|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.90.39.24
In Reply to: RE: Would Class AB be closer to X or Y-axis of that graph? posted by FenderLover on August 17, 2016 at 06:15:24
The graph I showed in the previous post is the dynamic curve.The dynamic curve takes into consideration the voltage across the tube, the idle current and the load impedance.
The graph above shows how a dynamic curve is derived.
This one happens to be of a not so linear tube with a poorly chosen operating point.
The plate voltage moves farther as the tube is driven to more current vs. when the tube is driven to less current. That is what creates harmonic distortion.
What one wants to do is operate the tube so the plate voltage moves in an equal way up and down from the idle point for the same amount of grid bias change.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/17/16Follow Ups:
Gonna need to break out the old College Physics book to translate.
8^)
Does Class "A" push-pull pentode amps require the kind of global feedback that Class AB1 pentode amps need to reduce distortion?
Thanks!
IMO a Class A PP pentode amp would not require as much NFB as a Class AB PP pentode amp because the tubes a Class A pentode amp are running in a more linear way than the tubes in a Class AB pentode amp.
Having said that, the real reason a PP pentode amps needs NFB is to lower the plate resistance of the pentodes so the damping factor is improved.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
A preamp with only minimal swing requirements could happily run Class A at numerous operating points along the load line whereas a power amp is much more limited. Or are you saying that any amp must be biased to either the lowest distortion (nother can of worms) or approximately half supply (also with it's issues)?
Naz
"A preamp with only minimal swing requirements could happily run Class A at numerous operating points along the load line..."
When you say Class A do you mean A. running the tube in the most linear way possible or B. running the tube so that plate current never cuts off?
A tube can't be running in the most linear way possible and at the same time be operating in the non-linear cutoff region.
A tube operating in the non-linear cutoff region (but not cutting off) meets some people's definition of Class A.
My whole point is that Class A is more than that.
Class A is (according to all the examples in all the books) operating a tube only in the most linear part of the curve.
If, in your preamp tube example, the preamp tube's operation is staying in the most linear part of the curve anyway, why not bias it in the middle of the most linear part of the curve? What would be the harm in that?
As you point out, a power tube is more limited. If we want some power and we want the tube to operate only in the most linear part of the curve we pretty much have to bias it in the middle of the most linear part of the curve.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Throwing load lines and distortion into the mix further confuses your definition of Class A. Eg, I know you like CCS loading and for good reason; a more horizontal load line and lower distortion as a result. But this makes it all the less necessary to run an operating point anywhere near the "mid point" unless max swing is necessary. Indeed, a near horizontal load line virtually kills of off a substantial part of your argument.
To include a particular requirement for an operating point in order to meet the definition of Class A also creates more problems than it solves. A set of plate curves may look a lot more linear over a part of the range and indeed may well be for a particular load line if high swing is necessary. However, expand the set of curves and what appears to be the non-linear portion at low plate current suddenly looks very much the same as at high current over the chosen amplifying range. Many designs successfully run tubes in this so called non-linear area and achieve very low distortion and great sound. I'm talking mainly Phono stages here, running plate currents less than 10% of what would be considered to be reasonable current for the particular tube used.
Naz
Naz, I don't necessarily disagree with what you have said but all the examples of Class A in the books follow what you call my "particular requirement[s]".
"Throwing load lines and distortion into the mix further confuses your definition of Class A."
Loading for Class A is different than loading for Class AB.
It is my understand that the higher load impedance (to lengthen and straighten the linear part of the curve) is part of the Class A package.
From all the reading I've done, Class A is not just one thing, it's a whole package with the intent of achieving the least harmonic distortion possible (while still leaving some power in a power stage).
There is no single definition of Class A. If I had to pick one that best describes Class A it would be running the tube only in the most linear part of the curve.
I think if you disagree with that you are disagreeing with most of the technical books.
