|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
108.70.59.148
In Reply to: RE: All that is not a clear, concise definition... posted by Tre' on August 16, 2016 at 21:16:30
The problem with what you present is that in the context of a real-world device you can not unambiguously define "linear" w/o also introducing an arbitrarily chosen measure of linearity "goodness" such as absolute deviation from perfect, % distortion etc. Words such as "substantial", "most linear" (implying imperfect absolute linearity), and "approximately" are open to interpretation and thus not ideal as part of a definition although they may useful as parts of descriptions and characteristics.
As an example of the inherent ambiguity in the stuff you cited, consider the plot used to graphically demonstrate "linearity". What's to prevent someone from extending the "linear" region by +/-10%? To my eye everything is still linear. Maybe you'll agree, maybe you won't - who's correct ? How about to the point just before the line begins to curve? But where exactly does the line begin to curve? How far away from that curve point must one be to remain in the linear region? Without an agreement on what constitutes acceptable deviation from perfect linearity, a measure of linearity goodness, you might pick one point and I might pick another. IOW, the plot and accompanying text are ambiguous and ambiguity really isn't compatible with a definition.
Follow Ups:
"How about to the point just before the line begins to curve?"
A closer look will show that the most linear part of the curve is really just the middle part as shown in the previous graph.
By the point where the line visibly curves in the first graph we have, in fact, already left the linear part of the curve.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Whilst it may be customary and preferable to design Class A for lowest distortion, that may not be, and in fact mostly won't be half supply voltage.
A designer may chose slightly different operating points for Class A for many different reasons but meet all the criteria of Class A conduction. If such circuits cannot be described as Class A then what what Class do they fall into?
Science and Engineering doesn't allow for huge areas of ambiguity. I was taught that conduction determined the Classes, not linearity or any other arbitrary, non concise parameter.
My 2C,
Naz
Hopefully it was conduction *at rated/maximum power*...:)
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
"Hopefully it was conduction *at rated/maximum power*...:)"
I know what you are saying Doug, and for purposes of accurately describing an amplifier's design I agree 100%.
But from a purely engineering/physics standpoint an amp can operate in class A at low outputs while being unable to maintain/sustain it when more power and therefore larger voltage swings are required.
Does that make it a "class A amp"? IMHO no - but that doesn't prevent the amp from operating in class A under certain conditions. This reminds me of the debate about amplifier power back in the 60s/70s - just because the FTC decided RMS power was what could be advertised doesn't mean that peak power was nonexistent.
Can't run Class A from an AB operating point.
That is the thing, you can't go saying it's A just because it makes one of the criteria for it. Gotta be the whole 9 yards...or it is something else.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
"Can't run Class A from an AB operating point."
Exactly right!
--------------------------
Buy Chinese. Bury freedom.
How does one differentiate between different possible operating levels. Lets throw the arbitrary concept of maximum power out the door and consider this to be a preamp design. Can a preamp design only be called class A if it is biased to Tre's set of rules?
From a technical perspective I see no issues with a circuit operating class A up to a given level and then class AB above that level. The issues arise here when the ambiguity comes into play and since the transition period from A to AB can vary wildly from design to design. I see the need to further define the grey areas and am fine with doing this on a case by case basis but to amend a technical definition with subjective and ambiguous terms is just bad practice.
dave
again with the cherry picked set of conditions...
Class A is defined by conduction at max power. Cut off and it is an AB design. Could be deep AB, but is still AB.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
What is max power in a linestage or other circuit than a power amplifier?
Plus at some point all class A circuits are going to cut off and I think the issue here is using a softly defined terms. (max power)
dave
"Plus at some point all class A circuits are going to cut off and I think the issue here is using a softly defined terms."
Part of the definition of Class A is how it's driven (loaded too, but that's a whole other discussion).
With Class A the input signal is limited to keep the operation within the linear part of the operating curve. That's just part of the package.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Where it printed that you must limit the input voltage in order to achieve Class A operation?
By extension of that concept, any amp you can clip is not a Class A design.
dave
It doesn't have to be printed, it's obvious.Tremaine 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."
With too much input signal the tube operation will go beyond the most linear portion of the dynamic characteristic [curve] and that little distortion will turn into a lot of distortion.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/18/16
I think the big difference here is between a general or first order definition being soft and ambiguous and a more precise model being the same.
The more precise the definition the less ambiguity allowed which is why we create more complex models in the first place and these models must be held to higher standards. Further defining anything with more imprecise terms is counterproductive and it is often these same people with complex definitions based on subjective terms espousing 5 decimal place accuracy to justify their opinions.
Tremaine uses the phrase "most linear" and while that seems like a strong statement, it is actually impossible to define without adding lots of other information / assumptions. Luckily his writings were directed to peers / students where a conceptual understanding of what he is saying doesn't take a huge leap of faith. Fast forward to today where on the interwebs, anyone can be an expert and we have a very different situation.
dave
Soft or ambiguous as it may be, I am quite comfortable ruling out any portion of an AB amp's power output as Class A. I am pretty sure the folks responsible for this were from somebody's marketing department...but given the target of such marketing should also bear some responsibility, I'll lay it on both groups.
For a linestage in particular, I would see no issue waiving the requirement for operation in the middle of the possible load line...could be the dose of 2HD available by running near cut off is desirable...LOL
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Where's the "most linear part" of this curve?
I think we have squeezed all the juice out of this orange. :-)
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
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: