

**Transistors (BJTs) are CURRENT amplification devices.

No need to emulate a trained galah.

**It seems you've forgotten this point.




unless we try to restrict the current change in the device, and
then the
gain is

enormous and subject to the perils of the high collector
resistance and circuit slowness due to capacitances, and the non
linear input resistances
of
other
devices in the signal path.

A class A single bjt voltage amp stage with a single MJE340 with
another
for a
constant current source
collector load can have a gain of maybe 1,000, or like a pentode
with a CCS load
in its anode circuit.
If external loops of shunt NFB are applied around either device,
then the
linearity can be quite good for
either device, because for a gain reduction of 100, from 1,000 to
10, the
thd is
also reduced about
100 fold.

The linearity is dependant on the NFB applied, and not on the
basic VOLTAGE
linearity
of the device without NFB, which for a bjt is relatively poor, and
poorest
when
considerable
current change occurs in the device.

**Transistors (BJTs) are CURRENT amplification devices.

Gee, I think I heard that parrot again.

**Until you figure that out, I need to repeat it.


Maybe you could confirm what i am saying if you
went right back to basics with a breadboard and a CRO
and examined the basic workings of the basic devices that mankind has
invented over the last 100 years.

**Sure. Right after you recall that BJTs are current amplification
devices and cannot be used in identical topologies to tubes. They
operate in a fundamentally different way and need to be treated
accordingly.

[Prune]

Eeyore wrote in
:
***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Graham



Once again, that says nothing about the BJT's use in audio unless you use a
metric that actually correlates with perception as determined in blind
testing. THD does not, and AES papers have been published stating that
result.

[Andre Jute]

robert casey wrote:

Andre Jute wrote:

How does negative feedback work?

Negative feedback is simply a negative voltage fed back from the output
to the input amplifying device to offset part of the harmonic
distortion which is present as a positive voltage. It costs nothing
except a loss of gain and a few side effects such as phase shift and
possible instability which are well known in the mathematical
literature and more or less easily guarded against depending on the
level of NFB.

'Wow!' those meeting NFB for the first time will now say, 'Something
for free! I'll grab some of that for my amp.' Hey, I said it, and I am
a professional intellectual, by definition an infinite skeptic. NFB is
a thing of beauty that will draw you in. It is like an electronic
Marxism which admits of no contrary arguments because it has subsumed
them all into The Holy Measurements. To question The Measurements is to
commit heresy. You need to be of strong mind to resist the
blandishments of such a universal panacea and of strong stomach to
withstand the hysterical assaults of the lesser engineers defending
their holy grail. (And when you do get hold of a superior engineer to
explain NFB to you, you need to be high-domed indeed because suddenly
NFB can turn very intricate.)

Unfortunately NFB doesn't come without a price. It levies a cruel
charge on the perceived quality of the sound. Negative feedback is what
gives all those 'blameless' transistor and big PP tube amps their
chillingly unnatural sound.

Then how did NFB come to be such a panacea in amplifier design?

Your guess is as good as mine. Hi-fi design is not prestige work...
The exceptions to this rule are
normally audio enthusiasts in charge of their own small audio
manufactories with niche markets; those who grow larger from this base
follow the mainstream mantra of "mo' NFB give lowa' THD" because the
marketing channels demand it from them if they wish to grow. At this
point they usually cease to offer anything different, only the
exclusivity of a very high price. (I know, because a sub-board I
designed for a supplier to the trade turns up in so many very expensive
amps with so many different big names neatly silkscreened on it... it
strikes me as the sort of detail a real designer, as distinct from a
marketer, would take under his own control.) Those very few makers who
will sell you an ultrafi amp without any NFB operate even tinier shops,
usually one man and a cat, just hanging on.

The mechanism by which NFB wrecks your sound



2. In real life, as distinct from simplified theory, NFB adds artifacts
of its own. Remember, it is a loop. The signal starts at the input and
is amplified by devices until it reaches the output. From the output a
part of the signal called the negative feedback is fed back to the
input. Here a loop is completed and the combination, less distorted,
reaches the output again, a part of the combination is fed back,
endlessly. The artifacts we want to consider here are created by the
fed-back residue of harmonic distortions adding to both the fundamental
and the distortions already created by the amplifier, then some portion
of the sum of the original and the feedback distortion is fed back
again and added on, until the ooh-ah bird flies up its own fundament.
It looks marginally less disgusting as a recursive mathematical formula
with lots of nested parenthetical parcels of noise being loaded onto
your music. But it is a monkey on the back of your sound, with a
smaller monkey on the back of the first monkey, a still smaller monkey
on the back of the second monkey, and so on ad infinitum. These
additive artifacts are all higher harmonics and the more dominant ones
are all odd. Suppose, for the sake of simplicity, a superbly designed
ultrafidelista amp with some second harmonic and zero odd harmonics
before NFB. Add NFB and the second harmonic will be lowered but the
recombinant new loop now contains newly added intermodulation effects
between the fundamental and the residual second harmonic, and that is
third harmonic. In the next cycle a small but nasty dose of fifth
harmonic that wasn't there before is added by interaction between the
still residual second harmonic and reduced newly added third harmonic.
In short, the artifacts NFB adds to the distortion mix are all of the
most harmful kind. But, say the proponents of NFB, so what? Every time
the loop cycles the added artifacts are smaller, even if there are more
of them... The whole affair starts to smell of trying to argue with a
Marxist who simply declares any inconvenient truth 'an anomaly'. (If
this sounds like a mess from which you should run a mile, you have come
to the right conclusion. Start running now. It gets worse.)

I've had similar thoughts on NFB. That the fed back correction signal
suffers the same distortion in the amp as it tries to cancel the
original distortion. Thus more higher harmonics, and more intermod.

Every DIYer who studies on it eventually comes to this conclusion.

3. We thus arrive at a situation where distortion has been lowered
by NFB but where the most disturbing odd harmonic distortions are still
present to some measure, with the added disadvantage that new and
extremely disturbing artifacts of higher harmonic distortions have been
created by the very process of using negative feedback to lower
distortion. Regardless of the absolute level of THD, or the volume
setting, the mix of harmonics has been adversely affected and now
includes a higher proportion of third and higher harmonics than before
NFB. Let me say that again: after NFB, third and higher harmonics will
make up a greater part of the distortion than before.

Seems the industry needs to establish a new distortion measurement
standard, not THD but one that gives differing weights to say the 2nd,
the 3rd etc harmonics.

That too. But my case really is that THD doesn't predict musical joy,
so, now that it is so easy to get low levels of THD, we should find a
new measurement altogether, nothing to do with THD, or THD just a
small, weighted part of it, which predicts pleasure in music.

My ears seem to accept and even like the 2nd
harmonic up to 45dB down, but the 3rd must be at least 10dB lower. Your
Ears May Vary (YEMV). But an amp may need to be better than this to
avoid intermod distortions in complex music.

Traditionally a 30dB separation was considered adequate. I have a old
book by Farl J Waters which proceeds from this 30dB separation to give
rule of thumb shortcuts for everything, starting with the power supply
caps. A lot of people still use those shortcuts without knowing what
they result in.

4. Low volume levels perforce account for 99 per cent of audiophile
listening because we all have families or neighbours, and we would like
to keep our ears. ... At any lower level perceived
interference of this harmonics cocktail with the music will increase in
inverse proportion to the volume level. At low volume levels the
artifacts generated by NFB will by their nature as higher harmonic
distortions be disproportionately far more disturbing. At these normal
listening levels 0.75 per cent of second harmonic distortion may be
below the threshhold of perception for sophisticated listeners, whereas
tiny amounts of third and higher odd harmonic distortions grate.

Maybe I missed something, but I don't follow. Unless you're talking
about crossover distortion in class B amps.

No, not at all. This is a complicated interdisciplinary argument I'm
trying to stuff into a short par. Try this:

a) Most people listen at low level.

b) The lower the listening level, the greater the perceived effect of
NFB artifacts. At higher volume levels NFB artifacts can become less
obvouus.

c) Therefore NFB artifacts interfere most with the volume band most
commonly in use.

It follows from the argument above that ultrafidelista should choose an
intrinsically linear topology and device which does not require added
negative feedback to 'linearize' the output. The intrinsically linear
device is the thermionic tube in either its triode form or as a pentode
hogtied to work as a triode, which can be a most pleasing alternative
both economically and sonically. The topology is often single-ended
operation...

The ultrafidelista, who are as keen on silent amps as anyone else,
accept this small potential difficulty because it is the lesser evil
compared to NFB. Unbelievers (largely unwashed, according to reports)
sneer that ultrafidelista like this approach because of the 'added
euphonics', which is bow-wow techie talk for the warmth a big chunk of
second harmonic lays on a zero negative feedback single-ended
amplifier. But competent design can easily reduce the level of second
harmonic to below the level of perception without the need for NFB and
its deleterious after-effects. In any event, it is your amplifier. You
paid for it. You have a right to tune it as you please. The key thing
is to get rid of NFB and to understand why you did it.

