New Posts New Posts RSS Feed - Amplifier Classes (ABC)
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Welcome to the Graham Slee Audio Products Owners Club

 

Open to all owners plus those contemplating the purchase of a Graham Slee audio product wishing to use our loaner program: join here
Subscribe to our newsletter here (Rules on posting can be found here)


Amplifier Classes (ABC)

 Post Reply Post Reply
Author
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 9117
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Topic: Amplifier Classes (ABC)
    Posted: 20 Mar 2020 at 11:17am
This new topic examines the amplifier classes and tries to re-educate the hi-fi enthusiast about the fear of class-B...


This image is from Basic Electronics, volume 2, page 30
Not simple enough for Google-Bot to understand...
Back to Top
BAK View Drop Down
Senior Member
Senior Member
Avatar

Joined: 14 Mar 2010
Location: Kentucky, USA
Status: Offline
Points: 1010
Post Options Post Options   Thanks (0) Thanks(0)   Quote BAK Quote  Post ReplyReply Direct Link To This Post Posted: 20 Mar 2020 at 3:15pm

The class AB push-pull output circuit is slightly less efficient than class B because it uses a small quiescent current flowing, to bias the transistors just above cut off as shown in Fig. 5.5.1, but the crossover distortion created by the non-linear section of the transistor’s input characteristic curve, near to cut off in Class B, is overcome. In class AB each of the push-pull transistors is conducting for slightly more than the half cycle of conduction in class B, but much less than the full cycle of conduction of Class A

As each cycle of the waveform crosses zero volts, both transistors are conducting momentarily and the bend in the characteristic of each one cancels out.

The above is found at: https://learnabout-electronics.org/Amplifiers/amplifiers55.php

Footnote: (A little-known fact...)

When Class B is properly biased, it can have better distortion levels and much less chance of "thermal run-away" than Class AB.


Bruce
AT-14SA, Pickering XV-15/625, Technics SL-1600MK2, Reflex M, Lautus, Technics SH-8066, Dynaco ST120a, Eminence Beta 8A in custom cabs;; Using Majestic DAC
Enjoy Life Your Way!
Back to Top
Fatmangolf View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 Dec 2009
Location: Middlesbrough
Status: Offline
Points: 5311
Post Options Post Options   Thanks (0) Thanks(0)   Quote Fatmangolf Quote  Post ReplyReply Direct Link To This Post Posted: 23 Mar 2020 at 9:11pm
Good points well made.
Jon

Open mind and ears, whilst owning GSP Genera, Accession, Elevator EXP, Solo ULDE, Proprius amps, Cusat50 cables, Lautus digital cable, Spatia cables and links, and a Majestic DAC.
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 9117
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 25 Mar 2020 at 9:39pm
Apologies for the pregnant pause in taking this topic forward, but I've been a bit more busy than usual. It will be continued.
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 9117
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 27 Mar 2020 at 3:44am
I'm not going to guesstimate the length of time the signal spends at 0V as it passes through it every half cycle because it is incredibly small. Each degree at 1kHz takes less than 0.000003 seconds, but the size of zero is far shorter than 1 degree, and the more you think about it, the smaller it gets. Perhaps it's at the speed of light?

During that time, no current flows, and it can be class-A or class-B, and still, no signal current can flow to the load.

What we are interested in is how the signal negotiates the distance and time between no conduction and conduction, and conduction is where one of the transistors is doing the pulling (how can it push?).

Class-B is where one transistor is cut-off - is not conducting at all - and the other is. Class-A is where one transistor is pulling while the other transistor is consuming that pull while sharing it with the load.

Thinking about that last statement, class-A cannot exist in a "push-pull" arrangement. It can only be class-AB or maybe class-AAAAAB, but not fully class-A.

Therefore, all "push-pull" outputs must be class-B with differing amounts of bias depending on what the designer wants to achieve.

Pure class-A is a misnomer, if push-pull. Where does the half-cycle leave the number of A's and become a B?

By now, you should see that an "overlap" exists. Where it transitions from overlap to no overlap, there exists a change of distortion properties. If that change point is audible, the change point will be discernible, simple!

In which way will it be audible? It will be audible where the harmonic distortion distribution changes.

In an audio-worthy op-amp, that point tends to be at near clipping, and as such, its linear portion - most of the supply voltage - it will deliver a clean output.