I didn't write the following;
The Radio Handbook 14th Edition page 98 "......Such an amplifier is normally operated in the center of the grid-voltage plate-current transfer characteristic and gives an output wave-shape which is a substantial replica of the input wave-shape."
The Radio Handbook 38th Edition page 65 A Class A amplifier-- "...is one operated so that the wave shape of the output voltage is the same as that of the signal voltage applied to the grid."
Tremaine Audio-Cyclopedia 2nd Edition page 523 "The quiescent plate current is selected for a value in the most linear portion of the plate-current characteristic."
Page 543 "What is a class A amplifier? An amplifier in which the grid-bias voltage is set to approximately one-half the cutoff voltage to obtain linear operation."
Page 604 "When operating as a single tube, class a, little distortion exists because the grid signal operates in the most linear portion of the dynamic characteristic."
And I didn't create the graph above.
In what way is what I'm saying out of line with the text above or the graph above?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"It is my understand that the higher load impedance (to lengthen and straighten the linear part of the curve) is part of the Class A package.
From all the reading I've done, Class A is not just one thing, it's a whole package with the intent of achieving the least harmonic distortion possible (while still leaving some power in a power stage)."
That's incorrect. If the definition of Class A includes linearity at all, it is only from the perspective that operating a single-ended stage at the midpoint between cutoff and saturation will generally provide low distortion results. At the time this class designation was assigned to its respective operating condition, reducing distortion to the levels we expect today was not a consideration. What you're reciting is the modern use of Class A in an audiophile implementation. It carries an emphasis on reduced distortion - at the expense of output power - that is not a component of the original definition of Class A operation.
--------------------------
Buy Chinese. Bury freedom.
"What you're reciting is the modern use of Class A in an audiophile implementation. It carries an emphasis on reduced distortion - at the expense of output power - that is not a component of the original definition of Class A operation"
But can't it be seen as the natural extension of the original intent?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
It seems reasonable to call this a refinement of Class A operation for modern use. It is otherwise unrelated to the intent of those who created the classification.
--------------------------
Buy Chinese. Bury freedom.
The term, 'driver' used to be assigned to the stage preceding the finals is set up to deliver grid current. Now it is just the stage preceding the finals. Not an improvement in language use IMO...:)
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Even without grid current the Miller capacitance still needs to be driven and that takes some current.
Just saying.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
or the essence of what you say as being typically the main attributes of Class A. But taking a line here or there out of a handbook is out of context in that they are not the lines that define Class A, they are merely attributes typical of the class and its configuration as applied to SE power amps in most of your examples.
But Class A is a definition applied to all amplifiers and (for example) there is no way that it's necessary to set an operating point anywhere near "one half of the cutoff voltage in a low signal application to achieve performance worthy of Class A even by your high standards:)
The best way I can get my point across is to ask you this question. If I design an amplifier that under normal operating conditions remains in conduction with no grid current, but I do not set the operating point anywhere near the "midpoint" or "the most linear part of the curve" yet achieve the main goal (even using your argument) of low distortion, how would YOU classify it?
Dave also made a valid point about clipping. Does this means that any amplifier capable of overload cannot be classified as Class A?
Naz
"The best way I can get my point across is to ask you this question. If I design an amplifier that under normal operating conditions remains in conduction with no grid current, but I do not set the operating point anywhere near the "midpoint" or "the most linear part of the curve" yet achieve the main goal (even using your argument) of low distortion, how would YOU classify it?"
I think I would just ask you why you didn't set the operating point at the midpoint.
I answered Dave.
A Class A amplifier driven beyond the linear part of the curve is a Class A amplifier that is being abused. Why would anyone want to do that?
I think this horse is dead.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
What Class is it in, it must be in something?
I can give you many reasons why I may not bias for the midpoint in a low signal amp and I'll bet you may too from what I've read.
I also don't appear to have been clear enough in making my points re plate curves and their "apparent" most linear part or CCS loading, which renders the point almost superfluous anyway ... sigh!
