"Class A" nonsense

[Rhodesplyr]

A friend of mine proudly me today that he had a "Class A" guitar amp that got 18W from a pair of EL84s in push-pull.

Now, as far as I know, that's not possible. It's possible to get 18W from a pair of EL84s running Class AB1 like most push-pull tube output sections.

Who started all this marketing nonsense about "Class A" amps? I can see the motivation to impress buyers.....

David

[Chuck H]

This hardly reqires a response (based on your original premise of marketing and appearant understanding), but I can understand your frustration. Fact is, most el84 amps that are cathode biased idle at, or close to 100% dissapation. MFG's like to call this class A because the term has mojo. But when these amps are pushed the cathode voltage rises disproportionate to the sag in the B+ supply and the amps end up in AB1 (more +V at the cathode is the same as more -V at the grid). So... Even though these amps are technically biased class A they end up in AB1 during operation. A distinction that has long been ignored, probably for marketing purposes. Even the Vox AC30 ends up in AB1 when overdriven. But try telling a non amp building original AC30 owner that his amp isn't class A.?. Be ready to fight about it

Chuck

Edit: Just to clarify... In order to be an actual class A amp the AC30, most 18 watters, some Matchless amps, etc. would need to reduce the power tube grid signal voltage to avoid pushing these amps into AB1. Technically these amps are biased class A. But IMHO that hardly counts since an amp being played is not idling.

I recently designed an amp that may go into production. It is el84 based, cathode biased and idles at or near 100% dissapation... We may decide to call it class A for marketing, even though it's really not IMO.

[Robert M. Martinelli]

Hi,
you're absolutely correct, 17-18 W is about the most you can get out of a couple of EL84s ( AB1 P-P ).

This subject has been dealt with very thoroughly in the past, so I encourage you to do a little search and read all that's been said about it.

Unfortunately, some people don't know what they're talking about when they mention the various classes of operation. The misunderstanding about "class A" most likely comes from the fact that people thinks that biasing a tube at 100% ( idle - no signal ) automatically means we are in class A. This is not the case - the class of operation is solely determined by looking at how long the tube(s) conduct with respect to the signal's period.

If the SIGNAL flows through a tube for a whole period, 360 degrees ( or 2*Pi ) we are in true class A, and only one tube can suffice in reproducing the input waveform ( more tubes can be paralleled to increase the power output ).

If the SIGNAL flows through a tube exactly for 180 degrees ( or Pi ) we are in class B, and at least two tubes are required to reproduce the input waveform ( but zero-crossing distortion will be present ).

If the SIGNAL flows through a tube for more than 180 degrees but less than 360 degrees ( or between Pi and 2*Pi ) we are in class AB1, and as above, at least two tubes are needed ( the tubes are biased to get rid of cross-over distortion )

If the SIGNAL flows through a tube for more than 180 degrees but less than 360 degrees AND some ( control ) grid current is flowing we are in class AB2.

Part of the above misunderstanding is related to the "cathode bias" method, that, over the years, has come to be ( erroneously ) considered synonymous of "class A", but, as you rightfully noted, some manufacturers have been guilty of claiming their amps as "class A" for marketing reasons, thus generating more confusion on an already confused matter.

The Vox AC30 is the best example of this misunderstanding I can think about.

It' s often referred to as being a "class A" amp, while it's a cathode-biased "class AB1" amp. The four EL84s are biased ( by means of a 51 Ohm cathode resistor ) at about 125% ( idle ) ( some 47 mAmps with some 310 VDC dropping through the tubes ), still , looking at the waveforms, it appears clear that the amp operates in "class AB1".

Should the skeptics need further proof to confirm the above statement :

1 - The output power of an AC30 is between 34 and 36 Watts RMS. ( four EL84s operating in "class A" would yield only some 20 Watts RMS ).

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.

Cheers

Bob

[Enzo]

It isn't of course all the manufacturer's fault, as Bob mentions, a lot of people find out their amp is cathode biased and think that means ClassA.


