A Test. How much Voltage (power) do your speakers need?

[Pano]

Originally Posted by Flaesh

p.s.: woofers are equalised about +13dB @ 30Hz. so there is more voltage..

Good point. The test is simple an does not take into account EQ. If you have a boost somewhere then your voltages at that frequency will be higher. Your 30Hz boost will need about 4.5X more voltage than you measured.
Conversely, if you had a cut at 120Hz or 220Hz, your readings would be off.


Bill, thanks for your results and the good info. I'm surprised about the interaction with the meter, what type is it?

Originally Posted by wrankin

I am also not sure if the "sample music" I listened to was normalized to the same level as the tones.

It probably isn't, but it does not matter. Here's why:
Once you have set your level, it's set. Because digital playback has an absolute maximum level (a ceiling) we know the voltage can never go higher than that.
The test signal is 12dB lower than that ceiling. So the max voltage possible is 4X what you measure from the test tone. Make sense?

However, if you wanted to know the average level, then matching the test tone to the music is important. Looking at a lot of recordings, I see that most loud sections have an average level of -10dB to -12dB. Of course they aren't all like that, but those levels are typical. And that's one of the reasons I chose -12dB for the test signals.

If you have software like Goldwave, you can open the music file and read its average level. Read the whole file, or just a selected part of it. Very handy tool.

[Richard Ellis]

9.42 VAC @ 120 Hz, 8.99 VAC @ 220 Hz.
Carver TFM-15CB driving my DIY MTM, dual 6.5", Four Ohm, Trick Piezo.
Room size 5M by 5.5M.....A rather "live" room, ceramic floor, cemented walls, no particle board nor drywall ..........A decent volume, turned up just shy of onset of distortion harmonics.

__________________________________________________ ______Rick.......

[forr]

This kind of test needs a voltmeter with a voltage-peak-hold function (I own four having it...) connected across the driver or the loudpseaker and some records having modulation peaks reaching the full scale deviation (FSD).

Yesterday, I discovered this program
ocenaudio
It can provide statistics on a record :
Peak amplitude; Minimum, Maximum and Average levels in dB (awfully called "Powers").

From there, you can obtain a lot of informations about the voltages applied to your speakers, if they may ever suffer from thermal compression, if your amplifiers ever clip, etc...

This could help to extend this disccussion which deserves it :
Amplitude and frequency distributions wanted

[AndrewT]

Forr,
this kind of test does not need a special voltmeter.
The test signal is a constant amplitude steady sinewave chosen to be in the reasonably accurate range of cheap average reading DMMs and scaled to read out in Vac.

[Pano]

Yep. Exactly what Andrew said.

But thanks for the link, I'll check it out.

[theAnonymous1]

I get 1.88V at just under neighbor aggravating levels, which is a little lower than I would like most of the time. Good thing my amp is capable of 200Wrms @ 8ohm.

Speakers are Phase Technology PC80's (88dB/2.83V/1m).

[forr]

Originally Posted by AndrewT

Forr,
this kind of test does not need a special voltmeter.
The test signal is a constant amplitude steady sinewave chosen to be in the reasonably accurate range of cheap average reading DMMs and scaled to read out in Vac.

Using musical signals, knowing peaks and average voltages applied to the amplifiers and loudspeakers (and drivers in multiways) gives a larger and more detailed picture of the whole than measuring volts of continuous low frequency sinewaves according to the proposed test procedure. I think that the peak hold function can now be found on some low cost multimeters.

[weltersys]

Originally Posted by Pano

You using 120V at -12dB on speakers in your living room, Art?

I did not include a 100V+ scale because most folks here aren't trying to fill a stadium. Home systems are the aim of this test.
But your points are valid and I appreciate the data points, thanks.

I find 110 dB of LF adequate in my living room, and don't have room for a 45 x 22.5 x 26.5 speaker, and my Technics tuner sub output can probably only do 30 volts or so.
The rather efficient (around 101 dB 1 watt 1 meter) Keystone tapped horn sub has been duplicated by PASC, he is considering replacing his present subs in his home stereo with that speaker.

Lots of people get used to 120 dB of bass in car and concert systems and want to reproduce that in the home, without cabin gain it takes a lot of horsepower to do it.

The 120V test was done in my shop in an area around 20 x 16 x 9, and was off scale (over 126 dB) on my dB meter at one meter. I actually don't find 130 dB at 30-40 Hz peaks uncomfortable, though I don't listen above 110 dBC, 90 dBA in my home stereo.

My shop seems to saturate around 120 dB at low frequencies if listening in the middle portion, increasing voltage does not increase level after a point.
Wobbly walls and phase cancellation.

Of more usual home stereo interest, the Acoustic Research AR2 at 10.9 volts was pretty anemic, only about 100 dB at one meter, a pair in my shop (or living room) may have done about 95 dB maximum at normal listening distance.
I think the AR2 was a 4 ohm speaker with a 87 dB rating, 10.9 v is around 30 watts, and that was as much as it could take without exceeding 10% distortion at low frequencies.
They definitely ran out of gas below levels I like with some types of bass heavy music.

Art

[weltersys]

Originally Posted by forr

Using musical signals, knowing peaks and average voltages applied to the amplifiers and loudspeakers (and drivers in multiways) gives a larger and more detailed picture of the whole than measuring volts of continuous low frequency sinewaves according to the proposed test procedure. I think that the peak hold function can now be found on some low cost multimeters.

Unfortunately, the integration time on multimeters is far to slow to read musical peaks with any detail or accuracy.

[Pano]

Originally Posted by theAnonymous1

...just under neighbor aggravating levels, which is a little lower than I would like most of the time

LOL
Thanks for posting.

Originally Posted by weltersys

Unfortunately, the integration time on multimeters is far to slow to read musical peaks with any detail or accuracy.

I agree. It would be nice if you could do this, but I would not trust the meter on music. An Oscilloscope, yes.

However: If you think you need to measure peak values, you have missed a fundamental part of the test. We already know where the peaks are, they are 12dB higher than the test tone. That's 4X the voltage. Where else could they be?

The test tone could be at any other level, as long as you know what that level is, you know where the peaks will be.