Why? They got told to build an amp. That given the whole deck of requirements, it was going to be an AB amp that could not afford a solid supply. I believe the term used to deride the appropriate folks would be, 'bean counter'.
Now the bean counter may have been right, and such an amp would not have been sale-able at the required price point. For that I think you can judge who is to blame, yes?
cheers,
Douglas
Not the case if the output stage is configured as a LTP with a CCS tail. In this case the signal current can only circulate plate to plate and cathode to cathode. It will not work with one tube cut off. With class B prohibited, it will only operate in case A. In a normal PP amp as above, when operating in class A there is no guarantee of AC balance between the two halves of the circuit. The stronger tube will send an imbalance current from its plate through the PS picking up distortions and colorations there on its way back to its cathode. A LPT circuit enforces AC balance limited by the impedance of the CCS.
As an aside I have PPP KT-120 mono-blocks that output 37W of class A triode power (and a whole lot of heat).
and there in lies the problem. Class A is not defined only by 360-degree conduction. Add at full power, at its distortion limit and you approach completeness.
An AB amp operates in AB...the likely marketing derived, incomplete definition of Class A is not the whole story.
cheers,
Douglas
Quite often.
Unfortunately I spent about 20 years of my engineering career dealing with, and trying to balance the marketing guys, the bean counters, and other engineers against the laws of physics.
It is impossible to tell a marketing person that you can't design a VHF police radio that puts out 5 watts of RF in a cell phone sized form factor with a cell phone battery. We actually had to waste half a year to build them a prototype. The battery lasted about an hour max, and the antenna would burn you from the RF.....new marketing team.....new problems. After several years of this, it became mandatory for the marketing people, and several other "committee" members to spend some time in the field with cops, firemen, and paramedics to see the real world where the products get used.
It's a fine line to walk trying to live in the world of "design by committee." The results will ALWAYS be a compromise, and so will the ad copy. The rules are far stricter with "mission critical" radio equipment where lives are at stake if it doesn't work right. Fortunately many of the critical specs are determined by the FCC and APCO (Association of Police Communications Officers).
The world of high end HiFi have no such watchdogs.
The definitions of class A, class AB, class B, and class C can be found in any number of documents from the 1930s onward. Try checking Western Electric, Bell Labs, ARRL, RCA, Eimac, etc. for these definitions.
Whatever 'audio' marketing tells you, the original definitions apply.
Changing the definitions does not allow for a level playing field for all players
(Oh, that IS what marketing often is).
The class of an amplifier is defined by the angles of conduction, regardless whether the circuit is single ended, push pull, etc.
Using anything but class A for single ended will not be pleasing to most listeners (*).
2 anti-phase single ended that a combined at the output in another phase reversal,
may be defined as a special case push pull, whether they use a push pull transformer, or 2 single ended transformers. This will allow the tubes to go into cutoff without causing the case listed above (*)
Take a push pull amplifier that can operate in class A at lower signal levels (no tube in cut off), but that can also operate at higher signal levels in AB (tubes alternately in cut off at signal extremes).
To argue over whether that amp is a class A or a class AB amplifier seems pointless.
Actually, there have been marketing firms in the past that have rated Both the class A power out, and the class AB power out of a push pull amplifier model.
As to grid current, 1 is no appreciable grid current (only capacitive current), and 2 is appreciable grid current (conduction current, not just capacitive current).
Normally the 1 is assumed when no number is present (unwritten):
A =A1 A2
AB = AB1 AB2
B = B1 B2
C is the exception. C is most often used for RF. Most of those RF tubes that are used in C Do draw grid current in most real applications. So although there is no written 1 or 2 suffix, the 2 is assumed for class C.
The Control Grids of some RF tubes are rated for Hundreds of Watts dissipation. And also the Screen Grids of some RF tubes are rated for Hundreds of Watts dissipation.
Try that with your favorite audio tube (KT88, etc.)
I once built an audio amplifier using the screen grid of a 4-65 as the "plate". The real plate was disconnected.
The screen of the 4-65 RF tube is rated for 10 Watts, the same as a type 45 triodes 10 Watt plate dissipation rating.