Can we prove any of this scientifically?

We have already. All of this is the technical subtext to my longtime
contention that what the ultrafidelista hear and love is not a directly
heated triode sound as is claimed by many enthusiasts but a Class A1,
ZNFB sound. (Admittedly, as we have seen above, the right sound is
virtually guaranteed with a ZNFB DHT SE amp of conservative provenance
but may have to be developed the hard way with more economical or
higher-power contenders.) In comparative ABX tests conducted over a
number of years, I found that professional musicians, certified golden
ears, choose the triode-linked Class A1 PP ZNFB EL34 whenever it is
present in the test over all other contenders including SE 300B and
'blameless' high-NFB silicon.

I can understand the preference for single ended class A ZNFB amps, but
what is special about DHT vs indirectly heated ones?

A certain amount of internal or natural feedback! My case is only about
*added* feedback, what the designer proposes, not what the triode god
has already disposed.

More generally, DHT, and even the indirectly heated power triodes, are
mostly used only in SE, which is of course strictly Class A, and then
in expensive amps where it is not de riguer to hog the last watt from
the tube, so mostly in Class A1. And Class A1 is the peculiar sound
people have come to associate with SET. But you can get the same sound
with ZNFB trioded pentodes in PP as long as you stick strictly to Class
A1.

But transistor amps won't work at all without NFB!

It should be possible to build single ended transistor amps, General
Electric used to do it in table radios (not hifi of course, but the
topology should work for hifi given careful selection of the parts)
using a transistor that feeds an audio output transformer like those
used for tube radios. Take a look at the audio section of
http://www.geocities.com/wa2ise/radios/ssaf.gif
A hifi amp would of course use a proper power supply and not be "hot
chassis".
Of course this topology is not efficient in terms of power consumption,
but I'll take that hit.

Sure. Nelson Pass published the ZEN design for people who want to try
SE amps in silicon. But you don't need to go to that expense just to
get a taste. I've published an inexpensive opam circuit for
experimenters that can be forced into SE operation and therefore Class
A for a little way. You need sensitive speakers but if you're talking
about going into SE amps, you probably already have those. The
articles, circuit, board layout and a photo of the power supply are
he

http://members.lycos.co.uk/fiultra/K...dre%20Jute.htm
KISS 191A by Andre Jute An inexpensive opamp amplifier with minimum
parts count that sounds good and has further potential

http://members.lycos.co.uk/fiultra/K...dre%20Jute.htm
KISS 191B by Andre Jute Developing the amp in KISS 191A to sound like a
300B for a fraction of the cost

http://members.lycos.co.uk/fiultra/K...20mGBschem.jpg
KISS 191C mGBschem.jpg The opamp circuit for the miniGainBrick Zip
LM675

http://members.lycos.co.uk/fiultra/K...%20mGBmatr.jpg
KISS 191D mGBmatr.jpg A layout suggestion for the miniGainBrick Zip
LM675

http://members.lycos.co.uk/fiultra/K...%20NoBleed.jpg
KISS191E NoBleed.jpg Living dangerously!

However, SE and Class A1 are not efficient in tubes either. Efficiency
just isn't a consideration for DIYers.

Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

[Prune]

"Andre Jute" wrote in
ups.com:

I think you are very mistaken about Hawksford, as any artifacts in the
feedback version of error correction is with the error being fed back,
not the signal itself. Moreover, in feedforward error correction, there
is no possibility for the sort of recursion artifacts you mentioned
before.

Nothing wrong with NFB in general engineering; I'd hate to drive a
modern car without power steering. But in audio NFB must be used with
the greatest delicacy, or preferably not at all, for it adds artifacts
that make people edgy.

I think there may be something to this. But:

That these artifacts are outside present
measuring systems does not make them absent, it merely makes those who
claim that the answer has been found for all time into fools.

The burden of proof is on the ones claiming they exist, however.

Some people like the Lee papers I've mentioned elsewhere are trying to
find better metrics, and I'll add the following papers (these don't
concern specifically whatever these effects of NFB are, but are relevant
in general):

ERROR ACTIVITY AND ERROR ENTROPY AS A MEASURE OF PSYCHOACOUSTIC
SIGNIFICANCE IN THE PERCEPTUAL DOMAIN, Hollier, M.P., Hawksford, M.O.J.
and Guard, D.R., Proc. IEE of Vision, Image and Signal Processing, vol.
141, no. 3, pp 203-208, June 1994

SYSTEM MEASUREMENT AND IDENTIFICATION USING PSEUDORANDOM FILTERED NOISE
AND MUSIC SEQUENCES, Hawksford, M.O.J. JAES, vol. 52, no. 4, pp.
275-296, April 2005

DISTORTION IMMUNITY OF IMPULSE RESPONSE MEASUREMENTS USING MAXIMUM
LENGTH SEQUENCES, Dunn, C. and Hawksford, Institute of Acoustics,
Reproduced Sound 7, Proc. Institute of Acoustics, vol. 13, pt 7, pp
191-206, Nov 1991

ALGORITHMS FOR ASSESSING THE SUBJECTIVITY OF PERCEPTUALLY WEIGHTED
AUDIBLE ERRORS, Hollier, M.P., Hawksford, M.O.J. and Guard, D.R. JAES,
vol. 43, no. 12, pp 1041-1045, December 1995

And that's just a sampling.

[Trevor Wilson]

"Prune" wrote in message
4.76...
Can you get some comments in on FETs here, any of you guys? There are
quite a number of audio amplifiers that use both JFETs and MOSFETs so I
think it warrants some discussion.

**Power FETs are as rare as rocking horse ****, so there is no need to
consider them. MOSFETs are very linear, when operated at elevated currents.
Below their 'knee' they exhibit very poor linearity (about 10 times the
distortion of any other device). Above the knee, linearity varies with the
device, but the best is still not quite as good as the best BJTs. Where they
score, big time, is their freedom from second breakdown. They are reasonably
hard to damage and can deliver high frequencies without the problems
associated with BJTs.

IMO, the fatal flaw of all MOSFETs is their negative tempco of gm, which
tends to cause them to sound 'strangled'. Except those MOSFET amps which
operate at high bias currents (and temperatures), of course. Pass, et al.

JFETs are OK as low level devices, but possess such massively wide parameter
spreads such that careful design needs to be employed.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[Prune]

"Bret Ludwig" wrote in news:1153946550.397874.157650
@m73g2000cwd.googlegroups.com:

You , and what army?

I'm the only army he needs ;P

[Prune]

"Trevor Wilson" wrote in
:

IMO, the fatal flaw of all MOSFETs is their negative tempco of gm,
which tends to cause them to sound 'strangled'. Except those MOSFET
amps which operate at high bias currents (and temperatures), of
course. Pass, et al.

Thanks, I'm planning to build a Pass XA clone, which is one reason I asked
this question.

JFETs are OK as low level devices, but possess such massively wide
parameter spreads such that careful design needs to be employed.

What about Erno Borbely's designs? I don't have the technical ability to
evaluate them, but reviews I've heard from people that have built his all-
JFET circuits are all positive.

[Eeyore]

Prune wrote:

Eeyore wrote in
:
***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Graham

Once again, that says nothing about the BJT's use in audio unless you use a
metric that actually correlates with perception as determined in blind
testing.

I do tend to agree that this subject is worthy of more study for sure.


THD does not, and AES papers have been published stating that result.

That wasn't THD btw - it's basic linearity of the device.

However 0.002% is -94dB.

That means that any non-linearity of the above will be ** below the threshold
of audibility ** until an SPL of 94dB is reached. Inaudible simply from first
principles.

I'm tempted to analyse them for distortion too. EWB does a Fourier analysis so
you can see the individual harmonics.

Graham

[Eeyore]

Prune wrote:

Regarding voltage nonlinearity, why not use current amplifiers, with
current feedback?

Not terribly practical.

Many speakers work fine when driven with current
output,

Actually they don't !

and for a source, well, most DAC chips already have current
output.

Simply a consequence of current methodology. ( not current as in amps btw )


So simply do the whole system current mode throughout, and
solid state should perform excellently. No? Why not?

I guess you're not a designer if you have to ask that. It's a bit like - where
to even begin explaining !

Graham

[Eeyore]

Prune wrote:

Can you get some comments in on FETs here, any of you guys? There are
quite a number of audio amplifiers that use both JFETs and MOSFETs so I
think it warrants some discussion.

That'll be one of my next modelling exercises.

