"Class A" nonsense

[jmaf]

Originally Posted by Robert M. Martinelli

2 - In a true "class A" amp the current swing is almost non-existent ( current through the output tube(s) is almost constant ), while in an AC30 the current swing through the output tubes is significant and very visible.

Robert, I think your entire post was right, up until that point. Voltage swing is a consequence of tube current flowing through the output transformer. If you have zero current swing then you have a silent amp, no?

[hasserl]

Originally Posted by Old Tele man

...ah, not really true, because almost ALL musical instrument amplifiers are rated with around 5% THD because of the "nature of the (rock-n-roll) beast" and the fact that 5% THD is the amount that is "...barely perceptible..." to most listeners.

...so, the wording "...maximum undistored power output..." cannot really apply to musical amplifiers, which is our subject. They (manufacturers) know there's 5% THD in the mix, which lets them: (a) claim slightly more output power and (b) account for the manner in which most (but certainly not all) amplifiers are actually used and operated.

...(tongue-in-cheek observation: ah, but that is a very good kind of 5% THD in the hands of most blues players!)

But just because manufacturers fudge on the power rating doesn't justify changing the description.

Originally Posted by Enzo

And by that undistorted standard, there would be plenty of guitar amps that would have to have a zero watt rating, as there is no level at which the signal is undistorted.

I'm not aware of any, even the mighty 5150 has some level of undistorted output. It may generate distortion in the preamp fairly early, but in that case use the clean channel.

But that's a good point, how do you set the power rating of an amp that distorts the signal purposely in the preamp? Should it be the max undistorted power output as measured at the power amp only?

[Rhodesplyr]

Originally Posted by hasserl

But that's a good point, how do you set the power rating of an amp that distorts the signal purposely in the preamp? Should it be the max undistorted power output as measured at the power amp only?

I'm not sure even maximum undistorted power output works for me, though I do think the definition should be generally confined to the output stage.

Given something like two output tubes in a standard push-pull arrangement, the circuit would be rated for a certain amount of power output at the point where both tubes were still conducting, but one was very close to cutoff, while the other was drawing almost 100% of the B+. The reason I say that is because I think one of the advantages of Class A is that it makes minimal instantaneous current demands on the power supply. However, when one tube starts drawing more current than both tubes at idle, the operating points begin to shift, and demands are made on the energy storage of the power supply.

Take, for example, the Leslie 32H amp. Stock, it has four 5881s in parallel push-pull, and all share one 75 Ohm cathode resistor--with no cathode bypass capacitor. Up to a certain point, there is little/no AC on the cathodes because when one side's going positive, the other side is going negative. The sum is zero--assuming matched tubes....

However, when it crosses into AB1, suddenly, the presence or absence of a cathode bypass capacitor becomes significant because it tends to stabilize the cathode voltage and reduce the appearance of negative feedback across the cathode resistor. In fact, I sometimes add a cathode bypass capacitor to this amp, and by choosing the value carefully, I can tune how it behaves at that point. By making the cathode bypass cap (& the B+ storage) larger, I can extend the clean power out of the amp--to a point.

A "Class A" amp, to my way of thinking, shouldn't cross that line where a cathode bypass capacitor is needed (in push-pull), but as others have pointed out, it would be tricky to impose some kind of brick wall to prevent a circuit designed for Class A operation being driven beyond that.

I think understanding what happens when an amp moves from Class A to AB1 is important for techs and designers, but not to players out to buy an amp. If the amp is a head, I'd rather see that it has multiple output taps for different speaker impedances than the claim that it's "pure Class A."

[Robert M. Martinelli]

@ Chuck

Some 25 W RMS clean ( undistorted ), some 34 W when fully "cranked".

@ Steve

I have re-read my posts and can't find anything wrong, ( with the exception of my English, still in need to be improved ) the way I defined the class and the operating conditions of my AC30 seems OK to me, the output tubes are on ( signal-wise ) for more than 180 degrees ( Pi ) but appreciably less than 2*Pi, and there is a significant current swing ( unlike in class A amps ) both these facts place the AC30 in "AB1 territory" IMHO.

