Myth Busters: 1000W amp is only twice as loud as a 100W amp

[cotdt]

Originally Posted by Amit_112dB

Considering 10db is twice as loud, here is another fact

10% THD is 50% distortion, in terms of loudness

assume 100 w amp, 10% THD means 10w noise out of 100w
10w is 10db less than 100w
So 10% THD is 50% loudness distortion

10% THD is barely perceptible... try it for yourself. At 5% THD, you won't even pass the blind test.

[Telstar]

Originally Posted by FE3T

If i remember correct a 3db gain is what you need notice a gain in loudness.

This is BS, I clearly notice about 1dB steps if not less.

[djk]

10% THD would be 20dB down.

"This is BS, I clearly notice about 1dB steps if not less. "

Depends on the circumstances.

I made a stepper attenuator for an HF driver that went in 1/4dB steps. While the effect was clearly audible, no one could say if I had raised or lowered the HF by 1/4dB, though all would say that the sound had "gone out of focus" with the change.

[JMFahey]

Sorry but distortion meassurements compare Power, not Voltage, so 10% (distortion/noise/whatever) is 10% the power of the full signal, = -10dB.

[djk]

My HP distortion analyzer must be defective then.

[AndrewT]

Originally Posted by djk

My HP distortion analyzer must be defective then.

HP know what they are doing, Your analyser is OK.

[Charles Darwin]

Originally Posted by djk

"he passive crossover ensures that much of the harmonics are delivered to the treble and mid drivers."

And the clipping is largely at the lower frequencies and the harmonics feeding the tweeter are trivial (assuming a good crossover).

"Clipping destroys treble drivers. "

You still have not proved this, and direct examination of failures does not show damage from excess long-term average-power.

The primary failure mode seen on most tweeters is mechanical failure of single-strand lead-out wires. In other words, they don't burn up, they break.

Cone motion increases at a 12dB/oct rate as the frequency goes down. Program material can increase at a 6dB/oct rate going down to about 250hz or so. Unless the tweeter has an 18dB/oct crossover the lead-out wire can fracture. Most speakers having problems had either a 6dB or 12dB crossover, and a single-strand lead-out wire. I was astonished the first time I put a spectrum analyzer on a tweeter after repair. Playing program material it looked like the crossover was around 500hz, not the 4.5Khz it showed with pink noise. This was a 6dB crossover, and the single-strand lead-out wire tweeters dropped dead on a regular basis.

Feel free to come over and I can show you exactly that with my analogue synth.
If the eventual failure is burn up or breakage due to excess movement is neither here nor there. The point is the tweeter died because of being fed too large a signal by the clipping amp.

[Charles Darwin]

Originally Posted by cotdt

10% THD is barely perceptible... try it for yourself. At 5% THD, you won't even pass the blind test.

I did the THD test on the Klippel site with my speakers and could fairly easily hear -45dB THD (0.5%).

[JMFahey]

An analog synth, or even using an oscillator set to, say, 6KHz sinewave and then driving the amplifier to clipping , simply is NOT a realistic test signal , it can't be used as "proof" as to what a Music program , (*any* music program by the way, from Classic to Punk to Death Metal) can do when overdriving a Power Amp connected to a multiway Speaker cabinet, which is the point of this discussion.

OF COURSE, if you drive a tweeter coil with 100W RMS sinewave (clipped or not) at a frequency which passes easily through the crossover, it will burn the VC before mechanical stress breaks the wire.
So what?
Not what we are discussing here.

[Charles Darwin]

I have no intention to start off with a 6kHz wave.
I'd use a 100Hz sine and clip it or 50Hz or 20Hz, it doesn't really matter as long as it is well below the crossover point so that none of the original, unclipped signal would ever reach the tweeter.