Graham

[Prune]

robert casey wrote in news:OQTxg.531$0e5.203
@newsread4.news.pas.earthlink.net:

Seems the industry needs to establish a new distortion measurement
standard

There are a number of attempts at better metrics, but all experimental and
far from bein under consideration for a standard. I've already mentioned
several papers by Lee and Hawksford that examine the failings of existing
metrics and propose alternative measurements with better perceptual
correlation. As a scientist, I do believe that ultimately we can measure
electronically any possible type of audible distortion (well except the
subjectivists' psychological bias ;P), but we are not there yet.

[Trevor Wilson]

"Prune" wrote in message
4.76...
"Trevor Wilson" wrote in
:

IMO, the fatal flaw of all MOSFETs is their negative tempco of gm,
which tends to cause them to sound 'strangled'. Except those MOSFET
amps which operate at high bias currents (and temperatures), of
course. Pass, et al.

Thanks, I'm planning to build a Pass XA clone, which is one reason I asked
this question.

**You might be better building a decent (push pull) BJT design. You can get
away with a whole bunch less bias current and still get better performance.
Until someone builds some better MOSFETs, I don't see any point in using
them. Back in the late 1980s and easrly 1990s, MOSFETs were far more viable
devices. Unfortunately, they were also much more expensive too. Nowadays,
their prices have fallen substantially, but so has the price of BJTs. OTOH,
BJT performance has improved dramatically, whilst MOSFETs have not.


JFETs are OK as low level devices, but possess such massively wide
parameter spreads such that careful design needs to be employed.

What about Erno Borbely's designs? I don't have the technical ability to
evaluate them, but reviews I've heard from people that have built his all-
JFET circuits are all positive.

**I built a few many years ago. All sounded pretty decent.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[Patrick Turner]

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Eeyore wrote:

Patrick Turner wrote:

Trevor tries to correct Jute's assertions about the perils of NFB but
tells
bigger lies to posture in a worse position.

There is no peril whatever with NFB when intelligently and correctly
applied.

Graham

I happen to agree with you here. I use NFB as I see fit.

But Trevor tried to say there would be little reason why one couldn't
build
a bjt amp without NFB.

**That would be a lie. Here are my EXACT words. Do pay close attention.

ALL AMPLIFIERS USE NFB. EVERY SINGLE ONE.

Understand?


Since he said bjts are so darn linear, then why not dispense entirely with
the NFB?

**Impossible. ALL AMPLIFIERS USE NFB. EVERY SINGLE ONE.

But you are quite torpedoed intellectually.

You said bjts are linear.

I said then there was no need for NFB.

Patrick Turner





--
Trevor Wilson
www.rageaudio.com.au

--
Posted via a free Usenet account from http://www.teranews.com

[Patrick Turner]

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

**It's not misleading. It's fact.

Nothing is a fact until all sides agree it is, and you have to proove what
you
say.

**Just look at the curves.

The current in/out curves for bjts are linear.

But the base-emitter voltage versus collector current is not linear.

large amount of series voltage NFB or series current NFB is required to
linearize the voltage
amplifications in bjts.

Patrick Turner.

[Andre Jute]

Thanks for the explanation about Hawksford and the references, Prune.

I'm not mistaken about Hawksford (I was one of the few to make the
effort to understand what he was talking about when he first burst on
the scene). It is far worse. I am indifferent to Hawksford. I don't
want a gazillion fixes to my sound. I have discovered a lovely natural
sound, with the noise level far below my perception, from easily
understood and generally available devices, if a little expensive, that
do not require any fix, not one. So Professor Hawksford can apply his
fixes to transistors, which need to be fixed, god knows, and good luck
to him. My music, my devices, my amps are already good. I don't fix
what ain't broke, and I don't replace proven-good gear with this week's
dicey fashion-victim's must-have. I leave that to young guys like you
feeling their oats and to middle-aged professors like the sainted
Malcolm laying down their markers for a retirement full of fat
consultancy fees.

Andre Jute
Insufferably content. Retro and smug with it. Invariably proven right
in the long term.

Prune wrote:
"Andre Jute" wrote in
ups.com:

I think you are very mistaken about Hawksford, as any artifacts in the
feedback version of error correction is with the error being fed back,
not the signal itself. Moreover, in feedforward error correction, there
is no possibility for the sort of recursion artifacts you mentioned
before.

Nothing wrong with NFB in general engineering; I'd hate to drive a
modern car without power steering. But in audio NFB must be used with
the greatest delicacy, or preferably not at all, for it adds artifacts
that make people edgy.

I think there may be something to this. But:

That these artifacts are outside present
measuring systems does not make them absent, it merely makes those who
claim that the answer has been found for all time into fools.

The burden of proof is on the ones claiming they exist, however.

Some people like the Lee papers I've mentioned elsewhere are trying to
find better metrics, and I'll add the following papers (these don't
concern specifically whatever these effects of NFB are, but are relevant
in general):

ERROR ACTIVITY AND ERROR ENTROPY AS A MEASURE OF PSYCHOACOUSTIC
SIGNIFICANCE IN THE PERCEPTUAL DOMAIN, Hollier, M.P., Hawksford, M.O.J.
and Guard, D.R., Proc. IEE of Vision, Image and Signal Processing, vol.
141, no. 3, pp 203-208, June 1994

SYSTEM MEASUREMENT AND IDENTIFICATION USING PSEUDORANDOM FILTERED NOISE
AND MUSIC SEQUENCES, Hawksford, M.O.J. JAES, vol. 52, no. 4, pp.
275-296, April 2005

DISTORTION IMMUNITY OF IMPULSE RESPONSE MEASUREMENTS USING MAXIMUM
LENGTH SEQUENCES, Dunn, C. and Hawksford, Institute of Acoustics,
Reproduced Sound 7, Proc. Institute of Acoustics, vol. 13, pt 7, pp
191-206, Nov 1991

ALGORITHMS FOR ASSESSING THE SUBJECTIVITY OF PERCEPTUALLY WEIGHTED
AUDIBLE ERRORS, Hollier, M.P., Hawksford, M.O.J. and Guard, D.R. JAES,
vol. 43, no. 12, pp 1041-1045, December 1995

And that's just a sampling.

[Patrick Turner]

robert casey wrote:


How does negative feedback work?

Negative feedback is simply a negative voltage fed back from the output
to the input amplifying device to offset part of the harmonic
distortion which is present as a positive voltage. It costs nothing
except a loss of gain and a few side effects such as phase shift and
possible instability which are well known in the mathematical
literature and more or less easily guarded against depending on the
level of NFB.

'Wow!' those meeting NFB for the first time will now say, 'Something
for free! I'll grab some of that for my amp.' Hey, I said it, and I am
a professional intellectual, by definition an infinite skeptic. NFB is
a thing of beauty that will draw you in. It is like an electronic
Marxism which admits of no contrary arguments because it has subsumed
them all into The Holy Measurements. To question The Measurements is to
commit heresy. You need to be of strong mind to resist the
blandishments of such a universal panacea and of strong stomach to
withstand the hysterical assaults of the lesser engineers defending
their holy grail. (And when you do get hold of a superior engineer to
explain NFB to you, you need to be high-domed indeed because suddenly
NFB can turn very intricate.)

Unfortunately NFB doesn't come without a price. It levies a cruel
charge on the perceived quality of the sound. Negative feedback is what
gives all those 'blameless' transistor and big PP tube amps their
chillingly unnatural sound.

Then how did NFB come to be such a panacea in amplifier design?

Your guess is as good as mine. Hi-fi design is not prestige work...
The exceptions to this rule are
normally audio enthusiasts in charge of their own small audio
manufactories with niche markets; those who grow larger from this base
follow the mainstream mantra of "mo' NFB give lowa' THD" because the
marketing channels demand it from them if they wish to grow. At this
point they usually cease to offer anything different, only the
exclusivity of a very high price. (I know, because a sub-board I
designed for a supplier to the trade turns up in so many very expensive
amps with so many different big names neatly silkscreened on it... it
strikes me as the sort of detail a real designer, as distinct from a
marketer, would take under his own control.) Those very few makers who
will sell you an ultrafi amp without any NFB operate even tinier shops,
usually one man and a cat, just hanging on.