BTW, congratulations on your sight/soldering ability! - At the moment I have problems in properly soldering 2N3055s, let alone something smaller....

Cheers

Bob

[Robert M. Martinelli]

Originally Posted by jmaf

Robert, I think your entire post was right, up until that point. Voltage swing is a consequence of tube current flowing through the output transformer. If you have zero current swing then you have a silent amp, no?

Hi jmaf,
Maybe due to my bad English, I expressed myself in a wrong manner....I meant that in a class A amp the current draw does not change ( only ideally, it changes a bit in reality ) between "idle" and "working" conditions.

In class "AB1" the current swing is way more significant and always very evident.

My apologies, I'm working hard to improve my English, but it seems I'm miserably failing.....

Cheers

Bob

[Merlinb]

Originally Posted by balijukka

this requires that the output is not clipped, won't it.

jukka

Yep, and that's the only universally logical way to define class of operation.

[balijukka]

Originally Posted by Merlinb

Yep, and that's the only universally logical way to define class of operation.

That's the way my teachers taught me, and how I thought for many years.
If it clips, it's a clipper, not an amplifier.

But since they (damned buggers) dug up that IRE thing, and then I found the same in Reich's book, and then I even found this one from one site. :

Electrical Engineering Glossary Definition for Class A
Glossary Term: Class A

Definition

The simplest type of amplifier, class A amplifiers are those in which the output transistors conduct (i.e. do not fully turn off) irrespective of the output signal waveform. This type of amplifier is typically associated with high linearity but low efficiency.


I don't mind anymore if someone uses hazy terminology as long as I understand what they mean. And if I don't, I just ask clarification.

IMHO Marketing people use whatever they like if it sells. The dumb ones are the ones that buy. Unfortunately geetarr players are very closely related to the bass players who are famous for their rapier like intelligence.

jukka

[Merlinb]

Originally Posted by balijukka

Electrical Engineering Glossary Definition for Class A
Glossary Term: Class A

Definition

The simplest type of amplifier, class A amplifiers are those in which the output transistors conduct (i.e. do not fully turn off) irrespective of the output signal waveform. This type of amplifier is typically associated with high linearity but low efficiency.

That definition has not been thought through, and makes little sense. Just about every class A amp ever made would have to be reclassified as not-class A on the grounds that they could potentially be driven to cut-off. For audio purposes only sine-wave exitation is a universally valid parameter, due to its relationship with power (rms) output.

[hasserl]

Here's a couple of thoughts from Aiken on the topic:

A class A amplifier is defined as one which is biased to a point where plate current in all the output devices flows for the entire 360 degrees of an input cycle, at the full, unclipped output of the amplifier.

(emphasis mine)

The people who say a class AB amp is "class A at lower volumes" are technically wrong, but for the right reasons. If you were to define class A as being only conduction for a full 360 degree phase angle, you would be correct. However, there is more to the definition of amplifier classes than that.

The defining factor in a determining whether or not an amplifier is class A, class AB, or class B *has* to be made at the full output before clipping, otherwise, the class definitions have no meaning whatsoever. It is indeed, a very black and white thing

As you can see, there is indeed such a thing as a "true class AB" amplifier, just as there is a "true class A" amplifier, and the class definitions are not at all ambiguous, except to those who don't understand them, or choose to ignore them for marketing advantage.

One more thing: What if you push the class A or class AB amplifier into clipping? Does it then become a class AB/ B, C, or D amplifier? No, of course not. It is simply the same class amplifier it was to begin with, but driven into clipping. A class A amplifier driven to clipping is still a class A amplifier by definition. This is why amplifier classes are defined the way they are. Otherwise, the class designations would have no meaning. Any amplifier can be driven beyond it's limits into a fully-clipped square wave output (unless it is limited), but that doesn't make it a class D switching amplifier, now does it?