In fairness to an amp, I don;t consider an amp NOT class A just because it is possible to drive it into AB land. Just as my mom's car doesn't become a race car just because I tried to drive faster than the guy next to me. If the amp is designed to be used in the class A mode, and it functions in class A most of the time when used normally, that seems good enough. Sometimes my amp functions as a doorstop too.

Your basic class AB amp functions part time in class A by that 360 degree standard. What if I cram 200v of signal into my amp's input, does that make it a class C amp? Surely it can;t pass the majority of that signal.

I guess I draw a distinction between the class of an amp - as in its design - versus the class it operates in moment by moment.

OOps, you managed to push a grid to cutoff... can;t call it a whatever anymore. That seems too hard core. I am not disagreeing with Rhodes' premise, but I am not fanatical either.

[PaoloJM]

It isn't of course all the manufacturer's fault, as Bob mentions, a lot of people find out their amp is cathode biased and think that means ClassA.


In fairness to an amp, I don;t consider an amp NOT class A just because it is possible to drive it into AB land. Just as my mom's car doesn't become a race car just because I tried to drive faster than the guy next to me. If the amp is designed to be used in the class A mode, and it functions in class A most of the time when used normally, that seems good enough. Sometimes my amp functions as a doorstop too.

Your basic class AB amp functions part time in class A by that 360 degree standard. What if I cram 200v of signal into my amp's input, does that make it a class C amp? Surely it can;t pass the majority of that signal.

I guess I draw a distinction between the class of an amp - as in its design - versus the class it operates in moment by moment.

OOps, you managed to push a grid to cutoff... can;t call it a whatever anymore. That seems too hard core. I am not disagreeing with Rhodes' premise, but I am not fanatical either.

To make sense of all of this in my own head it tend to define an amps Class of Operation as the Class it operates in at it's stated power output (Vrms).

Under extremely small drive signals even a Class B amp may meet the criteria for Class A, as it's unlikely that they have absolute zero idle current.
Equally I've scoped my Harley Benton GA5, which is very definitley Class A single ended, and it goes into cut off for part of the signal when the volume is above 7 ish where it's pumping out 7-8 watts. Is it Class AB1 so?

To me Class is only relevent at the amps stated power output.

[mooreamps]

single ended, and it goes into cut off for part of the signal when the volume is above 7 ish where it's pumping out 7-8 watts. Is it Class AB1 so?

I going to say technically, no. Any power amp, in guitar amp land, I would still call that class A. Now, having said that, let me write some comments about classes of operation.

These classes of operation come from radio and transmitter land, where all the signals past the first RF amp gain stage tend to be consistent in amplitude. Therefore, there is "less" confusion if a gain stage is operating in class A, or any of the other classes. If you have a radio frequency transmitter operating, lets say, Class B grounded gird, it will always operate Class B, even regardless of the level of the drive signal.

Now, in guitar amplifier land, the input signal can vary so much, you can have an audio power amp, such as the AC-30, biased as Class A push/pull, but the input signal can become high enough to drive the PA into Class A/B.


-g

[Old Tele man]

...EIA rules specify the class of operation is determined at the condition of maximum claimed power output, not at idle.

...EIA is the standard, any other way is simply "...fudging..."

[balijukka]

Here we go again
We had a huge discussion about this at AX84 last year.
I remember in the end someone found the old operation class specs which say that the parameters include the input signal when determining the class of operation. Meaning that the operation class can change during operation from A to AB if the input changes. The classes are operation classes, not amp classes.
This leads to that there is no such thing that class A amp. (unless you make some kind of limiter into the input that prevents the amp from cutting off)
At least this was the situation back then (1930's)
This was as far we got, then the discussion died.
I'm still not sure if we agreed on anything at all.
Can you point me to that EIA specification.

jukka

[Robert M. Martinelli]

...EIA rules specify the class of operation is determined at the condition of maximum claimed power output, not at idle.

...EIA is the standard, any other way is simply "...fudging..."

A Heart felt +1 ! Succinct and straight to the point!

( much better than I could ever do with my wordy posts )

This is especially true about the Vox AC30 I used as an example in my previous post.....I have heard ( and busted their bubble ) many friends of mine saying "the AC30 is class A" or "it works in class A at lower volumes", just because they don't know neither the differences between the various classes of operation nor the standards to comply with when testing or determining the class of operation.

