Efficiency at max output does not tell the whole story.
Power dissipation for the class B P/P configuration ( zero idle current ) is a bit trickier:
when the load is receiving about 40% of its maximum power, the output devices will be at their hottest and will be dissipating approximately 20% of the maximum load power (or about half of the power delivered to the load at that point). Thus, if a class B amplifier is rated to produce a maximum load power of 100 watts, the output devices will get their hottest when the load is receiving 40 watts, and each output device will be dissipating 20 watts. The outputs will be dissipating less power when the load is at maximum ...
Some bias P/P tube outputs with idle currents causing max or near max tube dissipation in an attempt to stay in class A as long as possible with rising signal.
Again tube dissipation actually drops at max output, but may in fact exceed max dissipation limits somewhere between no signal and max signal.
This is usually not an issue because of the crest factor of music and speech but very noticeable for those who do sine testing.
Power dissipation for the class B P/P configuration ( zero idle current ) is a bit trickier:
when the load is receiving about 40% of its maximum power, the output devices will be at their hottest and will be dissipating approximately 20% of the maximum load power (or about half of the power delivered to the load at that point). Thus, if a class B amplifier is rated to produce a maximum load power of 100 watts, the output devices will get their hottest when the load is receiving 40 watts, and each output device will be dissipating 20 watts. The outputs will be dissipating less power when the load is at maximum ...
Some bias P/P tube outputs with idle currents causing max or near max tube dissipation in an attempt to stay in class A as long as possible with rising signal.
Again tube dissipation actually drops at max output, but may in fact exceed max dissipation limits somewhere between no signal and max signal.
This is usually not an issue because of the crest factor of music and speech but very noticeable for those who do sine testing.
Squeezing blood out of a Turnip gets blood on your shirt.
Maximum power, perhaps means 10% Distortion.
Have fun listening to it!
What kind of a contest is this thread becoming?
Maximum power, perhaps means 10% Distortion.
Have fun listening to it!
What kind of a contest is this thread becoming?
Thank you for the reply. I do understand that the tubes not designed expressly for audio output, would have no need to give output power. OTOH, some of those tubes have made fine output tubes nonetheless. I think the 6AQ5 was one of the tubes that would be in this category. Concerning your last sentence, YOU can easily calculate those things. LOL You forget, I'm very new to this and have never had any bench experience at all. I'm hoping to change all of that but at the moment, I'm still 'in school' learning from whom I can. You know, when I look at the curves on the datasheet, my eyes just glaze over. They are next to meaningless to me. It's like reading Chinese. No offense intended to anyone.
I have yet to read all of the answers here but this one seems top notch. Being able to hear about max efficiencies does help. One thing that is obvious, is that not all tubes operate at max efficiency. OTOH, some outdo the theorectical, I can only guess, by being underrated to begin with. I don't like ending a sentence with a preposition but I will leave it as is. lol Thank you.
Agreed. That is normally the definition.
Then as someone else says, efficiency is not constant from zip to max Pout but it is pretty irrelevant as long as plate dissipation is within tube limits which is normally the case for datasheet suggested operative conditions.
@6A3sUMMER, nope! You can get 100W out of a pair of EL34 with 5% THD at max Pout without any feedback. The actual problem to solve is distortion at lower level as being a class B amplifier THD is not much lower. This can be done brilliantly with bifilar primary that erases cross-over. It's the same trick of the McIntosh without the complications of cross-coupled screen-grids and other stuff. Just bifilar primary and pentode output stage. With regulated g2 supply, THD is even a bit less! Then with just 10% cathode feedback (resulting in about 7 dB in this case), THD drops down to about 1% at full power and much less at lower level.....
I hope I can post soon (mid October I hope) the actual mono amplifiers with their big 11K output transformers. Good news for all is that the common JJ EL34 does the job and, IMHO, is the closest to the original Philips also sound wise. Curiously, with other EL34's, even expensive ones, power is a bit less or significantly less for ANY application. JJ just got this tube right.
Having said this, the 60W option with 450V plate voltage or the 75W option with 500V plate voltage in my opinion are more practical and convenient.
Thank you all for your replies. Naturally, I'm still confused but I have learned some things here. Don't feel bad if I didn't answer your replies directly. BillyBob
45,
Just remember, the 100W output from a pair of EL34 tubes is the tube's output power, Before the loss of the output transformer.
Any reader/designer/builders of you . . . when you build your next EL34 100 Watt amplifier, which does not have negative feedback, please let me know the % distortion you get.
