For a push-pull amplifier, the pull-up transistor is on for 0 degrees through 180 degrees, and the pull-down transistor is on for 180 degrees through 360 degrees. Is there any difference in the way the circuit performs when it is operating in Class A vs. Class B? It would seem that in both cases, the two transistors still have to turn on and off at their respective times. So, what would the advantages be to operate the push-pull design in Class A? Don't you still have the crossover distortion and dead time?
For a balanced amplifier, would the best design for Class A operation be two single-ended circuits, one for the in-phase full cycle and the other for the inverted-phase full cycle, so that all transistors are on all the time? No crossover distortion, no dead time that you get with push-pull. Of course, you have more power capability with push-pull, but let's assume the power supplies for each design are the necessary size to produce the same output into 8 ohms, say, 100 watts RMS at 0.01% THD+N.
For a balanced amplifier, would the best design for Class A operation be two single-ended circuits, one for the in-phase full cycle and the other for the inverted-phase full cycle, so that all transistors are on all the time? No crossover distortion, no dead time that you get with push-pull. Of course, you have more power capability with push-pull, but let's assume the power supplies for each design are the necessary size to produce the same output into 8 ohms, say, 100 watts RMS at 0.01% THD+N.
Class A is when the transistors don't turn off...by definition.
With the same rail voltages and load you will get the same output power whether it is in Class A or Class AB...Just power supply requirements and transistor power dissipation go up...sometime significantly.
Hope that clears up some confusion
With the same rail voltages and load you will get the same output power whether it is in Class A or Class AB...Just power supply requirements and transistor power dissipation go up...sometime significantly.
Hope that clears up some confusion
Actually, what I am referring to is the sound quality. Since the two transistors responsible for each half of the full wave have to turn on and off with push-pull, is there any sound quality difference with push-pull operating in class A? The definition of not turning off in class A does not happen with push-pull, since the pull-up transistor turns on at 0 degrees and off at 180 degrees, then the pull-down transistor turns on at 180 degrees (minus the dead time) and off at 360 degrees. The on all the time definition only applies to a single-ended circuit, when one transistor creates the full positive and negative portion of the wave form. I am trying to understand what sound advantages there are with push-pull operating in class A. Are there any?
In a class A amp you can think of it as being two SE amps connected - ONE on EACH RAIL. Both "sides" are on all the time, they never turn off or go through zero, unless the amp is driven beyond the current capability set by the quiescent current + the max "on" swing" in which case it can lapse into class AB, and then the devices can be driven through "zero".
The advantage of push pull class A over AB is the lack of crossover distortion, and since both devices are on, the resulting harmonics (distortion) tend to be somewhat different in a favorable way.
If ur asking about PP Class A vs. SE, that is a slightly different discussion.
So, PP Class A: devices stay on all the time, positive and negative signal swings.
PP Class AB: devices are on slightly, and one side alternately will cut off with fairly small signal input (once the bias is exceeded)
_-_-
Balanced input is another story.
Bridged SE amps can be run with balanced input, but this is not the same as a PP Class A amplifier.
Look at the F5 (Pass designed) amplifier here - it is a very nice simple PP Class A amplifier
The advantage of push pull class A over AB is the lack of crossover distortion, and since both devices are on, the resulting harmonics (distortion) tend to be somewhat different in a favorable way.
If ur asking about PP Class A vs. SE, that is a slightly different discussion.
So, PP Class A: devices stay on all the time, positive and negative signal swings.
PP Class AB: devices are on slightly, and one side alternately will cut off with fairly small signal input (once the bias is exceeded)
_-_-
Balanced input is another story.
Bridged SE amps can be run with balanced input, but this is not the same as a PP Class A amplifier.
Look at the F5 (Pass designed) amplifier here - it is a very nice simple PP Class A amplifier
Last edited:
So, for push-pull in class B, the pull-up transistor operates only in the + part of the waveform and the pull down transistor operates only in the - portion, correct? If the push-pull amp is operating in class A, both the pull-up and pull down transistors are reproducing the entire waveform, I.e., 0 degrees through 360 degrees, right? I thought this would burn up the amplifier if both transistors were on at the same time.
Might make the room too hot !! 😀
8 of the amp(lets) (below) , let me shut off the heat to the living room !!
HUGE heatsinks + 2.4KVA of trafo !
PS - each was biased to 2.4A (always on -true class A)
OS
Attachments
sdinfo, no the amp ought not burn up. But it does eat a lot of current = a lot of heat = big heatsinks for most Class A amps.
Anatoliy, are you saying ur output stage prevents turn-off, or that you can bias it into Class A and have current limiting?
_-_-
Anatoliy, are you saying ur output stage prevents turn-off, or that you can bias it into Class A and have current limiting?
_-_-
Hard to see in the posted sim, because the devices seem to be hitting "0".
