Home » Audio » Thermionic Emissions » Class A, AB1, B, C Operation/Modes (How Class A, AB1, B, C Modes Work and Their Strengths and Weaknesses)
|Class A, AB1, B, C Operation/Modes [message #96335]
||Sun, 05 February 2023 20:40
Registered: May 2020
How Class A, AB1, B, and C Operation Work and Differences|
I thought it would be informative to discuss the differences between
Class A, AB1, B, and C operation. By doing such, one will more fully
understand how each component in your system operates.
Knowledge is power and the more you understand, the less chance
of being misinformed. I am going to keep this discussion as simple as
possible for our newbie friends. I will not cover every detail nor every proof.
Caveat: Lets leave out transformers from our discussion.)
Note: It might be good idea to print out all the figures at the bottom
of this post to examine while reading.
So let's get started.
What is a sine wave? A sine wave is:
"a curve representing periodic oscillations of constant amplitude as
given by a sine function. Also called sinusoid."
The sine wave is a constantly varying voltage. Figure 1 has a
pictorial of a sine wave, the wavy line. 120 vac at the wall outlet is a
sine wave, and voltage.
So is music made up of sine waves? The answer is yes.
Although looking at a musical signal with an oscilloscope might look
haphazard, with sharp peaks, those sharp peaks are simply very
high frequencies. Even a solo instrument's signal might look haphazard
due to natural harmonics from the instrument.
It will be easier to understand the different classes of operation if we use
a single frequency sine wave as pictured in Figure 1. The entire input
signal wavy line is a complete sine wave, 360 degrees. Half of a sine
wave is 180 degrees. One fourth of a sine wave is 90 degrees, one eighth
of a sine wave is 45 degrees etc.
Class A operation.
Suppose we have a single vacuum tube and we have it just drawing
current (idle current, Point Q of fig. 1) with no signal present. Now we apply
the input signal to the tube's grid and the output appears as X and Y
output in fig. 1. Notice X and Y look the same as the input sine wave
So what happened? The input voltage applied to the tube grid
controls the current flowing through the tube. In Class A, the current
flows through the tube all the time, the entire sine wave input signal,
360 degrees. That is very good. Again, that is also the classic definition of
Class A operation, or mode.
Virtually all phono stages, pre-amplifiers, input and phase
splitters in amplifiers are operated Class A. The following
tube stage presents a fairly constant load. That is good news.
Let us continue for tubes operated in Class AB1, B, and C.
Will all operations work in linear audio applications?
Class AB1, B, and C are defined as operating a single tube when the
current through the tube can be stopped, cut off, meaning 0 ma.
(ma is milliamps, or thousandths of an ampere.) So what is the
difference between AB1, B, and C operations?
First, we need to see something significant in fig 1, Class A operation.
It has to do with the tube's idling current in fig 1, the Q point, which is
set to 65 ma, half way between 0 ma and maximum 130 ma. in our
example. Notice we can go 65 ma. to 0 ma. and 65 ma to 130 ma.
Above and below are equal. So X and Y are equal output and mimic
the input signal. This current variation allows the tube to remain
conducting current the entire input sine wave voltage cycle, 360 degrees.
Again, this is Class A operation. Understanding Class A operation allows
us to understand Class AB1, B, and C operation more fully.
Let's bypass fig 2, AB1, for now.
Let's jump to fig 3, Class B operation/mode. Notice Q point is different.
It is not 65 ma but now 0 ma idling. We still have the same exact value
input signal, but only X appears at the output, Y being absent.
Only half the input signal is at the output. What happened?
Q point is set at 0 ma. As the signal goes positive,
more current flows through the tube, so X output appears.
However, how can we go less than 0 ma. current as the input signal
voltage goes negative? We cannot. Thus no Y output signal voltage.
Only ½ of the input signal appears at the output (180 degrees).
This is a classic example, definition of Class B operation.
Class B presents severe distortion to the input signal, and is generally
used in RF and industry. It can be used in audio if we go Push Pull, but it will produce crossover distortion, higher distortion in general, so is mostly used in PA systems where fidelity is not important.
Fig. 2, AB1 operation is between A and B, fig. 1 and fig. 3 respectively.
Let us check out fig. 2, AB1 operation. Once again we have our input
signal sine wave, and X and Y output voltage. However, we have only
some Y output sine wave signal present. Notice, however, the tube's idle
current, Q is between our Class A and Class B Q points, 65 ma and 0 ma
In our AB1 example, the idle current is set to 55 ma. Ok, as the input
signal is increased from no signal, X and Y output rise equally, Class A operation, until the negative input signal causes the tube current to
reach 0 ma. At that point the tube cannot go less than 0 ma current, so Y signal cannot continue to follow the negative input signal.