Anyway, it seems that we can't all agree on this and I respect everyones right to hold their own beliefs.
Naz
Don't take this personal but I would call it a poor attempt at Class A.
"I also don't appear to have been clear enough in making my points re plate curves and their "apparent" most linear part [see graph above, it's pretty clear to me what part is linear and if you move too far from the center it's not so linear anymore] or CCS loading, which renders the point almost superfluous anyway ... sigh!"
A horizontal load line won't help this tube much. Draw a horizontal load line and then derive a dynamic curve. It's going to be pretty curved with no straight part. (maybe a little between -14 and -18) The spacing starts large on the left and gets skinnier as you go to the right.
"Anyway, it seems that we can't all agree on this [that's clear] and I respect everyones right to hold their own beliefs."
I see some of those beliefs as flying in the face of what is explained in the technical books.
Oh well.
Take care. Nothing personal......
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
> > Don't take this personal but I would call it a poor attempt at Class A < <
Class A none the less?
> > A horizontal load line won't help this tube much < <
Well, actually it will but that's not the point. Are you saying that this tube can't ever be be run in Class A?
Have to be very careful with curves as they can look much worse depending on the scale. Believe it or not the curves look much the same if we expand the region 300-400V up to a max of only 5mA.
Cheers!
Yes, still Class A but not one that follows any of the examples in the books.Is there something wrong with my eyes?
For a 8 volt peak grid swing I get a plate swing of 142.5 one way and a plate swing of 122.5 the other and then there's the problem of the plate resistance increasing as we go to the right.
Let's try this instead.
I get 120 and 125.5 and the plate resistance hasn't changed much at all.But in the end you win. Your example is Class A but bad Class A.
Why in the deuce would we want to spend time talking about a bad Class A?
What would be the point in that?
I pledge from now on I will try to preface my talk of Class A with the word "good" if you will try to remember to preface your talk of what I call bad Class A with the word "bad". OK? :-) (with the understanding the "bad" just means that it doesn't follow the examples in the books and "good" just means that it does)
Or how about when you see me talking about Class A just remember that when Tre' talks about Class A he's always meaning Class A done the way the books show it (silly Tre').
So now we've come full circle. The book Micheal posted the link to says, on page 68 that a Class A amplifier is "An amplifier operated as shown in fig. 306 in which the output waveshape is a faithful reproduction of the input waveshape, is known as a Class A amplifier."
Fig 306 is for all intents and purposes the same as the figure below.
You'll just have to excuse me, I choose to stand on that.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/19/16
I accept that linearity varies from one tube to another. My only points are that linearity will always be better by varying degrees with a more horizontal load line and that to the eye, scale can be deceiving when viewing plate curves.
On the latter, here's a link to a graph of plate curve for a 6SN7 at low voltage and low plate current taken from Stephie's Site. It's not what I was looking for but it will tell the story. This region on a larger scale would normally look non linear. You can Sim curves to explore the effect yourself.
It's clear that good linearity can be gained at many different operating points where power is not a prerequisite.
Naz
"It's clear that good linearity can be gained at many different operating points where power is not a prerequisite."
I think I've left myself misunderstood.
While I think choosing the most linear operating condition is in keeping with the intent of Class A that is a different thing than choosing not to bias in the middle of whatever dynamic curve your chosen load impedance provides.
The sim tool linked below can show you what I mean.
The outside left arrow will change the load impedance and idle current.
The inside left arrow will change the load impedance only and the top white arrow will change the bias point (quiescent plate voltage and current) along that load line.
Note, keep the yellow arrow (grid swing) somewhat minimized to start with or it will limit what you can do with the other controls.
Watch the HD in the readout below as you make your adjustments.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
that there can be many operating points that will provide Class A operation unless max power or swing is a prerequisite.
Whilst I agree there will always be an operating area of lowest distortion, equally you could pick an area that would normally provide slightly higher distortion but achieve better overall results by making the load line more horizontal as your Sim will show. Great tool BTW!
