 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
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: mkuller is correct * posted by b.l.zeebub on July 27, 2018 at 14:21:06
If you mean that a Class A/B circuit operates in Class A until it's pushed then you are both wrong.Class A and Class A/B are two different types of circuits.
They are not the same, not at any power level.
It's an easy way to explain Class A/B (saying that it's Class A up to a point then it becomes Class B) but it just not technically true. It is neither Class A or Class B. It's Class A/B.
If you can show me an authoritative reference that says otherwise then do so.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/27/18 07/27/18Follow Ups:
This post reminded me of the new Vitus Audio integrated amp, that can operationally be switched from a 25WPC class A amp to a 300WPC class AB amp.
I'd love to know just how much needs to be switched around to accomplish this. My suspicion is that there are lower voltage power transformer taps for the output stage (maybe everything else is on its own set of windings) and a very different feedback arrangement for each setting.
There is definitely a difference between designing a 25WPC class A amp and a class AB amp that won't go into cutoff until you push past 25WPC. Rail voltage is higher in the class AB amp (Tre has pointed out the drawbacks to raising rail voltage and decreasing collector current) and there is a need for some serious feedback to keep the crossover distortion at bay which just isn't a requirement for the class A design.
If you can show me an authoritative reference that says otherwise then do so.
I trust both Nelson Pass and John Curl in explaining their products that way.
Do either of them make Class A only amplifiers?Ask yourself why?
Let me put it this way,
If you have an A/B amp and you think that the sound (while the amp is in "A" mode) is the best those transistors can do then go ahead and be happy in your ignorance and I'll just leave you alone.
I came to this thread trying to help people gain a deeper understanding but like they say "no good deed goes unpunished".
I'm out.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/28/18
Do either of them make Class A only amplifiers?
One does.
Ask yourself why?
I know why.
If you have an A/B amp and you think that the sound (while the amp is in "A" mode) is the best those transistors can do then go ahead and be happy in your ignorance and I'll just leave you alone.
All of Nelson Pass' class A amps (XA line) have enough capability to leave class A and produce more power . I'm quite happy in his ignorance. LOL!
If it can "leave Class A" to produce more power then it's not a Class A amplifier. It's just a Class A/B amplifier with higher idle currents (which is a good thing but still not true Class A)."I'm quite happy in his ignorance."
Nelson is not ignorant, you just don't understand what Nelson wrote.
Right here Nelson is saying the same thing I'm saying.
"Higher bias doesn't just move the Class A transition to higher ground - it has a profound influence on the amplifier at all power levels. It lowers the distortion at low levels as well as high levels, as seen in the distortion vs power curves for an amplifier with the bias set at different levels."
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/28/18 07/28/18
(nt)
Obviously, Pass is not a class A only snob.
Nelson Pass makes Class A amps too, in his First Watt lineup.
What Tre' is saying is that although Class AB amps can technically operate in Class A mode within the constraints of bias, they will not be operating optimally in Class A.
Say you compare a Class AB amp that is Class A to 10W to a Class A amp rated at 10W. The Class A amp will have superior distortion specifications because it would be optimized for Class A, with higher bias as needed.
So a Class AB amp operating in Class A mode will have inferior performance compared to a Class A amp. If quality Class A performance is demanded, then a Class A amp is required.
First off, what are we optimizing for? Maximum Class A output? Minimum distortion at low input signal levels? Minimum distortion at higher input signal levels? Thermal stability? Bandwidth? Something else?
Biasing the device so that the Q point is in the middle of load line results in the maximum Class A output, but that isn't necessarily the optimum for anything else.
Suppose you want to pick a different Q point, one that gives you a little better linearity at low signal levels. If this happens to be at a higher bias than the midpoint, you're increasing quiescent current dissipation and reducing maximum output. Some designers might make that trade, but most probably won't. On the other hand, suppose this optimum Q point is lower than the midpoint. If the circuit is single-ended, you're reducing maximum output. If the circuit is push-pull, you're reducing maximum output in Class A , but potentially increasing maximum output overall because one device can still amplify when the other hits cutoff.
That's just one possible rationale for choosing a Q point. Other factors that might play into the decision are thermal stability, limits on quiescent current dissipation imposed by the thermal design and operating environment, longevity, distortion at higher signal levels, and in the case of transistors, bandwidth/response time. And most of all, what sounds the best to the designer.
Class A and Class AB are modes of operation, not types of amplifiers. You can't categorically say that one amp which is operating up to 10W in Class A is better than another amp operating up to 10W in Class A.
I do understand what you're getting at.All I will say is picking the Q that gives the least harmonic distortion would mean placing the idle point in the dead center of the most linear part of the dynamic curve.
That's what I would call a Class A design.
I'm not a transistor guy but the above is how Class A is described in the tube text books and WRT designing for the least inherent (before applying NFB) harmonic distortion, transistors should be no different.
