[John Byrns]

In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
This discussion isn't a matter of knowing much of anything, if I read
Henry correctly he is not saying that the operation of the triode doesn't
involve NFB, or that it does for that matter, all Henry is saying is that
there are several alternate ways to describe the operation of the triode
when it is viewed as a black box.

That is one of my conclusions, but not necessarily the most important
one. The main purpose of my postings has been to dissect the triode
NFB claims, and to debunk them -- carefully and critically. Doing so
requires both knowledge and analysis skill.

It sounds like I misread your posts, and you are actually claiming that
the triode tube absolutely doesn't have internal negative feedback, not
simply that it may or may not have internal negative feedback?


Regards,

John Byrns

[Henry Pasternack]

"John Byrns" wrote in message ...
It sounds like I misread your posts, and you are actually claiming that
the triode tube absolutely doesn't have internal negative feedback, not
simply that it may or may not have internal negative feedback?

My purpose has been to demonstrate that Patrick's arguments to the
effect that a triode absolutely *does* have NFB are inconsistent and
invalid. I believe it is possible to model a triode (in at least two ways)
in terms of negative feedback, but I think such models are contrived
and serve no useful purpose.

Practically speaking, no, after thinking hard about this for the past few
days, I really don't feel that a triode "has" internal negative feedback.
Negative feedback is not a "thing". It's an abstraction and the extent
to which a device "has" it really depends on how well that abstraction
fits the physical characteristics of the device. It's not necessarily a
black-and-white thing. In this case, I don't think the fit is very good.

-Henry

[John Byrns]

In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
It sounds like I misread your posts, and you are actually claiming that
the triode tube absolutely doesn't have internal negative feedback, not
simply that it may or may not have internal negative feedback?

My purpose has been to demonstrate that Patrick's arguments to the
effect that a triode absolutely *does* have NFB are inconsistent and
invalid. I believe it is possible to model a triode (in at least two ways)
in terms of negative feedback, but I think such models are contrived
and serve no useful purpose.

Practically speaking, no, after thinking hard about this for the past few
days, I really don't feel that a triode "has" internal negative feedback.
Negative feedback is not a "thing". It's an abstraction and the extent
to which a device "has" it really depends on how well that abstraction
fits the physical characteristics of the device. It's not necessarily a
black-and-white thing. In this case, I don't think the fit is very good.

I don't agree that the feedback model is contrived and serves no useful
purpose, the triode model I use in spice is based on a feedback
equation. Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could be
onto something.


Regards,

John Byrns

[Henry Pasternack]

"John Byrns" wrote in message ...
I don't agree that the feedback model is contrived and serves no useful
purpose, the triode model I use in spice is based on a feedback
equation. Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could be
onto something.

OK.

-Henry

[Ian Iveson]

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Could you post it please? It is not too many lines of text, I assume.

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could
be
onto something.

Who was that please?

thanks, Ian


"John Byrns" wrote in message
...
In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
It sounds like I misread your posts, and you are actually
claiming that
the triode tube absolutely doesn't have internal negative
feedback, not
simply that it may or may not have internal negative feedback?

My purpose has been to demonstrate that Patrick's arguments to the
effect that a triode absolutely *does* have NFB are inconsistent
and
invalid. I believe it is possible to model a triode (in at least
two ways)
in terms of negative feedback, but I think such models are
contrived
and serve no useful purpose.

Practically speaking, no, after thinking hard about this for the
past few
days, I really don't feel that a triode "has" internal negative
feedback.
Negative feedback is not a "thing". It's an abstraction and the
extent
to which a device "has" it really depends on how well that
abstraction
fits the physical characteristics of the device. It's not
necessarily a
black-and-white thing. In this case, I don't think the fit is very
good.


Regards,

John Byrns

[Patrick Turner]

Henry Pasternack wrote:

"John Byrns" wrote in message ...
It sounds like I misread your posts, and you are actually claiming that
the triode tube absolutely doesn't have internal negative feedback, not
simply that it may or may not have internal negative feedback?

My purpose has been to demonstrate that Patrick's arguments to the
effect that a triode absolutely *does* have NFB are inconsistent and
invalid. I believe it is possible to model a triode (in at least two ways)
in terms of negative feedback, but I think such models are contrived
and serve no useful purpose.

I see NFB in every triode when its set up to allow the NFB to work;
ie, the control grid is the input terminal and the anode voltage is the output
and free to move on a load.
The two voltages of input and output sum inside the tube to be a virtual grid
as the real controlling element of the anode voltage, or current, or both.

When the grid is at a fixed voltage in reference to the cathode, and input is to the anode,
then the triode acts as a diode.
Only one arm of the shunt NFB network exists, ie, from anode to space charge
so the anode voltage determines the anode current but in conjunction with a fixed
Eg.
If the grid is removed to make a diode, one cannot say an interactive loop of NFB exists
but still the anode "feeds back" a voltage field which controls Ia.

NFB is working when there is the interaction of input and output to form a third
net or summed effect, either by way of series NFB as used so commonly in global NFB loops,
or shunt NFB used less commonly.

So when you say I say that triodes do have NFB absolutely, one has to consider where
the triode is being used. I don't see NFB when its used as a diode.



Practically speaking, no, after thinking hard about this for the past few
days, I really don't feel that a triode "has" internal negative feedback.
Negative feedback is not a "thing". It's an abstraction and the extent
to which a device "has" it really depends on how well that abstraction
fits the physical characteristics of the device. It's not necessarily a
black-and-white thing. In this case, I don't think the fit is very good.

But the interaction of the voltage fields in a triode are real, and therefore are a thing,
and a thing to consider, no?

I leave you to consider.

Patrick Turner.


-Henry

[Patrick Turner]

John Byrns wrote:

In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
It sounds like I misread your posts, and you are actually claiming that
the triode tube absolutely doesn't have internal negative feedback, not
simply that it may or may not have internal negative feedback?

My purpose has been to demonstrate that Patrick's arguments to the
effect that a triode absolutely *does* have NFB are inconsistent and
invalid. I believe it is possible to model a triode (in at least two ways)
in terms of negative feedback, but I think such models are contrived
and serve no useful purpose.

Practically speaking, no, after thinking hard about this for the past few
days, I really don't feel that a triode "has" internal negative feedback.
Negative feedback is not a "thing". It's an abstraction and the extent
to which a device "has" it really depends on how well that abstraction
fits the physical characteristics of the device. It's not necessarily a
black-and-white thing. In this case, I don't think the fit is very good.

I don't agree that the feedback model is contrived and serves no useful
purpose, the triode model I use in spice is based on a feedback
equation. Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could be
onto something.

Regards,

John Byrns

Maybe 2 years ago when NFB in triodes was considered and discussed here
I came up with a few formulas myself.
Alas i don't have copies of what i said at the time.

What formulas are you refering to?

Patrick Turner.

[Henry Pasternack]

"Patrick Turner" wrote in message ...
[Baloney deleted]

Maybe Ian was right.

-Henry

[Peter Wieck]

Andrew Jute McCoy created of whole cloth:
So a deputation
of the directors went immediately to Drucker's office to make him an
offer he couldn't refuse. My chairman was a huge fellow in his early
thirties with short, very blond hair. The rest of us were muscular
young men (high-profile sporting achievements were essential to rising
in our organization -- that's how we met most of our clients) of
definitely Aryan stock. Drucker was a Hungarian, a refugee from the
Nazis...

1. Drucker was Austrian.
2. He left Germany by choice in 1933 based on conscience, not fear. The
sort of conscience you lack.
3. He would have destroyed the likes of you and your imaginary gang
without thought, and certainly without fear.

On this, you need to be very careful, Mr. McCoy. My father and Mr.
Drucker were contemporaries, went to law school together (Drucker was
one term ahead), left Germany on the same transport at the same time,
and remained much better than acquaintances until my father's death in
1973.

