(John Atkinson) wrote:

(Nousaine) wrote in message
.net...
"normanstrong" wrote:
Let's suppose your speaker is rated up to 100 watts, and it crosses
over from woofer to tweeter at 2000Hz. Since music contains very
little content above 1000Hz it is unlikely that a 100W signal will
contain enough power above 2000Hz to damage your tweeter, which might
very well be limited to about 10 watts.

Now imagine a change: a 1000Hz note is played at full output. That
will be 100W, most all of it going into the woofer. Now we continue
increasing the input so that what comes out of the amplifier consists
of a 1000Hz square wave having the same peak value as the original
100W output. The amplifier is now outputting a 200W signal, half of
which is the original 1000Hz sine wave, and the rest consisting of all
the odd harmonics (3000, 5000, 7000, etc.) What you have here is 100
watts worth of signal being fed directly to the tweeter. Since the
tweeter itself cannot handle anything like 100 Watts, it burns out.

This scenario contains some interesting assumptions doesn't it? First
it assumes the power supply of the 100 watt amplifier can deliver 200
watts simply by driving it into clipping?

Only if the sinewave is so overdriven that a completely square waveform
results.

So exactly how does being driven into clipping make the output devices and
power supply capable of delivering power that is un-tappable under other
operating conditions?

This may be unlikely but it is not impossible. And for the same
output voltage the squarewave does indeed offer twice the power.

In
reality, the clipped amplifier will be giving out more power for the same
output voltage, the amount of which depends on the degree of clipping.

But you're assuming it ALL goes to tweeter frequencies. What are the
frequencies at which amplifiers are driven into clipping?

And it assumes that the harmonic structure of clipping distributes
equal energy above the fundamental.

For a fully squared waveform, I believe this is the case. For a clipped
but not fully squared signal, all that can be said is that the extra power
represented by the harmonics in the scenario that Norm was referring to
is indeed fed to the tweeter,

Maybe.

which might not be able to handle it.

That's the prime issue here. Tweeters are cooked by too much power over a given
period. What clipping, as you call the party effect, does is raise the average
power level over time. It doesn't, by itself, represent a danger to tweeters.

Having seen speakers with tweeters blown by relatively low-powered
amplifiers driven continuously into clipping (the so-called "Party Test"),
I think it not improbable.

John Atkinson
Editor, Stereophile

Drive the speaker to the same SPL level with a larger amplifier and you'll fuse
that tweeter in exactly the same time.

(John Atkinson) wrote:



(Tom Nousaine) wrote in message
...
(John Atkinson) wrote:
(Nousaine) wrote in message
tt.net...
"normanstrong" wrote:
Let's suppose your speaker is rated up to 100 watts, and it crosses
over from woofer to tweeter at 2000Hz. Since music contains very
little content above 1000Hz it is unlikely that a 100W signal will
contain enough power above 2000Hz to damage your tweeter, which might
very well be limited to about 10 watts.

Now imagine a change: a 1000Hz note is played at full output. That
will be 100W, most all of it going into the woofer. Now we continue
increasing the input so that what comes out of the amplifier consists
of a 1000Hz square wave having the same peak value as the original
100W output. The amplifier is now outputting a 200W signal, half of
which is the original 1000Hz sine wave, and the rest consisting of all
the odd harmonics (3000, 5000, 7000, etc.) What you have here is 100
watts worth of signal being fed directly to the tweeter. Since the
tweeter itself cannot handle anything like 100 Watts, it burns out.

This scenario contains some interesting assumptions doesn't it? First
it assumes the power supply of the 100 watt amplifier can deliver 200
watts simply by driving it into clipping?

Only if the sinewave is so overdriven that a completely square
waveform results.

So exactly how does being driven into clipping make the output devices
and power supply capable of delivering power that is un-tappable under
other operating conditions?