[tekhed7]

Originally Posted by Merlinb

To decide if an amp is class A, you need only ask a simple question:
If a device goes into cut-off, does the output power continue to increase?

If the answer is no then we have class A. This is why single-ended amps can only be class A (or C, which doesn't really apply to us). As soon as it goes into cut-off (or grid-current limiting) the output signal stops increasing, we cannot squeeze any more out of it.

If the answer is yes then we have class AB or B. One device may be in cut-off, but another device continues to provide output power. This is why Class AB or B amp always need at least two devices.

I think many here seem to forget that it's more that an amp OUTPUT STAGE can be "capable" of operating in a certain "Class" rather than BEING a certain class amp.
Any Fender Champ 6V6 will operate in Class A mode when the guitar is mildly strummed, but if you were to push the output STAGE so hard that it's cut off almost 1/2 the time, then it's running nearly Class B.
But since tube Output Stages, unlike transistor ones, don't really have a clean cutoff point, in terms of plate current vs. grid voltage, the boundries get fuzzy.
The point of running Class A vs., say, AB2 is you won't have any ["cross-over"] distortion due to the handoff of the signal between the side that's pushin' & the one that's pullin'. But that gets tough when your often pushing the power supplies to their limit as you push the guitar amp into clipping.
The only way to totally avoid "cross-over"... as I see... it is to go buy an Ovation...
or an old Martin (acoustic).

[Merlinb]

Originally Posted by tekhed7

I think many here seem to forget that it's more that an amp OUTPUT STAGE can be "capable" of operating in a certain "Class" rather than BEING a certain class amp.
Any Fender Champ 6V6 will operate in Class A mode when the guitar is mildly strummed, but if you were to push the output STAGE so hard that it's cut off almost 1/2 the time, then it's running nearly Class B.

The fact that a class A amp can be driven to cut-off does not change its class- it remains class A. If you were to bias a single-ended valve at/near cut-off, it is still class A, but with exceedingly high or infinite input sensitivity (it becomes a half-wave rectifier). A single valve is incapable of delivering any more power once it clips, that is why a single device is always class A by defninition (although some textbooks don't seem to have written their definitions very clearly).

[Steve Conner]

Well, in RF applications, a single-ended amp biased beyond cutoff is called Class-C. The tube only amplifies the positive half-cycles, and the negative ones are filled in by the resonance of the plate tank circuit.

It's highly non-linear (the tube is basically a switch, either cut off or fully on) so is only suitable for FM, or AM with high-level modulation of the plate voltage.

Personally I like to stretch the analogy to audio, and say that a heavily overdriven Champ is running in Class-C too. :-)

In Merlin's example you would find that the supply current draw increased with drive, therefore it isn't Class-A. Unless he were to advertise a rated power of 2 milliwatts or whatever it can make before cutoff.

[Enzo]

hasserl

I'm not aware of any, even the mighty 5150 has some level of undistorted output. It may generate distortion in the preamp fairly early, but in that case use the clean channel.

I admit to a little hyperbole there, yes. Especially since the powr amp only is what gets rated class A for this discussion. But in an amp like the 5150, ther is no clean channel. the schematic calls it a clean channel, but it is only clean by comparison to the other chanels. PV calls it a "rhythm" channel.

[Robert M. Martinelli]

tekhed7 wrote : I think many here seem to forget that it's more that an amp OUTPUT STAGE can be "capable" of operating in a certain "Class" rather than BEING a certain class amp. ( end quote ).

Standards exist for a precise purpose; the standards say that the class of operation has to be determined by the amp's behavior at maximum output power ( thus specifying a precise and well defined measuring point ); according to this statement, it has no sense saying that a certain amp can change its class of operation at different volumes ( output power levels ) IMHO. Once the measuring point has been set, the results have to be referred to that point alone.