Class C audio will not be enjoyed by most listeners, with or without grid current.
I will let others talk about class D. That is a whole other subject.
Whatever 'audio' marketing tells you, the original definitions apply.
Changing the definitions does not allow for a level playing field for all players
(Oh, that IS what marketing often is).
The class of an amplifier is defined by the angles of conduction, regardless whether the circuit is single ended, push pull, etc.
Using anything but class A for single ended will not be pleasing to most listeners (*).
2 anti-phase single ended that a combined at the output in another phase reversal,
may be defined as a special case push pull, whether they use a push pull transformer, or 2 single ended transformers. This will allow the tubes to go into cutoff without causing the case listed above (*)
Take a push pull amplifier that can operate in class A at lower signal levels (no tube in cut off), but that can also operate at higher signal levels in AB (tubes alternately in cut off at signal extremes).
To argue over whether that amp is a class A or a class AB amplifier seems pointless.
Actually, there have been marketing firms in the past that have rated Both the class A power out, and the class AB power out of a push pull amplifier model.
As to grid current, 1 is no appreciable grid current (only capacitive current), and 2 is appreciable grid current (conduction current, not just capacitive current).
Normally the 1 is assumed when no number is present (unwritten):
A =A1 A2
AB = AB1 AB2
B = B1 B2
C is the exception. C is most often used for RF. Most of those RF tubes that are used in C Do draw grid current in most real applications. So although there is no written 1 or 2 suffix, the 2 is assumed for class C.
The Control Grids of some RF tubes are rated for Hundreds of Watts dissipation. And also the Screen Grids of some RF tubes are rated for Hundreds of Watts dissipation.
Try that with your favorite audio tube (KT88, etc.)
I once built an audio amplifier using the screen grid of a 4-65 as the "plate". The real plate was disconnected.
The screen of the 4-65 RF tube is rated for 10 Watts, the same as a type 45 triodes 10 Watt plate dissipation rating.
Class C audio will not be enjoyed by most listeners, with or without grid current.
I will let others talk about class D. That is a whole other subject.
Last edited:
The class first appeared when first PP stage was designed. It was called "Class B", while previous amps were called "Class A", to stress that 2 devices work alternately, delivering more power and saving battery life.
Later "Class AB" was defined, for higher idle current, with gm doubling on low volume that is most critical for listening. Later A2 was defined, that assumed presence of control grid current.
Class C was used in transmitters, when power tubes were biased to cut off at idle.
Class D was called an amp that used pulse width modulation of a carrier signal, with saturating output devices, that gave the highest possible efficiency due to low power drop on power devices that switch from cut-off to saturation.
However, some other possibilities exist. I developed class A+B and class A+C amps. In class A+B amp output devices were always conducting the current, but upon demand their current amplification factor was increased. Class A+C works in class A, but 2 additional class C devices gradually add current upon demand. The last approach was also patented by Peter Walker and called "Current Dumping".
Later "Class AB" was defined, for higher idle current, with gm doubling on low volume that is most critical for listening. Later A2 was defined, that assumed presence of control grid current.
Class C was used in transmitters, when power tubes were biased to cut off at idle.
Class D was called an amp that used pulse width modulation of a carrier signal, with saturating output devices, that gave the highest possible efficiency due to low power drop on power devices that switch from cut-off to saturation.
However, some other possibilities exist. I developed class A+B and class A+C amps. In class A+B amp output devices were always conducting the current, but upon demand their current amplification factor was increased. Class A+C works in class A, but 2 additional class C devices gradually add current upon demand. The last approach was also patented by Peter Walker and called "Current Dumping".
I thought the suggestion that these technical terms have been well defined for years, and that whatever the historical definition was should be the convention was indeed a powerful suggestion.
In RDH4 I could not find a reference to class A, only to class B which appeared to perhaps be synonymous with PP amps.
In Crowhurst's Basic Audio Vol 2, there is a long discussion of Class A, Class B, and Class AB tube PP amps. Class A indeed refers to a PP amp that can achieve full power with both tubes conducting current all the time.
In RDH4 I could not find a reference to class A, only to class B which appeared to perhaps be synonymous with PP amps.
