Measuring amplifier output power without a scope.

My vintage scope broke down (probably PSU caps are dead), but I need measure the power output of my DIY half power Bassman (with 6v6).

I had an idea, of taking a multimeter, setting it on AC peak volt metering, and increasing the 1kHz output into a 8ohm 100W resistor, until the voltage stops to increase. When the distortion starts to kick in, you shouldn't see increase in peak voltage, right?

Then I could measure RMS value, and calculate the power as usual.
I guess this would only work if the clipping is symmetrical...

Or any other approaches that work reasonably good? (accuracy of -+1W is good enough).

> until the voltage stops to increase

IF it is truly a peak detector, that does work. (If it is an averaging meter with markings 1.4 times bigger, it is still an average.)

When I had a passive needle-meter and a VTVM (peak-to-peak) both on the same output, I could clearly see the p2p meter top-out while the averaging meter continued to rise.

But tube amps don't clip real-sharp like transistor amps. They may rise some even when clipping is audible. They may over-shoot and show false peaks.

Ears are sharp tools. Hook your dummy load, then also hook a clean speaker with about 100r-1K of series resistance across that. You will hear a small version of what the amp puts out. Gross distortion will be perfectly audible. Problem is calibrating your ear to discount the mild distortion that simple tube amps make at medium levels (below clipping). Wind it up until really bent, wind-back to pretty-clean, and explore the edge of distortion.

If you know what the amp sounds like when over-driven happily, you can probably find that same spot at the lower level of this monitor speaker.

Two-6V6 does not need 100W dummy load.

Guitar amps are NOT critical items and +/-1W on a 10W-20W amp is excessive precision; I'd only look to be sure I hadn't built a 3-Watt amp with 20w worth of parts.

it might be nice to have a sheet of paper with volts vs watts into a fixed load so you do not have to round up a calculator every time you need to know power, so if

Voltage=Current * Resistance  (Ohm's Law) and Power=Voltage * Current

rearrange Ohm's Law:  Current=Voltage/Resistance and plug this into the power formula:

Power=Voltage * (Voltage/Resistance)  so

Power = Voltage^2 / Resistance              or P=E^2/R for geeks 

so for 8 ohms,

Power=V^2 / 8    16 Ohms: P=V^2/16 and  P=V^2/4 (4 ohms)

don't need peak  volt meter, speaker idea is cool, 
or just  advertise square  wave power like everybody else except Marantz 

usually sq wave power is close to sine wave power, maybe 5 to 10 percent more for sq wave but who is counting with a complex guitar signal full of peaks and valleys, only way to measure that power is with a piece of Nichrome wire in an oil tank and measure heat rise, 
and there should not be much assymetry in a P=P amp,

we use an old school Weston AC volts analog meter from the junk pile  which has had the back plate pasted with a hand  drawn wattage scale , this was done while using the DMM to read volts which was translated to watts which was written on  the paper scale while the Weston was also hooked up.
now we have a dedicated power meter which does not  tie up the DMM,  no batteries required and we need not remember formulas nor do we need a calculator. when we flick the switch for another 8 ohm-100 watt resistor to come into play as a parallel resistor we simply double the power written on the scale for this four ohm load.

if your lazy click open a spread sheet and build a table of values, then print it out>

> usually sq wave power is close to sine wave power

I call "different".

On a large SE amp I could get 13W pretty-clean, 17W pretty-bent sine-like, and 23W clipped to the max, based on eyeballing a 'scope cross-checked with a meter.

For a "prefect" amplifier, square would be twice sine, which meshes with my 13 and 23W numbers. (SE tube doesn't make excellent squares.)

your right,

the reason we did not see a ton of difference is the sine wave gets the tops flattened so even though it still resembles a sine wave, it is getting ready to turn into a sq wave,

they have integrals for this but they are lost in the cobwebs, probably a piece wise integral and a sine integral,

lets see, for a 1 volt signal at 1 hz, we would have a value of 1 from 0 to pi radians for the sq wave,

a sine wave integrates into cosine, and would have a value of ? from zero to pi radians,

now if we could just get an amp or fuzz box to put out a sawtooth wave...

Thank you for your advise.
I hooked up the high power 8 ohm resistor to the output with a speaker and 470 ohm resistor across it. I increased the volume of the amp just until the sine got a little raspy.

And just as I expected, the power output is way too low.
It seems the output is only 6W with 6v6 tubes and 12W with 5881 tubes.
The bias seems to be spot on - about 40mA for each 5881 tube with plate voltage of 440V (so it's about 17.5W of plate dissipation)

What could be the problem? Pretty sure it's not about the preamp stage, otherwise there wouldn't be increase in wattage when I changed the tubes to the original 5881 configuration. Also not about power tubes or the rectifier, since I swapped them. Something wrong with the phase inverter? Tried changing the phase inverter tube also - still the same result.
All the voltages are within tolerance of the original schematic.

At least the half wattage concept works  ;D

EDIT: I created a new topic for further troubleshooting.