You lie the way others breath.

Peter Wieck
Wyncote, PA

[Patrick Turner]

Henry Pasternack wrote:

"Patrick Turner" wrote in message ...
[Baloney deleted]

Maybe Ian was right.

-Henry

Its not often I would ever delete someone's post while bracketing
to tell folks that all of what someone said was baloney.

But if Ian was right, I'd not be at all worried, since its such a rare occasion.

But he was right to ask for a formula used for the NFB in a triode, and he didn't
say categorically that there was no NFB in triodes, as he has before.

As soon as someone mentioned simulation programs and people's models of triodes
incorporating a formula to express the NFB he wanted to see if there was any validity.

But If I ever say something about NFB and express the concept, he
almost automatically acts to show he is incapable of apreciating concepts,
and calls me a bully and BS artist et all because I challenge his stance.

He tries to rule over us, but I cannot be ruled by anyone.

Have a great day contemplating NFB in triodes, whether or not it exists,
and how the NFB may or may not be expressed by formulas.

I'll give you a tip.
All the electrostatic activity in vacuum tubes, actions of grids anodes and other complex interactions
and positioning of space charges on electron streams et all
has exhaustively been worked out and written about fairly conclusively by around 1935.
Just think of a 6L7 multigrid F converter used in radio; man, what complexity!
This "stuff" found in old books may or may not have been noticed by writers of tube
simulation programs, to allow them to accurately model a triode
given its basic cathode, grid, screen, anode dimensions, positionings, and grid wire spacing.
All actions by all this stuff are covered by formulas.

Guys who designed tubes didn't just guess their way towards a successful 300B design,
or a successful 6V6 design.

Modern folks used to instant easy understanding without having to place themselves in the position
of a tube designer in a factory lab are flumoxed and bewildered by the simplest of math.
But respect the math they must, and perhaps if they cannot handle concepts
without the math like I have tried to share with triode NFB then perhaps with the math they may see
there is more to a triode than it being some kind of vacuum diode or strange sort of resistor.

Neither resistors or diodes amplify anything, so what relevance is there if they contain NFB?
I don't think two prong devices can have NFB, and I think one needs
4 terminals, one of which is shared by two circuits, the input grid circuit and output anode-load-PS circuit

The shared terminal is the cathode.


Just my 2c.

Patrick Turner.

[John Byrns]

In article , "Ian Iveson"
wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Could you post it please? It is not too many lines of text, I assume.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

The triode model has three essential elements. The first is a
transconductance amplifier, or voltage to current converter, which drives
the output terminal of the model. The input to the voltage to current
converter is taken from the output of a three halves power function which
simulates the inherent distortion producing non linearity of the triode,
this comes from child's diode law. The final element in the model is a
summer which creates the negative feedback and feeds the input of the
three halves power law function. The inputs to the summer are the
voltages at the input and output terminals of the triode model, scaled by
appropriate constants. Note that the voltage to current converter is
connected so that on a voltage basis it provides the phase inversion
necessary for the feedback to be negative rather than positive. It's that
simple, and the negative feedback in this model should be obvious to even
the most dimwitted among us, although that doesn't imply one way or the
other whether a real triode actually has internal negative feedback like
this model does.

I am eagerly awaiting Henry Paternack's application of his EE degree to
create a model where everything is done in the forward path, and there is
no feedback involved, it should be something to behold.

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could
be onto something.

Who was that please?

It was a long time ago so I haven't a clue at this point who it was. All
I can say is that this news group was the first place I saw the idea, and
while I saw it here towards the end of the previous century, it wouldn't
surprise me if the idea is considerably older, and predates the internet
by quite a few years. You can find some old references to the idea using
Google, although the first mention of it in this news group may be hard to
pin down without a fairly extensive search effort.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

[HP]

"John Byrns" wrote in message ...
Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

[Description deleted]

This model is too simple to account for all but the most basic distortion
mechanisms in the tube.

I am eagerly awaiting Henry Paternack's application of his EE degree to
create a model where everything is done in the forward path, and there is
no feedback involved, it should be something to behold.

Your sarcasm is unnecessary. Many people have worked on this problem.
You can find much information about it on the web, for example, he

http://digilander.libero.it/paeng/sp...cuum_tubes.htm

There is a fellow named Mark Kelly who posts frequently on the AudioAsylum
Tube/DIY forum. Mark is a vintner, a math fiend, and a very smart guy. He
has worked extensively on modeling the distortion in the 300B. If you are
interested in some authoritative information on this subject, I suggest you look
him up.

-Henry

[John Byrns]

In article ,
"HP" wrote:

"John Byrns" wrote in message
...
Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

[Description deleted]

This model is too simple to account for all but the most basic distortion
mechanisms in the tube.

Perhaps my choice of words was poor, I did not mean to imply that the
simple triode model I described included anything except the most basic
triode non linearity. My point was that this model makes use of
negative feedback.

I am eagerly awaiting Henry Paternack's application of his EE degree to
create a model where everything is done in the forward path, and there is
no feedback involved, it should be something to behold.

Your sarcasm is unnecessary.

What you actually mean is that you can't deliver any evidence to support
your claim.

Many people have worked on this problem.

If many people have actually worked on this problem, as you claim,
please point me to some results that show how to create even a simple
triode model that includes the basic triode non linearity and does not
rely on negative feedback in the model.

You can find much information about it on the web, for example, he

http://digilander.libero.it/paeng/sp...cuum_tubes.htm

A quick scan of this site shows no information on triode models that
don't make use of negative feedback, I will have to go through the site
in more detail later to see if your elusive forward path only model is
actually described there.

There is a fellow named Mark Kelly who posts frequently on the AudioAsylum
Tube/DIY forum. Mark is a vintner, a math fiend, and a very smart guy. He
has worked extensively on modeling the distortion in the 300B. If you are
interested in some authoritative information on this subject, I suggest you
look
him up.

I have little or no interest in the more subtle distortion mechanisms in
triode tubes, this subthread is about internal negative feedback in
triodes and that is what I am addressing.

You made an implicit claim that the triode tube can be modeled
mathematically without using negative feedback, and you also mentioned
Child's law. All I am asking is that you describe a triode model that
takes account of Child's law, and does it all in the forward path, or is
that too much to ask?


Regards,

John Byrns


-Henry

[Henry Pasternack]

"John Byrns" wrote in message ...
[Deleted]

Shame on you, Byrns.

-Henry

[Ian Iveson]

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Who knows? He was hereabouts a while ago. I may have misrepresented
him...he may simply have claimed that my models would be better with
feedback. He was cheeky like that.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

SPICE is not hard, John. It is a simple sequence of algebra plainly
stated in text. The only complication is the way negative values are
excluded. We can all handle that I'm sure. If you show your actual
model, which is only a few lines long and probably shorter than your
wordy description, I can explain the spice part to anyone who doesn't
follow it.

Then we can see whether it is really true that

...the negative feedback in this model should be obvious to even
the most dimwitted among us...

Obviously you *describe* the model in terms of feedback. You did so
before I queried your statement. Further description from you in terms
of feedback doesn't really serve to back up your claim, does it?

...It's that
simple, and , although that doesn't imply one way or the
other whether a real triode actually has internal negative feedback
like
this model does.

Eh? But you said

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they
could
be onto something.

You have lost your own thread, obviously. Perhaps you are hoping to
avoid the question of where the dummy summing node is in a real
triode?

I could describe a resistor model as follows:

The input voltage to the resistor is applied to two terminals and used
as one input to a summing node. The other input to the summing node is
the voltage dropped across the two terminals by the output current,
according to ohm's law. The final output current through the resistor
is driven by a transconductance amplifier with high gain which takes
its input from the summing node.