Which thread, which Steven?
We had the one some years back - trying to recall the amp/designer - where no one could decide if it actually turned off or not, and there were quite a few built, iirc... and then someone offered up a paper or another source that showed that preventing switch off was worse or no better than switch off because of the non-linearity introduced in the "crossover region", iirc.
Can't quite dredge up the right name, it had been a commercial amp, and the designer had come on and offered up some of his old PCBs...
_-_-
Which thread, which Steven?
We had the one some years back - trying to recall the amp/designer - where no one could decide if it actually turned off or not, and there were quite a few built, iirc... and then someone offered up a paper or another source that showed that preventing switch off was worse or no better than switch off because of the non-linearity introduced in the "crossover region", iirc.
Can't quite dredge up the right name, it had been a commercial amp, and the designer had come on and offered up some of his old PCBs...
_-_-
I think it was Steve Dunlap who had that output stage that caused so much controversy.
edit: Yes here it is - The Krill. Is that what you were thinking of?
edit: Yes here it is - The Krill. Is that what you were thinking of?
Last edited:
Ostripper wrote: "Might make the room too hot !!
8 of the amp(lets) (below) , let me shut off the heat to the living room !!
HUGE heatsinks + 2.4KVA of trafo !
PS - each was biased to 2.4A (always on -true class A)
OS"
Your oscilloscope trace shows full waveforms, but they are out of phase with each other, indicating this is a balanced design, rather than a push-pull design.
8 of the amp(lets) (below) , let me shut off the heat to the living room !!
HUGE heatsinks + 2.4KVA of trafo !
PS - each was biased to 2.4A (always on -true class A)
OS"
Your oscilloscope trace shows full waveforms, but they are out of phase with each other, indicating this is a balanced design, rather than a push-pull design.
Last edited:
Push-pull class A is just a class AB output stage operated at a higher idle current, so that both transistors stay "on" well past the zero V crossing. Class AB does not have "dead time" but it does have a voltage range where the positive and negative transistors are both conducting, the class A region. Higher idle current expands the class A region.
Crossover distortion might be characterized as gross distortion, as in class B where the transistors are both off near zero V, or the more subtle increase in THD in class AB caused when one side is turing off while the other is turning on. Class A has the lowest distortion, and the advantage of push-pull class A/AB is the very low distortion at low output levels while still being capable of delivering higher output.
For push-pull class A to be valid, the idle current would have to be such that both transistors are on for the full cycle. Otherwise, where would be the defined point of operation where the transistors are on at the same time for a specified amount of time, but not for the full cycle, and stated as "now it is class A operation"?
Class A means the transistors are on at the same time for the full cycle. Class AB could mean the transistors are on at the same time for part of the cycle. Class B is where they are not on at the same time, and you get dead time and crossover distortion.
Class A means the transistors are on at the same time for the full cycle. Class AB could mean the transistors are on at the same time for part of the cycle. Class B is where they are not on at the same time, and you get dead time and crossover distortion.
You get "crossover" type distortion whenever one rail's transistors cease operating, it's just not at the "zero" crossover point. It is commonly referred to as Gm doubling". Or halving, depending on how you look at it...
_-_-
_-_-
Yes, it does not turn off, just current limiting follows by driving requirements. It is such a sneaky symmetric opamp with unity gain, and "short" tails in diffpairs that are getting even shorter when it simulates "clas AB". You may see on the right, how currents change with signal.
When it clips it sounds like a tube amp clipping.
Since the OP says he's referring to sound quality let me put forward a somewhat heretical notion here - the SQ is really nothing much to do with switching on/off of the output stage. Distortion is already low enough even in classB amps not to be audible, assuming its not gross high-order stuff.
What makes a SQ difference is the noise induced onto the PSU rails. In classA its less han for classB - but there's one form of classA where its practically zero. That being balanced, CCS-loaded SE where the output transistors operate as virtual shunts and a constant current is drawn (by virtue of the two CCSs) no matter the signal or load.
Anyone having any listening experience of such a classA amp? They're at best only 25% efficient so not particularly popular.
What makes a SQ difference is the noise induced onto the PSU rails. In classA its less han for classB - but there's one form of classA where its practically zero. That being balanced, CCS-loaded SE where the output transistors operate as virtual shunts and a constant current is drawn (by virtue of the two CCSs) no matter the signal or load.
Anyone having any listening experience of such a classA amp? They're at best only 25% efficient so not particularly popular.
Member
Joined 2009
Paid Member
It's twice as power-efficient as single ended. And you can design it as a heavily biassed Class AB so that you have more room on peaks instead of early clipping.
An interesting option, which I like but can't run Class AB (only A), is the JLH 1969 Class A 10W.
Also look at Look up Hiraga Class A 30W, or PASS F5.
- Status
- Not open for further replies.