This part of this thread started when I perceived that people thought that the so called "Class A" part of the operation of a Class A/B design would be the same as true Class A. It isn't and I did my best to show that. Along the way someone linked an article by Nelson Pass where he, in fact, was saying the same thing as me.
The point is, with the "Class A" part of an A/B output stage, the transistors are not biased in the dead center of the most linear part of the dynamic curve (they are instead biased and therefore operating in a less linear part of the dynamic curve) so the performance can not, will not be up to the same standard as a circuit where the transistors are biased in the dead center of the most linear part of the dynamic curve. (all other things being equal)
The concept of the above is what I was trying to get across.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/30/18
Amplifiers are not my strength, but here's how I look at it.
Class AB is a superset that includes class A and class B.
Single-ended amplifier circuits are inherently class A.
Push-pull amplifier circuits are inherently class AB, but their operation can be confined to class A or class B.
Biasing to the center seems like an obvious starting point for a single-ended circuit. It may not be optimal in some respects, but it does maximize the usable output, which is usually one of the goals.
But I think there is more flexibility with push-pull. If it's acceptable to operate in class AB when high output is required, then the designer is free to choose the Q point based on other criteria. Hypothetically, if you take a push-pull amplifier circuit that is configured to run almost exclusively in class A, you could double the supply voltage while adjusting component values to keep the Q point the same. Now you have an amplifier that can run in class AB up to 4x the class A output power (provided the power supply and devices are up to it).
When it comes to solid state amplifiers, I think it would be hard to find a "Class A" amplifier that doesn't cross into class AB before reaching the clipping point when driving lower impedance loads.
I think that people who buy Class AB amps have speakers that require a bit of power for good sound, so they compromise a bit on distortion for power.
Whereas people who buy low power Class A amps have speakers that are quite efficient with benign impedance. These people are after the best amplifier performance possible, taking advantage of the sonic superiority of Class A in spite of the power consumption and inefficiencies.
So the theoretical limits of Class A are not an issue if the amps are not pushed to their limits. I have low power Class A single ended and push-pull transistor amps and they are paired with efficient speakers that have benign impedance.
It's horses for courses. If you have efficient speakers with benign impedance, power them with a low power Class A amp for best sound. If you have inefficient speakers, power them with a Class AB (or B or D) amp, unless of course you can justify a high power Class A amp that runs very hot, burns a large amount of electricity, weighs hundreds of pounds, and costs a big bundle of money.
Got it.
Mine was more a concept piece.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
although Class AB amps can technically operate in Class A mode within the constraints of bias, they will not be operating optimally in Class A.
Fair enough. I see that higher bias levels continue to reduce distortion even after the stage "reaches" class A (fig 3).
I don't know what to say . I feel like you will think I'm picking on you.
I am not. I am only trying to help.
A circuit is either Class A or it's not. It doesn't "reach" Class A.
** A side note to help clarify. It could be said that a Class A/B amplifier operates in "Class A mode" until it "reaches" the "Class B mode" and that would be an easy way to explain it to someone who's not technically inclined.
That being said, A Class A amplifier is Class A from the smallest output to the largest output it is capable of and then it totally falls apart. So there is no "reaching" Class A with a Class A amp. It just is Class A until it's over driven.
Figures 1 through 3 are showing the output of one of the two transistors in a push pull circuit when those transistors are driven by a sine wave.
Figure 1 shows the output wave form of a transistor that has no bias current (Class B) when driven by a sine wave. The bottom half the wave form is missing.
Figure 2 shows the output wave form of a transistor that has some idle (bias) current (Class A/B). The bottom half of the wave form is mis-shapen but at least there is some bottom half wave form.
Figure 3 shows the output wave form of a transistor that has full idle (bias) current (Class A). This is the idle current that it takes to get the transistor to give a bottom half wave form that looks as much like the top half wave form as the transistor is capable of. This is spoken of as running the device in as "linear a fashion as possible". Some books use the term that Class A is when you run the device only in the "most linear part" of the operating curve.
Another note for clarity, The more inherently linear the device is to start with, the closer one gets to the goal. With a triode vacuum tube the bottom half can look exactly like the top half (within limits). With a transistor, not so much.
P.S. Missing or mis-shapen wave forms or parts of wave forms = harmonic distortion.
I hope you take this in the spirit of good will. The spirit in which I intend it.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I feel like you will think I'm picking on you.
You made statements not supported by fact. To wit:
If it can "leave Class A" to produce more power then it's not a Class A amplifier.
That's a direct quote from Pass.
I hope you take this in the spirit of good will. The spirit in which I intend it.
What I now understand is there are degrees of performance once you enter the state where the devices no longer switch - and that continues to improve with higher levels of bias.