That is, the current rises until the voltage dropped across the
resistor equals the input voltage.

But a resistor doesn't have internal feedback in the proper
engineering sense. The example is often used to demonstrate how
feedback theory can be used for *anything*, regardless of whether the
feedback is real or imaginary. Load regulation is not sufficient to
imply feedback. Necessary but not sufficient: it's a common source of
confusion.

Perhaps your SPICE model is the same as mine. Load regulation, yes.
Feedback, no. Unfortunately the few remaining dimwits amongst us
remain unenlightened, and still don't know the difference.

cheers, Ian

[Andre Jute]

John Byrns wrote:
In article , "Ian Iveson"
wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

[snip]

Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

Flipper was a young man with great clarity of thought and superior
logic who graced us with his presence for a while. But I think he
decided he could learn more electronics from a book faster than he
could in the poisoned atmosphere on RAT. Flipper is exactly the sort of
experimenter we should move heaven and earth to keep. But of course the
usual clowns, spraying on their worthless little piece of wasteland
like squabbling cats, ganged up to make things unpleasant for him.

Andre Jute

[John Byrns]

Ian, all I claimed was that the triode spice model I use is based on a
negative feedback equation, I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Whe

Ipk is the output current into the plate or output terminal of the model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the model
relative to the cathode or common terminal


Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.


Regards,

John Byrns


In article ,
"Ian Iveson" wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Who knows? He was hereabouts a while ago. I may have misrepresented
him...he may simply have claimed that my models would be better with
feedback. He was cheeky like that.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

SPICE is not hard, John. It is a simple sequence of algebra plainly
stated in text. The only complication is the way negative values are
excluded. We can all handle that I'm sure. If you show your actual
model, which is only a few lines long and probably shorter than your
wordy description, I can explain the spice part to anyone who doesn't
follow it.

Then we can see whether it is really true that

...the negative feedback in this model should be obvious to even
the most dimwitted among us...

Obviously you *describe* the model in terms of feedback. You did so
before I queried your statement. Further description from you in terms
of feedback doesn't really serve to back up your claim, does it?

...It's that
simple, and , although that doesn't imply one way or the
other whether a real triode actually has internal negative feedback
like
this model does.

Eh? But you said

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they
could
be onto something.

You have lost your own thread, obviously. Perhaps you are hoping to
avoid the question of where the dummy summing node is in a real
triode?

I could describe a resistor model as follows:

The input voltage to the resistor is applied to two terminals and used
as one input to a summing node. The other input to the summing node is
the voltage dropped across the two terminals by the output current,
according to ohm's law. The final output current through the resistor
is driven by a transconductance amplifier with high gain which takes
its input from the summing node.

That is, the current rises until the voltage dropped across the
resistor equals the input voltage.

But a resistor doesn't have internal feedback in the proper
engineering sense. The example is often used to demonstrate how
feedback theory can be used for *anything*, regardless of whether the
feedback is real or imaginary. Load regulation is not sufficient to
imply feedback. Necessary but not sufficient: it's a common source of
confusion.

Perhaps your SPICE model is the same as mine. Load regulation, yes.
Feedback, no. Unfortunately the few remaining dimwits amongst us
remain unenlightened, and still don't know the difference.

cheers, Ian

[Peter Wieck]

Andrew Jute McCoy blathered:
Its usual bunch of crap.

Tell us more about Peter Drucker.

Peter Wieck
Wyncote, PA

[John Byrns]

In article . com,
"Andre Jute" wrote:

John Byrns wrote:
In article , "Ian Iveson"
wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

[snip]

Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

Flipper was a young man with great clarity of thought and superior
logic who graced us with his presence for a while. But I think he
decided he could learn more electronics from a book faster than he
could in the poisoned atmosphere on RAT. Flipper is exactly the sort of
experimenter we should move heaven and earth to keep. But of course the
usual clowns, spraying on their worthless little piece of wasteland
like squabbling cats, ganged up to make things unpleasant for him.

Andre Jute

I can't locate any of Flipper's posts, was (s)he partial to using the no
archive header?


Regards,

John Byrns

[Patrick Turner]

John Byrns wrote:

In article , "Ian Iveson"
wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Could you post it please? It is not too many lines of text, I assume.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

The triode model has three essential elements. The first is a
transconductance amplifier, or voltage to current converter, which drives
the output terminal of the model. The input to the voltage to current
converter is taken from the output of a three halves power function which
simulates the inherent distortion producing non linearity of the triode,
this comes from child's diode law. The final element in the model is a
summer which creates the negative feedback and feeds the input of the
three halves power law function. The inputs to the summer are the
voltages at the input and output terminals of the triode model, scaled by
appropriate constants. Note that the voltage to current converter is
connected so that on a voltage basis it provides the phase inversion
necessary for the feedback to be negative rather than positive. It's that
simple, and the negative feedback in this model should be obvious to even
the most dimwitted among us, although that doesn't imply one way or the
other whether a real triode actually has internal negative feedback like
this model does.

It would be nice to see a sketch of the model John

For those who don't know what a "3 halves power function" is I suggest that
the input voltage is
fed into an equivalant distortion generator at the front of the triode model
where the resulting
real input voltage is the square root of the voltage cubed, and perhaps
multiplied by a constant.

Its amazing how low the triode distortion is when you have a CCS load.

Patrick Turner.



I am eagerly awaiting Henry Paternack's application of his EE degree to
create a model where everything is done in the forward path, and there is
no feedback involved, it should be something to behold.

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they could
be onto something.

Who was that please?

It was a long time ago so I haven't a clue at this point who it was. All
I can say is that this news group was the first place I saw the idea, and
while I saw it here towards the end of the previous century, it wouldn't
surprise me if the idea is considerably older, and predates the internet
by quite a few years. You can find some old references to the idea using
Google, although the first mention of it in this news group may be hard to
pin down without a fairly extensive search effort.

Regards,

John Byrns

Surf my web pages at, http://users.rcn.com/jbyrns/

[Henry Pasternack]

"John Byrns" wrote in message ...
Henry Pasternack would like to make that the subject in order to avoid
having to present equations describing triode operation without using
negative feedback, in accordance with his claim that triode operation
can be described without reference to negative feedback.

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.

I don't know why you have to be so confrontational.

Obviously you can model this as a transconductance inside a feedback
loop. But you could also say (where 'd' represents partial differentiation):

gm = dIp / dVgk

and

rp = dVpk / dIp

I'll leave it as a trivial exercise for the reader to do the calculus to find
out the actual formulas.

Since there are no cross-products, gm doesn't depend on Vpk, and rp
doesn't depend on Vgk. So the whole thing is separable into a trans-
conductance in parallel with a resistor (the Norton model) -- as it must
be since the two models are equivalent by definition. And there is no
feedback, either.

The large-signal characteristics can be found by solving the differential
equations, or you can determine them (either in real life or on paper) by
plotting Vpk as a function of Ip with Vgk held constant (giving Rp), and
Ip as a function of Vgk with Vpk held constant (giving Gm).

This is obvious, right?

-Henry

[Henry Pasternack]

"Ian Iveson" wrote in message ...
Perhaps your SPICE model is the same as mine. Load regulation, yes. Feedback, no. Unfortunately the few remaining dimwits amongst
us remain unenlightened, and still don't know the difference.

Why, thank you, Ian.

-Henry

[Patrick Turner]

John Byrns wrote:

Ian, all I claimed was that the triode spice model I use is based on a
negative feedback equation, I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Whe

Ipk is the output current into the plate or output terminal of the model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the model
relative to the cathode or common terminal

Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.

Regards,

John Byrns

OK,

Let's write that formula like we may in the real world.