So for clarity I'm saying that the operating point cannot be singled out in the definition of Class A because several factors dynamically interact, meaning that similar distortion characteristics can be obtained at various operating points, within the limits of swing and power we are designing for of course.
To include subjective statements like lowest distortion into the definition would be absurd. Eg, in an SE amp one of the biggest factors determining distortion is the load line so how would one determine whether an SE amp met the criteria of Class A by your expanded description?
Anyway, one thing we both agree on is that best practice should always be used in order to achieve the very best result. So I'll leave it at that.
Thanks for a civil discussion that I'm sure will be of some benefit to some others, particularly newbies.
Cheers
"A preamp with only minimal swing requirements could happily run Class A at numerous operating points along the load line"
So let's take a V8 motor and fill two cylinders on one side with concrete. Now it's a V6, right? It certainly meets the minimalist requirements some here are attempting to apply to amplifiers.
To your comment - no one is discounting a reasonable modification of the operating point as being included in the classification. But operating an AB1 power amp so it never goes to cutoff doesn't qualify. That's not reasonable. An amplifier must be operating near the midpoint of its curve to be Class A. That description is sometimes omitted for the sake of brevity, but it's just as important to the definition as any of the other requirements. Anyone who doesn't like this criteria is welcome to hold a séance and argue the point with those long-dead engineers.
--------------------------
Buy Chinese. Bury freedom.
If a V8's Management system turns of 4 cylinders when not needed what would YOU call it?
It's mode of operation at any given time is what describes it most accurately but only while it is in that mode.
We can argue endlessly about the engine's configuration, capabilities, where it spends the most of it's time etc etc and that's one of the problems ... or opportunities for marketers:)
To me, the engine example is akin to hardware and software/firmware in a microcontroller. Another analogy is SE and PP. This is the configuration, (the hardware, V8 or V6). But PP could run in various classes either permanently or conditionally, akin to the programming in the V8 that can turn cylinders off.
Naz
Your V8/V6 analogy with cemented cylinders is pretty weak. A much better one would be a V8 with cylinder deactivation technology. In this case one could ask the question: If the engine is never pushed hard enough to activate the last two cylinders is it a V8 or a V6? . I know how most would answer this question.
Regarding classes, you state operation "near" midpoint. What is "near"? How far can one deviate from near before it's no longer A? The problem with these definitions is ambiguity and there seems to be a reluctance to address it.
Good thing the question wasn't, 'when is a V8 a V4?' because the side, and entirely correct answer would be when Honda builds one to avoid the cylinder count limit...32 valves, 8 connecting rods, 8 spark plugs and 4 combustion chambers...LOL
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
intensional definitions
extensional definitions
demonstrative definitions
precising definitions
ostensive definition
"In classical thought, a definition was taken to be a statement of the essence of a thing."
Just like there is more to the term "definition" than might meet the eye, there is more to Class A than just not reaching cutoff.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"The problem with these definitions is ambiguity and there seems to be a reluctance to address it."The ambiguity is minor and has no impact on our understanding of the definition. There was no need to address it 50 years ago, and none now. Only those who want to materially change the fundamental meaning and intent of "Class A" have a problem with this.
--------------------------
Buy Chinese. Bury freedom.
Edits: 08/18/16
If the intent is to operate the tube in as linear a way as one can then the details aren't all that important.
If the intent is to operate the tube in the non-linear cutoff region while still claiming Class A because "the current never cuts off" the whole intent of Class A is being ignored and some of the potential benefits of Class A (over class AB) are being thrown away.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I have to agree with Naz's line of though here. Adding ambiguous subjective terms to a technical definition is poor practice typically imposed by the marketing division.
All of the semantics being bantered about here are somewhat pointless among those educated in the field but can be counterproductive when put in the hands of someone new to the art or someone with an agenda. It reminds me of the Class A SE Fuzz Box I have been working on :-)
dave
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: