Hello,

I think it is kind of odd to have "class A" and "pull-push" in the same
sentence. Can someone elaborate on this?

"Class A" means the output device is always conducting, and one device
is enough (right?). "Pull-push" is applicable to class B or class AB,
where the devices are conducting in half cycles or a little more.

Roland

In article ,
Roland wrote:
Hello,

I think it is kind of odd to have "class A" and "pull-push" in the same
sentence. Can someone elaborate on this?

"Class A" means the output device is always conducting, and one device
is enough (right?). "Pull-push" is applicable to class B or class AB,
where the devices are conducting in half cycles or a little more.

Roland

Class A means that the output devices are always conducting. This can be
done single-ended or push-pull. Single-ended designs are horribly inefficient
(up to 12.5% when a resistor is used between output and the negative rail,
25% with a current source) compared to push-pull (up to 50%). Push-pull
can cancel even ordered harmonic distortions.

Class B implies push-pull. When one device is on the other is off - each
is only used for half the waveform. Non-linearities as the devices approach
shut-off make this non-viable for hi-fi.

Class AB also implies push-pull. There is some overlap between the
devices so that one device is providing substantial output before the
other shuts off and cross-over distortion is minimized.

In tube designs A1/AB1 indicate the grid is always negative w.r.t. the
cathode so no current flows, A2/AB2 the opposite.

--
a href="http://www.poohsticks.org/drew/"Home Page/a
Life is a terminal sexually transmitted disease.

In article ,
Drew Eckhardt wrote:

Class B implies push-pull. When one device is on the other is off - each
is only used for half the waveform. Non-linearities as the devices approach
shut-off make this non-viable for hi-fi.

For audio, this is true. Some types of RF amplifier run Class B (or
even Class C) single-sided, and depend on the output tank circuitry to
filter out the resulting distortion.

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

"Roland" in om...
Hello,

I think it is kind of odd to have "class A" and "pull-push"
in the same sentence. Can someone elaborate on this?


Actually, it is a valuable technique used currently here and there,
including in high-performance signal-path analog ICs, some of them related
to audio (many of them not so, but still concerned with some of the same
distortion issues). Ah, if only I had the time to elaborate. Time is of
the essence, alas.

For one concrete example of where the technique pays: if your basic class-A
amplifying element is real (nonideal) and exhibits any kind of even-order
distortion (where the distortion components behave like, for instance, a
square or fourth power of the desired signal), then class-A push-pull tends
configurationally to cancel these distortion components. Two similar
amplifying elements are driven with anti-phase inputs and then their outputs
subtracted. Even-order distortion components subtract out to zero. (This
principle by the way is not limited to memoryless, or instantaneous,
nonlinearities; it embraces also the "even" Volterra kernels that
characterize lumped frequency-dependent nonlinearities.)

The technique was concretely useful in recent years in CMOS and BiCMOS
analog IC design work. Time is, again, of the essence, but that is the gist
of it. (If you have experience with my occasional postings in the past 20+
years you may have faith of documentary support for assertions here once
again, but I am badly pressed right now. Anyway anybody can go learn about
such things who is willing to do the work.)

-- Max

Max Hauser wrote:

"Roland" in om...
Hello,

I think it is kind of odd to have "class A" and "pull-push"
in the same sentence. Can someone elaborate on this?


Actually, it is a valuable technique used currently here and there,
including in high-performance signal-path analog ICs, some of them related
to audio (many of them not so, but still concerned with some of the same
distortion issues). Ah, if only I had the time to elaborate. Time is of
the essence, alas.

For one concrete example of where the technique pays: if your basic class-A
amplifying element is real (nonideal) and exhibits any kind of even-order
distortion (where the distortion components behave like, for instance, a
square or fourth power of the desired signal), then class-A push-pull tends
configurationally to cancel these distortion components. Two similar
amplifying elements are driven with anti-phase inputs and then their outputs
subtracted. Even-order distortion components subtract out to zero. (This
principle by the way is not limited to memoryless, or instantaneous,
nonlinearities; it embraces also the "even" Volterra kernels that
characterize lumped frequency-dependent nonlinearities.)

The technique was concretely useful in recent years in CMOS and BiCMOS
analog IC design work. Time is, again, of the essence, but that is the gist
of it. (If you have experience with my occasional postings in the past 20+
years you may have faith of documentary support for assertions here once
again, but I am badly pressed right now. Anyway anybody can go learn about
such things who is willing to do the work.)