"What I now understand is there are degrees of performance once you enter the state where the devices no longer switch - and that continues to improve with higher levels of bias."
Yes, but only to a point. The bias level can only go as far as full Class A otherwise the device will run into saturation (ie; non-linear behavior much like cutoff but at the other end of the dynamic operating curve).
Only the center part of the of the operating curve is linear. It becomes non-linear at each end. The idea is to bias in the middle of the linear part, ie; Class A.
Please note that I have made no attempt to address trick circuits such as the one Nelson mentions in the link.
"In 1991 Pass Labs developed a hybrid class topology which paralleled a push-pull Class A output stage with a current source which biased it into single-ended Class A. The Aleph 0 amplifier operated as a single-ended Class A amplifier to its output rating of 75 watts into 8 ohms, and at currents beyond that it continued to deliver current as a push-pull Class A circuit."
My comments only address the behavior of a basic single ended or push pull amplifier not a trick circuit that somehow combines the two as described above by Nelson.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
.
Have Fun and Enjoy the Music
"Still Working the Problem"
There's so much audiophile entrenched "understanding" on this topic that the goal post has been moved over the decades. It seems you've caught yourself in a similar predicament other times as well. :)Any amplifier with a push-pull output stage is not Class-A. Only single-ended amplifiers are (or can be) Class-A...or "true" Class-A, if that floats your boat.
Anything else is some sort of "hybrid" configuration that doesn't fit the classic definition.Dave.
Edits: 07/28/18
Neither am I.I was just trying to explain what Nelson explained in the link.
A Class A/B amp operating in the so called Class A Mode* (*Nelsons word, not mine) is not the same as true Class A.
"Higher bias doesn't just move the Class A transition to higher ground - it has a profound influence on the amplifier at all power levels. It lowers the distortion at low levels as well as high levels, as seen in the distortion vs power curves for an amplifier with the bias set at different levels."
Nelson said it and you don't question him, I say it and you think I don't know what I'm talking about.
Whatever!
Like I said in another post on this thread, I entered this thread trying to explain these things to you all. No one what's to believe me. Then someone posts a link to Nelson Pass saying the same thing.
If you don't want to understand what Nelson and I are trying to explain to you, then neither Nelson or I can help you.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/28/18 07/28/18
What Nelson Pass says:
That link proves my point.
"Higher bias doesn't just move the Class A transition to higher ground - it has a profound influence on the amplifier at all power levels. It lowers the distortion at low levels as well as high levels, as seen in the distortion vs power curves for an amplifier with the bias set at different levels."
Nelson Pass is using the term "Class A" loosely when he talks about the "Class A" transition in a Class A/B amplifier.
When the transistors are biased to full Class A, then we are talking about true Class A and it makes a difference as Nelson says.
That is the difference I was trying to point out.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
So that the first 9?dbw is class-A. I think that's the first 8 watts!?Just one running at 16Wpc in UL and 20 in pentode, I guess they must move out of A into AB fairly often? on peaks into easy 8 ohm load 91db speakers in the near-field and a medium-sized l-shaped room. On large force works with lots of percussion and drums.
Yes?
But I bi-amped the spheres/ One amp one in pentode 20wpc into the mid-bass WR driver no LPass, and one in triode 6?wpc into the filtered big domes.
Giving me perhaps an other ~3db headroom over 16wpc. And, yes the system played a good bit louder without stress.
They are both now in pentode mode and have matching PSU storage. I will be experimenting with them into QUAD 63s before I decide whether to sell one or both.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Edits: 07/27/18
For a solid state amplifier with a push-pull output stage I think it is fair to equate a "class A/B" amplifier where the peak load current is never larger than the bias current with "class A".
Regards
13DoW
I understand that almost everyone does that but that doesn't make it technically correct.
If you were to build a Class A push pull amplifier using the same output transistors you would bias them differently for more reasons than just more so called "Class A" power.
They would be biased so that the first watt would be more inherently linear.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Maybe we are cross purposes but class A is all about output current. Nothing to do with the first watt. The linearity of the output stage is determined (to the first order) by the change in current. For class A that is going to be almost zero to twice the bias current as the load current varies by plus/minus the bias current (more than that is class AB)
13DoW
You miss understood me.
I only used that (the distortion at 1 watt) as one example of the difference between Class A/B (even when the Class A/B amplifier is in the so called "Class A Mode") vs true Class A. There are other reasons.
As Nelson Pass says, ""Higher bias doesn't just move the Class A transition to higher ground - it has a profound influence on the amplifier at ALL POWER LEVELS. It LOWERS THE DISTORTION at LOW LEVELS as well as HIGH LEVELS, as seen in the distortion vs power curves for an amplifier with the bias set at different levels."
Note, "Class A Mode" is a phrase used by Nelson Pass for a Class A/B amplifier operating at low power levels when neither output device reaches cutoff.
