How to measure watts?

[dano12]

....say from the output of a LM386 or other small wattage amps?

I googled and found a confusing array of mystery text.

Is there an easy way to do this?

Edit: to clarify, if

P= I x E (power in watts = amps times volts)

Where do I measure amperage and where do I measure voltage?

[frequencycentral]

If your neighbours come around to complain.................it's more than 1 watt.

[dano12]

If your neighbours come around to complain.................it's more than 1 watt.

Lol, excellent response.

[Boogdish]

I would think you'd want to measure it on the leads going to your speaker, since you're measuring how many watts are going into it.

[R.G.]

You measure it at the speaker leads.

The simpler thing to do is to measure the voltage output when you feed the amp a low-frequency sine wave and compute from there as though it was a resistive load.

The power into a resistive load of R ohms is P = (V^2)/R where that reads "V-squared", and it is AC volts RMS. Your DMM, unless it's the hugely expensive true-RMS kind, reads approximately volts rms on the AC scale. High frequencies can confuse the cheaper meters that actually measure rectified average DC volts and scale them to look like the RMS of a sine wave, so it only works well for low frequency sine waves; the intended application for these meters is measuring AC power line volts, and they're pretty good at that. Anything under about 500Hz sine wave works OK.

So you measure the AC volts out and measure (or listen) to get it just before it starts distorting, and then compute the square of the voltage divided by either 4, 8, or 16. That's the power out into the various resistive loads.

Speakers are not noticeably resistive loads, but they vary from speaker to speaker and type to type, so the world gave up and just presumes resistive for testing amplifiers.

[waltk]

R.G.

Thanks for the simple explanation.  Suppose I have:

a true RMS meter (Fluke 87III),
a simple test oscillator (homemade with ~1Khz, and three output levels),
and a decent LCR meter (that measures Z at 100hz, 120hz 1Khz, 10Khz)

... how would your answer change if I wanted a more accurate measurement?

[frequencycentral]

If your neighbours come around to complain.................it's more than 1 watt.

Lol, excellent response.

...and you can add 1 watt for every house along the street that comes round to complain. Ok not as technical as R.G.s response. But far more 'rock n roll'.

[Ronsonic]

R.G.

Thanks for the simple explanation.  Suppose I have:

a true RMS meter (Fluke 87III),
a simple test oscillator (homemade with ~1Khz, and three output levels),
and a decent LCR meter (that measures Z at 100hz, 120hz 1Khz, 10Khz)

... how would your answer change if I wanted a more accurate measurement?

You would still run the amp into a known load, give it signal and increase the level to the maximum before distortion, measure the AC voltage, square that and then divide by the load resistance. A 'scope would make it easier to determine when it starts to clip. You may have to fiddle between the gain and tone controls and the output level of your sig gen to get max undistorted output.

Myself, I rarely use 1K Hz, there just isn't a lot of music up there. We inherited that as a standard from ancient communications and radio gear. 200 - 400Hz looks a lot more like a guitar signal.

[waltk]

You would still run the amp into a known load, give it signal and increase the level to the maximum before distortion, measure the AC voltage, square that and then divide by the load resistance. A 'scope would make it easier to determine when it starts to clip. You may have to fiddle between the gain and tone controls and the output level of your sig gen to get max undistorted output.

Hi Ronsonic (and apologies to dano for butting in on your thread),

Thanks for the reply.  I guess I was wondering whether there would be any advantage to measuring the actual impedance of the speaker at a given frequency and using that as your "known load".  Would that change the equation (dividing by Z instead of R)?

I've always thought it would be cool to have a 'scope, but with my current level of understanding, the sheer number of knobs on them is a bit intimidating.  I have a lot to learn before that would be useful.

[R.G.]

I guess I was wondering whether there would be any advantage to measuring the actual impedance of the speaker at a given frequency and using that as your "known load".  Would that change the equation (dividing by Z instead of R)?

Then you would have an accurate answer - for one speaker, at one frequency. Since the Z varies with frequency - a lot! - and since speakers, even nominally the same type/number are different, you'd have to measure every speaker. Audio people gave up on that one a long time ago, and just assume the speaker is resistive. We know it's not, but it's better than a 600 page report on "what power does your amp put out?".

[petemoore]

I have a lot to learn before that would be useful.
 Getting around a scope isn't terribly 'thick' for what you want to do, from what I understand.
 They can be hard to ballpark. SS amps which require a certain ohmage load might sound much less loud than a tube amp of <1/5th the wattage rating IME, even when at min-ohm loads, that's no way to tell 'how loud' even.
 Brands tend to be 'linear' in amp naming, a '10' amp will be smaller than a '40' and 40 smaller than 100.
 There was much talk of English and American watts back when ['70's], around here, but that faded out.
 I just go by the # of and type of tubes to get a ballpark idea.
 Most of the SS amp ratings seem to be somewhat congruous with one amp to the other, the 'min-ohmage' deal can make a big difference in how much you can actually get out of any particular amp though.
 It's an 'average' thing to discuss for the most part, frequency, picking strength attack and many other things can cause the output to rise and fall...very far.
 It becomes fairly immediately evident [such as experiencing being drowned out, or not being able to get clean sound], and a signifigant increase in available watts is the solution.
 Which is no help at all for explaining how to measure watts...