Iak = constant x square root of ( [ µ x Vgk ] + Vak ) cubed.

you say u is the gain of the triode but don't you mean µ, the amplification
factor of the triode?
And isn't the µ the resultant amplification factor after assuming NFB to be
effective, since
µ would be perhaps 100 times or very many times greater without anode to
space charge
feedback?

µ is a negative figure, since the triode is an inverting amplifier.
Vak also must be a negative voltage.

How is Vak calculated? surely it must take in the RL connected and so would
be
equal to µ x RL / ( RL + Ra )?


so we then would have for a 1/2 6SN7 example, with 1V input Iak = c x square
rt ( [ -20 x 1 ] - 16.66 ) cubed.

16.66 is the gain of a 1/2 6SN7 with RL = 50k at 4mA of dc, and where Ra is
approx 10k.

So Ia = c x sq.rt -36.66 cubed.

This formula you have offered us and acompanied by a statement that FB
summing is
within it does not seem at all easy to use or
be informative in any way.

One must be able to easily apply it.

The standard FB equation is :-

Gain with NFB = Gain without NFB / [1 + ( gain without NFB x ß ) ]

Let gain with NFB = A',
Gain without NFB = A.
The fraction of the output fed back in series or in shunt with
the input is ß.
The terms are compatible with those in RDH4!!!!

So A' = A / [ 1 + ( A x ß ) ]

Let us assume we have a very high load connected to the 6SN7
so that its gain without its NFB would be 1,000.


In the case of a triode where we may say µ is 1,000 without NFB, and that the
ratio
of distances between space charge and grid and between anode and space charge
is 1 : 20,
so that ß = 0.05,

then A' = 1,000 / [ 1 + ( 1,000 x 0.05 ) ] = 1,000 / 51 = 19.6.

Gain approaches the 6SN7 µ when a high value load is connected.

Its not necessary to know exactly what the amplification factor might me
without
the NFB; maybe its 10,000, so,
A' = 10,000 / [ 1 + ( 10,000 x 0.05 ) = 10,000 / 501 = 19.96.

If the RL was say 50k, then gain without NFB assuming Ra without NFB to be
huge
is A = gm x RL.
gm is taken as being the small signal transconductance with a fixed Ea
so that there is no anode NFB, and at the dc idle condition
of say Ea = 150V, and Ia q = 3mA, so gm if measured would be 2mA/V.
So with 50k load and without NFB, gain would be 0.002 x 50,000 = 100.

So applying yhe FB equation,

A' = 100 / [ 1 + ( 100 x 0.05 ) ] = 100 / 6 = 16.66.

Notice that when i calculated the 6SN7 gain with 50k above, using A = µ x RL
/ ( RL + Ra ) I got 16.66 for 50k.

The standard gain formula for ALL vacuum tubes incorporates the µ in the
presence of whatever NFB exists.

The standard FB formula can also be used to calculate gain, and the amount of
NFB applied.

If the triode has no load, ie a CCS loading only, then it could be argued its
internal applied
NFB becomes immense, since the applied NFB = 20 log of ( A / A' ),
and where A would be the µ without NFB.
The µ of a triode is the amplification factor with NFB applied, or the ratio
of electrode
distances.
In a pentode the µ is indeed very high, since anode voltage effects on Ia or
upon the
space charge in conjunction with voltage effects from the grid are extremely
low;
the screen effectively increases the anode to space charge distance
enormously.

So hence we see that triodes loaded with CCS inded have extremely linear
behaviour
at small signal levels because their NFB is maximally applied.
But with pentodes, this is n't the case and gain is immense and linearity
not so wonderful unless we apply exterally what is applied within a triode,
and then it should be possible to give a pentode the same gain as a triode
( based on using either triode of pentode with the same gm )
but with greater linearity because the NFB is not applied via a path which is
subject to a constant
x sq.root of a voltage cubed.

Perhaps I may not to make any sense to you in what i say, but I found your
formula
not to shed any light on FB application, wheras the derivations of mine from
RDH4 and other sources do
shed some light or provoke thought.

Patrick Turner.

[Patrick Turner]

Henry Pasternack wrote:

"Ian Iveson" wrote in message ...
Perhaps your SPICE model is the same as mine. Load regulation, yes. Feedback, no. Unfortunately the few remaining dimwits amongst
us remain unenlightened, and still don't know the difference.

Why, thank you, Ian.

-Henry

The "load regulation" is due to the action of the NFB.

Ian evades the issue of NFB in triodes, then calls anyone who dis-believes what
he thinks to be dimwits, if not bullies etc.

Patrick Turner.

[John Byrns]

In article ,
Patrick Turner wrote:

John Byrns wrote:

In article , "Ian Iveson"
wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Could you post it please? It is not too many lines of text, I assume.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

The triode model has three essential elements. The first is a
transconductance amplifier, or voltage to current converter, which drives
the output terminal of the model. The input to the voltage to current
converter is taken from the output of a three halves power function which
simulates the inherent distortion producing non linearity of the triode,
this comes from child's diode law. The final element in the model is a
summer which creates the negative feedback and feeds the input of the
three halves power law function. The inputs to the summer are the
voltages at the input and output terminals of the triode model, scaled by
appropriate constants. Note that the voltage to current converter is
connected so that on a voltage basis it provides the phase inversion
necessary for the feedback to be negative rather than positive. It's that
simple, and the negative feedback in this model should be obvious to even
the most dimwitted among us, although that doesn't imply one way or the
other whether a real triode actually has internal negative feedback like
this model does.

It would be nice to see a sketch of the model John

I'm not in the mood to try an ASCII sketch right now so you will have to
make do with the following equation which describes the elements of the
sketch mathematically.

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Whe

Ipk is the output current into the plate or output terminal of the model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the model
relative to the cathode or common terminal


Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.

For those who don't know what a "3 halves power function" is I suggest that
the input voltage is
fed into an equivalant distortion generator at the front of the triode model
where the resulting
real input voltage is the square root of the voltage cubed, and perhaps
multiplied by a constant.

Its amazing how low the triode distortion is when you have a CCS load.

It's not amazing at all if you look at the above equation. When the
output current is constrained to be a constant by the use of a "CCS
load", then the equation forces Vpk to be equal to -u * Vgk, which is a
linear relation.


Regards,

John Byrns

[John Byrns]

In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
Henry Pasternack would like to make that the subject in order to avoid
having to present equations describing triode operation without using
negative feedback, in accordance with his claim that triode operation
can be described without reference to negative feedback.

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.

I don't know why you have to be so confrontational.

One reason is because I am annoyed by the way you wave your EE degree
around as if it guarantees that your opinions are correct without the
necessity for any further proof. In reality an EE degree is no
guarantee that the holder isn't clueless. This is especially true of
the EEs that have received their degrees in the last 30 years or so, who
frequently have majored in computer engineering and don't have a clue
about traditional electronics.

Obviously you can model this as a transconductance inside a feedback
loop. But you could also say (where 'd' represents partial differentiation):

gm = dIp / dVgk

and

rp = dVpk / dIp

I'll leave it as a trivial exercise for the reader to do the calculus to find
out the actual formulas.

That sounds like code meaning that you actually don't know how to
develop the actual formulas for a model without feedback.

Since there are no cross-products, gm doesn't depend on Vpk, and rp
doesn't depend on Vgk.

It isn't at all obvious to me that condition holds true when the model
includes Child's diode equation, as mine does.

So the whole thing is separable into a trans-
conductance in parallel with a resistor (the Norton model) -- as it must
be since the two models are equivalent by definition. And there is no
feedback, either.

While that is certainly true for your simple linear model, what happens
when a nonlinear function is included in the model, as in my model?