Not sure I follow you Tom. The peak voltage remains remains the same but
squaring the waveform increases the RMS voltage hence the power delivered
into the load. In a real-world amplifier, the increased current demand
will cause the supply rails to droop a little so you won't get exactly
twice the power, but the difference doesn't change Norm's argument: that
in the case he describes fully clipping a 100W amplifier results in more
power being fed to the tweeter in the frm of odd-order harmonics..

But you're assuming it ALL goes to tweeter frequencies. What are the
frequencies at which amplifiers are driven into clipping?

I was specifically addressing Norm's scenario, where the test frequency
is below the crossover frequency and its third harmonic and higher are
above. Clip a 100Hz tone int the same speaker and the first 4 odd
harmonics will now be within the woofer's passband, resulting in less
stress on the tweeter.

And it assumes that the harmonic structure of clipping distributes
equal energy above the fundamental.

For a clipped but not fully squared signal, all that can be said is
that the extra power represented by the harmonics in the scenario that
Norm was referring to is indeed fed to the tweeter, which might not be
able to handle it.

That's the prime issue here. Tweeters are cooked by too much power
over a given period. What clipping, as you call the party effect, does
is raise the average power level over time. It doesn't, by itself,
represent a danger to tweeters.

It does in the scenario described by Norm. The relatively "pink" energy
distribution of normal music allows a speaker designer to get away with
using a tweeter with less long-term power handling capability than the
woofer. But "whiten" the musical spectrum by, say, driving the amplifier
into clipping, and the tweeter can be overloaded in a relatively short
period of time.

Drive the speaker to the same SPL level with a larger amplifier and
you'll fuse that tweeter in exactly the same time.

You are overlooking the fact that the clipping changes the crest factor
of the music Tom. A fully clipped 100W amplifier frys the tweeter in
Norm's speaker whereas a 200W amplifier driven to its maximum output
voltage into the same speaker without clipping will not destroy the
tweeter because the short-term transients will not exceed the unit's
short-term thermal capability.

You need to describe at the music's spectral content, its crest factor
(peak:mean ratio), and the time it takes for the tweeter to thermally
overload (as well as the time cnstant of your SPL meter) before you can
say something like "Drive the speaker to the same SPL level with a larger
amplifier and you'll fuse that tweeter in exactly the same time."

All I am saying is that my experience aligns with Norm's hypothesis:
clipped amplifiers tend to destroy tweeters; unclipped, high-power
amplifiers tend not to.

John Atkinson
Editor, Stereophile

My experience does not. The most likely scenario for frying a tweeter is an
accidental high power high frequency signal (MLS mostly with an untended gain
setting and an accidental input switch) delivered to the speaker system.
Otherwise small amplifiers never seem to endanger tweeters even with a 50-watt
Bryston or Stewart driven into heavy clipping at 10-100 Hz.

Using a larger amplifiers (larger Bryston or Crown Macro-Tech for example)
endangers tweeters in a much more imposing manner.

Back in the 70s I had a pair of Dahlquist DQ-10s with a tweeter fuse. 10 watt
amplifiers amost never took out a tweeter fuse while my 250-watt Heathkit
AA-1640 seemed to think that tweeter fuses were clay pigeons.

For those who weren't familiar with those speakers they had severe dynamic and
extension limitations at low frequencies and users often cranked the hell out
of them trying to get some bass out of them. But as you turned the gain up the
upward spectral shift turned them into screamers.

For me, this pointed out to me the real need for a subwoofer (which the
companion Dahlquist product was NOT.) But even after adding a subwoofer and a
larger amplifier the most easily damaged part of the system was the tweeter
fuse. And open fuses became more common with a bigger ampliifer.

This, by the way, is one of the endearing characteristics of active speakers
.... precice amplifier sizing for any given piece of the system. IME Active
systems are among the most reliable of speaker systems (although well designed
speakers in general are incredibly reliable as a class.)

What I mean is that during any gathering I can hand the remote to anybody and
leave the room without worry that drivers will be fried while I'm gone.

But I don't see any 300-watt tweeter amplifiers used to "protect" the driver.
That's more easily accomplished with a properly sized amplifier and good
engineering.