The op-amp's emitter resistors will have been made comparatively large in value compared to a power amp, and as such, will not go into thermal runaway, as in a power amp if we are not careful.

The op-amp can, therefore, be considered class-A for all intents and purposes, and so any attempts at forcing the class-A output into class-A is futile, and will only result in increased distortion. But that is a story for another time and place.

To prevent a meltdown situation or the need to use fans - which is OK at considerable sound pressures, but no good in-home listening - class-B bias offers relatively cool, reliable running.

To be continued
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 9117
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 28 Mar 2020 at 1:06am
To explain bias spreading, I have included a diagram from Nuts & Volts by Ray Marston.



Q2 and Q3 have approx. 0.6V base-emitter voltages, and D1 and D2 drop approx. 0.6V each.

If wire links replaced D1 and D2, the signal swing from Q1 would have to go 0.6 positive before Q2 started to conduct, and 0.6V negative before Q3 began to conduct.

There would be crossover distortion at the zero-crossing of the signal, as shown in the plot below.



Due to misinformation, many believe this illustrates a class-B output: they are wrong.

D1 and D2 provide bias such that the upper and lower transistors conduct either side of zero to produce a smooth transition.

The standing current through Q2 and Q3 can be anything slightly above zero, and provided emitter resistors R4 and 5 are around 1/10th the load, you will not see any flats on an oscilloscope display. It will merely transition as an undistorted sine-wave.

The label "crossover distortion" is highly misleading. It should say "gross crossover distortion."

No matter how undistorted the transition looks on an oscilloscope, there is crossover distortion, but not because of any flattening: it is merely not quite an exact representation. It is very close, and so close that a pure sine wave laid on top of it, would reveal nothing to the naked eye.

The visually undistorted sine wave with no flats at zero-crossing is what class B does, and also what any number of A's before the B does.

To be continued
Not simple enough for Google-Bot to understand...
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Telling it as it is

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 9117
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 29 Mar 2020 at 12:40am
The voltage holding/spreading the bases apart should be 1.2 volts (2 x 0.6V) for the transistors to just start to conduct, and so there is no step either side of 0V.

That takes care of each transistor's base-emitter junctions, but what about the emitter resistors?

For current to flow, there must be, by ohms law, a voltage developed across each emitter resistor.

And, because it does, it can be considered as class-aB or class-AB1, but it is still really class-B.

Any current flowing at all in the emitter resistors will have the effect of the two resistors being paralleled, in series with the load, at low output.

As the signal grows toward high output, more current is drawn through one half of the output stage than the other, and it will appear to the load that there is only one resistor in series with it.

It is where the resistance "changes" that the distortion characteristics change, and if the current flowing is not optimised, the crossover distortion is heard.

The portion of the output signal where both emitter resistors are seen in parallel is known as "gm-doubling," and if this were a common emitter stage, the gain would double.

It is not, however, common emitter. It is an emitter follower, so the gain cannot be greater than 1. It can only be 1 or less, and, will be nearer to 0.9, this being due to the higher than 100% negative feedback of an emitter follower.

Therefore "gm-doubling" is simply a concept to explain the "thickening" of signal current around the crossover point.

Pure class-A is a deceiving name for a class-B amplifier, biased on heavily. It would be far better to call it "high-bias."

This "thickens" the crossover area until the signal reaches a much louder swing, where it is hoped the transition distortion won't be heard.

The problem here is heat, and stabilising against thermal runaway. It will be heavily dependant on ambient temperature and supply voltage. The chances are that the distortion product change will happen at different loudness between one sample and another.

It is often reported that high-bias amplifiers sound bright or sibilant with some music, which can translate as being thin on busy sections such as in pop and rock. High bias amplifiers seem to have found a niche with orchestral listeners because of the "added interest."

Oliver's criterion suggests the optimum emitter resistor voltage to be 26mV - the same as transistor thermal voltage - but source resistance and transistor beta can result in that being far too high.

It is not just resistance and specified beta, but impedance, which changes with frequency, and beta-gain bandwidth, that also varies with frequency. This led Oliver to set a lower voltage of 13mV, but even that can be too high.

There are also mismatches between complementary transistors (NPN and PNP), which can necessitate different values of emitter resistance between the top and bottom half.

This explanation is intentionally simplified and really only brushes the surface, but hopefully explains that there is more to it than asking, "is it class-A?"
Not simple enough for Google-Bot to understand...
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.

This page was generated in 0.063 seconds.