8Ω vs. 4Ω

[Gnobuddy]

How much power would be lost in an AB763 circuit running 2 6L6s in push pull driving a 44:1 transformer by putting in an 8Ω load instead of the designed-for 4Ω load?

I don't have the faintest idea, nor can I think of any way in which one could go about calculating this. I think this is one of those cases where an actual signal generator, dummy load, AB763 amp, and workbench are needed to find the answer.


-Gnobuddy

 

[Kay Pirinha]

...plus a THD analyzer to get a reference point...
Best regards!

 

[Gnobuddy]

...plus a THD analyzer to get a reference point...
Best regards!

I've never had the luxury of a THD analyzer, but I can't believe I forgot to add an oscilloscope to my list! (Slaps self on the forehead.)


-Gnobuddy

 

[Kay Pirinha]

Howsoever. One needs to define a reference to which the output power relates. If we compare for instance an output power @ 4 ohms/10 % THD and 8 Ohms/1 % THD, this would be the same as a comparison between apples and oranges.
Best regards!

 

[Gnobuddy]

Howsoever. One needs to define a reference to which the output power relates.

Understood, but the traditional way (long before THD analyzers were even remotely affordable) was to look at the speaker output on an oscilloscope, turn up the signal level until the tips of the sine just started to flatten, then back off a hair. That was your maximum "clean" output.

That's certainly nowhere near as precise as using a THD analyzer, but this isn't NASA instrumentation we're talking about, just an old valve guitar amp! And the low (or no) negative feedback design means THD climbs very gradually with increasing power, unlike modern solid-state amps, where there is a very sharp knee in the THD vs output power curve.


-Gnobuddy

 

[PRR]

I was thinking about this issue earlier, and I think it almost makes no sense to define an output impedance for a device that's being hustled from cut-off to saturation every half-cycle. There isn't "an" output impedance, there is an impedance that varies continuously during every audio cycle, and varies a lot....

Assume only moderate (not exact) cleanliness.

What is the "most" Volts *and* Amps your tube can put to the load?

Taking a random 6L6GC, no-load you can swing near 360V and dead-short you can swing 350mA. Max power is at the knee and, for 400V on plate and screen, you can swing 310V and 320mA. So 1K per plate or 4K plate-to-plate, essentially Fender's loading.

If you double the load you can now swing near 350V but current falls to 175mA. If you halve the load you can swing near 340mA but only 170V.

So the best-power (large signal) impedance is near 1K.

That assumes the power supply does not sag. All practical supplies sag. Even extravagant regulated supplies have limits when you suck real hard. This adds to the large-signal output impedance. This is easier to see in simple transistor amps, where 10% nominal regulation amounts to another Ohm in series with the load, which starts to bite for 4r and 2r loads.