Should every manufacturer be left free to set his own standard, the resulting "anarchy" would lead to an even greater confusion ( look what's happened already with the "PMPO" rating, used shamelessly by some "consumer electronics" manufacturers )

I respect every other opinion on the matter, however I'm going to keep on sticking to the standards.

( and, just for the record, I absolutely don't care about the class of operation of an amp - if it sounds good, it's good.... )

Cheers

Bob

[Merlinb]

Originally Posted by Steve Conner

Well, in RF applications, a single-ended amp biased beyond cutoff is called Class-C. The tube only amplifies the positive half-cycles, and the negative ones are filled in by the resonance of the plate tank circuit.

It's highly non-linear (the tube is basically a switch, either cut off or fully on) so is only suitable for FM, or AM with high-level modulation of the plate voltage.

Personally I like to stretch the analogy to audio, and say that a heavily overdriven Champ is running in Class-C too. :-)

Class C requires a tuned load since the 'output signal' is the modulating signal- that's what you're ultimately trying to get out. Just making something a half-wave rectifier doesn't make it class C.

In Merlin's example you would find that the supply current draw increased with drive, therefore it isn't Class-A.

Class has nothing to do with current draw from the power supply, only with current flowing in the device itself. A single ended valve is non-linear and will always increase its average current draw from the power supply, somewhat. This doesn't change its class either. Again we come back to the simple question which answers all this: if the device clips, do we continue to get output signal? Class A = no
Class B = yes
Class C = yes

[Steve Conner]

Well then, by your own admission the overbiased single-ended circuit is Class-B or C. The negative grid excursions will clip early, but the positive excursions will keep on developing more and more power. The supply current will increase, and so will the current in the device and the (extremely distorted) audio output.

This is not just an abstract point: I've seen people build single-ended guitar amps like this. It's what ends up happening if you use a PT that's too high voltage for the job. The plate and screen voltages end up too high, so you have to apply a lot of negative grid bias to stop the tube from overdissipating, and this makes it clip very asymmetrically. The result is a supply current that increases sharply under drive. Would you still consider a bodge job like this to be Class-A? I'd call it Class-AC.

Don't even get me started on Class-OMG and Class-WTF.

[Merlinb]

Originally Posted by Steve Conner

Well then, by your own admission the overbiased single-ended circuit is Class-B or C. The negative grid excursions will clip early, but the positive excursions will keep on developing more and more power.

I think you misunderstand me. When the valve hits cut-off the output signal clips and stops increasing, that's the point I make.
Also, if you drive the positive input peaks to the point of grid-current clipping, again the output signal clips and cannot increase further. The output signal exactly follows the current conduction of the device, for the whole cycle. No change in current = no change in output sugnal.

The result is a supply current that increases sharply under drive. Would you still consider a bodge job like this to be Class-A?

Yes of course. As I said earlier, it simply becomes a single ended, Class A amp which has higher input sensitivity than if it were 'centre biased'.

[Steve Conner]

Originally Posted by Merlinb

I think you misunderstand me. When the valve hits cut-off the output signal clips and stops increasing, that's the point I make.

No, it doesn't. In the case of this cold-biased SE amp, the output signal gets more and more distorted with increasing drive after the negative peaks hit cutoff, but the output power carries on increasing until the positive peaks hit grid current.

[jmaf]

Originally Posted by Robert M. Martinelli

Hi jmaf,
Maybe due to my bad English, I expressed myself in a wrong manner....I meant that in a class A amp the current draw does not change ( only ideally, it changes a bit in reality ) between "idle" and "working" conditions.

In class "AB1" the current swing is way more significant and always very evident.

My apologies, I'm working hard to improve my English, but it seems I'm miserably failing.....

Cheers

Bob

Robert, I lived 1 year in Roma and I can tell you, if you see me try some italian, you'll deserve more than apologies for the assassination of your native language.

From my perspective, as a portuguese speaker, your english is perfect, I'm here to learn from the more senior members and your posts are extremely didactic.