-- Max



Would the "bridging" of two class A mono amps (necessitating the
inversion of signal on one amp and sharing of common rails, etc.)
be of a class A push-pull configuration? If so, would it produce
a similar effect in reducing even-order distortion?

- Jeff

On Fri, 09 Jul 2004 15:05:22 GMT, Jeff Wiseman
wrote:



Max Hauser wrote:

"Roland" in om...
Hello,

I think it is kind of odd to have "class A" and "pull-push"
in the same sentence. Can someone elaborate on this?


Actually, it is a valuable technique used currently here and there,
including in high-performance signal-path analog ICs, some of them related
to audio (many of them not so, but still concerned with some of the same
distortion issues). Ah, if only I had the time to elaborate. Time is of
the essence, alas.

For one concrete example of where the technique pays: if your basic class-A
amplifying element is real (nonideal) and exhibits any kind of even-order
distortion (where the distortion components behave like, for instance, a
square or fourth power of the desired signal), then class-A push-pull tends
configurationally to cancel these distortion components. Two similar
amplifying elements are driven with anti-phase inputs and then their outputs
subtracted. Even-order distortion components subtract out to zero. (This
principle by the way is not limited to memoryless, or instantaneous,
nonlinearities; it embraces also the "even" Volterra kernels that
characterize lumped frequency-dependent nonlinearities.)

The technique was concretely useful in recent years in CMOS and BiCMOS
analog IC design work. Time is, again, of the essence, but that is the gist
of it. (If you have experience with my occasional postings in the past 20+
years you may have faith of documentary support for assertions here once
again, but I am badly pressed right now. Anyway anybody can go learn about
such things who is willing to do the work.)

-- Max

Would the "bridging" of two class A mono amps (necessitating the
inversion of signal on one amp and sharing of common rails, etc.)
be of a class A push-pull configuration? If so, would it produce
a similar effect in reducing even-order distortion?

You are confusing Class A with single-ended. Bridging a pair of
single-ended amps would be a futile exercise, as they are already
deficient in load-driving ability. OTOH, it would indeed reduce
second-order distortion, if the pair were well-matched. Much easier
(and much cheaper) to build a Class A push-pull amp from scratch!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Roland wrote:
Hello,

I think it is kind of odd to have "class A" and "pull-push" in the
same sentence. Can someone elaborate on this?

"Class A" means the output device is always conducting, and one device
is enough (right?). "Pull-push" is applicable to class B or class AB,
where the devices are conducting in half cycles or a little more.


No. Push-pull indicates a bipolar power supply and output devices in pairs
attached to each polarity supply. They can be happily biased into class A
where they are boh conducting all the time.

Otherwise with Class A we'd be stuck with output coupling capacitors or
transformers, which are not desirable.

geoff

Geoff Wood wrote:

Roland wrote:

Hello,

I think it is kind of odd to have "class A" and "pull-push" in the
same sentence. Can someone elaborate on this?

"Class A" means the output device is always conducting, and one device
is enough (right?). "Pull-push" is applicable to class B or class AB,
where the devices are conducting in half cycles or a little more.



No. Push-pull indicates a bipolar power supply and output devices in pairs
attached to each polarity supply. They can be happily biased into class A
where they are boh conducting all the time.

Otherwise with Class A we'd be stuck with output coupling capacitors or
transformers, which are not desirable.

geoff


I think you probably understand better than you have stated. Push-pull
certainly does NOT require a bipolar power supply. You hint at why in
your second paragraph.

--
The e-mail address in our reply-to line is reversed in an attempt to
minimize spam. Our true address is of the form .

Stewart Pinkerton wrote:

On Fri, 09 Jul 2004 15:05:22 GMT, Jeff Wiseman
wrote:
stuff deleted

Would the "bridging" of two class A mono amps (necessitating the
inversion of signal on one amp and sharing of common rails, etc.)
be of a class A push-pull configuration? If so, would it produce
a similar effect in reducing even-order distortion?

You are confusing Class A with single-ended. Bridging a pair of
single-ended amps would be a futile exercise, as they are already
deficient in load-driving ability. OTOH, it would indeed reduce
second-order distortion, if the pair were well-matched. Much easier
(and much cheaper) to build a Class A push-pull amp from scratch!


I guess the thing that kinda intrigued me was the potential of
taking a bridgable stereo amp and having it's harmonic
distortions improve by bridging it into a mono amp.

Just thinking...
:-)

- Jeff

"CJT" wrote in message
...
I think you probably understand better than you have stated. Push-pull
certainly does NOT require a bipolar power supply. You hint at why in
your second paragraph.