In Crowhurst's Basic Audio Vol 2, there is a long discussion of Class A, Class B, and Class AB tube PP amps. Class A indeed refers to a PP amp that can achieve full power with both tubes conducting current all the time.
The polarization of an electronic device defines the amplifier class.
So, a tube or a bjt or a mosfet amplifier operates in class A when the theoretical efficiency eta (Pout/Pin) is <=25%.
Class A with an impedance as load acts like a push pull class A, both exhibit an eta <=50%.
Class B eta is <=78.5%.
Class AB has no standard eta (it depends on Ibias); of course it will be somewhere between 50 and 78.5%.
That's it.
EDIT:
So, when your class AB amp at very low power exhibits an efficiency of 50%, and at high level an efficiency of, say, 70%, his average efficiency belongs to class AB.
But technically it acts like a class A push pull when at 50% of efficiency.
And as class AB when >50%.
In the end:
1 - it is always class AB;
2 - it is class A (eta 50%) and class AB (eta >=50%) and class B (eta 78.5%).
1 = 2.
EDIT 2:
Not always class AB amps operates in class A too.
It depends, you know, by efficiency.
So, when a manufacturer claims its amp is class A until, say, 4W and then class AB, it is not a lie.
It's always class AB, because the global efficiency is somewhere between 50 and 78.5%.
A class AB not so 'hot' could never operate in class A, but if its efficiency is <=78.5%, it is still a class AB amp.
corrected some acronyms.
So, a tube or a bjt or a mosfet amplifier operates in class A when the theoretical efficiency eta (Pout/Pin) is <=25%.
Class A with an impedance as load acts like a push pull class A, both exhibit an eta <=50%.
Class B eta is <=78.5%.
Class AB has no standard eta (it depends on Ibias); of course it will be somewhere between 50 and 78.5%.
That's it.
EDIT:
So, when your class AB amp at very low power exhibits an efficiency of 50%, and at high level an efficiency of, say, 70%, his average efficiency belongs to class AB.
But technically it acts like a class A push pull when at 50% of efficiency.
And as class AB when >50%.
In the end:
1 - it is always class AB;
2 - it is class A (eta 50%) and class AB (eta >=50%) and class B (eta 78.5%).
1 = 2.
EDIT 2:
Not always class AB amps operates in class A too.
It depends, you know, by efficiency.
So, when a manufacturer claims its amp is class A until, say, 4W and then class AB, it is not a lie.
It's always class AB, because the global efficiency is somewhere between 50 and 78.5%.
A class AB not so 'hot' could never operate in class A, but if its efficiency is <=78.5%, it is still a class AB amp.
corrected some acronyms.
Last edited:
No Cutoff, at full designed power and load.
We can always force cutoff by increasing the drive or load. But that's not what designers do.
Grid current and "A" have nothing to do with each other. We got then numbers for that.
The very first push-pull amps, we would call "A". It took a year to really grasp that idle current could be run WAY lower, class B in theory but always AB in practice.
Class C is brutal (and distorted) work, and historically only used with tuned circuits. In that environment, grid current is the usual thing to make the most of the tube dollar. Doug Self has a family tree (not his!) of how these class "G" etc amps can be decomposed into A B and even C portions. The high-volt peaks are covered by stages which are surely "C", though not the energy-storing in/out tanks familiar in radio work.
We can always force cutoff by increasing the drive or load. But that's not what designers do.
Grid current and "A" have nothing to do with each other. We got then numbers for that.
The very first push-pull amps, we would call "A". It took a year to really grasp that idle current could be run WAY lower, class B in theory but always AB in practice.
Class C is brutal (and distorted) work, and historically only used with tuned circuits. In that environment, grid current is the usual thing to make the most of the tube dollar. Doug Self has a family tree (not his!) of how these class "G" etc amps can be decomposed into A B and even C portions. The high-volt peaks are covered by stages which are surely "C", though not the energy-storing in/out tanks familiar in radio work.
There were some Single Ended RF amplifier Tube output stages that operated in class AB.
The single tube conducted more than 180 degrees, and less than 360 degrees of the RF
carrier. The output tank circuit acts like a flywheel, so the output was a RF carrier sine wave, not a truncated/rectified sine wave.