The mechanism by which NFB wrecks your sound



2. In real life, as distinct from simplified theory, NFB adds artifacts
of its own. Remember, it is a loop. The signal starts at the input and
is amplified by devices until it reaches the output. From the output a
part of the signal called the negative feedback is fed back to the
input. Here a loop is completed and the combination, less distorted,
reaches the output again, a part of the combination is fed back,
endlessly. The artifacts we want to consider here are created by the
fed-back residue of harmonic distortions adding to both the fundamental
and the distortions already created by the amplifier, then some portion
of the sum of the original and the feedback distortion is fed back
again and added on, until the ooh-ah bird flies up its own fundament.
It looks marginally less disgusting as a recursive mathematical formula
with lots of nested parenthetical parcels of noise being loaded onto
your music. But it is a monkey on the back of your sound, with a
smaller monkey on the back of the first monkey, a still smaller monkey
on the back of the second monkey, and so on ad infinitum. These
additive artifacts are all higher harmonics and the more dominant ones
are all odd. Suppose, for the sake of simplicity, a superbly designed
ultrafidelista amp with some second harmonic and zero odd harmonics
before NFB. Add NFB and the second harmonic will be lowered but the
recombinant new loop now contains newly added intermodulation effects
between the fundamental and the residual second harmonic, and that is
third harmonic. In the next cycle a small but nasty dose of fifth
harmonic that wasn't there before is added by interaction between the
still residual second harmonic and reduced newly added third harmonic.
In short, the artifacts NFB adds to the distortion mix are all of the
most harmful kind. But, say the proponents of NFB, so what? Every time
the loop cycles the added artifacts are smaller, even if there are more
of them... The whole affair starts to smell of trying to argue with a
Marxist who simply declares any inconvenient truth 'an anomaly'. (If
this sounds like a mess from which you should run a mile, you have come
to the right conclusion. Start running now. It gets worse.)

I've had similar thoughts on NFB. That the fed back correction signal
suffers the same distortion in the amp as it tries to cancel the
original distortion. Thus more higher harmonics, and more intermod.

The original thd in any amp is again distorted when fed to the second
differential input
to the amp. Many more artifacts are generated which have significant levels
when
open loop bandwidth is poor, the amount of NFB is below 14 dB, and open loop
distortion is high
at say 10%.

But where the open loop bw is wide, NFB is below 14 dB, and open loop thd
low, then
the effects of second order generation of artifacts is largely inaudible.
This is the case with most SET amps without loop FB and used in their
designed for region of power.

But where you have a pentode or bjt amp with attrocious open loop bw, NFB is
less than 14 dB,
and open Dn is high, then don't expect marvellous music.
In this case some say no amount of NFB will make the outcome pleasing.
But very few amplifiers have 10% open loop Dn at 1 watt where they are used
mostly.
I myself am not allergic to NFB and have built some superlative SS amps using
lots of NFB.
I'm doing a pair of 845 amps without NFB at the moment.

All the silly arguments about NFB would cease if we had the best samples of
different genres
of amplifers assembled together for a listening test under fair test
conditions.




3. We thus arrive at a situation where distortion has been lowered
by NFB but where the most disturbing odd harmonic distortions are still
present to some measure, with the added disadvantage that new and
extremely disturbing artifacts of higher harmonic distortions have been
created by the very process of using negative feedback to lower
distortion. Regardless of the absolute level of THD, or the volume
setting, the mix of harmonics has been adversely affected and now
includes a higher proportion of third and higher harmonics than before
NFB. Let me say that again: after NFB, third and higher harmonics will
make up a greater part of the distortion than before.

Seems the industry needs to establish a new distortion measurement
standard, not THD but one that gives differing weights to say the 2nd,
the 3rd etc harmonics. My ears seem to accept and even like the 2nd
harmonic up to 45dB down, but the 3rd must be at least 10dB lower. Your
Ears May Vary (YEMV). But an amp may need to be better than this to
avoid intermod distortions in complex music.

4. Low volume levels perforce account for 99 per cent of audiophile
listening because we all have families or neighbours, and we would like
to keep our ears. ... At any lower level perceived
interference of this harmonics cocktail with the music will increase in
inverse proportion to the volume level. At low volume levels the
artifacts generated by NFB will by their nature as higher harmonic
distortions be disproportionately far more disturbing. At these normal
listening levels 0.75 per cent of second harmonic distortion may be
below the threshhold of perception for sophisticated listeners, whereas
tiny amounts of third and higher odd harmonic distortions grate.

Maybe I missed something, but I don't follow. Unless you're talking
about crossover distortion in class B amps.



It follows from the argument above that ultrafidelista should choose an
intrinsically linear topology and device which does not require added
negative feedback to 'linearize' the output. The intrinsically linear
device is the thermionic tube in either its triode form or as a pentode
hogtied to work as a triode, which can be a most pleasing alternative
both economically and sonically. The topology is often single-ended
operation...

The ultrafidelista, who are as keen on silent amps as anyone else,
accept this small potential difficulty because it is the lesser evil
compared to NFB. Unbelievers (largely unwashed, according to reports)
sneer that ultrafidelista like this approach because of the 'added
euphonics', which is bow-wow techie talk for the warmth a big chunk of
second harmonic lays on a zero negative feedback single-ended
amplifier. But competent design can easily reduce the level of second
harmonic to below the level of perception without the need for NFB and
its deleterious after-effects. In any event, it is your amplifier. You
paid for it. You have a right to tune it as you please. The key thing
is to get rid of NFB and to understand why you did it.

Can we prove any of this scientifically?

I have clients who preferred my triode amps with 12dB of NFB.

Then what about Rout?

Loop NFB must be used with UL and pentode amps.....





We have already. All of this is the technical subtext to my longtime
contention that what the ultrafidelista hear and love is not a directly
heated triode sound as is claimed by many enthusiasts but a Class A1,
ZNFB sound. (Admittedly, as we have seen above, the right sound is
virtually guaranteed with a ZNFB DHT SE amp of conservative provenance
but may have to be developed the hard way with more economical or
higher-power contenders.) In comparative ABX tests conducted over a
number of years, I found that professional musicians, certified golden
ears, choose the triode-linked Class A1 PP ZNFB EL34 whenever it is
present in the test over all other contenders including SE 300B and
'blameless' high-NFB silicon.

I can understand the preference for single ended class A ZNFB amps, but
what is special about DHT vs indirectly heated ones?



But transistor amps won't work at all without NFB!

It should be possible to build single ended transistor amps, General
Electric used to do it in table radios (not hifi of course, but the
topology should work for hifi given careful selection of the parts)
using a transistor that feeds an audio output transformer like those
used for tube radios. Take a look at the audio section of
http://www.geocities.com/wa2ise/radios/ssaf.gif
A hifi amp would of course use a proper power supply and not be "hot
chassis".
Of course this topology is not efficient in terms of power consumption,
but I'll take that hit.


Many early car radios in the 1950s had 5 watt class A SE transistor
amplifiers.
They were easy to build, and all had loop NFB since the bjts have very high
output resistance compared to the speaker load they drive.

Patrick Turner.

[Prune]

Eeyore wrote in
:

I guess you're not a designer if you have to ask that. It's a bit like
- where to even begin explaining !

Well, Krell makes systems that are current mode throughout, DAC-preamp-
power amp (not sure about speaker, maybe the final stage of the amp is a
transimpedance one).

So clearly there are designers that don't have a problem with the idea.
But I'd rather ask here rather than the Krell people, as I wouldn't expect
them to be open with technical discussion.

[Prune]

"Andre Jute" wrote in news:1153971297.062170.131540
@m79g2000cwm.googlegroups.com:

So Professor Hawksford can apply his
fixes to transistors, which need to be fixed, god knows

Works fine with tubes. smoking_amp at the diyaudio forums has used it
successfully to improve linearity of various tube stages, and compared it
(favorably) with other things like nested feedback loops etc.

As for the references I posted, they have nothing to do with error
correction (that Hawksfor is also an author is coincidental, he just has
too many publications), but simply coming up with ways to measure
distortion that would correlate well with what people perceive, without the
time and cost of extensive blind testing.

[Patrick Turner]

Eeyore wrote:

Patrick Turner wrote:

Eeyore wrote:

If you take a single transistor amplifier circuit and do the same, let's say at ~
1mA Then for a typical small-signal silicon device the gm will be 40mA/V ( basic
physics ). Much greater than any toob, note !

If you restrict the signal level to say +-10% of Iq then the change in gm will be
from ~ 36 to 44 mA/V ( also +- 10% ) . Not exactly an *enormous change* !

But if the R load of the bjt was say 5k6 in a typical circuit with a 12V rail supply,

Who said anything about 12V ?

Because this is a very typical supply voltage in consumer electronics.

Nowdays hardly anyone uses discrete bjts; it all opamps on double sided boards
with even lower supply voltages.
Measured thd is incredibly low because of huge amounts of NFB applied around the opamps....


then 10% Ic change gives +/- 0.1mA, or 0.39vrms of output.
If the change in gm is 36 to 44, then the gain change over a swing of 0.39Vrms is 5%,
and I dare not calculate the 2H but it will be large.

The use of a triode with 56k load and with a +/-10% in Ia won't result in anything like
the
huge gm change you have in your bjt and the thd will be far less, even at 10 times the
output voltage.

Ok, I've done a couple of experiments in EWB.