Yes, especially when you consider the term was invented in the valve era.

TonyP.

A lot of solid state amps had single ended power supplies and used a
coupling cap. McIntosh uses a bipolar supply and an autoformer in
their solid state amps and with good results-I wish someone would
build such a solid state amp with more modern SS design. Mac is
characterized by old design and low build cost, less than most people
think.

In article ,
(Sam Byrams) wrote:

A lot of solid state amps had single ended power supplies and used a
coupling cap. McIntosh uses a bipolar supply and an autoformer in
their solid state amps and with good results-I wish someone would
build such a solid state amp with more modern SS design. Mac is
characterized by old design and low build cost, less than most people
think.

Rat Shack/Optimus used autoformers in some of their car stereos to break
the 20W barrier without a switching power supply. It worked great as
long as there were no low frequencies to saturate the autoformers.

TonyP wrote:
"CJT" wrote in message
...
I think you probably understand better than you have stated.
Push-pull certainly does NOT require a bipolar power supply. You
hint at why in your second paragraph.

Yes, especially when you consider the term was invented in the valve
era.

Bipolar refers to pos and neg wrt ground. Nothing to do with bipolar
transistors.

I cannot imagine a reason for creating a push-pull circuit without a biplor
supply. All that biasing you would need to get the operating point near
the centre, for what benefit ?


geoff

On Sat, 17 Jul 2004 18:40:23 +1200, "Geoff Wood"
-nospam wrote:

TonyP wrote:
"CJT" wrote in message
...
I think you probably understand better than you have stated.
Push-pull certainly does NOT require a bipolar power supply. You
hint at why in your second paragraph.

Yes, especially when you consider the term was invented in the valve
era.

Bipolar refers to pos and neg wrt ground. Nothing to do with bipolar
transistors.

I cannot imagine a reason for creating a push-pull circuit without a biplor
supply. All that biasing you would need to get the operating point near
the centre, for what benefit ?

The benefit of having reduced even harmonic distortion, of course. I'm
not aware of *any* push-pull valve amp which goes to the considerable
expense of having two power supplies.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Stewart Pinkerton wrote:
On Sat, 17 Jul 2004 18:40:23 +1200, "Geoff Wood"
-nospam wrote:


TonyP wrote:

"CJT" wrote in message
...

I think you probably understand better than you have stated.
Push-pull certainly does NOT require a bipolar power supply. You
hint at why in your second paragraph.

Yes, especially when you consider the term was invented in the valve
era.

Bipolar refers to pos and neg wrt ground. Nothing to do with bipolar
transistors.

I cannot imagine a reason for creating a push-pull circuit without a biplor
supply. All that biasing you would need to get the operating point near
the centre, for what benefit ?


The benefit of having reduced even harmonic distortion, of course. I'm
not aware of *any* push-pull valve amp which goes to the considerable
expense of having two power supplies.

I believe some oscilloscope amplifiers did, but I'm not aware of any
consumer audio amplifiers that did, and I would be especially surprised
to find any consumer audio power amplifiers that did.

--
The e-mail address in our reply-to line is reversed in an attempt to
minimize spam. Our true address is of the form .

"Geoff Wood" -nospam wrote in message
...
TonyP wrote:
Yes, especially when you consider the term was invented in the valve
era.

Bipolar refers to pos and neg wrt ground. Nothing to do with bipolar
transistors.

Yep, and the number of valve amps with bipolar power supplies is very small
indeed.

I cannot imagine a reason for creating a push-pull circuit without a
biplor
supply. All that biasing you would need to get the operating point near
the centre, for what benefit ?

Apart from PP valve amps, quasi complimentary transistor amps were common in
the sixties.

TonyP.

"Geoff Wood" -nospam wrote in message

TonyP wrote:
"CJT" wrote in message
...
I think you probably understand better than you have stated.
Push-pull certainly does NOT require a bipolar power supply. You
hint at why in your second paragraph.

Yes, especially when you consider the term was invented in the valve
era.

Bipolar refers to pos and neg wrt ground. Nothing to do with bipolar
transistors.

I cannot imagine a reason for creating a push-pull circuit without a
biplor supply.

It was done all the time, through the late 1960's

The two popular work-arounds we

(1) Output transformer
(2) Output coupling capacitor

All that biasing you would need to get the operating point near the
centre, for what benefit?

There really isn't much to the biasing. Two resistors and a capacitor.

Single supplies are used to avoid the expense of the second power supply. A
lot of first and second generation SS power amps had single-ended power
supplies and output coupling capacitors. I had a Heath AR-15 that had a
single supply and output coupling caps.