So what good is it if X becomes larger while Y? What about adding
a second output tube which mirrors the first tube, except it
handles the negative portion of the input signal, increasing in
current as the signal goes more negative. Then X and Y output sine wave
mirrors the input sine wave signal. They naturally blend together.
That is called Push Pull.
So is there any advantage in designing Push Pull?
IF designed properly, efficiency is much higher than class A,
much more power output with the same or less distortion. One
can also eliminate the inherent negatives of a class A output stage.
See below *.
However, a push pull stage is much more difficult to design.
But the nice thing in AB1 mode is that both output tubes operate in
Class A mode at the same time until each output tube reaches its 0 ma
For example, a 6L6GC, beam power tube in AB1 mode can produce
55 watts rms output in Class AB1 operation. However, both output
tubes are operating at least 15 watts in Class A mode before sliding
into AB1 mode. That is conservative ratings.
In triode mode, we can figure half the power output of beam power mode,
so at least 7.5 watts Class A operation of both tubes before sliding into AB1
Even at 1 watt Class A output, a typical speaker can at least peak
into the mid 80s spl, depending upon the efficiency of one's speakers.
And the harmonic distortion is extremely low. My whole KT88 amp
produces only 0,05% at 1 watt output, with no global negative feedback.
Ok, we have discussed Class A, AB1, and B operation. Let us check out
fig. 4, Class C operation.
The first thing one notices is that Q idle is below 0 ma. How can that be?
Notice the perforated line to Q. What is actually pictured is the grid bias
is so negative that less than half, in fact, a very small portion of the input
signal is even large enough to cause current flow through the tube. Thus X
appears to be small and Y does not exist at all. A larger, huge input signal
must be present to obtain lots of power output in Class C mode.
The plus is that the efficiency can reach 80%, but the minus is that
the distortion is gigantic. Class C operation is usually used in radio
frequencies (RF) and Industrial applications.
So what have we learned?
A. Class A is used in virtually all small signal applications since the load is relatively constant.
B. Class AB1 Push Pull and A are used in most output applications.
C. Class B is used as Push Pull, almost exclusively for PA systems.
D. Class C is never used in linear analog audio designs.
E. There is a smooth blending in properly designed Push Pull stages.
F. In Class AB1, both output tubes X and Y run Class A until each
tube reaches 0 ma. on positive and negative peaks of the
sine wave cycle.
G. The output impedance/damping factor remains virtually constant over
the entire sine wave with Push Pull. Class A single ended amplifiers
are a different story. See below *.
H. There is no signal gap between output tubes, nor crossover distortion
until approaching Class B mode/operation.
I. 120 hz power supply hum is mostly cancelled.
* For a single output tube amplifier, different considerations apply.
For instance, we want the amplifier's output impedance (Z) to be
constant with varying power output and over the entire signal cycle,
360 degrees. To accomplish this, the tube's plate resistance (Ra) must
However, Fig. 5 shows the Ra line of a typical single
ended triode tube varies/curves substantially as the current changes.
Of course as the current changes, the output power also changes.
At peak power output, the damping factor varies from
maximum damping of that SET amplifier design
to virtually no damping. Pull remains virtually constant
under the same conditions.
There are other pros and cons that we might discuss later.
I hope this has helped in understanding how Class A, AB1,
B, and C work.
One can check out:
RCA Tube Manual
RCA Radiotron Designers Handbook
Semiconductor and Tube Electronics by James G Brazee
|Re: Class A, AB1, B, C Operation/Modes [message #96340 is a reply to message #96338]
||Tue, 07 February 2023 16:13
Registered: May 2020
I run mine Class AB1, and triode for years. Some day will try UL. |
At the volume I listen it is mostly if not all Class A mode as you Go.
At one watt output, the amplifier distortion is ~0,05%, so still quite
I have been running sophisticated listening testing and found my mono
blocks do not alter the sound, when no load. However, my load is a
variable cone type speaker so what do I do?
The trick is to match the amp to speaker with the correct total gauge
speaker wire. I don't worry too much about self inductance.
This is fun,,, ya right. Each system will be different.
It takes a lot of time, so beware.
As it turns out, I am running all copper, 10 strands of 18 gauge, 6 feet
long in parallel for each leg to the speaker, adn the other speaker.
Total gauge is ~9.2 gauge if I remember correctly. Yours will vary.
If I run 9 strands or 11 strands in one leg, the sound is not optimumal
in my design. It sounds either too thin or too dull.
Anyone can start testing in their own system for optimum sound.
I would start with hardware store doorbell wire as a starter.
I used double wire in jacket.
Make sure the wire/cable length to one speaker matches the leg length
in the other speaker. Otherwise you may have to add or subtract
a strand for optimum matching of both speakers. Self inductance will
also be different, but maybe low enough to not matter.
Replacing a strand of regular wire with Jenalabs wire will certainly
alter the sound. Right now, I am using one strand of Jenalabs
18 gauge wire in each leg with nine regular strands of hardware store
wire (99.9% pure).