One of the nice things about simulation is that you can do experiments using models that
would simply not be feasible in the real world and get astonishingly accurate results.

I simulate nothing, and make only real world comparisons.



My criteria were simple. To accomadate the 12AX7 I made B+ = 300V, Ia = Ic = 2mA and Rl =
50k ohms.

This gives Va = 200V which I'm sure you will recognise is helpful to a triode compared to
using a lower voltage.

I also made an equivalent circuit using the same criteria using an MPSA42 300V small signal
silicon bjt http://www.fairchildsemi.com/ds/MP/MPSA42.pdf using a suitable bias source
and adding an emitter R to give the same voltage gain as the 12AX7.

So, the DC conditions are identical, as is the gain.

Yes but the 12AX7 has no external NFB. The MPSA42 has a lot of linearizing current NFB.

The MPSA without ANY loop NFB is not as linear as a triode.



The 12AX7 gave a voltage gain of 48.2x
For an input voltage swing of +/- 0.1 V the voltage swing at the anode was +4.8046 and -
4.837 V
This translates to a linearity of 0.336%

For the MPSA42 the voltage gain was in fact 48.8x ( keeping to preferred values )
The voltage swing at the collector was +4.882 and - 4.885 V
This translates to a linearity of 0.034 % ( or 10 times better )

I also ran a test using an npn/pnp compound pair with all other parameters the same and the
linearity was ***0.002%***

***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Such contests run on a non level playing field tell us all nothing except that your'e
good at conductiong silly sporting events.

The bjt is very much dependant on NFB loops because of its intrisically non linear
character.

Using differential amplifiers with triodes with CCS loads will give you extremely low thd
without any current or other NFB.

In my 300w SS amps I have symetrical npn and pnp diff amps in parallel
at the input, and a complementary pair in common emitter for the driver of the mosfet outputs.
http://www.turneraudio.com.au/solids...s1mosfets.html
The supply is +/- 80V, and there is about 4% ths at about 40Vrms max of Vswing.
This is a lot worse than a pair of PP triodes in a driver can achieve, maybe 0.1% at 150Vrms
a-a.
http://www.turneraudio.com.au/schem-...ut+output.html

The global NFB applied in the mosfet amp reduces all this to less than 0.01% at 300 watts.
The thd falls about proportionately to Vo, so that at 1 watt where Vo = 2.83V the thd =
0.001%

The linearity is a function of applied NFB and gain, but the bjt devices
ain't linear when examined without NFB.
The mosfet output stage I have also would produce about 5% thd at 300 watts
with common source mode but in common drain mode, ie, source follower,
this Dn, mainly 3H plus other odd order junk is less than 1%, and less than the drive amp Dn,
so the output stage Dn does not dominate the THD production as occurs in a tube amp.
Both drive amp and output stage Dn is reduced equally by the global NFB.

You have not won.

SS devices are less linear than triodes, but ARE adequately linear and DO HAVE
oddles of gain which allows NFB to make them linear.


Most tube amps don't need more than 20dB of NFB in the form of
output stage NFB and global NFB.

Some SET amps don't need any global NFB because they are substantially linear
without NFB.

I have heard plenty of systems to confirm my beliefs and tolerance of
SET amplifiers.



Patrick Turner.






Graham

[Trevor Wilson]

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Eeyore wrote:

Patrick Turner wrote:

Trevor tries to correct Jute's assertions about the perils of NFB
but
tells
bigger lies to posture in a worse position.

There is no peril whatever with NFB when intelligently and correctly
applied.

Graham

I happen to agree with you here. I use NFB as I see fit.

But Trevor tried to say there would be little reason why one couldn't
build
a bjt amp without NFB.

**That would be a lie. Here are my EXACT words. Do pay close attention.

ALL AMPLIFIERS USE NFB. EVERY SINGLE ONE.

Understand?


Since he said bjts are so darn linear, then why not dispense entirely
with
the NFB?

**Impossible. ALL AMPLIFIERS USE NFB. EVERY SINGLE ONE.

But you are quite torpedoed intellectually.

You said bjts are linear.

**More linear than Triodes, in fact.


I said then there was no need for NFB.

**Irrelevant. ALL AMPLIFIERS USE NFB. EVERY SINGLE ONE.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[Eeyore]

Prune wrote:

Eeyore wrote in
:

I guess you're not a designer if you have to ask that. It's a bit like
- where to even begin explaining !

Well, Krell makes systems that are current mode throughout, DAC-preamp-
power amp (not sure about speaker, maybe the final stage of the amp is a
transimpedance one).

So clearly there are designers that don't have a problem with the idea.
But I'd rather ask here rather than the Krell people, as I wouldn't expect
them to be open with technical discussion.

I suspect that their 'current mode' is a marketing term.

Graham

[Patrick Turner]

Prune wrote:

Regarding voltage nonlinearity, why not use current amplifiers, with
current feedback? Many speakers work fine when driven with current
output, and for a source, well, most DAC chips already have current
output. So simply do the whole system current mode throughout, and
solid state should perform excellently. No? Why not?

Most speakers and their crossover filters are designed to only give a flat
response
with a voltage source of low impedance below 1 ohm.

A current source will result in a very non-flat response since speaker Z
varies along the band
and the filters won't work as planned.

Many speakers may appear to work with current sources but critical listeners
will soon
realise something is wrong.

Patrick Turner.



Patrick Turner wrote in
:



Arny Krueger wrote:

"Patrick Turner" wrote in message
...


Arny Krueger wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

Unfortunately for your arguments this is not the case with
regard to voltage
gain.
Transistors need only a small range of turn on voltage applied
between their
base and emitter,
and their gm varies enormously between being on the threshold of
turn on
of
collector-emitter current and saturation.

It's almost like in Turner-world there are no such things as
emitter resistors.

I am forced to used Re to linearize bjts....

So what? Re also helps with bias stability.

I am not complaining about having to use NFB.
I have readily used it in the solid state amps I have designed and
built.

I am merely saying that NFB is necessary because bjts are not very
linear, and collector resistance is much higher than the loads to
which an amp is connected.

Its negative FB that makes a transistor circuit with fairly non linear
devices linear.
People have said here that SS devices are linear voltage amplifiers.
They are not.
But they do have a lot of gain and NFB can easily be used to provide
the linearity
that is not built into the device.







Only enormous amounts of externally applied NFB corrects the
inherent voltage
non-linearity of all BJTs where there is substantial current
change as in
the case of all
power transistors.

It's almost like in Turner-world there are no such things as
emitter followers.

This is a huge amount of series voltage NFB....

If NFB is such a sonic terror, why did almost every tubed audio amp
and preamp of the days of tubes have it?

A huge % of consumer electronics avoided NFB in preamps because that
meant having gain which could be thrown away, needing extra tubes.
NFB was indeed commonly used in phono amps with two halves of a 12AX7
in countless amps.
But the tone controls and line stages rarely used NFB.
Power amps nearly always had NFB to linearize pentode/tetrode power
tubes and reduce Rout.
I didn't say NFB was a terror, I just said lots of NFB is used in SS
amps.

But triode amps can easily be built without global NFB loops and still
give low thd and
great sound.




Try setting up any bjt in common emitter mode without any
emitter resistor
and
apply a voltage from a very low impedance to the base.

Proving once again that if you do something really stupid, you
will get suboptimal results.

I am not here to discuss whether the bjt amp can sound well, or
measure well.
Both are known to be possible so it is believed by most folks.

It's almost like in Turner-world sonic transparancy is the
ultimate sin.

I get all the transparancy I want.

So do I. Only, I'm obviously harder to please than you are, Pat.

You are impossible to please.

Patrick Turner.



Why does Arny suggest that i sin with ultimate badness?

I think I just got KROOGERED!

Figure of speech.

Without external loops of NFB including the emitter follower
connection which is a high amount of applied series voltage NFB,
the BJT is a hopeless
voltage amp unless we try to restrict the current change in the
device, and then the gain is
enormous and subject to the perils of the high collector
resistance
and
circuit slowness due to capacitances, and the non linear input
resistances
of
other devices in the signal path.

Proving once again that if you do something really stupid, you
will get suboptimal results.

.....Another statement of BS, dislocated from the preceding
text....

Not BS, just accepted practice, even in the heyday of tubes, back in
the 60s.

[Patrick Turner]

Prune wrote:

Can you get some comments in on FETs here, any of you guys? There are
quite a number of audio amplifiers that use both JFETs and MOSFETs so I
think it warrants some discussion.

Do a search on Google about j-fet and mosfet basics first.

Patrick Turner.



"Trevor Wilson" wrote in
:


"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...
Modern transistors are more linear than even the best Triodes.

You have said this misleading statement before.

**It's not misleading. It's fact.

Nothing is a fact until all sides agree it is, and you have to proove
what you
say.