The large-signal characteristics can be found by solving the differential
equations, or you can determine them (either in real life or on paper) by
plotting Vpk as a function of Ip with Vgk held constant (giving Rp), and
Ip as a function of Vgk with Vpk held constant (giving Gm).

This is obvious, right?

Yes, that is obvious, but it does not explain how you might transform
those measurements into equations that describe the operation of the
vacuum triode without incorporating negative feedback into those
equations. There are several problems with doing this, including the
unwelcome effects that the grid voltage affects Rp, and the plate
voltage affects the Gm.

You have yet to deliver on your claim that the operation of the vacuum
triode can be described without reference to negative feedback, except
in the simplest case where the triode is modeled without incorporating
any non linearity in its operation.


Regards,

John Byrns

[Henry Pasternack]

"John Byrns" wrote in message ...
One reason is because I am annoyed by the way you wave your EE
degree around as if it guarantees that your opinions are correct without
the necessity for any further proof.

I do no such thing.

That sounds like code meaning that you actually don't know how to
develop the actual formulas for a model without feedback.

The formulas are trivial, but the proof doesn't depend on them.

It isn't at all obvious to me that condition holds true when the model
includes Child's diode equation, as mine does.

While that is certainly true for your simple linear model, what happens
when a nonlinear function is included in the model, as in my model?

It all works out. You just need to learn about partial differential equations.

Yes, that is obvious, but it does not explain how you might transform
those measurements into equations that describe the operation of the
vacuum triode without incorporating negative feedback into those
equations. There are several problems with doing this, including the
unwelcome effects that the grid voltage affects Rp, and the plate
voltage affects the Gm.

See the previous comment.

You have yet to deliver on your claim that the operation of the vacuum
triode can be described without reference to negative feedback, except
in the simplest case where the triode is modeled without incorporating
any non linearity in its operation.

See the previous comment.

I don't like you, Byrns, because your main purpose is to try to hurt me. I
have no interest in teaching you. If you are really interested in the answer
to this question, my earlier posting outlines the approach, and the solution
is a straightforward exercise. You can work it out for yourself.

This discussion between you and me is over.

-Henry

[John Byrns]

In article ,
"Henry Pasternack" wrote:

"John Byrns" wrote in message
...
One reason is because I am annoyed by the way you wave your EE
degree around as if it guarantees that your opinions are correct without
the necessity for any further proof.

I do no such thing.

That sounds like code meaning that you actually don't know how to
develop the actual formulas for a model without feedback.

The formulas are trivial, but the proof doesn't depend on them.

It isn't at all obvious to me that condition holds true when the model
includes Child's diode equation, as mine does.

While that is certainly true for your simple linear model, what happens
when a nonlinear function is included in the model, as in my model?

It all works out. You just need to learn about partial differential
equations.

Yes, that is obvious, but it does not explain how you might transform
those measurements into equations that describe the operation of the
vacuum triode without incorporating negative feedback into those
equations. There are several problems with doing this, including the
unwelcome effects that the grid voltage affects Rp, and the plate
voltage affects the Gm.

See the previous comment.

You have yet to deliver on your claim that the operation of the vacuum
triode can be described without reference to negative feedback, except
in the simplest case where the triode is modeled without incorporating
any non linearity in its operation.

See the previous comment.

I don't like you, Byrns, because your main purpose is to try to hurt me. I
have no interest in teaching you. If you are really interested in the answer
to this question, my earlier posting outlines the approach, and the solution
is a straightforward exercise. You can work it out for yourself.

This discussion between you and me is over.

Yes, it is obvious from the above that you haven't a clue, to use an
expression that my wife is fond of using, you are showing yourself to be
an "empty suit".


Regards,

John Byrns

[Ruud Broens]

"John Byrns" wrote in message
...
: In article . com,
: "Andre Jute" wrote:
:
:
: ...the triode model I use in spice is based on a feedback
: equation.
:
: Really? Flipper claimed his model has feedback too.
:
: Who is " Flipper"?
:
: [snip]
:
: Regards,
:
: John Byrns
:
: Surf my web pages at, http://users.rcn.com/jbyrns/
:
: Flipper was a young man with great clarity of thought and superior
: logic who graced us with his presence for a while. But I think he
: decided he could learn more electronics from a book faster than he
: could in the poisoned atmosphere on RAT. Flipper is exactly the sort of
: experimenter we should move heaven and earth to keep. But of course the
: usual clowns, spraying on their worthless little piece of wasteland
: like squabbling cats, ganged up to make things unpleasant for him.
:
: Andre Jute
:
: I can't locate any of Flipper's posts, was (s)he partial to using the no
: archive header?
:
:
: Regards,
:
: John Byrns

seemingly
you can find the thread with rudy + hybrid
as i've quoted most, it is easy to follow the thread
website is up
http://flipperhome.dyndns.org/

cheers,
Rudy

[Ian Iveson]

John wrote

...all I claimed was that the triode spice model I use is based on a
negative feedback
equation...

Your claims are shifting all the time, as I pointed out in the part of
my post you ignored.

Ipk = c * (u * Vgk + Vpk) ^ 1.5

This is a "feedback equation" only in your imagination.

The engineering definition of feedback, as opposed to your loose and
muddled confusion with common usage, does not follow from the form of
the equation.

"Feedback equation" has no particular meaning, AFAIK. If you mean it
is an equation arising from the analysis of a control system with
feedback, then the control system and the feedback must come first.
Just because your equation has something in common with the form of
some equation that might arise from such an analysis does not make it
a "feedback equation" in the sense you seem to be trying to use it.

As in the past, I advise you, and others here who have clearly still
not done so, to get a basic student text on control systems theory. I
have given a few examples in the past, particularly in my definitive
statements on the matter.

The definitions are easy to find. They are the same in every case.
They are the same definitions implied throughout RDH (check the
references to standard works on control system theory contained
therein), and the same that underpin all the frequency and transient
analyses that so many people seem to have learned by rote, and hence
misapply at every opportunity.

The form of your equation indicates load regulation. That is, a system
whose output is affected by its load. For example a transformer. Or
indeed any real amplifier or other source *without* feedback.

Indeed, the more the feedback, the less, usually, the load regulation.
A perfect opamp with infinite gain and 100 percent feedback would have
an output perfectly insensitive to load, ie it would have no load
regulation at all.

Ian


"John Byrns" wrote in message
...

Ian, , I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about
how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I
also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack
would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative
feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION


Whe

Ipk is the output current into the plate or output terminal of the
model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the
model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the
model
relative to the cathode or common terminal


Note that the "+" represents the summing node for the negative
feedback.

Note also that the actual gain of the model is negative as a result
of
the way Ipk is defined, this is required for the feedback to be
negative.


Regards,

John Byrns


In article ,
"Ian Iveson" wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Who knows? He was hereabouts a while ago. I may have misrepresented
him...he may simply have claimed that my models would be better
with
feedback. He was cheeky like that.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

SPICE is not hard, John. It is a simple sequence of algebra plainly
stated in text. The only complication is the way negative values
are
excluded. We can all handle that I'm sure. If you show your actual
model, which is only a few lines long and probably shorter than
your
wordy description, I can explain the spice part to anyone who
doesn't
follow it.

Then we can see whether it is really true that

...the negative feedback in this model should be obvious to even
the most dimwitted among us...

Obviously you *describe* the model in terms of feedback. You did so
before I queried your statement. Further description from you in
terms
of feedback doesn't really serve to back up your claim, does it?

...It's that
simple, and , although that doesn't imply one way or the
other whether a real triode actually has internal negative
feedback
like
this model does.

Eh? But you said

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they
could
be onto something.

You have lost your own thread, obviously. Perhaps you are hoping to
avoid the question of where the dummy summing node is in a real
triode?