I just asked because I have this mental model of an output transformer and I tried to visualize a constant current draw in class A. I understand your point perfectly now, thanks. Alas, this whole discussion taught me a lot, I guess that's why I lurk on this forum most of the time I can.

[tekhed7]

With all due respect, a Class A amp might not be designed to put out Any real power. It may just be a voltage amp. This whole idea of an Audio amplifier as a Whole being of a certain class is a manufacturer's creation, as far as I'm concerned... standards be damned.
Again, an amp STAGE can only be truely defined to be operating "in a certain class" by virtue of the device's measured operating parameters AT THE TIME OF MEASUREMENT.
As has been pointed out, it came out of the radio world where input signals are far more consistent & lack the wild dynamics of [processed] guitar signals.
Ya got your Class A your AB1, AB2, Class B & C. Anything else... (Class D, G, H...) is all from the imagination of the audio world's sales dept's, & as such is defined & redefined over the decades.
I'm sticking with the ARRL handbook definitions, thankyou very much. But, still, this thread has been very educational.

[balijukka]

Originally Posted by Robert M. Martinelli

Standards exist for a precise purpose; the standards say that the class of operation has to be determined by the amp's behavior [U]at maximum output power[/U

Can you please show us this standard. Who? Where? When?

jukka

[Old Tele man]

Class A = 360° conduction
Class B = 180° conduction
Class C < 180° conduction

...Class AB is anything between Class A (360°) and Class B (180°); suffix "1" denotes no grid current, suffix "2' denotes grid current.

[Robert M. Martinelli]

Originally Posted by balijukka

Can you please show us this standard. Who? Where? When?

jukka

Hi jukka,

EIA, to name one ( cited already by Old Tele Man ), BTW you're missing two "W"s, namely What and Why.

I've tried to explain them already in my previous posts, so I'm not going to try to describe them once more, FWIW I've been re-reading the whole thread and my posts, and I don't feel my statements can be considered "wrong". I just tried to point out that, if we don't want the discussion to go on forever, we first have to agree to take the measurements/state the class of operation at one point, or to refer to some standard, otherwise the whole discussion is pointless IMHO.

Cheers

Bob

[Robert M. Martinelli]

Originally Posted by Old Tele man

Class A = 360° conduction
Class B = 180° conduction
Class C < 180° conduction

...Class AB is anything between Class A (360°) and Class B (180°); suffix "1" denotes no grid current, suffix "2' denotes grid current.

Hi OTM,

As usual, just perfect - concise and straight to the point - I needed almost one page to say the same thing - but I totally agree.
Cheers
Bob

[Chuck H]

OK... HOLY F!SH F#$K ALREADY.

What bias class an amplifier is operating in is a seperate issue from it's indicated bias class of operation because an amp only has to correctly operate within it's indicated bias class at it's indicated output power. (period)

There seem to be a few posters that think an amp that ends up in class AB1 (or even B or C for single ended amps) should not be called class A, and that's just not true. If the amp does function properly in class A up to it's rated output then it's a class A amp. The adjustable input level (or at least pre distortion level) control on a guitar amp is essential. To attenuate this control such that the highest possible input signal could not drive the amp beyond class A would render an amp useless for many instruments and playing conditions that provide a lower input signal. If the amp ends up in AB1, B or C at a higher wattage than is indicated by the MFG then the amp is being operated improperly and it should be expected that it will not perform to it's indicated specs. Not that it's a bad thing to do this, or even unintended by the MFG. But it is secondary to the bias class of the amp.

Chuck

[Robert M. Martinelli]

Originally Posted by Chuck H

OK... HOLY F!SH F#$K ALREADY.

Chuck

Hi Chuck,
sorry if I go "off topic" for a moment, but I take the chance to learn something new ( or to improve my English ) every time I got the chance......can you explain to me the meaning of the above phrase? Is it a "slang" expression?
Thanks
Cheers
Bob

[Merlinb]

Originally Posted by Steve Conner

No, it doesn't. In the case of this cold-biased SE amp, the output signal gets more and more distorted with increasing drive after the negative peaks hit cutoff, but the output power carries on increasing until the positive peaks hit grid current.

