precise power meter

[pfloyd_is_god]

I’m planning on building an accurate power meter for my guitar amp I’m building. I’m sure one can’t rely on p=v^2/r for a speaker, as the impedance is going to change with frequency. So what I’m planning on doing is taking a current signal and a voltage signal from the speaker, multiplying them, and integrating this for an rms output. However, as a speaker is an inductive load, the current and voltage signals will probably be out of phase, so I’ll integrate the voltage and current signals first, and then multiply them, is this mathematically accurate?

Secondly, what should I be looking for in terms of ADC specs? How much resolution do you need to accurately measure power output? I’m sure you don’t need 16 bits, and the same for sampling rate, I have my doubts that the high end of the spectrum adds much to power.

Attached is the basic analogue circuitry I’m aiming for. Will the current sensing resistor, R1 degrade the performance of the system much, other than reducing output power? One could also put it before the speaker, after the amp, and use another differential amplifier to take the voltage drop.

any help is apreciated, cheers

[rpapps]

Hi
Why not sense the current inductively using a current transformer?

Pass the speaker lead through a ferrite toroid a few times and wins as many turns of a small diameter wire as you can fit on the toroid as a secondary. Now secondary voltage is proportional to primary current. Pass a known current through primary and measure secondary voltage and you'll have your scaling factor.

You might like to look at analog multipliers as well. Would save a lot of digital hardware.
Just a thought.
Rob

[destroyer X]

That music is one excellence.... magnificent, astonishingly, incredible... a classic of our days.

Not so simple as good...no , not, non, não, niet, nichst, nein!

regards,

Carlos

[pfloyd_is_god]

Forgot to attach the image....

Why not sense the current inductively using a current transformer?

From what I gather, current sensing transformers are designed for applications which require limited accuracy but high current capacity, whereas a current sensing resistor will be more precise, but will be less practical when dealing with high current, p=I^2*r. The voltage drop across a 10 mOhm resistor for my amp shouldn’t go much over 80 mV, that’s about 700 mW of power dissipation, I can live with that.

I could go for an analogue solution, but where’s the fun in that?, I'm most likely going to be using an MCU for other reasons anyway.

Please, correct your text, Pink floid (floyd) is not good

Are you seriously insulting the Floyd?

[runebivrin]

I think Carlos read "good", rather than "God" in your user name, and he's saying they're better than just "good".

Reun

[djk]

Is there a point to all this?

Why don't you make something useful, like a circuit that measures voice coil temperature rise?

Speakers go into power compression and eventally burn up due to the coil getting hot.

Put a 0.05 ohm sense resistor in the ground side and amplify the voltage. Divide down the voltage from the amplifier hot side until equal. Drive a set of differential comparators. The output of the comparators could be an LED ladder, say 5 green, 3 yellow, and 2 red.

If this is for a 'class', you could make it a lot more complicated by using a µP.

[Upupa Epops]

Whatabout phase shift in voltage sensing ?

[destroyer X]

And you texan...hehe...

have ten steps, turn and shot!

having ink pistols please!

I love Pink Floyd...he will be the Classic Music in near future.

carefull with brazilians.... we will turn in the first step!!!!

regards,

Carlos

[pfloyd_is_god]

like I said, I'm going to use an micro processor, and I’m going for a little more precision than a 3 stage bar graph.

What about phase shift in voltage sensing ?

Hence the integrate voltage and current before multiply, I’m fairly confident this will work. So does anyone have any ideas on how much dynamic range and bandwidth I need?

[jackinnj]

if you want to acurately measure power consider using and AD737 or AD536 True RMS Converter -- they do the multiplying, square - rooting etc. for you. The AD737 is 200mV max input and requires an input attenuator, the AD536 doesn't.