Single-ended supplies are still widely used in portable applications,
including car radios. One battery!

In the end people realized that with an AC-powered single supply you needed
1 big cap for the power supply and 2 smaller ones for the output coupling
capacitors, while with split supplies you only needed two smaller ones (one
for each power supply).

Stewart Pinkerton wrote:
On Sat, 17 Jul 2004 18:40:23 +1200, "Geoff Wood"

The benefit of having reduced even harmonic distortion, of course. I'm
not aware of *any* push-pull valve amp which goes to the considerable
expense of having two power supplies.

...l.and in the transistor case, then putting that reduced harmonic
distprtion signal through a bloody great electrolytic capaciotr ?!!!


geoff

Arny Krueger wrote:
"Geoff Wood" -nospam wrote in message

Single supplies are used to avoid the expense of the second power
supply. A lot of first and second generation SS power amps had
single-ended power supplies and output coupling capacitors. I had a
Heath AR-15 that had a single supply and output coupling caps.

A centre-tapped transformer (versus single sec) and two capacitors (instead
of one)expensive ? OK, a higher voltage-rated bridge too....

Single-ended supplies are still widely used in portable applications,
including car radios. One battery!

In the end people realized that with an AC-powered single supply you
needed 1 big cap for the power supply and 2 smaller ones for the
output coupling capacitors, while with split supplies you only needed
two smaller ones (one for each power supply).

How did it take so long to figure that out ?

geoff

"Geoff Wood" -nospam wrote in message
...
Stewart Pinkerton wrote:
On Sat, 17 Jul 2004 18:40:23 +1200, "Geoff Wood"

The benefit of having reduced even harmonic distortion, of course. I'm
not aware of *any* push-pull valve amp which goes to the considerable
expense of having two power supplies.

..l.and in the transistor case, then putting that reduced harmonic
distprtion signal through a bloody great electrolytic capaciotr ?!!!

Having dual power supplies eliminates the need for a blocking capacitor in a
transistor amp. The tube amps used a single power supply and a center-tapped
output transformer.

"Geoff Wood" -nospam wrote in message

Arny Krueger wrote:
"Geoff Wood" -nospam wrote in message

Single supplies are used to avoid the expense of the second power
supply. A lot of first and second generation SS power amps had
single-ended power supplies and output coupling capacitors. I had a
Heath AR-15 that had a single supply and output coupling caps.

A centre-tapped transformer (versus single sec) and two capacitors
(instead of one)expensive ? OK, a higher voltage-rated bridge too....

Single-ended supplies are still widely used in portable
applications, including car radios. One battery!

In the end people realized that with an AC-powered single supply you
needed 1 big cap for the power supply and 2 smaller ones for the
output coupling capacitors, while with split supplies you only needed
two smaller ones (one for each power supply).

How did it take so long to figure that out ?

That's really a good question. The first SS amps came out around 1963
(Acoustech 1), while the AR15 came out in 1967. Both had single polarity
power supplies, and output coupling caps By 1970 split supplies were pretty
much the rule, and output coupling caps were pretty much gone.

I suspect that the output coupling caps were chosen to have a secondary
function as speaker protectors. If the output stage blew, its output
terminal would often be effectively shorted to the output of the power
supply. If there was no speaker fuse, that was usually the end of the
woofer, as well. As the power supply for a modest 50 wpc amp was about 80
volts, there was some chance that crossover caps in series with upper range
drivers would be overwhelmed by the high voltage. They might go, taking the
upper range drivers with them.

The early SS amps were pretty fragile and likely to have serious problems.

On Sun, 18 Jul 2004 11:22:24 +1200, "Geoff Wood"
-nospam wrote:

Stewart Pinkerton wrote:
On Sat, 17 Jul 2004 18:40:23 +1200, "Geoff Wood"

The benefit of having reduced even harmonic distortion, of course. I'm
not aware of *any* push-pull valve amp which goes to the considerable
expense of having two power supplies.

..l.and in the transistor case, then putting that reduced harmonic
distprtion signal through a bloody great electrolytic capaciotr ?!!!

Indeed - which will cause only a *tiny* amount of distortion - despite
what you might have read from some overenthusiastic 'high enders' like
Walt Jung! Note also that McIntosh produce single-supply push-pull
transistor amps with no output capacitor. Of course, they use the much
*worse* solution of a transformer!! Has to be better (and cheaper!) to
use a bipolar supply in that case...................
--

Stewart Pinkerton | Music is Art - Audio is Engineering