**Just look at the curves.






Unfortunately for your arguments this is not the case with regard
to voltage
gain.
Transistors need only a small range of turn on voltage applied
between their
base and emitter,
and their gm varies enormously between being on the threshold of
turn on
of
collector-emitter current and saturation.

**Transistors are controlled by CURRENT, not Voltage. The CURRENT
gain of modern transistors is extremely linear, over a wide range of
Collector currents and Voltages.

Current relationships between base and collector currents are
substantially
linear.

**EXTREMELY linear.

The voltage relationship is not.

**Irrelevant.

You don't have the concepts correctly in your mind.

**Sure I do. Modern BJTs are extremely linear amplification devices.





Only enormous amounts of externally applied NFB corrects the
inherent voltage
non-linearity
of all BJTs where there is substantial current change as in the
case of all
power transistors.

**Transistors (BJTs) are CURRENT amplification devices.

Sure, we all know that, but their voltage linearity is intrinsically
poor.

**Er, that's because they're current amplifiers. And in that sense,
they are exceptionally linear.






Try setting up any bjt in common emitter mode without any emitter
resistor
and
apply
a voltage from a very low impedance to the base. The bjt is
disgustingly
non
linear
compared to a triode using no external applied NFB.

**Transistors (BJTs) are CURRENT amplification devices. Apply a
current to
the Base and measure the change in Collector current. Get back to me
when you've done this.

The current applied to the base is propelled by the applied voltage.
No voltage = no current.

**Duh. No current flow = no Voltage. What's your point?


Set up a small bjt in bread board as a an SE device in class A with
zero NFB and
record the
voltage linearity.

**Points:
* No amplifier on the planet uses zero NFB. Why would you want to
attempt to build one?
* Se amplifiers are dumb. Push pull makes more sense.

Its terrible.
Lots of NFB is needed to linearize the bjt, which is easy, because
they have a
lot of gain...


Even most pentodes which do not have internal electrostatic NFB
like a triode
can easily
out perform a bjt.

**Not even close. Pentodes are vastly less linear than even average
BJTs. Triodes, at least, are reasonably linear.

Your ignorance of tube behaviour is breathtaking.

Try reading RDH4 about the comparison of triode connected pentodes
and the same tube running as a pentode.
At low levels the pentode was more linear than the triode.

**Are we discussing POWER amps, or something else?





I am not here to discuss whether the bjt amp can sound well, or
measure well.
Both are known to be possible so it is believed by most folks.

**Me either. I'm here to correct Jute's lies.


Without external loops of NFB including the emitter follower
connection which is a high amount of applied series voltage NFB,
the BJT is a hopeless
voltage amp

**Transistors (BJTs) are CURRENT amplification devices.

No need to emulate a trained galah.

**It seems you've forgotten this point.




unless we try to restrict the current change in the device, and
then the
gain is

enormous and subject to the perils of the high collector
resistance and circuit slowness due to capacitances, and the non
linear input resistances
of
other
devices in the signal path.

A class A single bjt voltage amp stage with a single MJE340 with
another
for a
constant current source
collector load can have a gain of maybe 1,000, or like a pentode
with a CCS load
in its anode circuit.
If external loops of shunt NFB are applied around either device,
then the
linearity can be quite good for
either device, because for a gain reduction of 100, from 1,000 to
10, the
thd is
also reduced about
100 fold.

The linearity is dependant on the NFB applied, and not on the
basic VOLTAGE
linearity
of the device without NFB, which for a bjt is relatively poor, and
poorest
when
considerable
current change occurs in the device.

**Transistors (BJTs) are CURRENT amplification devices.

Gee, I think I heard that parrot again.

**Until you figure that out, I need to repeat it.


Maybe you could confirm what i am saying if you
went right back to basics with a breadboard and a CRO
and examined the basic workings of the basic devices that mankind has
invented over the last 100 years.

**Sure. Right after you recall that BJTs are current amplification
devices and cannot be used in identical topologies to tubes. They
operate in a fundamentally different way and need to be treated
accordingly.

[Eeyore]

Patrick Turner wrote:

Eeyore wrote:

Patrick Turner wrote:

Eeyore wrote:

If you take a single transistor amplifier circuit and do the same, let's say at ~
1mA Then for a typical small-signal silicon device the gm will be 40mA/V ( basic
physics ). Much greater than any toob, note !

If you restrict the signal level to say +-10% of Iq then the change in gm will be
from ~ 36 to 44 mA/V ( also +- 10% ) . Not exactly an *enormous change* !

But if the R load of the bjt was say 5k6 in a typical circuit with a 12V rail supply,

Who said anything about 12V ?

Because this is a very typical supply voltage in consumer electronics.

Well, not really. Even in the bad old days of single supply rails, 30V was common.


Nowdays hardly anyone uses discrete bjts; it all opamps on double sided boards
with even lower supply voltages.
Measured thd is incredibly low because of huge amounts of NFB applied around the opamps....


then 10% Ic change gives +/- 0.1mA, or 0.39vrms of output.
If the change in gm is 36 to 44, then the gain change over a swing of 0.39Vrms is 5%,
and I dare not calculate the 2H but it will be large.

The use of a triode with 56k load and with a +/-10% in Ia won't result in anything like
the
huge gm change you have in your bjt and the thd will be far less, even at 10 times the
output voltage.

Ok, I've done a couple of experiments in EWB.

One of the nice things about simulation is that you can do experiments using models that
would simply not be feasible in the real world and get astonishingly accurate results.

I simulate nothing, and make only real world comparisons.

Because of it's nature you'd be hard pressed to do this one easily in the real world.

There's nothing wrong with simulation.


My criteria were simple. To accomadate the 12AX7 I made B+ = 300V, Ia = Ic = 2mA and Rl =
50k ohms.

This gives Va = 200V which I'm sure you will recognise is helpful to a triode compared to
using a lower voltage.

I also made an equivalent circuit using the same criteria using an MPSA42 300V small signal
silicon bjt http://www.fairchildsemi.com/ds/MP/MPSA42.pdf using a suitable bias source
and adding an emitter R to give the same voltage gain as the 12AX7.

So, the DC conditions are identical, as is the gain.

Yes but the 12AX7 has no external NFB. The MPSA42 has a lot of linearizing current NFB.

The MPSA without ANY loop NFB is not as linear as a triode.

But that's not how you use it ! And there's no loop here, just an emitter R.

The point is entirely irrelevant. To get a gain of 48x from a silcon bjt you only *need* ~ 2.5V
supply ! That's equally unlikely in the real world.


The 12AX7 gave a voltage gain of 48.2x
For an input voltage swing of +/- 0.1 V the voltage swing at the anode was +4.8046 and -
4.837 V
This translates to a linearity of 0.336%

For the MPSA42 the voltage gain was in fact 48.8x ( keeping to preferred values )
The voltage swing at the collector was +4.882 and - 4.885 V
This translates to a linearity of 0.034 % ( or 10 times better )

I also ran a test using an npn/pnp compound pair with all other parameters the same and the
linearity was ***0.002%***

***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Such contests run on a non level playing field tell us all nothing except that your'e
good at conductiong silly sporting events.

Denying the bjt the advantage of a 300V B+ is also making the playing field not level.


The bjt is very much dependant on NFB loops because of its intrisically non linear
character.

Using differential amplifiers with triodes with CCS loads will give you extremely low thd
without any current or other NFB.

And your CCS load is ? a BJT !


In my 300w SS amps I have symetrical npn and pnp diff amps in parallel
at the input, and a complementary pair in common emitter for the driver of the mosfet outputs.
http://www.turneraudio.com.au/solids...s1mosfets.html
The supply is +/- 80V, and there is about 4% ths at about 40Vrms max of Vswing.
This is a lot worse than a pair of PP triodes in a driver can achieve, maybe 0.1% at 150Vrms
a-a.
http://www.turneraudio.com.au/schem-...ut+output.html

The global NFB applied in the mosfet amp reduces all this to less than 0.01% at 300 watts.
The thd falls about proportionately to Vo, so that at 1 watt where Vo = 2.83V the thd =
0.001%

I tend to linearise locally first so no stage produces that much thd.


The linearity is a function of applied NFB and gain, but the bjt devices
ain't linear when examined without NFB.

You also wouldn't typically use a bjt in a similar configuration to a tube though. Nor undeed have
to ac couple everywhere !


The mosfet output stage I have also would produce about 5% thd at 300 watts
with common source mode but in common drain mode, ie, source follower,
this Dn, mainly 3H plus other odd order junk is less than 1%,

That sounds very similar to my own results too.


and less than the drive amp Dn,
so the output stage Dn does not dominate the THD production as occurs in a tube amp.
Both drive amp and output stage Dn is reduced equally by the global NFB.

You have not won.

