Question about tube power dissipation and signal output

[dougman0988]

I'm sure I'm overthinking this, but lets say I have an all-tube amp with 12ax7's which have 1 watt max power dissipation per triode.  Does this mean I can never (safely) get more than 1 watt out of my speaker?
 
I understand tubes use the most power when idle (no signal present).  So if I bias each tube for 0.5 watts, like 200vdc at the plate with 2.5mA current, I cant get a signal more than that much power out of them, correct?

[iccaros]

in a well biased amp your full signal should equal zero and you would have a net effect of not drawing more current than when at idle, In a class A you may draw less current at one side of the swing, but you are dissipating the same power (voltage and current) to net effect of no change.

In any case a 12ax7 is a horrible output tube IMO, While a 12au7 has less gain, it can dissipate more current giving you a bigger swing which gives you more output.. you could get 2watts if you had the perfect load.

[R.G.]

Do you mean you're trying to drive speakers with 12AX7's?

The relationship of power output to output device dissipation is not straightforward. With semiconductors in Class B push-pull, the output devices dissipate at a maximum about 40% of the maximum power out to the speaker. I'll have to go look up the numbers if you need closer than this handwaving. It's been a while.

In Class A, you can only get out as audio a fraction of the power dissipated in the output tubes. If you're doing single ended, this is down in the 12% range, meaning the output devices waste 88% or so of the power sent to the amp and the speakers get 12%. Class A done perfectly in a push-pull amp can never be better than 50%, meaning the output devices waste as heat a power equal to the power sent to the speakers.

Class B push-pull can get up to 70-80% efficient, meaning the output devices dissipate 20-30% of the total power, the rest goes to the speakers.

That's for semiconductors. Tubes are worse; triodes are especially worse.You'll never get as much as 1W out of a 12AX7 to the speakers. You'll have to go push pull to get more than some tens of milliwatts. Triodes will never use the full voltage swing of the power supply feeding them. In transistor terms, they can't saturate to a high current and a low voltage because their internal plate impedance is so high. This is where the additional heat is generated. This very issue was one of the reasons for the development of pentodes and power beam tubes: they use their power supply voltage and current much more efficiently, letting you get more AC output with less DC wasted power.

[PRR]

> 12ax7's which have 1 watt max power dissipation per triode.  Does this mean I can never (safely) get more than 1 watt out of my speaker?

In single-ended "clean" audio, you can't get sine output power even half your device dissipation.

You also must load appropriately. Triodes are simple. One section of 12AX7 is a 60K plate resistance, good power loading is 120K to 300K.

There are NO 120k power output transformers. Either the wire-size is too-too fine, or there's no bass at all.

If there were, you would use like 420V supply, 120K load, 2.4mA idle current, get about 420V * 120K/(60K+120K) = 280V swing peak, 200V sine RMS, 200^2/120K= 0.33 Watts output.

Note that 420V is well over the rated plate voltage of 12AX7.

With available output transformers such as 25K, and nearly-legal supply like 360V, one section could make 0.2W.

> like 200vdc at the plate with 2.5mA current

Best SE load I can find there is 22Vrms 0.82mArms 0.018 Watts audio out; sweetly this does "work" with available 25K OT.

This all assumes capacitor grid coupling and zero grid current. Anything else is a headache. If you allow and can supply grid current then a Class A2 mode around 200V could output over 0.1W, BUT the grid current gets greater than the plate current. So you need something fatter than 12AX7 to drive your 12AX7. Or an interstage transformer, which costs more than any of the many small POWER tubes which will easily make small or large Watts. 12FX5 is made to pull loads with just 110V. 6Y6 is a much bigger tube but has a 200V rating for 6 Watts.

[PRR]

1 Watt of device(s) dissipation will theoretically support:

Class A (SE or P-P)- 0.5 Watts clean sine output (1 Watt if signal is ALWAYS "full sine")

Class B- 3.5 Watts clean sine out (you never get here with real devices)