Cheers

Bob

[PaoloJM]

...EIA rules specify the class of operation is determined at the condition of maximum claimed power output, not at idle.

...EIA is the standard, any other way is simply "...fudging..."

Who is the EIA?
It seems that I am in agreement with them whoever they are.

[Steve Conner]

The Electronics Industry Association (of America) I guess. Other standards apply in other countries, such as DIN or IEC.

I agree with this method of rating, even though it throws up some oddities. For instance, the Vox AC30 could be made into a true Class-A amp by just renaming it to the AC22.

It's also interesting to consider single-ended amps. Class-B operation requires two devices in push-pull, so a single-ended amp goes straight from class-A to class-C when it's overdriven.

[balijukka]

For instance, the Vox AC30 could be made into a true Class-A amp by just renaming it to the AC22.
.

LOL, that's a good one

jukka

[balijukka]

Here's a jewel from Rick Campbell :

A couple more definitions for the list:

"True Class A" -- Same as "Class A" but costs twice as much.

"Pure Class A" -- Same as "True Class A" but costs ten times as much and contains components made of Unobtanium.

jukka

[Zoe_N_Iain]

The way I always think of it is simple. See what maximum power you can get out of 2 valves in parallel single ended operation. Now connect them in push-pull instead (with an appropriate load impedance, of course). Does the resulting power amp put out exactly the same power as, and no more than, the PSE one? No matter how hard you drive it? If yes, it's class A. If no, it's AB1 to some varying degree. Simples.

[Steve Conner]

So it should have been the AC12?

[Jazz P Bass]

Conner
What was the application on the TPS11165?
Quote: "The TPS61165 is available in a space-saving,
· 2mm ´ 2mm QFN package with thermal pad."
Thats a small IC!

[Robert M. Martinelli]

Hmmm.....
Let's put it in another way....
You'll all forgive me if I keep on using my '64 AC30 as an example, but, having scoped the waveforms in the past, and having kept the results, I can give some real-life figures.

With a signal generator set to 1 Khz ( Sine ) and adjusting the signal's amplitude so that the amp doesn't clip, I've verified that, over the 2*Pi period, each couple of tubes stays some 6/5*Pi on, some 3/5*Pi off ( IDLING at some 48 mAmps, but OFF signal-wise nonetheless, and that's the thing to consider while evaluating the class of operation ) then on for the remaining 1/5*Pi. Being the tube(s) on for more than Pi, but appreciably less than 2*Pi, the class of operation is undeniably "AB1".

To keep the tubes out of the ( signal ) cutoff region at all times so that the SIGNAL flows continuously through the tubes ( thus achieving true "class A" operation ) the tubes' operating point should be considerably moved, and in these new conditions it would only be possible to get some 20 W RMS out of four EL84/6BQ5 ( BTW that's exactly four times the power an EL84 SE output stage yields with a single tube, as ZnI noted ).

Cheers

Bob

P.S. ( off topic ) I'd like to build a fixture for my digital camera to take snapshots from the screen of my TEK2335, has someone tried to do something similar already? I suck as a tube amp tech, but I confess I'm even worse as a photographer....

[hasserl]

...EIA rules specify the class of operation is determined at the condition of maximum claimed power output, not at idle.

...EIA is the standard, any other way is simply "...fudging..."

The Electronics Industry Association (of America) I guess. Other standards apply in other countries, such as DIN or IEC.

I agree with this method of rating, even though it throws up some oddities. For instance, the Vox AC30 could be made into a true Class-A amp by just renaming it to the AC22.

It's also interesting to consider single-ended amps. Class-B operation requires two devices in push-pull, so a single-ended amp goes straight from class-A to class-C when it's overdriven.

The "maximum claimed power output" should probably be "maximum undistorted power output". That would disqualify the AC30, and would prevent the situation of just claiming lower power output.