While it may be called "Class A Mode" it is not the same as true Class A and that was my point from the beginning. A point that Nelson makes better than I.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I caught up with the rest of this thread and think I understand where you are coming from. Single-ended circuits have to be class A in order to swing load current symmetrically. Push-pull circuits with transformer coupling can use two single-ended stages in anti-phase to swing load current symmetrically. This arrangement can be class-AB and will not be as linear as a single-ended class-A stage with the same output specification for the reasons you have stated.In solid state power amplifiers the voltage gain is created using class-A circuits. The output stage is most often a pair of followers connected together to source and sink load current to and from the load. This stage can be biased for class-B, class-AB or class-A operation. Let's have a look at an example of a class-A amplifier designed for 50Wrms into 8 ohms. The peak voltage swing of the output stage must be +/-28.3V and the peak load current will be +/-3.54A. So, for class-A the output stage must be biased at 3.6A with +/- 30V supply voltage rails (rounding up). The output stage will be able to source or sink 3.54A without either follower device shutting off (i.e. with no current flowing through it). The idle dissipation of that output stage will be 216W. In operation as the output voltage rises the pull-up follower will supply current to the load and the difference between the bias current and the load current flows through the pull-down follower e.g. at 25Wrms the peak load current is 2.5A so 3.6A flows through the pull-up follower, 2.5A flows out to the load and 1.1A is left to flow through the pull-down follower. When the peak load current reaches 3.54A no current flows in the pull-down follower. For an 8ohm load that happens at our maximum output voltage so the load current cannot be higher. If the load resistance was really 6 ohms then with a 28.3V peak output voltage the peak current into the load will be 4.72A. The pull-up follower is biased for 3.6A so it will turn-on more to conduct an extra 1.12A and all of that current will be delivered to the load. No current will flow in the pull-down follower and we have left class-A operation.
Now let us make another amplifier with double the output voltage swing but leave the bias current unchanged. The supply voltage rails will now be +/- 60V and with the 3.6A bias current the idle dissipation will be 432W. This amplifier is now a 400Wrms into 8ohms class-AB amplifier that will operate in class-A for the first 50W (in 8ohms). To the first order the behavior of these two amplifiers up to 3.54A peak load current is identical. This is the point that I, and I think others in this thread, have been making.
Hope that helps
13DoW
Edits: 07/30/18 07/30/18 07/30/18 07/30/18
Are you saying that the transistors in a push pull transistor circuit with +/- 60 volts, biased at 3.6A, have the same linearity characteristics as the transistors in a push pull transistor circuit that has +/- 30 volts, biased at 3.6A?With a tube, the operating point (voltage, idle current and load impedance) determines the linearity. Changing the voltage across the device while leaving the idle current and load impedance the same will change the linearity of the device.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/31/18
All active devices have a non-linear transfer characteristic from Vin to Iout - for tubes and FETs this the predominantly square law and for bipolar transistors (BJTs) it is predominantly exponential. This is the primary cause of distortion.
For transistors, the change in transfer characteristic caused by changing the voltage across the device (drain-source or collector-emitter) is modeled by a term called the Early Voltage that acts as a resistor in parallel to the transistor and defines its inherent output resistance. But, this is a linear effect so does not contribute to distortion. The linearity of the output followers in my hypothetical amplifier example will not change as the supply voltage goes from +/-30V to +/60V. There could be second order effects if that change mandated picking different transistors to handle the additional voltage. Higher voltage handling FETs will have lower transconductance and higher voltage BJTs will likely have lower current gain that could increase distortion but, again, this would be a second or third order effect.
I am not a tube guy but I would expect them to behave in a broadly similar way (though I don't think anyone makes a tube power amp with cathode follower outputs). Maybe the tube equivalent of the early voltage is a non-linear effect?
Regards
13DoW
"The linearity of the output followers in my hypothetical amplifier example will not change as the supply voltage goes from +/-30V to +/60V."Is that because they are followers?
BTW The Circlotron circuit uses bridged cathode followers as an output stage.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/01/18
Do any of you guys actually design amplification circuitry, other than for your own personal enjoyment?
I'm asking because I'd like to know if any of you are pros in this field.
Yes. Not for the audio frequencies, though - RF.
I guess if you count control circuits that operate at low frequencies, they could be considered as audio. But, sound reproduction is not their intended purpose.
I meant to add a footnote to my last post that I am a semiconductor designer not an audio electronics designer so take my inputs as you will. But basic principles are true in both disciplines except that in the audiophile world there is little correlation between objective performance and subjective preference so knowing those principles/not knowing them may not help/hinder.
Regards
13DoW
I have designed and built tubed microphone preamps use in professional recording studios. Does that count?Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/02/18
Exactly! Well explained.
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