SS devices are less linear than triodes, but ARE adequately linear and DO HAVE
oddles of gain which allows NFB to make them linear.

Most tube amps don't need more than 20dB of NFB in the form of
output stage NFB and global NFB.

Some SET amps don't need any global NFB because they are substantially linear
without NFB.

I have heard plenty of systems to confirm my beliefs and tolerance of
SET amplifiers.

The point of my simulation is to show that a single stage of amplification using a transistor
using the same voltages, currents and gain can perform more linearly than a triode. In that
respect it outperforms the tube hands down.

The compound configuration is simply stunning and is more typical of real practice using solid
state too.

Graham

[Andre Jute]

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must
smoking bad dope, Wilson.

Try again.

Andre Jute
Habit is the nursery of errors. -- Victor Hugo

[Trevor Wilson]

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must
smoking bad dope, Wilson.

**Then please provide a complete list of (power) transistors which you have
direct experience with. Let's see how up to date your knowledge of silicon
devices actually is. Please cite the curves which show how non-linear these
devices are.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[Andre Jute]

Trevor Wilson wrote:

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must be
smoking bad dope, Wilson.

**Then please provide a complete list of (power) transistors which you have
direct experience with. Let's see how up to date your knowledge of silicon
devices actually is. Please cite the curves which show how non-linear these
devices are.

--
Trevor Wilson
www.rageaudio.com.au

1. Oh, dear, no sense of humour at Rage Audio. Typical silicon slime
characteristic.

2. I am not on trial for my knowledge of transistors. I don't do
transistors except to prove a point (my 675 amp) or as pseudo-choke
loads for more respectable devices.

3. I do tubes. This is a tube conference. You're a guest here and you
stink like three day-old fish.

4. I make music with devices that work without kludges. The music they
makes suits me fine. If I want advice on how to improve my music, I
certainly won't seek it from a choleric mouthfoamer like you.

5. Are these transistor curves you're demanding from me (I'm at a loss
to understand why you believe I should possess anything so
incriminating -- read the thread title!) before or after the
transistors have been kludged half-right with excessive NFB and left
contaminated with nervewracking high-order artifacts?

6. You have nothing useful to say to me. Please have the last word. I
shant' be bothering with you any more.

Andre Jute

[Trevor Wilson]

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must be
smoking bad dope, Wilson.

**Then please provide a complete list of (power) transistors which you
have
direct experience with. Let's see how up to date your knowledge of
silicon
devices actually is. Please cite the curves which show how non-linear
these
devices are.

--
Trevor Wilson
www.rageaudio.com.au

1. Oh, dear, no sense of humour at Rage Audio. Typical silicon slime
characteristic.

2. I am not on trial for my knowledge of transistors. I don't do
transistors except to prove a point (my 675 amp) or as pseudo-choke
loads for more respectable devices.

**Then don't make idiotic comments about how linear Triodes are. If you
don't know how good modern transistors are, then you can't tell the world
that Triodes are better. Sure, A good Triode is better than a 2N3055, but it
can't match a modern transistor.



3. I do tubes. This is a tube conference. You're a guest here and you
stink like three day-old fish.

**My rights to exist here are precisely equivalent to yours. I have every
right to expose your lies/stupidity, whenever I see fit.


--
Trevor Wilson
www.rageaudio.com.au



--
Posted via a free Usenet account from http://www.teranews.com

[Prune]

Eeyore wrote in
:

I suspect that their 'current mode' is a marketing term.

It's not. Someone had posted schematics of a preamp at diyaudio a while
back. It was a current amplifier. They are also pretty clear about it in
the documents.

[Patrick Turner]

Eeyore wrote:

Patrick Turner wrote:

Eeyore wrote:

Patrick Turner wrote:

Eeyore wrote:

If you take a single transistor amplifier circuit and do the same, let's say at ~
1mA Then for a typical small-signal silicon device the gm will be 40mA/V ( basic
physics ). Much greater than any toob, note !

If you restrict the signal level to say +-10% of Iq then the change in gm will be
from ~ 36 to 44 mA/V ( also +- 10% ) . Not exactly an *enormous change* !

But if the R load of the bjt was say 5k6 in a typical circuit with a 12V rail supply,

Who said anything about 12V ?

Because this is a very typical supply voltage in consumer electronics.

Well, not really. Even in the bad old days of single supply rails, 30V was common.

For power amps yeah, but few preamp makers thought there was any need for 30V.
30V is of course much better because the swing from an SE device could be up to
about 14V peak, or 10vrms, and thus where only a volt was needed there was a good chance the THD
was very low because of all the NFB, but never because the devices were linear.



Nowdays hardly anyone uses discrete bjts; it all opamps on double sided boards
with even lower supply voltages.
Measured thd is incredibly low because of huge amounts of NFB applied around the opamps....


then 10% Ic change gives +/- 0.1mA, or 0.39vrms of output.
If the change in gm is 36 to 44, then the gain change over a swing of 0.39Vrms is 5%,
and I dare not calculate the 2H but it will be large.

The use of a triode with 56k load and with a +/-10% in Ia won't result in anything like
the
huge gm change you have in your bjt and the thd will be far less, even at 10 times the
output voltage.

Ok, I've done a couple of experiments in EWB.

One of the nice things about simulation is that you can do experiments using models that
would simply not be feasible in the real world and get astonishingly accurate results.

I simulate nothing, and make only real world comparisons.

Because of it's nature you'd be hard pressed to do this one easily in the real world.

There's nothing wrong with simulation.

No, its easy to set up a bread board. Its lazy not to do so.



My criteria were simple. To accomadate the 12AX7 I made B+ = 300V, Ia = Ic = 2mA and Rl =
50k ohms.

This gives Va = 200V which I'm sure you will recognise is helpful to a triode compared to
using a lower voltage.

I also made an equivalent circuit using the same criteria using an MPSA42 300V small signal
silicon bjt http://www.fairchildsemi.com/ds/MP/MPSA42.pdf using a suitable bias source
and adding an emitter R to give the same voltage gain as the 12AX7.

So, the DC conditions are identical, as is the gain.

Yes but the 12AX7 has no external NFB. The MPSA42 has a lot of linearizing current NFB.

The MPSA without ANY loop NFB is not as linear as a triode.

But that's not how you use it ! And there's no loop here, just an emitter R.

Don't talk rubbish. The EMITTER R IS A LOOP of series negative current feedback,
one that raises the already high collector resistance so that R out is determined by the
value of the R load connected.

The emitter R forces the combination of bjt and Re to act as a purer constant
signal current powering the load. Nothing wrong with it; its good practice.




The point is entirely irrelevant. To get a gain of 48x from a silcon bjt you only *need* ~ 2.5V
supply ! That's equally unlikely in the real world.

The 12AX7 gave a voltage gain of 48.2x
For an input voltage swing of +/- 0.1 V the voltage swing at the anode was +4.8046 and -
4.837 V
This translates to a linearity of 0.336%

For the MPSA42 the voltage gain was in fact 48.8x ( keeping to preferred values )
The voltage swing at the collector was +4.882 and - 4.885 V
This translates to a linearity of 0.034 % ( or 10 times better )

I also ran a test using an npn/pnp compound pair with all other parameters the same and the
linearity was ***0.002%***

***** 0.002% ***** !!!!! BJT wins !

Eat your hearts out tubies !

Such contests run on a non level playing field tell us all nothing except that your'e
good at conductiong silly sporting events.

Denying the bjt the advantage of a 300V B+ is also making the playing field not level.

Trying to run a 12AX7 on a field with 12V supply is equally stoopid.

But much consumer electronics routinely use between 200 and 350V for
tube preamps and input stages of power amps and between 10V and 20V for bjt circuits.




The bjt is very much dependant on NFB loops because of its intrisically non linear
character.

Using differential amplifiers with triodes with CCS loads will give you extremely low thd
without any current or other NFB.

And your CCS load is ? a BJT ! Indeed, but the CCS could be another tube, but the bjt CCS is far
cheaper.



In my 300w SS amps I have symetrical npn and pnp diff amps in parallel
at the input, and a complementary pair in common emitter for the driver of the mosfet outputs.
http://www.turneraudio.com.au/solids...s1mosfets.html
The supply is +/- 80V, and there is about 4% ths at about 40Vrms max of Vswing.
This is a lot worse than a pair of PP triodes in a driver can achieve, maybe 0.1% at 150Vrms
a-a.
http://www.turneraudio.com.au/schem-...ut+output.html

The global NFB applied in the mosfet amp reduces all this to less than 0.01% at 300 watts.
The thd falls about proportionately to Vo, so that at 1 watt where Vo = 2.83V the thd =
0.001%

I tend to linearise locally first so no stage produces that much thd.

