Amp Wattage Measurement - Approximations

[HaroldBrooks]

Saw a video several months back and a gent claims you can measure the A/C voltage from your speakers with a multi-meter, measure the ohms of the speaker(s), and then use Ohms law to approximate the power (at any level of distortion you like).

I recently applied this technique to one of my amps and while playing I could keep it right around 15 volts without too much trouble (sometimes higher, sometimes lower), so at least in terms of my meter, and for what I was playing (single notes in the key of B, the average was 15 volts.

So using ohms law I would come up with 15vac / 8ohm = 1.875amps and then, 1.875 amps x 15vac = 28 watts (with a highly distorted signal in this case).

Or just use a shortcut of squaring 15vac divided by ohms ; (15^2) / 8 = 28 watts

One thing I got from all this is that wattage is a moving target, and depending on what you play it's quite variable, not to mention the efficiency of the speakers and cabinet you use. So SPL can vary tremendously regardless of what the wattage spec 'says' the amp should be doing.

Is there any validity to this method of measurement as an approximation, or is it all rubbish unless you measure things with a scope, a dummy load, and with a specific frequency ?

Thanks for any input !

Here's the video from Gerald Weber.
https://www.youtube.com/watch?v=1b2jQWK8xlQ

[Leo_Gnardo]

Thanks for keeping the punchline 'til the end.

[Old Tele man]

Just remember, it's AVERAGE power in Watts, not RMS power!

Root-Mean-Square (RMS) is merely the mathematical conversion technique that's applied to sinusoidal VOLTAGE and CURRENT to derive their "average" equivalent value(s). And, thus, an average value will consist typically of values above and below that average value.

[Justin Thomas]

But... But... But what was the answer???

I have a much simpler way of measuring power that only requires my guitar, an amp, & me.
"I can play this in my bedroom without waking anyone...." = GARBAGE!
"This works for jamming in the living room..." = 25-50W.
"Hey, I can invite my rowdy drummer buddy over!" = 50-100W.
"What? WHAT? SPEAK UP!!!" = Holy Crap that's awesome!

In all honesty, I've never had to measure output power. Granted I'm just a tinkerer who can usually help your old tube amp work right, but my ears do good telling me whether an amp is "working" or not.

Jusrin

[Helmholtz]

The standard method requires:

- a load resistor corresponding to nominal output impedance
- a continuous sine input signal of around 100mV and 400Hz
- an AC voltmeter (preferably true RMS), simple AC voltmeters mostly give viable RMS readings at 400Hz sine but not much higher; they can't deal with multi-frequency or non-sinusoidal (distorted) signals - while a good true RMS meter might.

RMS (sine wave) output power is given by Vrms*Vrms/Rload. A scope is used to determine the onset of clipping. You may avoid the scope by connecting a speaker (or headphones) wired in series with a 470 Ohm resistor and a 0.1µ cap over the load resistor and detect the onset of clipping by ear.

It is no good idea to:

- use a guitar signal as its not continuous/sinusoidal and contains many frequencies which is likely to give a wrong meter reading
- use a speaker load as its impedance changes with frequency and speakers typically are not able to stand full sine power.

[HaroldBrooks]

But... But... But what was the answer???

I have a much simpler way of measuring power that only requires my guitar, an amp, & me.
"I can play this in my bedroom without waking anyone...." = GARBAGE!
"This works for jamming in the living room..." = 25-50W.
"Hey, I can invite my rowdy drummer buddy over!" = 50-100W.
"What? WHAT? SPEAK UP!!!" = Holy Crap that's awesome!

In all honesty, I've never had to measure output power. Granted I'm just a tinkerer who can usually help your old tube amp work right, but my ears do good telling me whether an amp is "working" or not.

Jusrin

I too am more impressed by an amps actual performance, and less by just wattage, as I know speaker efficiency, cabinets, and the chosen frequency curve exert a tremendous influence on the attainable "Loudness" and perceived volume...

