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

[Pano]

If you have trouble understanding this test, or you have questions, please read:
TEST FAQ
Update July 3rd 2012:
Although the title of the tread contains the word "test" this is better thought of a s a survey or poll. The word "test" has caused a good bit of confusion and misunderstanding.
The test tone provided is not used to set you system levels, it is used to measure your system levels. You will set the levels by ear, or by pink noise if you wish.
What you will be doing is measuring your actual listening voltages and reporting them here. The test tone allows you to do that with very good accuracy using only digital playback and a voltmeter.

---

100Also have a look at Archimago's blog for a very clear explanation of this test, along with good photos and link to more info.

---

100
Here is a simple test to determine what voltage your speakers need to play their loudest. From that you can determine how much power you need.
Knowledge is good and this test will tell you just how much voltage (or power) you really need. Efficient speakers in a small room may need only a volt or two. Less efficient speakers in a big room may need a dozen volts or more.
I've included a poll so that we can see what's typical. If you want to also post your speaker details and room details, that would be nice.

For the moment, I'll leave this as a voltage test, we will talk about power later. Those of you who know Ohm's law can figure it out for yourselves. Amp power tends to be an emotional issue, so we'll post the results in volts for now. Since most amps are voltage devices anyway, it's not a bad way to measure. And a voltage measurement removes the speaker impedance from the equation.

All you need for this test is some sort of digital playback (CD, iPod, Squeezebox, computer, etc) and a volt meter with a low AC scale. Any scale 20V or lower should be OK. Meters with only 200V AC scale won't be accurate enough for this test.

Next post has the test tones and instructions.

 

[Pano]

The Test Tones

Attached you will find 4 test tones files in MP3 format. They are simply sine waves in 44.1K/16 bit format. The tones have an RMS value of -12dBFS. There is 120Hz for mains with woofers and 220Hz for systems that use a subwoofer. EDIT: Recently added, mono file for subs with 10 secs each of 25Hz, 30Hz, 40Hz, 50Hz, 60Hz. Also included is a 20 sec sweep from 25Hz-120Hz. With this you can look for any peaks that EQ might cause in the low end.
Check these bass tones at a low volume first. Sine wave tests can be hard on bass drivers. Use caution before testing at loud settings.

• Download one of the files below.
• Unzip the MP3 and burn it to a CD or put it in whatever digital playback gadget you use.
• Select and play back a few dynamic music tracks as loud as you ever do (like rocking out after a few beers.) Take note of your volume settings.
• Use the highest setting you found when playing back music really loud and leave the volume there for the following measurement.
• Playback the test tone and measure the voltage at your speaker terminals. Measure at the amp or at the speaker, either is fine.
• Post the voltage you measured.

The principal of the test:
With digital playback, there is a maximum value "set in stone", so to speak. That value is 0dB. All other levels are below this. Knowing that, it's easy to figure out what the peak level is; it's 12dB more than your measured voltage. That means peak is 4X the voltage. That's as high as the voltage to your speakers will ever go.

Because you've played back a few CDs and determined the maximum level you use, the loudest level you ever want, you've established your peak voltage.
Playing back and measuring the test tones will tell you what that peak level is. The peak voltage will be 4X higher than your measurement.

We use a sine wave to test with (after setting the level) because it's easy for a voltmeter to measure accurately, unlike music. All the meters I've tested have been accurate at 120Hz and within a small fraction at 220Hz.

Thanks for taking the test. It will be interesting to see what voltages are used across a wide range of speaker types, room sizes and musical tastes.
Please post the voltage you measure and fill out the poll. Thanks!
**********************************************************************
EDIT. 3 YEARS LATER
For ease in figuring out what power amp you need to get enough voltage to your speakers, here is the guide from post #77

How do your test results relate to amplifier power? Here's a little trick for you:

Take your voltage measurement from the test tone
Square that number
The result is the RMS power rating (@8 ohms) that you need for your amp not to clip.

Easy!

For example:

You play the test tone and measure 3.5V AC
3.5x3.5=12.25
You need an amp rated at 12.25 watts RMS (minimum) to play your "loud enough" level and not clip.

Easy!

It's the law, Ohm's law.

 

[pinkmouse]

Right, a quick test on the second system shows that a good loud level in a 4x3x3m room takes 2.9V. However, this is taken on the 8ohm low mid driver of an efficient, (92dB), actively crossed over sub/sat system, so not typical I suspect.

 

[tuxedocivic]

So are we supposed to post our measured voltage, or the measured voltage x 8?

You said the peak voltage will be 4x's the measured voltage and then later said 8x's, but I'm pretty sure 12db makes it 8x's, right???

 

[Pano]

Please post your measured voltage. We'll do the math to figure out peaks and averages. Sorry if that was not clear.
The poll is for measured voltage from the -12dB.

12dB voltage is 4X. Thanks for catching the typo, I fixed it.
Peak will be 4X your measured (posted) voltage. Maximum RMS of a sine wave will be 2.83 x higher than the test tone.

 

[Pano]

Right, a quick test on the second system shows that a good loud level in a 4x3x3m room takes 2.9V.

Thanks Al. That sounds about right for 92dB and that size room.