You're still misunderstandng me. I'm talking about what the output signal does while the device is actually clipping, on a single half-cycle, not over several cycles.
In class A while the device is in cut-off the output signal clips and stops increasing- it can't do anything else. We can only get change in output signal while the device is conducting, i.e., 360 degrees of a cycle.

Just to make it clearer, in a class AB amp when one device goes into cut-off, the output signal does not clip, because there is another device that takes over and 'fills in the gap' as it were. We can get a continual change in output signal for less than 360 degrees current conduction in any one device.

[Steve Conner]

I'm not misunderstanding. I'm trying to make the point that output power with half a cycle missing is still output power, because "RMS" audio power is measured as a time-averaged quantity that has little to do with the actual instantaneous power at any moment.

You're certainly correct that the instantaneous output power wouldn't increase any more on those half-cycles where the tube was in cutoff. But the power on the positive half-cycles would keep increasing until it was limited by grid current.

Therefore, the supply current would increase, the mean audio power read on a wattmeter would increase, and the signal through a speaker would get subjectively louder. Until in the limit, we'd have a chopper driving the speaker with a square wave. Very distorted, but also very loud, theoretically 100% efficient, and quite musically valid.

If you pulled one tube out of a 50W marshall, gapped the OPT to stop it saturating, and cranked it right up, you'd still get at least 25 watts.

[Merlinb]

Originally Posted by Steve Conner

If you pulled one tube out of a 50W marshall, gapped the OPT to stop it saturating, and cranked it right up, you'd still get at least 25 watts.

Agreed- the average output power will increase when a cold-biased valve is driven hard, but that doesn't change it from being anything other than class A. Efficiency is still limited to a theoretical maximum of 50%, and that is only achieved through 360 degree conduction of the device. (And in any case, output power is not a measure of class of operation, conduction is)

[Steve Conner]

Well, I still disagree with that. Drive it hard enough and you're in Class-C. The efficiency is more than 50% and the conduction angle is less than 180 degrees.

Valve plasma globe conversion

[pdf64]

To drive it hard enough to do that though, you've taken it past the onset of clipping and therefore (as I understand it) outside the scope of audio amp operation class. As you demonstrate, driven beyond clipping, audio amp operation class is meaningless, or at least requires another definition.

[Merlinb]

Originally Posted by Steve Conner

Well, I still disagree with that. Drive it hard enough and you're in Class-C. The efficiency is more than 50% and the conduction angle is less than 180 degrees.

Valve plasma globe conversion

Relative to the input signal conduction is less than 360 degrees, yes, but relative to the output signal we still have 360 degree conduction (i.e., when the device stops conducting we stop getting output signal too).
As PDF64 points out, at such high distortion levels it is no longer meaningful to relate output signal to input signal in terms of audio class, we need to stick to current conduction.

[Chuck H]

This post wasn't directed at you Bob (or anyone in particular). I just tagged to the end of the thread. Lots of fun peripheral debates going on here, and they should continue as part of this thread, but I (for one) do think the particular characteristics that qualify an amp to be called class A (or not) has been settled.

So if the AC30 is supposed to be a 30 watt amp (as the name and popular conception imply) then it is not a class A amp. Niether is the AC15 I guess. These would both be AB1 amps with a hot bias.

Of course theres another debate to that. If these amps reach cutoff AFTER their true RMS output has been reached, could they still be considered class A??? regardless of their titles or any assumptions, perhaps this should be the criteria.

Uh oh, now I've started doing it?!?

Chuck

[Old Tele man]

...pulling one output tube from a push-pull circuit reduces the output power to one-FOURTH, not simply to one-HALF.

...just repeat the "jingle" "...one-half of one-half is one-fourth!" until you've memorized it and THEN you'll understand.

[pgambon]

Any amp with two output tubes (or more) that claims to be Class A is suspect. It seems usually they aren't.