But it's also nice to know if your amp is functioning within the realm of normalcy given it's design. I would feel depressed if I had an amp with two 6V6's or 6L6's that wasn't churning out half of what it should wattage wise, and want to know what was wrong so I could fix it.

So I was hoping the Weber method was a quick approximation, or something that could be used a comparison, not an absolute and not required to be all that accurate, but hopefully of some use.

[nevetslab]

But... But... But what was the answer???

I have a much simpler way of measuring power that only requires my guitar, an amp, & me.
"I can play this in my bedroom without waking anyone...." = GARBAGE!
"This works for jamming in the living room..." = 25-50W.
"Hey, I can invite my rowdy drummer buddy over!" = 50-100W.
"What? WHAT? SPEAK UP!!!" = Holy Crap that's awesome!

In all honesty, I've never had to measure output power. Granted I'm just a tinkerer who can usually help your old tube amp work right, but my ears do good telling me whether an amp is "working" or not.

Jusrin

Very few of us have a multimeter that has VERY FAST reading rates. Most economical Multimeters are 2-3 readings per second, to begin with. Using a Guitar as your signal source produces a complex waveform that has an initial peak envelope, then rapidly decays from that initial waveshape to those of the notes/harmonics that trail off, prior to striking more notes or chords. The loudspeaker used with the amp is NOT a fixed impedance, but has an impedance cuve that gets into the vicinity of the 'nominal' impedance in the range of 150Hz, with a broad Q that's in that impedance for about an octave (1/2-oct either side of 150hz, typically).

Now, if you had two FAST reading multimeters, one to read AC Volts, the other to read AC Amperes, then you wouldn't need to be concerned with the impedance curve. But, then your task is WHEN do you take the reading of both the AC Volts and AC Amps of the chord you struck on the guitar? You would still have to take those two numbers and multiply them to arrive at a power level. The meter reading rate is the basic problem, as is the envelope of the guitar 'signal'.

While I have an AC Power Analyzer that reads True RMS, that measures both AC Volts and AC Current, and displays the results in Watts (Valhalla 2101), it too has slow reading rates....also in the 2-4 readings per second. Point is, you'll never get a steady reading that's meaningful. Now, you could sit there and play the same chord over and over and decide to choose the highest reading. Doing this with your two multimeters, that may only be AVERAGE reading and not True RMS reading, trying to get the max readings on both while you're playing the same chord over and over is still a crap shoot. Neither instrument is reading at the same time, most likely, but still, you'll eventually arrive at some peak reading on both, which then have to be multiplied to arrive at that Average Wattage for that moment.

It's really a complex instrumentation problem. System Multimeters have reading rates upwards to 100 readings per second, and due to their sophistication, will also have TRUE RMS conversion, and can either store or output the data stream. And, can be sync'd so one is reading voltage, the other reading current at the same time. Feeding both into a multiplier circuit that can output the results in a calibrated DC level, scaled in Watts RMS, it can be displayed on a CRT as an instantaneous varying wattage over time. Very few of us have such facilities at our disposal.

Oh, you can get some numbers with just a multimeter reading the AC voltage off the loudspeaker, and compute what that number represents. As to it's accuracy.....best of luck.

[g1]

It is no good idea to:

- use a guitar signal as its not continuous/sinusoidal and contains many frequencies which is likely to give a wrong meter reading
- use a speaker load as its impedance changes with frequency and speakers typically are not able to stand full sine power.

I think the bolded part here may be the most critical reason why this method is not reliable. You need a verified load resistance.

Have a look at this graph. The lower curve (and numbers on the right) are impedance of the speaker.

[Helmholtz]

but has an impedance cuve that gets into the vicinity of the 'nominal' impedance in the range of 150Hz, with a broad Q that's in that impedance for about an octave (1/2-oct either side of 150hz, typically).

AFAIK, nominal speaker impedance is measured/defined at 400Hz. It often stays within around 20% between 200Hz and 1kHz and rises below and above. A 4x12 Marshall cab has an impedance of around 40 Ohm@150Hz rising to 65 Ohm at its 120Hz resonance.