I could describe a resistor model as follows:

The input voltage to the resistor is applied to two terminals and
used
as one input to a summing node. The other input to the summing node
is
the voltage dropped across the two terminals by the output current,
according to ohm's law. The final output current through the
resistor
is driven by a transconductance amplifier with high gain which
takes
its input from the summing node.

That is, the current rises until the voltage dropped across the
resistor equals the input voltage.

But a resistor doesn't have internal feedback in the proper
engineering sense. The example is often used to demonstrate how
feedback theory can be used for *anything*, regardless of whether
the
feedback is real or imaginary. Load regulation is not sufficient to
imply feedback. Necessary but not sufficient: it's a common source
of
confusion.

Perhaps your SPICE model is the same as mine. Load regulation, yes.
Feedback, no. Unfortunately the few remaining dimwits amongst us
remain unenlightened, and still don't know the difference.

cheers, Ian

[Andre Jute]

Ian Iveson wrote:
As in the past, I advise you, and others here who have clearly still
not done so, to get a basic student text on control systems theory. I
have given a few examples in the past, particularly in my definitive
statements on the matter.

This is quite incredible.

I can remember when Ian Iveson was totally baffled by the mysterious
appearance of the quiescent current on the tube; it had to be explained
to him that it doesn't happen automatically but that the designer
chooses the amount of current and then arranges it by his choice of
plate voltage and negative grid bias.

But now this selfsame ignorant clown Ian Iveson offers us "my
definitive statements" on the internal workings of triodes!

There are more marvels of unearned arrogance in Iveson's full post
below.

Andre Jute
Relentless rigour
Ian Iveson wrote:
John wrote

...all I claimed was that the triode spice model I use is based on a
negative feedback
equation...

Your claims are shifting all the time, as I pointed out in the part of
my post you ignored.

Ipk = c * (u * Vgk + Vpk) ^ 1.5

This is a "feedback equation" only in your imagination.

The engineering definition of feedback, as opposed to your loose and
muddled confusion with common usage, does not follow from the form of
the equation.

"Feedback equation" has no particular meaning, AFAIK. If you mean it
is an equation arising from the analysis of a control system with
feedback, then the control system and the feedback must come first.
Just because your equation has something in common with the form of
some equation that might arise from such an analysis does not make it
a "feedback equation" in the sense you seem to be trying to use it.

As in the past, I advise you, and others here who have clearly still
not done so, to get a basic student text on control systems theory. I
have given a few examples in the past, particularly in my definitive
statements on the matter.

The definitions are easy to find. They are the same in every case.
They are the same definitions implied throughout RDH (check the
references to standard works on control system theory contained
therein), and the same that underpin all the frequency and transient
analyses that so many people seem to have learned by rote, and hence
misapply at every opportunity.

The form of your equation indicates load regulation. That is, a system
whose output is affected by its load. For example a transformer. Or
indeed any real amplifier or other source *without* feedback.

Indeed, the more the feedback, the less, usually, the load regulation.
A perfect opamp with infinite gain and 100 percent feedback would have
an output perfectly insensitive to load, ie it would have no load
regulation at all.

Ian


"John Byrns" wrote in message
...

Ian, , I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about
how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I
also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack
would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative
feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION


Whe

Ipk is the output current into the plate or output terminal of the
model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the
model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the
model
relative to the cathode or common terminal


Note that the "+" represents the summing node for the negative
feedback.

Note also that the actual gain of the model is negative as a result
of
the way Ipk is defined, this is required for the feedback to be
negative.


Regards,

John Byrns


In article ,
"Ian Iveson" wrote:

John Byrns wrote

...the triode model I use in spice is based on a feedback
equation.

Really? Flipper claimed his model has feedback too.

Who is " Flipper"?

Who knows? He was hereabouts a while ago. I may have misrepresented
him...he may simply have claimed that my models would be better
with
feedback. He was cheeky like that.

Rather than posting the entire triode model, which requires some
understanding of the operation of spice, I will give you a verbal
description of the essentials of the triode model.

SPICE is not hard, John. It is a simple sequence of algebra plainly
stated in text. The only complication is the way negative values
are
excluded. We can all handle that I'm sure. If you show your actual
model, which is only a few lines long and probably shorter than
your
wordy description, I can explain the spice part to anyone who
doesn't
follow it.

Then we can see whether it is really true that

...the negative feedback in this model should be obvious to even
the most dimwitted among us...

Obviously you *describe* the model in terms of feedback. You did so
before I queried your statement. Further description from you in
terms
of feedback doesn't really serve to back up your claim, does it?

...It's that
simple, and , although that doesn't imply one way or the
other whether a real triode actually has internal negative
feedback
like
this model does.

Eh? But you said

Some years back when someone said the triode used NFB
internally, I thought of that equation and realized that they
could
be onto something.

You have lost your own thread, obviously. Perhaps you are hoping to
avoid the question of where the dummy summing node is in a real
triode?

I could describe a resistor model as follows:

The input voltage to the resistor is applied to two terminals and
used
as one input to a summing node. The other input to the summing node
is
the voltage dropped across the two terminals by the output current,
according to ohm's law. The final output current through the
resistor
is driven by a transconductance amplifier with high gain which
takes
its input from the summing node.

That is, the current rises until the voltage dropped across the
resistor equals the input voltage.

But a resistor doesn't have internal feedback in the proper
engineering sense. The example is often used to demonstrate how
feedback theory can be used for *anything*, regardless of whether
the
feedback is real or imaginary. Load regulation is not sufficient to
imply feedback. Necessary but not sufficient: it's a common source
of
confusion.

Perhaps your SPICE model is the same as mine. Load regulation, yes.
Feedback, no. Unfortunately the few remaining dimwits amongst us
remain unenlightened, and still don't know the difference.

cheers, Ian

[Andre Jute]

Henry Pasternack wrote:
"John Byrns" wrote in message ...
One reason is because I am annoyed by the way you wave your EE
degree around as if it guarantees that your opinions are correct without
the necessity for any further proof.

I do no such thing.

You're a liar, Pasternack. You use your diploma as bullyprod. Here is
an example of you, Henry Pasternack, trying to substitute your diploma
for knowledge from less than three days ago in this very thread:

"By the way, did you know I studied EE in Frederick Terman's
department?
He was professor emeritus, until he died, while I was at university.
Some
of my professors were his former students, including a legend or two
from
the "golden age" of tube radio, plus a good share of more modern
legends."

And even more ugly diplomaed-quarterwit sneering from Pasternack:

"I have to say, Patrick, you really don't know how to make a correct or

convincing engineering argument. And I do have the credentials to say

that. You are straining my patience."

Do you deny that those are your words, Plodnick?

That sounds like code meaning that you actually don't know how to
develop the actual formulas for a model without feedback.

The formulas are trivial, but the proof doesn't depend on them.

Then why bring them up at all? Either you know how to develop the
formula for a model without feedback or you don't. If you do, show us
that you are good for something except manure.

It isn't at all obvious to me that condition holds true when the model
includes Child's diode equation, as mine does.

While that is certainly true for your simple linear model, what happens
when a nonlinear function is included in the model, as in my model?

It all works out. You just need to learn about partial differential equations.

Apparently you, Henry Pasternack, haven't yet "learned about partial
differential equations", which is why you copy out basic principles
from a standard textbook and then pretend to know what they mean while
sneering that others don't have the math to continue the workings.

We have seen you in action too often before to be taken in, Pompass.

Yes, that is obvious, but it does not explain how you might transform
those measurements into equations that describe the operation of the
vacuum triode without incorporating negative feedback into those
equations. There are several problems with doing this, including the
unwelcome effects that the grid voltage affects Rp, and the plate
voltage affects the Gm.

See the previous comment.

Translation: Pasternack doesn't have an answer but is too suffused with
pride in his diploma to admit his ignorance.