[hasserl]

The way I always think of it is simple. See what maximum power you can get out of 2 valves in parallel single ended operation. Now connect them in push-pull instead (with an appropriate load impedance, of course). Does the resulting power amp put out exactly the same power as, and no more than, the PSE one? No matter how hard you drive it? If yes, it's class A. If no, it's AB1 to some varying degree. Simples.

Well, that may work for you, but I think you're the only one it works for. That's not the definition of operating class.

[Robert M. Martinelli]

....I've scoped my Harley Benton GA5, which is very definitley Class A single ended, and it goes into cut off for part of the signal when the volume is above 7 ish where it's pumping out 7-8 watts. Is it Class AB1 so?

Hi Paolo, I was re-reading this post of yours, and I think that by saying that your amp went into CUT OFF ( not conducting ) you actually meant that your GA5 was CLIPPING ( squashing/squaring the waveform ) with the volume above 7.

Being a SE design, your amp is "per se" "class A", as a single tube ( or more than one tube in parallel ) can't go into cutoff at any time ( it has to be on during the whole period to amplify the entire waveform appearing at its grid ).

Cheers

Bob

[Albert Kreuzer]

Hi Paolo, I was re-reading this post of yours, and I think that by saying that your amp went into CUT OFF ( not conducting ) you actually meant that your GA5 was CLIPPING ( squashing/squaring the waveform ) with the volume above 7.

Being a SE design, your amp is "per se" "class A", as a single tube ( or more than one tube in parallel ) can't go into cutoff at any time ( it has to be on during the whole period to amplify the entire waveform appearing at its grid ).

I think he is talking about asymmetrical clipping. But as we are speaking of undistorted output, this doesn't apply here.

You'll all forgive me if I keep on using my '64 AC30 as an example

I forgive you

Cheers,
Albert

[Robert M. Martinelli]

I think he is talking about asymmetrical clipping. But as we are speaking of undistorted output, this doesn't apply here.



I forgive you

Cheers,
Albert

Yeah, that's what I thought, I definitely think he was talking about clipping and not cutoff.

Thanks for your forgiveness, Albert....I' m always afraid of boring people by keeping on talking 'bout my beloved AC30

OK, now let's start talking 'bout my '66 Vox UL4120 instead

( just kidding )

Cheers

Bob

[Albert Kreuzer]

OK, now let's start talking 'bout my '66 Vox UL4120 instead

I'd like to, but that would be worth a thread of its own

Cheers,
Albert

[Old Tele man]

The "maximum claimed power output" should probably be "maximum undistorted power output". That would disqualify the AC30, and would prevent the situation of just claiming lower power output.

...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!)

[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.

[mooreamps]

It's also interesting to consider single-ended amps. Class-B operation requires two devices in push-pull,

Class B does "not" require two devices in push-pull.
Class B is the condition of bias, such that, the tube is biased right at cut-off. This is a bias condition much lower in voltage, than a stage biased for Class A.

so a single-ended amp goes straight from class-A to class-C when it's overdriven.

Again, Class C is the condition of bias, such that, the power tube is biased far below cut-off, which again the power tube is not operating at the same condition as being biased for Class A.

I don't know why this is just so hard for some to understand ; and ; has absolutly "nothing" to do with the level of the input drive signal.

-g

[balijukka]

[QUOTE=mooreamps;119174
has absolutly "nothing" to do with the level of the input drive signal.

-g[/QUOTE]

A direct quote from IRE standards 1938:

1E69. Class A Amplifier.* A class A amplifier is an amplifier in which the grid bias and alternating grid voltages are such that plate current in a specific tube flows at all times.
1E70. Class AB Amplifier. A class AB amplifier is an amplifier in which the grid bias and alternating grid voltages are such that plate current in a specific tube flows for appreciably more than half but less than entire electrical cycle.

according to this the input signal does matter.
The class would be a step funtion of the input signal.