The THD of the drive stage is of little consequence because of the amount of NFB
and the spectral content of the combined driver and output stage
is benign because the most of the output AB stage Xover Dn is removed in the
source follower connection. The 0.01% thd at 300W is mainly 3H with other H well down.

I could try linearizing the simple VAS more but it means more devices, and more FB loops
within the global one and then stability can suffer.
I'd then get maybe Dn 0.001% at 300W and Dn at 1 watt which is so low it is difficult to measure.
I doubt there would be the slightest sound change.
The input stages are also simple and produce such a small signal that
there is no need for any further linearization with current mirror loading of the diff pairs.




The linearity is a function of applied NFB and gain, but the bjt devices
ain't linear when examined without NFB.

You also wouldn't typically use a bjt in a similar configuration to a tube though. Nor undeed have
to ac couple everywhere !

I do use bjts and mosfets like one would use tubes in a tube circuit.

See

http://www.turneraudio.com.au/solids...no-mosfet.html

This has 4 active devices, one CCS tail for the input diff pair,
and cap coupling and an OPT.

3 FB loops exist.

Its sounds well.




The mosfet output stage I have also would produce about 5% thd at 300 watts
with common source mode but in common drain mode, ie, source follower,
this Dn, mainly 3H plus other odd order junk is less than 1%,

That sounds very similar to my own results too.

and less than the drive amp Dn,
so the output stage Dn does not dominate the THD production as occurs in a tube amp.
Both drive amp and output stage Dn is reduced equally by the global NFB.

You have not won.

SS devices are less linear than triodes, but ARE adequately linear and DO HAVE
oddles of gain which allows NFB to make them linear.

Most tube amps don't need more than 20dB of NFB in the form of
output stage NFB and global NFB.

Some SET amps don't need any global NFB because they are substantially linear
without NFB.

I have heard plenty of systems to confirm my beliefs and tolerance of
SET amplifiers.

The point of my simulation is to show that a single stage of amplification using a transistor
using the same voltages, currents and gain can perform more linearly than a triode. In that
respect it outperforms the tube hands down.

But your experiment uses the bjt with a large amount of current NFB.

Your experiment is thus quite flawed because you are not comparing devices
operating completely free of external loops of NFB.



The compound configuration is simply stunning and is more typical of real practice using solid
state too.

I have nothing against the use of whatever compounded topologies you dream up.

I do have something against people trying to tell me bjts are just as good voltage amplifiers as
triodes are.

The bjts are simply not linear voltage amplifiers unless their behaviour is much controlled by
oddles of NFB, and I am not averse to using bucket fulls of NFB around SS amps.

I once had a gang of fellows around to listen to some music and set up secret switch so the speakers
could be
switched from a class A PP tube amp with 4 x 6550 per channel to the 300W mosfet amp.
The source was from the same twinned outputs of the same preamp.
During the evening I switched to the mosfet amp and told everyone I'd made a big change.
They looked around at what i had there and couldn't find anything, and couldn't hear
a sound change. After 1/2 an hour and a few more tracks I owned up to the trick.
One guy then said the mosfet amp sounded a bit more "mosfet misty" .
Yeah, sure.
Nowdays I don't know if I'd get away with the same trick because my speakers are so much better
and I have a better CD player.

Changing from 0.02% tube thd to 0.00??% thd in the mosfet amps was entirely inaudible.

I could say both sounded OK because they were designed and built by the same person.

I have witnessed ppl who have listened to SE tube amps for the first time after listening to good PP
amps,
not knowing anything at all about them. Often they find the sound much more natural and detailed,
musical and
just plain right. I have never had it occur the other way around.

Despite the drift towards efficient near class B circuits with high NFB and low cost, the SE tube amp
continues to stalk the hi-fi market and finds buyers despite the weight, heat, cost and
occasional tube replacement.

I've auditioned AB comparisons with SET 8 watt amps against 100W/ch SS amps and found
the tubes sounded better and measured worse. I've seen guys lose when they traded their SS amps away
for ****y little SET amps.
Its the better sound that makes the impression to some folks.

The classic case of poor component matching was the guy with two Yamaha 2200 stereo amps
(total of 800 watts possible) for
biamping to 95dB efficient JBL monitors with 2 x 15" bass-mids and horn loaded tweets in each
ginormous
box in a room 10' x 12' with 8' ceiling.
Do I take much notice of what this dude says we ought to have? Well of course,
but I keep my thoughts private.

Patrick Turner.







Graham

[Patrick Turner]

Trevor Wilson wrote:

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must
smoking bad dope, Wilson.

**Then please provide a complete list of (power) transistors which you have
direct experience with. Let's see how up to date your knowledge of silicon
devices actually is. Please cite the curves which show how non-linear these
devices are.

But dear Trevor, you are not exactly experienced with power transistors now are
you
except by way of repairing SS amps?

I do believe your own admission about expeirence sums up your design and build
expereince as equal to
zero since 1976, a year when one Peter Stein began serious production in
Australia of fine
solid state ME Technology amplifiers which became as well known in Oz as
vegemite,
a nutricious yeast extract for sandwiches. Alas hardly anyone in the main two
Oz
audio clubs like ME amps. You once said you didn't need to design anything
because Peter
had it sussed.

The curves being referred to here are the base input current to output
collector current.
These are very linear in most bjts.

But the curves which are never shown in data sheets are the Vbase-emitter Vs
Collector current.

The trouble with news groups is that there are plenty of blokes wanting to say
things
for the sake of saying things, and winning arguments rather than facing the
realities of the
devices they like to favour.

The falsely educated man depends on an argument like a drunk needs a lamp post.

Seldom does the educated man utilise the lamp to shed light on the argument,
lest
he feel the pain of being wrong about the basics.

Patrick Turner.





--
Trevor Wilson
www.rageaudio.com.au

--
Posted via a free Usenet account from http://www.teranews.com

[Prune]

Patrick Turner wrote in
:

The curves being referred to here are the base input current to output
collector current.
These are very linear in most bjts.

But the curves which are never shown in data sheets are the
Vbase-emitter Vs Collector current.

So let me ask something akin to what I've asked elsewhere. Why not use
BJTs for current gain and triodes for voltage gain? Indeed, the
electrostatic headphone amplifier I was asking about some time ago does
exactly that (though also uses MOFETs for some of the voltage gain, and
I've yet to find help here about my question with replacing that stage with
triodes as well).

[Patrick Turner]

Trevor Wilson wrote:

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

"Andre Jute" wrote in message
oups.com...

Trevor Wilson wrote:

Jute continually refer to 2N3055
--
Trevor Wilson
www.rageaudio.com.au

No, I don't.

No such tube as the 2N3055 in my Compleat Mullard List of Really Good
Valves.

No such transistor as the 2N3055 in my Completely Up-to-date Germanium
Reference.

Just called my mate Bob Norris at Fairchild and he says you must be
smoking bad dope, Wilson.

**Then please provide a complete list of (power) transistors which you
have
direct experience with. Let's see how up to date your knowledge of
silicon
devices actually is. Please cite the curves which show how non-linear
these
devices are.

--
Trevor Wilson
www.rageaudio.com.au

1. Oh, dear, no sense of humour at Rage Audio. Typical silicon slime
characteristic.

2. I am not on trial for my knowledge of transistors. I don't do
transistors except to prove a point (my 675 amp) or as pseudo-choke
loads for more respectable devices.

**Then don't make idiotic comments about how linear Triodes are. If you
don't know how good modern transistors are, then you can't tell the world
that Triodes are better. Sure, A good Triode is better than a 2N3055, but it
can't match a modern transistor.

But you continue to make the idiotic comparison between triodes and bjts based
on entirely
false perceptions.

We already know the input / output VOLTAGE linearity of any appropriate triode
is better than a 2N3055.

We know that bjts have good current linearity.

Current linearity cannot be compared simply to voltage linearity.
Its like comparing oranges to asperagus

But Trevor does not know anything about the input / output voltage linearity of
bjts.

He thinks the current linearity is the same as voltage linearity.

In tubes, input / output current linearity cannot exist because there is no grid
current
unless the grid is run positively, and sometimes this is attempted, and the
voltage linearity
is not much disturbed, and the current linearity of tubes is quite poor.

In another 10 years maybe the Trevors of this world will learn to
be slower to shoot their mouth off about what they don't fully understand,
and then realize that understanding basics isn't the land of simplistic one
liners,
and that each and every statement about device operation is only quite valid
when a group
of operating conditions is clearly defined and kept in mind and correct context
when discussions occur.

Patrick Turner.






3. I do tubes. This is a tube conference. You're a guest here and you
stink like three day-old fish.

**My rights to exist here are precisely equivalent to yours. I have every
right to expose your lies/stupidity, whenever I see fit.

--
Trevor Wilson
www.rageaudio.com.au

--
Posted via a free Usenet account from http://www.teranews.com