You have yet to deliver on your claim that the operation of the vacuum
triode can be described without reference to negative feedback, except
in the simplest case where the triode is modeled without incorporating
any non linearity in its operation.

See the previous comment.

Translation: Pasternack doesn't have an answer but is too suffused with
pride in his diploma to admit his ignorance.

I don't like you, Byrns, because your main purpose is to try to hurt me. I
have no interest in teaching you. If you are really interested in the answer
to this question, my earlier posting outlines the approach, and the solution
is a straightforward exercise. You can work it out for yourself.

Pasternack's earlier contribution was five bob each way. To be fair, he
did say that, in his personal opinion, he didn't like the NFB
explanation. But, since even Plodnick must be aware that his opinion is
now worth nothing on RAT (and even less with his diploma!), he's not
bothering to defend it, he's just nastily slinging slime.

This discussion between you and me is over.

That's probably the smartest thing you've said this year, Plodnick.
Every time that you've crossed Mr Byrns, he has conclusively nailed
your hide to the wall. Remember your pathetic mistake of Zo = 1 * Rp?
(1)

-Henry

Run, rabbit, run.

Unsigned out of contempt for worthless scum

(1) And not to forget your dumb mistake with the chokes in "DC on the
core" which I used to wipe your little friend Michael LaFevre, maximum
cigarogem of the Magnequest Scum, when he was dumb enough to believe
that your diploma is a guarantee of knowledge. I remember that in that
instance too you ran away, Plodnick, leaving your foolish little friend
Creepy Mike to be exposed as a complete electronics ignoramus.

[Patrick Turner]

I wrote to the group who ignored me, and they bickered about academic
credentials,
and one guy got all hurt feelings and tried to end the discussion with another
who continued to
persue with his question which went unanswered.

I post again to see if anyone is awake, or have stopped being so childish as to
be offended by
personal posturing, and to see if they could proceed to stay on the topic of NFB
in triodes.........

Patrick Turner wrote:

John Byrns wrote:

Ian, all I claimed was that the triode spice model I use is based on a
negative feedback equation, I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Whe

Ipk is the output current into the plate or output terminal of the model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the model
relative to the cathode or common terminal

Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.

Regards,

John Byrns

OK,

Let's write that formula like we may in the real world.

Iak = constant x square root of ( [ µ x Vgk ] + Vak ) cubed.

you say u is the gain of the triode but don't you mean µ, the amplification
factor of the triode?
And isn't the µ the resultant amplification factor after assuming NFB to be
effective, since
µ would be perhaps 100 times or very many times greater without anode to
space charge feedback?

µ is a negative figure, since the triode is an inverting amplifier.
Vak also must be a negative voltage.

How is Vak calculated? surely it must take in the RL connected and so would
be
equal to µ x RL / ( RL + Ra )?

so we then would have for a 1/2 6SN7 example, with 1V input Iak = c x square
rt ( [ -20 x 1 ] - 16.66 ) cubed.

16.66 is the gain of a 1/2 6SN7 with RL = 50k at 4mA of dc, and where Ra is
approx 10k.

So Ia = c x sq.rt -36.66 cubed.

This is a difficult thing to undestand because we are asked to find a square root

of a negative number.

Formulas must be able to be applied by ordinary people and with full easy
understanding
and without confusion over signage of terms.

Then and only then are they useful in the real world to help us understand the
world,
and design better amplifiers. Otherwise we are wasting time.



This formula you have offered us and acompanied by a statement that FB
summing is
within it does not seem at all easy to use or
be informative in any way.

One must be able to easily apply it.

The standard FB equation is :-

Gain with NFB = Gain without NFB / [1 + ( gain without NFB x ß ) ]

Let gain with NFB = A',
Gain without NFB = A.
The fraction of the output fed back in series or in shunt with
the input is ß.
The terms are compatible with those in RDH4!!!!

So A' = A / [ 1 + ( A x ß ) ]

Let us assume we have a very high load connected to the 6SN7
so that its gain without its NFB would be 1,000.

In the case of a triode where we may say µ is 1,000 without NFB, and that the
ratio
of distances between space charge and grid and between anode and space charge
is 1 : 20,
so that ß = 0.05,

then A' = 1,000 / [ 1 + ( 1,000 x 0.05 ) ] = 1,000 / 51 = 19.6.

Gain approaches the 6SN7 µ when a high value load is connected.

Its not necessary to know exactly what the amplification factor might me
without
the NFB; maybe its 10,000, so,
A' = 10,000 / [ 1 + ( 10,000 x 0.05 ) = 10,000 / 501 = 19.96.

If the RL was say 50k, then gain without NFB assuming Ra without NFB to be
huge
is A = gm x RL.
gm is taken as being the small signal transconductance with a fixed Ea
so that there is no anode NFB, and at the dc idle condition
of say Ea = 150V, and Ia q = 3mA, so gm if measured would be 2mA/V.
So with 50k load and without NFB, gain would be 0.002 x 50,000 = 100.

So applying yhe FB equation,

A' = 100 / [ 1 + ( 100 x 0.05 ) ] = 100 / 6 = 16.66.

Notice that when i calculated the 6SN7 gain with 50k above, using A = µ x RL
/ ( RL + Ra ) I got 16.66 for 50k.

The standard gain formula for ALL vacuum tubes incorporates the µ in the
presence of whatever NFB exists.

The standard FB formula can also be used to calculate gain, and the amount of
NFB applied.

If the triode has no load, ie a CCS loading only, then it could be argued its
internal applied
NFB becomes immense, since the applied NFB = 20 log of ( A / A' ),
and where A would be the µ without NFB.
The µ of a triode is the amplification factor with NFB applied, or the ratio
of electrode
distances.
In a pentode the µ is indeed very high, since anode voltage effects on Ia or
upon the
space charge in conjunction with voltage effects from the grid are extremely
low;
the screen effectively increases the anode to space charge distance
enormously.

So hence we see that triodes loaded with CCS inded have extremely linear
behaviour
at small signal levels because their NFB is maximally applied.
But with pentodes, this is n't the case and gain is immense and linearity
not so wonderful unless we apply exterally what is applied within a triode,
and then it should be possible to give a pentode the same gain as a triode
( based on using either triode of pentode with the same gm )
but with greater linearity because the NFB is not applied via a path which is
subject to a constant
x sq.root of a voltage cubed.

Perhaps I may not to make any sense to you in what i say, but I found your
formula
not to shed any light on FB application, wheras the derivations of mine from
RDH4 and other sources do
shed some light or provoke thought.

Patrick Turner.

[Andre Jute]

John Byrns wrote:
Ian, all I claimed was that the triode spice model I use is based on a
negative feedback equation, I did not intend to start a discussion of
spice, the equation is given below. I did not make any claims about how
resistors might, or might not, be described. Nor did I claim that a
triode could not be described by equations that don't use negative
feedback, although I am interested in seeing how that is done. I also
did not make any claims about the accuracy of the simple distortion
mechanism incorporated in my equation, although Henry Pasternack would
like to make that the subject in order to avoid having to present
equations describing triode operation without using negative feedback,
in accordance with his claim that triode operation can be described
without reference to negative feedback.

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Whe

Ipk is the output current into the plate or output terminal of the model

c is a constant needed to make the equation work

u is the voltage gain of the triode

Vgk is the input voltage at the grid, or input, terminal of the model
relative to the cathode or common terminal

Vpk is the output voltage at the plate or output terminal of the model
relative to the cathode or common terminal


Note that the "+" represents the summing node for the negative feedback.

Note also that the actual gain of the model is negative as a result of
the way Ipk is defined, this is required for the feedback to be negative.