No negative feedback means: no Schade, no output plate to driver cathode, no Ultra Linear, no global around the output transformer, no output tube cathode to cathode, no output transformer secondary to output tube cathode, no screen feedback, etc.
I am waiting to have you tell me the dissipation of the plates, and the dissipation of the screens.
No cheating now, you have to measure those quiescent dissipations, and not exceed the data sheet maximum dissipations, plate and screen voltages, and other data sheet limits; and list the % distortion measurement of 100 Watts at the output transformer secondary (with no negative feedback).
I hope someone succeeds doing that.
Just remember, the 100W output from a pair of EL34 tubes is the tube's output power, Before the loss of the output transformer.
Any reader/designer/builders of you . . . when you build your next EL34 100 Watt amplifier, which does not have negative feedback, please let me know the % distortion you get.
No negative feedback means: no Schade, no output plate to driver cathode, no Ultra Linear, no global around the output transformer, no output tube cathode to cathode, no output transformer secondary to output tube cathode, no screen feedback, etc.
I am waiting to have you tell me the dissipation of the plates, and the dissipation of the screens.
No cheating now, you have to measure those quiescent dissipations, and not exceed the data sheet maximum dissipations, plate and screen voltages, and other data sheet limits; and list the % distortion measurement of 100 Watts at the output transformer secondary (with no negative feedback).
I hope someone succeeds doing that.
Last edited:
So what , you want to tell us that the datasheets are lying ? If is given 100W with 5% distortion is your fault if you can't get it .
Of course the output transformer is not included in the calculations , but using a "generous" one , the losses are minimal
Of course the output transformer is not included in the calculations , but using a "generous" one , the losses are minimal
Sorry, no, again. In your example 100 W output power relates to the efficiency, as always. Input power calculates to 100 x 100/78.5 = 127.4 W. Hence, waste is 27.4 W.
Best regards!
Yes, you're totally correct, I just had the biggest brain fart yesterday.
I asked the mods to strike out the text, but apparently they don't do that and have a rule with what is written is written.
(kinda strange imo)
So consider that post as unposted.
I did derive the correct formulas today and will post them later.
Yeah, absolutely no idea what I was thinking yesterday, omg
Anyway, here the correct way of calculating (hopefully I didn't make any major typos).
Little practical side note though, this can look a bit surreal.
Keep in mind that these efficiency calculations don't take other (practical) variables in mind.
Therefor you might end up with weird values.
For example a Class-A tube amplifier that is (close to) center biased, will mostly be around half of Pd for any practical purposes. (maybe that's where my brain went total weird).
Also a Class-B amplifier is not the same as a Class-AB in sense of efficiency.
Anyway, here the correct way of calculating (hopefully I didn't make any major typos).
Little practical side note though, this can look a bit surreal.
Keep in mind that these efficiency calculations don't take other (practical) variables in mind.
Therefor you might end up with weird values.
For example a Class-A tube amplifier that is (close to) center biased, will mostly be around half of Pd for any practical purposes. (maybe that's where my brain went total weird).
Also a Class-B amplifier is not the same as a Class-AB in sense of efficiency.
Requiring an 800 volt regulated supply and an 11k trafo are non-starters for most. When the supply sags to 750 volts you don’t get that full hundred anymore. And I’m nervous enough around 600 volts. An 800 V choke input supply is going to sit up around 1100 until the current comes up. Capacitor requirement goes from the ridiculous to the sublime pretty quickly. Two 450 volters isn’t going to cut it. Even doubling the rectifiers isn’t enough safety margin. Construction materials being equal, an 11k primary is more difficult to get 30-20k response from than a 5k. In order to hit the same spec you’ll spend much more. Makes one wonder why everyone fights using another pair of tubes. 150 watts out of two pair is far less trouble, and probably sounds better.
I don't need to remember. 60W, 75W and 100W is the actual power into 8R resistive load I get using 450V/5K, 500V/5K and 11K/800V, respectively.
According to the datasheet with 500V I should only get 70W. Datasheet are not the bible, they give typical figures, averages.
The quiescent anode current is about 25mA for 800V and 30-35 mA for 450-500V by memory. It is not really important as long as quiescent plate dissipation is within 20W. G2 current is around 3-4 mA with no signal and around 20-25mA at full power for the 3 cases. Because the g2 voltage stays at about 400V, the lower g2 current happens for the 800V case. It's all safe and happy running and more importantly sounds great!
