What determines amplifier wattage?

[tomjackson]

I know different valves have different maximum wattage outputs but what else determines the wattage?
I've seen many single ended amps with the same valve (EL34, for example) where the wattage ranges from about 7 to 12W. 

Also push-pull amps with 2 EL84 shown as between 12W and 18W.

So, if I had a single ended amp and knew the wattage with , say, a 6L6, could I determine the wattage with an EL34, without measurement?

[Twinfan]

I think the biggest change comes from the way of biasing the valve(s).  Cathode biasing gives less power than fixed biasing, and a different feel to the power section.

[HTH AMPS]

cathode versus fixed bias will change the efficiency of the output section - you get more headroom and wattage from a fixed bias amp.

the plate voltage also has a part to play - if you ever try an early 70s 50w Marshall head, they have a lower than average plate voltage (around 350v) and therefore break up really easy, with a thick tone but not that loud for a 50w amp.

lastly, the primary Z of the output transformer will affect the wattage (imo).  a classic example of this is the JTM45 where the KT66s are run at a higher primary Z than they'd like to see as per the data sheets.  this results in only around 35w from a pair of KT66s which could easily push out 50w run into the correct load.

[martinw]

Actually, the biggest variation is more to do with which version of power measurement the amp manufacturer uses or claims to have used.
There are different ways, and the range of difference can be huge. No method is invalid, including just estimating, as the definiton of "output power" for an amp is a bit nebulous.
Maximum clean power? Or fully clipped? Measured where? Measured with what? Which was calibrated when? And so on.
The most used method is to put a scope on the output from the output transformer, and measure the RMS AC voltage of a sine wave, just before it clips. Square this value then divide it by the nominal impedance of the load and that's your maximum clean power.
Of course this is artificial in several respects; guitars don't put out sine waves, impedance is not fixed but varies with frequency, the OT is not the end of the system (the speaker is a further element, being a transducer of varying efficiency) and so on.

So you can see it's a minefield. Each power claim should be accompanied by a lengthy description of how the mesurement was taken for it to have any validity.

[tomjackson]

I think I knew there would be no simple answer when I asked the question!

Perhaps a more useful indication of amp power would be the clean and maximum SPL at 30 paces (or somthing a little more scientific and standardised).

[jpfamps]

I suspect most amplifier wattages are determined by what sounds good to the marketing department, ie more watts = more sales.

The "standard" way amplifiers are measured is power delivered into a resistive load using a sine wave input (usually a 1 kHz) with a defined level of THD. It is important to use a resistive load as a reactive load, such as a speaker, will cause a phase difference between the current and voltage waveforms which complicates the power calculation.

Measure a few amps like this and you will find that many of the rated wattages are either a) not measured like this, or b) very optimistic!

The final wattage of an amp will also be determined by the B+ voltage and impedance matching (as explained above). A pair of EL84s run at class A with a B+ of 250 V into the correct load will give about 12 W maximum before clipping. Running the same valves in class AB with a higher B+ say 380 V with fixed bias and you probably get around 16-18 W.

Ultimately all this is fairly arbitrary, and you are best to make a judgement on an amp based on whether is works for you rather than trying to aquire a nominal wattage.