They could not use a class C amp there, because that particular single ended class AB amp was also doing AM modulation service. Class AB allowed a moderate % of AM modulation, with some reasonable degree of efficiency and modulation linearity.
Class C is not capable of doing linear AM modulation.
That was all discovered/designed with tubes.
As to solid state:
The early Analog cell phones used FM modulation, and class C RF amplifiers. Good
output efficiency.
The early Digital cell phones were all GSM, and could also use class C RF amplifiers.
Good output efficiency.
Then along came some digital cell phones with american digital modulation standard that required class AB RF amplifiers. I believe it was 8PSK (it had zero crossings in the vector diagrams).
Can't remember the name of the standard (IS-136?), but it was not very good output efficiency.
Then along came Qual Com digital cell phones, and OQPSK Modulation, and they could use class C RF amplifiers. Good output efficiency.
Then GSM cell phones evolved into EDGE, a special trick with more symbols, but no zero crossings in the vector diagram, so it was also good output efficiency.
Someone mentioned class D switching for high efficiency. Pulse width or Pulse Repetition rate modulation could be used to get audio out (and a Lot of low pass filtering is needed too).
The earliest form that even comes close to the efficiency of class D, but back then with tubes was class C.
A very high efficiency tube is the magnetron (a whole different RF story).
Just remember:
"Those who do not know history may not be able to repeat it" (even if they want to).
The single tube conducted more than 180 degrees, and less than 360 degrees of the RF
carrier. The output tank circuit acts like a flywheel, so the output was a RF carrier sine wave, not a truncated/rectified sine wave.
They could not use a class C amp there, because that particular single ended class AB amp was also doing AM modulation service. Class AB allowed a moderate % of AM modulation, with some reasonable degree of efficiency and modulation linearity.
Class C is not capable of doing linear AM modulation.
That was all discovered/designed with tubes.
As to solid state:
The early Analog cell phones used FM modulation, and class C RF amplifiers. Good
output efficiency.
The early Digital cell phones were all GSM, and could also use class C RF amplifiers.
Good output efficiency.
Then along came some digital cell phones with american digital modulation standard that required class AB RF amplifiers. I believe it was 8PSK (it had zero crossings in the vector diagrams).
Can't remember the name of the standard (IS-136?), but it was not very good output efficiency.
Then along came Qual Com digital cell phones, and OQPSK Modulation, and they could use class C RF amplifiers. Good output efficiency.
Then GSM cell phones evolved into EDGE, a special trick with more symbols, but no zero crossings in the vector diagram, so it was also good output efficiency.
Someone mentioned class D switching for high efficiency. Pulse width or Pulse Repetition rate modulation could be used to get audio out (and a Lot of low pass filtering is needed too).
The earliest form that even comes close to the efficiency of class D, but back then with tubes was class C.
A very high efficiency tube is the magnetron (a whole different RF story).
Just remember:
"Those who do not know history may not be able to repeat it" (even if they want to).
Last edited:
Yes it is.
That would be a meaningless definition since it means one circuit (or identical circuits) could be labelled with different operating classes depending on the whim of the person doing the labelling.
What becomes Class AB? The amplifier, or the mode of operation?Merlinb said:
If we say that an amplifier is Class AB then surely by definition that is a statement about the amplifier, not a statement about the amplifier plus a particular input signal? Putting a bigger signal in does not change the amplifier so it (the amplifier) remains the same device. As always, there is an exception: those amplifiers which change their bias point so much when signal is present that they really do change to a different class - rare in solid-state, but more common than people realise for valves?
Class AB should not have gm-doubling. In reality almost all 'Class B' amps are actually Class AB. Class B is a textbook idealisation but real devices don't work like that, so for practical purposes Class B means AB with low quiescent current and Class AB means AB with higher quiescent current - in both case gm-doubling should be avoided. Just to complicate things, there are the 'first watt' fans who like gm-doubling.Wavebourn said:
Alright then marco, if you'd be so kind as to offer a definition that describes the operation of the classic 2A3 SE use...and do please make mention of how it differs from the incomplete definition you propose for Class A.
cheers,
Douglas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.