(Jenalabs wuite is 6N pure, 99.9999% pure, and expensive.)
|Re: Class A, AB1, B, C Operation/Modes [message #96344 is a reply to message #96343]
||Tue, 07 February 2023 20:37
Registered: May 2020
gofar99 wrote on Tue, 07 February 2023 19:33|
Hi, All my amps are U/L. Nearly no global NFB (just 3 db for stability at way above band possible resonanaces). And yes indeed...everything matters. It took a long time to figure out how to best use my Martin Logan ESLs. They tend to be troublesome loads for some amps (none of mine though) and placement is critical.
Yes, I have heard Martin Logan's are a difficult load. i take it the
impedance varies wildly?
I agree Bruce, everything matters.
I am not running any global negative feedback, just a very small amount
of cathode/current feedback for the output tubes.
To help match woofer to full range driver, I am currenty using a 15"
piece of 18 gauge wire. I have 6" available, or no extra wire, or
other custom lengths if necessary.
You are so right concerning speaker placement.
Nice post Bruce. Appreciate your knowledge and input.
I guess back to Class A, AB1, B, and C amps.
Cheers and all the best Gents.
|Re: Class A, AB1, B, C Operation/Modes [message #96354 is a reply to message #96353]
||Thu, 09 February 2023 21:35
Registered: May 2020
gofar99 wrote on Thu, 09 February 2023 21:01|
Hi, MLs act like huge capacitors. They go from just over 4 ohms at low frequencies to 1 ohm at 20K. The slope of impedance can give many amps fits. It was why I added 3 db of frequency limited NFB to my amps. I figured they were about as tough a load as anything anyone would use. (some crossovers might be worse though) I have not had any misbehave without it...but testing shows a strong resonance point in most of the amps at about 70-85KHZ. So I start the slope at about 25-30K and it insures stability no matter what the amp sees on the output side. I have tested all of the various sizes with and without the NFB and have not been able to get any to mess up...still a couple of parts is cheap insurance. In the commercial versions and shown on the diy schematics I show NFB defeat switches and most folks can only say that the use of NFB cuts the gain by a few db. No surprise there...but in a blind test can't tell which one is which.
Yes, that is quite a change in load.
After discussing the matter with a Harvard Medical School chair years ago, one item I think we need to be cautious of is blind testing.
There are many confound variables that need to be addressed. Otherwise
the test will always be skewed toward no sonic difference. It then basically becomes a rigged test.
I used to test every day for weeks, months or longer, but addressing confounds. Now I have the ability to hear the change quite easily. I still test every day after a tweak is made, just to be sure though.
cheers and all the best.
|Re: Class A, AB1, B, C Operation/Modes [message #96446 is a reply to message #96354]
||Tue, 28 February 2023 19:04
Registered: May 2020
Something else to consider.|
First, pin 1 in a power cord is the ground wire, the 3rd prong of a
power plug. See Fig. 1 below.
I don't know if this has been mentioned earlier or in another post,
but when two or more components in a system have pin 1 connected to
the same outlet "terminal", there is a connection between the signal
ground of the components involved. See Pin1.jpg
In any case, musical information/signal current not only returns via
both left and right interconnect shields, but also through the pin 1
power cord ground wires, from component to component signal grounds.
This mixes the channels together to some extent, and is frequency
sensitive. There are all sorts of negative ramifications to the
musical parameters, such as sound stage, dynamics, frequency response
etc. (I know, the resistances and inductances seem small but I am
testing 1 part per million in my speaker crossovers, so it does
matter to some extent.)
As above, the mixing is non linear since we have two factors to
consider, resistance and inductance of the interconnect cable(ic)
shields and pin 1 power cord wires.
The ratio of the shield resistance to pin 1 resistance will not be
the same as the ratio of the shield inductance to pin 1 wire
There are solutions, but please be careful if/when implementing them.
1. Only have one component with pin 1 connected to ground. This
requires connecting all ics before plugging in any AC power plugs.
I do not accept any responsibility. You perform this
at your own risk of injury.
2. There is a second method, but I do not accept any
responsibility. You perform this at your own risk of injury.
It is installing multiple resistors, each high power, very
low ohmage between pin 1 and the component. The preamplifier is
the logical choice since the AC current draw is low, the rated
fuse is low.
(Amplifiers are higher current with higher amperage fuses,
so I would not install any resistors in one.
Do so at your own risk of injury.)
For instance, 3 twelve watt resistors in parallel, each resistor
4.5 ohms would result in 1.5 ohm total. The fuse should easily blow,
the resistor combo will be 30 watts rated. Even if one or two
resistors open, the fuse should easily blow first.
I cannot state this enough. Please be careful. I do not
accept any responsibility for any accidents or injuries.
Current Time: Fri Dec 08 00:27:06 CST 2023