 

[tuxedocivic]

Ah, ok. Thanks. Will try this later. I have insanely ineffecient speakers (82db, 8ohm), so it'll be interesting to see what I come up with.

 

[Pano]

Yes, I'm very interested to know! Thanks.

 

[weltersys]

Here is a simple test to determine what voltage your speakers need to play their loudest. From that you can determine how much power you need.
Meters with only 200V AC scale won't be accurate enough for this test.

I have tested a rather efficient (around 101 dB 1 watt 1 meter) four ohm tapped horn speaker (B&C 18SW115-4) with 120 volts at 60 Hz, a 200V AC scale worked fine at that level .

That speaker could have used more voltage to play louder, but the amp shut down after a few seconds each time I tried three times in a row, and I don't have an amp capable of more voltage swing than 120 volts.

An AR-2 hits about 10% distortion at 45 Hz with 10.9V.
A pair burn up quickly with 64V at 60 Hz applied .

I once burnt an Eminence Alpha 8 using about 35 volts, would that be too low to read accurately on a 200V scale ?

Art

 

[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.

 

[tuxedocivic]

I'm not sure it's doing it right for me, it sounded very quite. Like, my fridge??? I got 1.2V. But seems wrong. Even with 12db of headroom, it was much less than the average loudness of the music. Thoughts?

 

[Pano]

Thanks for trying it. Yeah, it might sound like your fridge. And it won't sound that loud if your system is clean. I was surprised at the low volume myself.

1.2V does seem very low for your 82dB speakers. It's close to what I get with 96dB beasts. Did you set your volume to its loudest normal setting? If in doubt, try the 220Hz tone.

 

[Radugazon]

Very good idea for a poll !

at the beer & rock party level, on a digital tri amp system, I read :

220 Hz : 2.02 V on the midbass, 0.13 V on the woofers
120 Hz : 1,77 V on the midbass, 0.87 V on the woofers

120 Hz sits at the cross over, grossly at -6dB on each output, so it's the 220 Hz that has to be retained.

The room core is 9 x 12 x 5 (meters). Listening distance is 7 meters, mids efficiency must be around 96dB. Drivers for both channels are 8x15" for the woofers and 10x12" for the mids, so the acoustical level can be considered as louder than the relaxed listening. The tones themselves were very loud, at the limit of being a nuisance.

 

[Pano]

Great, thanks. So I'll count you as basically 2V, as that's the highest level you found.
Those of you with active crossovers will find this test a bit more difficult as you can't just measure at "the amp". You have to measure each section. But I hope that it teaches you something about your levels.

I can supply tones at other frequencies, if needed. Above about 400Hz I don't know how accurate volt meters are. Some will be, some won't be. Lower tones are not a problem.

For those who own an oscilloscope, you can set your volume to loud and look for the peak voltages on music. Once you find the peaks, just divide by 4 to arrive at the -12dB of the test.

 

[tuxedocivic]

I figured out what was going on. The volume on WMP was turned down. My computer volume and iTunes volume was up, but WMP was down. Just turned it up and re-tried, I got 2.4V. Much better sounding.

So I've voted now. I didn't before because I was suspicious.

 

[Pano]

Cool! Glad you figure it out - and thanks for posting. That's less than I thought you'd get.

 

[Coconuts 500]

Just measured and I got over 9000 volts!

 

[Pano]

Electrostatic speakers, eh?

 

[Flaesh]

I've voted "2 or less".
I very rarely listen so loud. usually more less (big green smile were here. For some reason (???)emoticons not work in my posts.. "Disable smilies in text" is Off).
p.s.: woofers are equalised about +13dB @ 30Hz. so there is more voltage..

 

[wrankin]

Interesting test - a couple observations and then some data.

AV receiver did *not* like the additional capacitive load when I ran the meter in parallel with my 4-ohm speakers and went into some sort of oscillation feedback. Had to make all measurements with the speakers disconnected, just driving the meter.

I am also not sure if the "sample music" I listened to was normalized to the same level as the tones. I tried to pick music with a fair amount of dynamic range. All MP3s were played through my Tivo from my home music server.

Speakers - DIY 2-ways with Vifa M18 (4 ohm) woofer and D27 tweeter. The LEAP printout show it about 89dB efficient at 220Hz, but I'm not sure if that is per watt or at 2.83VAC.

Room - Large and odd shaped. Roughly 24x28 feet with a 24 foot ceiling peak. 30% of that space has a loft over it. There is also a open kitchen area to one side with an additional 300 ft^2.

Speaker front baffles are placed about 18 inches out from the wall with about 8 ft. of separation. Listening position for "laid back and loud" was about 15 ft. Critical listening tends to be done from around 8 ft.

Music selections were from "James Newton Howard & Friends" and Mannheim Steamroller's "Fresh Aire II".

General (normal family) listening levels are at -50dB on the AVR display, with a measured level for the 220Hz tone of .818 VAC.

Critical listening levels are around -40dB, with a measured output of 2.58 VAC.

Loud, but not tiring levels were around -34dB with 5.05 VAC output.

Loud "rock out" levels were at -30dB with 8.02 VAC.

Hope this helps,

-bill