Regards,

John Byrns

The difficulty with your model, whether in the form you offer it or in
the RDH nomenclature Patrick wants to standardize, is that it is
descriptive rather than explanatory. That is also what makes it the
best model we have for people who have ever worked on the bench with
triodes and wondered about the differences from pentodes, and certain
individual facets of triodes. However, since the negative feedback
(virtual or real makes no difference to the outcome) is subsumed in mu,
the amplification factor, we are left with the irritating niggle that
we haven't explained much, we have just described it, and found a
formula which with the help of a constant seems to give the correct
answers. The alternatives to your transconductance model, thrown out by
Pasternack with sneers attached about how the math will trip us up,
will eventually circle around to the same problem, that the NFB inside
the triode is subsumed inside one of the factors, so that at that point
the model again becomes descriptive rather than explanatory. I really
can't understand how Pasternack can fail to know something so obvious,
unless he has never worked those formulae through; perhaps he should
buy a copy of the RDH and learn the basics.

Of course, all of this is merely of academic interest. In practice we
work with tables or graphs for the triode which incorporate a
degenerative effect that only someone who doesn't own a soldering iron
can miss seeing loud and clear in any triode amp.

Still, I am surprised and disappointed that the wannabe theorists among
us (Iveson, Pasternack and so on), even without the bench and listening
time the rest of us have on tubes, have failed to ask themselves just
why it is that throughout the relevant chapters of the RDH we find in
the formulae different constants for triodes and pentodes...

The answer of course is that those old engineers made the same
observations we are making now, but on hand of vastly more data. We're
standing on their very broad shoulders.

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

[Peter Wieck]

Andrew Jute McCoy, whilst stalking Mr. Pasternack bleated:

More tripe.

Tell us more about Peter Drucker.

Peter Wieck
Wyncote, PA

[Peter Wieck]

Andrew Jute McCoy, while demonstrating fundamental lack of basic
knowledge bleated:

More crap.

And about Peter Drucker? Inquiring minds wish to know.

Peter Wieck
Wyncote, PA

[Peter Wieck]

Andrew Jute McCoy blathered:

More stupidity.

It ain't gonna get any better.... Its fantastical fabrications on Mr.
Drucker and his "Hungarian" heritage are typical of its truthfulness.

Peter Wieck
Wyncote, PA

[Chris Hornbeck]

On Thu, 05 Oct 2006 15:04:05 GMT, John Byrns
wrote:

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Like all of this discussion, a lot more interesting if
remotely connected to reality.

The discussion seems to not discriminate between small-
signal and large signal (note: there is *NO* large signal
model of a "triode", none[*]), between analogies and models,
or between levels of testability. Only levels of testosterone.

Nobody seems interested in my test requirements; can anyone
suggest a test acceptable to all?

Much thanks, as always,

*[This is the new fashion, following Andre's lead, of including
footnotes, however difficult on Usenet. Anything can be modeled,
although sometimes the simplest model is the thing itself.

A complete large-signal model of a triode would include
things like the electron cloud, field effects, geometry
of the bits-'n-widgits, some classical billiard ball physics
and some later relativistic physics (electrons in your old
telly's CRT achieve relativistic mass gains over 5%).

Plate curves have served us well since before we were born,
but a computer model *is* possible, for some particular device;
however useful, or applicable to the discussion.]

**[pps: The map is not the world, but the model *is* our
world.

Chris Hornbeck

[Chris Hornbeck]

On Wed, 4 Oct 2006 22:50:04 -0400, "Henry Pasternack"
wrote:

Clear, linear explanation snipped for bandwidth

This is obvious, right?

Maybe what's not obvious is that the small-signal linear
model is the crux of the biscuit. Large-signal non-linear
extrapolations are way, way down the road in this kind
of discussion.

No non-linear large-signal models can be extrapolated from
linear small-signal models(*). The twain don't meet.

But what *should* be obvious from your many posts is
that there is no linear model that includes feedback.
Maybe a non-linear model might? Out of my field,
unfortunately, so can't comment.

*[I'd maybe suggest starting to think about field-effect
devices as dI/dWhatever transconductance devices. A triode
is also just a tranconductance engine with a warped
pressure plate. Or something...]

**pps: Asterisks are so fun and so annoying; I'm starting
a twelve step propram tomorrow.

Much thanks, as always,

Chris Hornbeck

[Patrick Turner]

Chris Hornbeck wrote:

On Thu, 05 Oct 2006 15:04:05 GMT, John Byrns
wrote:

THE EQUATION

Ipk = c * (u * Vgk + Vpk) ^ 1.5

Like all of this discussion, a lot more interesting if
remotely connected to reality.

The discussion seems to not discriminate between small-
signal and large signal (note: there is *NO* large signal
model of a "triode", none[*]), between analogies and models,
or between levels of testability. Only levels of testosterone.

Nobody seems interested in my test requirements; can anyone
suggest a test acceptable to all?

The operation of a triode of any kind is usually always assumed to be
small signal
levels of below 1V at the anode unless otherwise stated.
Grid voltages are thus always less than a volt and
unless otherwise stated is is assumed that anode loads are more than Ra,

so that the operation which is observed, measured, ot theorized about is
substantially
linear and free of clipping or severe grid current.

The large signal operation where anode voltage swings may be thousands
of volts
is similar to low level operation but may involve substantial changes to
the small signal
op in terms of distortion and grid current.

In my example of the NFB in a 1/2 6SN7 it is assumed the Ea Eg swings
are less than volt.
The square root of something cubed within some equation which appeares
at the top of your pos
becomes somewhat irelevant, because over the span of a change in Ea of
1V, the
3/2 rule if applied does not make much difference to the observed
operation.
Everyone has ignored my post with an example of 1/2 a 6SN7 with all the
likely
real feedback equations applied to the triode. I guess my concerns and
mention of
well known standard FB equations to a 6SN7
points towards the very likely fact that NFB within a triode is quite
real.
and not something one can select to believe in or not.
The earth isn't flat either, and its roundness cannot be ignored lest
you have a restricted view of reality,
although if you are stranded in the desert in the middle of Oz one could
be forgiven for thinking
the earth was indeed very flat, and not see the roundness which is
obvious from the
uncomfortable view offered from a spacecraft in orbit.





Much thanks, as always,

*[This is the new fashion, following Andre's lead, of including
footnotes, however difficult on Usenet. Anything can be modeled,
although sometimes the simplest model is the thing itself.

A complete large-signal model of a triode would include
things like the electron cloud, field effects, geometry
of the bits-'n-widgits, some classical billiard ball physics
and some later relativistic physics (electrons in your old
telly's CRT achieve relativistic mass gains over 5%).

Plate curves have served us well since before we were born,
but a computer model *is* possible, for some particular device;
however useful, or applicable to the discussion.]

Anode curves have indeed served us well.
Someone makes a triode then runs a test with a tracer and
that's that. Then its easy to plot a load line and work out
an operation Q point, 2H, and Vswing, all without any idea that NFB
exists anywhere.
Since the FB exists within the 3 prongs of anode grid and cathode,
nothing can be done to
change it, re-arrange it, delete it, intensify it. Its just there.
µ, gm and Ra are given in the data and all 3 are parameters derived
as a result of the FB and electrode dimensions and distances.

And BTW, a 12AX7 can be made much smaller than it is if the RELATIVE
distances
between electrodes are maintained. But hardly anyone bothered to do it
because the damn thing got too fiddly to make and the teams of sheilas
on production lines worked fastest and best with sizes of vacuum tubes
at about the normal size that evolved.
maybe nano technology could allow much smaller versions of 12AX7.

We did have tiny nuvistors for a short time but they got blasted away by

the silicon bipolar junction transistor.

Patrick Turner.





**[pps: The map is not the world, but the model *is* our
world.

Chris Hornbeck