At least we all do know what sells best.

jukka

[balijukka]

That asterisk after class A, refers to 1 or 2 according to grid conduction.
too lazy to type it all.

jukka

[Chuck H]

We need to consider that input levels on amps such as guitar amps can vary greatly. Evidenced by the fact that Fender never intendeed their tweed to BF era amps be overdriven, but you can't turn them above four on the volume control without distortion when using a humbucker. So lets assume the input signal is the responsibility of the user and, for most designs, is attenuated before clipping by the amps volume control. Then we have to fall back to the amps indicated (rated) output power in RMS. No consideration of any other parameter including input signal needs to be examined in this case. If an amp does not drive the tubes into cutoff prior to when it's rated output power is achieved we can consider the amp class A regardless of weather it is pushed into AB at a greater output.

Just for giggles... Robert, how much does your AC30 put out before cutoff???

Chuck

[mooreamps]

I'm going to bury this once and for all.

Classes of bias for EL-84 {approx}

Class A -7.5 volts vdc
Class A/B 2 tubes in push/pull -11.7 volts vdc
Class B 2 tubes in push/pull -14.7 volts vdc
Class C 1 tube in grounded grid -30 volts vdc


These classes of operation were written ; as a tube operates within a radio frequency transmitter ; and not how they are used in a FKG guitar amp.'



-g

[Steve Conner]

Mooreamps: What you said about bias points is true enough. But a single-ended Class-B amp is unusable for audio, so I stand by what I said.

Bob: I think you have the operation of your beloved '64 AC30 wrong. This is what's supposed to happen in a push-pull Class-A amp:

1: The tubes must be set up such that their idle current is exactly half of the maximum current they can supply with the grids pushed as far positive as the driver can take them. This is achieved by adjusting load impedance, screen voltage, screen resistance and so on.

2: The push-pull drive now increases the plate current of one tube as it decreases the other. If our two tubes were idling at 48mA, then at the positive peak of the sine wave, one of them might conduct 40mA and the other 56mA. At the negative peak, the roles would reverse: 56mA and 40mA.

3: When the amp is producing its maximum output, then at waveform peaks one tube is passing 96mA and the other is exactly at the point of cutoff, passing 0. Any more drive results in clipping, because the tubes are fully on and fully off.

Note that at all times the sum of plate currents is 96mA. This is one way of recognizing a Class-A amp: the supply current doesn't change from idle to full output.

If we set the tubes up differently, such that their idle current was less than half of the maximum possible current, then we're in Class-AB. When one tube is cut off, the other one can still go further on.

For instance, a 50-watt Marshall (or a Vox AC50 ) idles at 34mA per tube, for 68mA total idle current. But it'll draw over 300mA when producing its rated 50 watts. At waveform peaks, one tube is fully on and passing about 450mA, while the other one is completely off.

The above discussion concerns unclipped power output only. All tube amps self-bias into a different class if you overdrive them hard enough.

The TPS61165 is for a white LED backlight driver board I'm building. I got a demo board for the chip, hooked the power up backwards, and blew it. I figured since it was already blown, I had nothing to lose by trying to change the chip out.

[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.

[balijukka]

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

jukka

[Rhodesplyr]

Here's a jewel from Rick Campbell :

A couple more definitions for the list:

"True Class A" -- Same as "Class A" but costs twice as much.

"Pure Class A" -- Same as "True Class A" but costs ten times as much and contains components made of Unobtanium.

jukka

I appreciate all the responses, but I think I like this one the best.

In my opinion, few end users are ever really going to know what "Class A" operation of a circuit is, much less why they should want it. If it weren't mentioned, they wouldn't notice. What irritates me is seeing the term turned into a marketing device. It certainly makes far more difference what kind of output transformer is used or what kind of speaker is hooked to it.

It reminds me of the THD figure wars among 1970s Japanese receivers or the Spinal Tap joke about the amp with a volume knob that goes up to 11.

Most of the classic Hi-Fi amps like Marantz 8Bs are Class AB1, and I think McIntosh amps approach Class B. It never stopped them from sounding good.

David

[Steve Conner]

I guess I agree. I recently rebuilt this old hi-fi amp, and it's very much AB1: it operates its EL34s in ultralinear mode with 400V B+ and 40mA idle current, for a measured power output of 30W/channel.

But to me it sounds just great. If I wanted low THD, I'd have used a solid-state amp instead.