# Efficiency of class A AB and B. What does it mean?

Discussion in 'Amp Tech Center' started by peteb, Jul 17, 2018.

1. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:
http://www.tungsol.com/html/faqs11.html

Per this Tung-Sol link it breaks down like this:

Class A: 20-35%

Class AB: 45-60%

Class B: 60-70%

Does everyone like these numbers? Or do you have different values?

What do they mean?

Power out over power in?

Where?

Power in must be idle plate dissipation in Watts calculated using DC volts and current.

Power out must be at the plate, AC volts and current on the plate load resistor.

The power out must be close to and higher by 10-20% to the power out measured at the speaker due to losses going thru the Output transformer.

Each is given as a range. Since the power out at the tube will be higher than the power out at the speaker, maybe a simple way to look at these numbers is the higher number may be what to expect at the tube, and the lower number might be what to expect at the speaker?

2. ### RecceFriend of Leo'sSilver Supporter

Posts:
3,392
Joined:
May 3, 2016
Location:
Northern Alabama
I am going to take two Advill and hope someone else understands what you are talking about. I use the plug them in adjust the knobs and listen to the sound coming out method. If it quits working take it to a tech.

stratoman1 and Ricky D. like this.
3. ### Axis29Poster ExtraordinaireAd Free Member

Posts:
6,715
Joined:
Jan 2, 2007
Location:
Beaumont, CA
I'd say we may have lost a little something in translation from Russian?

4. ### Art VanDelayTele-Afflicted

Posts:
1,209
Joined:
May 20, 2018
Location:
Arcata
Efficiency there must be related to life of tube because Class A would otherwise be the most efficient.

From my understanding of classes, class A is "always on", AB runs at lower voltage and comes to full power when a note is played, and class B means a complete on/off between played notes.

That said, class A would seem more efficient for nuanced playing while Class B would be more efficient in regards to power consumption.

viccortes285 and scooteraz like this.

Posts:
1,209
Joined:
May 20, 2018
Location:
Arcata
6. ### Art VanDelayTele-Afflicted

Posts:
1,209
Joined:
May 20, 2018
Location:
Arcata
What's crazy is Class D (outstanding for bass). I have an 800w Gallien Krueger the size of a hardcover book.

7. ### dan40Tele-Afflicted

Posts:
1,977
Joined:
Aug 19, 2015
Location:
Richmond Va
Class A wastes more than half of the tubes power as heat. This is why most designers went on to build class AB amplifiers as players wanted more and more volume. A 100 watt class A amp would be huge and heat a room on a cold winter day!

Wally likes this.
8. ### Old Tele manFriend of Leo's

Posts:
3,941
Joined:
May 10, 2017
Location:
Once banned always banned
Efficiency = POWER.out / POWER.in

Maximum Class B efficiency is PI/4 =0.785398 or 78.54%

...and don't forget that pentodes are slightly more efficient than triodes because they can "make use" of more of the available plate voltage.

Last edited: Jul 17, 2018
viccortes285 and peteb like this.

Posts:
1,977
Joined:
Aug 19, 2015
Location:
Richmond Va
10. ### elpicoTele-Afflicted

Posts:
1,154
Joined:
Sep 14, 2011
Location:
Vancouver BC
Efficiency is a comparison of how much power gets wasted in the tube vs how much power reaches the speaker. At idle efficiency would be zero because the tube is dissipating power but not sending any signal to the speaker. The amp must be passing signal to make a meaningful efficiency measurement. There's no standard for how much signal it should be passing when you make the measurement, but it's usually made near the max clean signal the amp can pass.

It's important to note that the amount of power the tubes dissipate while passing signal is NOT the same as their idle dissipation.

In a class A amp the tubes run hottest at idle, they're usually biased near 100% of their plate rating, but as they start passing signal they actually cool down because the speaker is starting to share some of that dissipation. In a perfect world the best a class A amp could do is 50% efficiency which would be achieved when the tube and the speaker split the dissipation evenly. In the real world this can't be achieved due to limitations of the tube and losses in the transformer.

A class B amp is the opposite, the tubes run coolest at idle then get hotter as you turn the volume up. Class B does give higher power output and efficiency than class A, but class B isn't used for audio amps due to the high distortion it generates.

Last edited: Jul 17, 2018
11. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:

Well

Amp efficiency, this is something that has some basis, something that can be worked with.

The Numbers commonly used for max plate dissipation and max power output are never agreed upon and are without a concise basis.

The numbers for tube efficiency appear acceptable at first glance.

The numbers are stated as approximate and are given as a range.

They are all in between 0 and 100 %.

There is an obvious progression from A to AB and then to B.

AB covers all the ground between A and B and so there should be no gaps between the classes.

Last edited: Jul 17, 2018
12. ### DougMPoster Extraordinaire

Posts:
5,196
Joined:
Jul 5, 2017
Location:
Honolulu, HI
In many SS audio (stereo) amps at low power even class AB amps operate in class A until a certain level is reached where they switch to class AB. Maybe someone with more knowledge of tube amps (both audio and guitar) could elucidate on whether that also occurs in tube power sections as well.

peteb likes this.
13. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:
The basics.