The fact that you are not able to do it doesn't mean that it magically becomes a rule especially because it is clear that you don't get how it works. Your objections have no foundations, sorry. Otherwise you would not ask the measurement of the output power without cathode feedback. The output power is optimized with cathode feedback ON (i.e. 5K is the total distributed load). So to do what you ask I should make another 5K plate-to-plate transformer. But the logic says that if distortion is just over 1% with 7 dB feedback, without it is very unlikely that it will be 10%. You are confused with SE amplifiers. No pentode PP amp without any fbk does 10% at nominal max output unless it is badly executed or the components are poor.
Be assured that this time I will post it.
Last edited:
It is only 400V on G2 to be regulated and I have not used any choke for the anode voltage so far, just plain good sized capacitors. The amp doesn't make any noise even if you stick your ear to the speakers because it is dynamically balanced (i.e. I don't care if quiescent anode currents of the output stage are slightly different, say 1-2 mA apart, because there is always a tiny airgap to take care of it). All bifilar transformers I have done have a bandwidth above 100KHz. The 5K transformer for the 60W type gets to 200KHz!
As I said, I find the 60W and 75W types more convenient and practical. I see the 60W as stereo integrated amp and the 75W as mono power amps.
You will not ear any difference respect to the 100W in terms of SPL. The 100W one was just a challenge. They do share a "family sound" as they are all done the "same" way.
Last edited:
I think it's best to define terms before launching into a long thread.
Trying to keep it simple here:
Power output = power delivered to the load (speaker or test load)
Power input = ??? -- I'd say for a tube amp, that would be voltage*current drawn from the plate supply (B+) and voltage*current drawn from the heater supply.
For a 2A3 PP Class AB1 amplifier, you're lucky to get 10W max continuous output delivered across your speaker load.
Each 2A3 will draw about 50mA plate current (Ip). Each 2A3 will be dissipating about 12.5W each, quiescent, so 25W there.
Each 2A3 draws 2.5V @ 2.5A of heater current (6.25W) so 12.5W for a pair.
That makes 25W + 12.5W = 37.5W input (quiescent) to get 10W power output into the speaker.
Remember that the current draw will rise at higher output levels.
10/37.5 = 26.67% efficiency.
That does not include the power dissipated by the driver stage(s). That's just the output stage.
Also, the power supplies will be less than perfect, so there will be losses causing just a bit higher power dissipation on the input side.
But you get the idea.
Trying to keep it simple here:
Power output = power delivered to the load (speaker or test load)
Power input = ??? -- I'd say for a tube amp, that would be voltage*current drawn from the plate supply (B+) and voltage*current drawn from the heater supply.
For a 2A3 PP Class AB1 amplifier, you're lucky to get 10W max continuous output delivered across your speaker load.
Each 2A3 will draw about 50mA plate current (Ip). Each 2A3 will be dissipating about 12.5W each, quiescent, so 25W there.
Each 2A3 draws 2.5V @ 2.5A of heater current (6.25W) so 12.5W for a pair.
That makes 25W + 12.5W = 37.5W input (quiescent) to get 10W power output into the speaker.
Remember that the current draw will rise at higher output levels.
10/37.5 = 26.67% efficiency.
That does not include the power dissipated by the driver stage(s). That's just the output stage.
Also, the power supplies will be less than perfect, so there will be losses causing just a bit higher power dissipation on the input side.
But you get the idea.
I was wrong again. I thought Pd was plate dissipation and not power dissipation. Though, it could be both?!?
Not sure if this answers your question as such but it is odd if your a beginner and a valves datasheet says valve X's maximum power dissipation is 25w but then goes on to say later in the datasheet it can do 100w in PP whatever. Measure it's cathode current etc and you'll find it's sitting at 17.5w. Turn the volume up to 3 pop some sounds on and measure the OP at the speaker and you'll find the outcome is 3w.
This is a bit like the small print in a spec sheet for a Foogle XS 100w amplifier saying that the 100w max is only at peak power or "RMS power" or only at 1khz. Get either a valve X PP output stage or the Foogle XS 100, stick a sign wave in and you might get a reading of 100w... but not for long. If you keep pushing either something might go pop or bang.
Like any specification it depends on quite a few other things, for instance we don't listen to sinewave's, then there's quiescent power conditions, speaker efficiency ad infinitum. hope that puts a different take on things.
This is a bit like the small print in a spec sheet for a Foogle XS 100w amplifier saying that the 100w max is only at peak power or "RMS power" or only at 1khz. Get either a valve X PP output stage or the Foogle XS 100, stick a sign wave in and you might get a reading of 100w... but not for long. If you keep pushing either something might go pop or bang.
Like any specification it depends on quite a few other things, for instance we don't listen to sinewave's, then there's quiescent power conditions, speaker efficiency ad infinitum. hope that puts a different take on things.