[J M Fahey]

GW is a confirmed trick man, just like that.
Or uses typical Politician tricks to catch audience... same thing.

His video is crap useless, all it tells you is that amp is "workung" ... which you can check yourself by ear.

Proper way to measure is to simplify, to reduce variables to minimum, not wildly varying elements inn the equation, so:

* with a plucked guitar you never know: frequency - amplitude - duty cycle while a continuous sinewave is well defined and, most important, repeatable.

* a speaker load has wildly varying impedance, while a resistor is stable and uniform across the frequency band.

* a sinewave has a defined peak to average valuie; a plucked guitar has *anything* depending on picking style.

* a sinewave has constant level, a guitar literally jumps all over the place.

If anything, and in a musical environment, a pressed down organ key will be acceptable.

Helmholtz trick to listen to clipping edge works well, in fact not even the series resistor is needed, since the small capacitor is high impedance enough in this test and all you need to hear is appartiton of buzz.
Personally I use a cheap piezo tweeter connected straight to load and kept near my ear.

That said, transistor amps have sharp defined clipping, they either have 0.01% distortion or 10% clipping just 1 volt away so this method works well; tube amps have 5% distortion even without visible clipping, and still sound relatively "sweet" , to boot Pentodes compress, period, that again blurs the frontier , so for tube amps it´s better to use scopes.

Since you wantb to "see" , not that much "measure" (with the scope that is), any software scope works fine.
I don´t mean those USB "digital scope boards" widely available today but simple software which turns your PC into a basic scope.
Only 2 problems: thay can not display DC levels nor go beyond 20kHz, ... because PC soundboards can not either, no big deal to *watch* aguitar soundwave.

I also published a simple and safe attenuator to connect speaker level signals to microphone/soundcard input, it works very well

search for scope-died-anyone-use-the-pc-based-ones at The Gear Page.

Can not straight link there because my IP is blocked on that board.

FWIW this is the deluxe multi range version, but the simplest one has just 2 resistors, 1 capacitor, 2 cheap diodes.

[HaroldBrooks]

Mr. Fahey, glad to see your back on-line.

Not to hijack my own thread, but I heard there was a huge power outage, not unlike the one on the east coast of the USA many years back.

[nevetslab]

AFAIK, nominal speaker impedance is measured/defined at 400Hz. It often stays within around 20% between 200Hz and 1kHz and rises below and above. A 4x12 Marshall cab has an impedance of around 40 Ohm@150Hz rising to 65 Ohm at its 120Hz resonance.

I don't recall the 400Hz Standard Test Frequency for defining nominal impedance. I think speaker mfgrs use the 0 phase frequency in the low midrange trough where impedance Q is broadest. Changes with each speaker type. I had stated 150Hz as the most recent impedance curve I had recently looked at & posted was that of an Ampeg BXE-410HL4. Here's a variation of speakers besides that one, three being free air, plus a Yamaha NS-10 speaker and that Ampeg bass speaker.



Regardless, in the exercise that was initially suggested, changing to a fixed resistor makes it far more predictable.

[Old Tele man]

Just to add some more mud to the waters: AMPS and SPKRS are (unfortunately) spec'd differently, for example, on many Fender schematics, it's 400Hz for bass amps and 1kHz for guitar.

[J M Fahey]

Mr. Fahey, glad to see your back on-line.

Not to hijack my own thread, but I heard there was a huge power outage, not unlike the one on the east coast of the USA many years back.

Thanks.
Luckily it lasted only 4 hours and apparently followed the exact same path as the US one: branches trying to cover for main line sections and dropping themselves.

[Helmholtz]

I think speaker mfgrs use the 0 phase frequency in the low midrange trough where impedance Q is broadest.

A speaker impedance has (at least) two zero phase frequencies. The lower one(s) at the bass resonance(s) and the next typically around 400Hz for a guitar speaker.
According to Wikipedia some manufacturers just use the middle frequency impedance minimum.
A Q-value can only be defined for a (bass) resonance, defining the half-value width of the peak.