Class A conducts for all 360 degrees of the cycle. It is never in cutoff. It idles at 100%

Class AB conducts for less than 360 degrees and more than 180 degrees. Goes into cutoff during part of the cycle and likes to idle at 50-70%.

Class B conducts for 180 degrees and is in cutoff for 180 degrees and idles at 0%.

Class AB and class B have the advantage over class A because they amplify only the top half of the wave so the wave amplitude, or vertical height can be doubled that of class A.

Class B has the advantage over class AB because each tube is in cutoff, meaning no current flow, for fully half the time. So it handles the half wave just like AB but B gets to fully rest for the full half cycle.

DougM likes this.
14. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:

They do say the amps shift between classes, but just to tie things down, let’s assume an amp is in either class A, B or AB at any one time.

DougM likes this.
15. ### DougMPoster Extraordinaire

Posts:
5,196
Joined:
Jul 5, 2017
Location:
Honolulu, HI
The fact that the tube goes into complete cutoff for half the cycle creates bad sounding crossover distortion, which is why class B isn't used in any audio circuits, whether for stereo or musical instruments. In class AB, in the half of the cycle where a tube isn't conducting the signal, it still never goes into complete cutoff, avoiding that crossover distortion, which is why it's called class AB. That was a very good description of the difference between AB and B.

Last edited: Jul 18, 2018
16. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:

Check

Where or how or what?

Pentode more then tetrode more than triode.

Ultra linear more than not

And a ss rectifier must improve efficiency.

17. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:

I am fairly certain it will cutoff in AB. If it doesn’t, then that is class A.

I’m glad you said the part “in class AB, the half of the cycle where the tube isn’t conducting...”

In AB there is not half of the cycle where it is not conducting. I think that is a common misconception that the PI or something selects one tube for half a cycle and then plays thru the other. The signal is applied to both tubes for 360 degrees, but in class AB, the tube goes into cutoff, zero current flow for more than zero degrees and less than 180 degrees.

18. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:
Elpico, thanks for posting. It was your knowledge in the other thread that got me thinking on this.

I agree heat dissipation is part of the equation, and it’s hard to measure.

Heat is part of it, but I agree more with your statement that a meaningful efficiency measurement must measure power out divided by power in. In this case I am thinking the power in must be the DC applied to the tube and the power out is the AC going out of the tube.

I agree with you that the speaker is a good place to measure power out. I think that it is a good approximation for power out of the tube, but not closer than say 10 % owing to losses going thru the OT.

19. ### petebFriend of Leo's

Posts:
4,208
Joined:
Apr 25, 2003
Location:

I wanted to look closer at the numbers, compare them with well known fender amps, starting with the single ended 6V6 because it is such a known quantity, so your comment is a great place to start.

TungSol says a tube in class A will have an efficiency of 20-35%.
Elpico says the class A effiecieny could ideally reach 50% ignoring losses in the transformer (10%)

Let’s put these two together and say a champ output tube stage including OT and speaker could be 40%.

One of the main questions to be looked at in this thread is the champ and the traditional 12 w max plate dissipation associated with the 6V6 tube, the established practice of dissipating 18 W, and what is the affect of dissipating 150% and how much extra volume and power is realized by increasing the bias from 12 Watts to 18.

The champ is known to put out 6 Watts max. It might me biased at 12 Watts and it might be biased at 18 Watts but it is always known to put out 6 Watts max regardless of bias.

Efficiency is power out over power in.

In the one case there is 6 W out / 12 Watts in. (50 %)

In the second case there is 6 Watts out / 18 Watts in (33%)

Actually these numbers don’t look to interesting, which makes the thread, well maybe the thread will show the numbers as right.

In the first case, 50% would be high according to tungsol and Elpico. And this is with the power out measured at the speaker not the tube making the efficiency of the tube over 50%. Maybe a champ biased at 12 W would not be able to put out 6 W to the speaker?

The second case looks pretty good. 33% is in range. The efficiency of the tube itself would be higher than 33%. Say the OT dropped the signal 10% from 6.6 to 6 W then 6.6/18 is 37%, still in range.

One of the central points of the THREAD to FOLLOW.

Let’s use the best numbers available.

The champ puts out 6 W, the 6V6 tube has max plate dissipation of 14 W
A 10% loss in the OT means the tube is putting out 6.6 W.

Efficiency of tube, OT and speaker:

6 W out put / 14 W input = 43%

Efficiency of tube:

6.6 W out / 14 W input = 47%

These numbers are not too surprising, a little higher than tungsol and basically in line with Elpico’s upper limit of 50%.

I agree that the upper end of A should coincide within the lower end of AB.

The point that wasn’t. I was thinking these calculations would show that the max plate dissipation for a 6V6 would have the be more than 14 W but the numbers don’t necessarily show that.

Last edited: Jul 18, 2018
20. ### DougMPoster Extraordinaire

Posts:
5,196
Joined:
Jul 5, 2017
Location:
Honolulu, HI
In a class AB push-pull circuit each tube handles either the positive or negative half of the waveform, but DC current still flows some, even when that tube isn't conducting audio. It never completely turns off. In a class B circuit the tube doesn't conduct any DC current when it isn't handling it's half of the waveform, it turns off, and so when it turns back on, gross crossover distortion occurs. Read Wikipedia's page on amplifier classes for a more complete description.

IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.