Possible to mechanically damage drivers rated for 100 watts with a 35 watt tube amp?

[rothwellaudio]

If a clipped signal is magically converted to DC, how does it get thru the crossover caps?

By being, effectively, pulsed. Imagine a greatly slowed-down version of what's going on: DC is applied to a capacitor, and the voltage on the capacitor rises as it charges up. When the capacitor is fully charged the voltage stops rising. The if the source of DC is removed the capacitor will discharge (assuming a discharge path exists, which is does in a crossover/speaker) and the voltage will drop back to zero. It's the charging and discharging current which flows through the speaker voice coil.

Yes, you're correct, DC will not get through a capacitor for longer than it takes to charge up the capacitor. However, because the clipped waveform is only producing effectively pulsed DC the speaker can still be damaged.
The same principle is used in switch mode power supplies and DC-DC converters. The DC is pulsed on and off and becomes effectively AC.

 

[spark1]

So, it's not DC.

 

[rothwellaudio]

So, it's not DC.

That depends on your definition of DC. You could take any waveform, say for example a sine wave, and say that it's DC in the sense that the top half of the wave causes current to flow in one direction only. The bottom half also causes current to flow in one direction only - but it's the opposite direction. So is it DC or AC? It makes sense to call it AC because the direction of current flow alternates.
However, if you just look at a portion of a waveform you could easily claim that that portion is DC because current flows in only one direction for the duration of that portion.
How much does it matter for frying speakers? The distinction is to some extent academic because DC, sine waves, and square waves will all fry speakers. However, since power is proportional to the area enclosed by the waveform, square waves have more power due to having a larger area and therefore can heat up voice coils more.

 

[fill35u]

Years ago, we did some serious experimentation here at AK, and examined the energy content of amp outputs, at various stages of clipping and non-clipping. We determined that low watt amps can indeed burn up midranges and tweeters of speakers rated for significantly more power than the amp was designed to produce. We did this with spectrum analysis.

Thread here, excellent reading:

https://audiokarma.org/forums/index.php?threads/clipping-amps-and-blowing-tweeters.570917/

Most of the work in that thread was using different *models*, the only real world evidence was anecdotal.

The fascinating mechabism that I had not considered prior to this thread is:

Rod Elliot is just plain wrong there. I've seen tweeters set on fire when an amp has gone into ultrasonic oscillation.

Yes, I've scoped *bad* amps that went into spurious oscillation only at high output levels(and even with no load!). But a well-designed amp in working condition should not only be immune to this, it should have built-in protection against excessive high frequency output. Any deficiency in this area would be revealed by thorough bench testing, so there's no excuse. OTOH, this site favors vintage and boutique designs, many of which I've found to be utterly lacking in protective circuitry.

Going the route of cone/dome tweeters is also a double penalty. They don't have the combination of sensitivity and power handling to achieve extremely loud ("drunk party") levels, and they're susceptible to damage from oscillating/overdriven amps. But compression drivers not only run about 10dB more sensitive, they also handle about ten times the power. That's about 20dB more max output. Drunk or not, if you turn up the volume too far on a CD/horn that's in the same living room, you will be instantly persuaded to turn it back down before anything can burn up...

Don't forget that there are protective devices and circuits that can be incorporated into the *speakers* to preserve the tweeters from the amplifier's proclivities.

 

[rothwellaudio]

Yes, I've scoped *bad* amps that went into spurious oscillation only at high output levels(and even with no load!). But a well-designed amp in working condition should not only be immune to this, it should have built-in protection against excessive high frequency output.

I wasn't disputing that, I was disputing "The 'Oscillations' Myth" from the Elliot Sound website. However, having read it again I have to admit that the author does concede that "damage to tweeters is possible". Initially I thought he was claiming that ultrasonic oscillation cannot damage speakers.

 

[spark1]

That depends on your definition of DC. You could take any waveform, say for example a sine wave, and say that it's DC in the sense that the top half of the wave causes current to flow in one direction only. The bottom half also causes current to flow in one direction only - but it's the opposite direction. So is it DC or AC? It makes sense to call it AC because the direction of current flow alternates.
However, if you just look at a portion of a waveform you could easily claim that that portion is DC because current flows in only one direction for the duration of that portion.
How much does it matter for frying speakers? The distinction is to some extent academic because DC, sine waves, and square waves will all fry speakers. However, since power is proportional to the area enclosed by the waveform, square waves have more power due to having a larger area and therefore can heat up voice coils more.

Yes....damage occurs when a driver is getting more power than it can handle, regardless of wave form. The fact that a signal is being clipped does not necessarily mean it will damage the driver.

But I would agree that clipping can result in the delivery of more power to the mids and tweeters than they would get if the signal were not clipped. I believe this to be primarily a function of the user continuing to add voltage (via the volume knob) after maximum amplification of the lower frequencies. The higher frequency portions of the signal may or may not be clipped in this situation, but in either case it is not the clipping that causes the damage...instead, it is this "additional " power (if it is surpasses the driver's power handling capabilities).

If it does not exceed those capabilities, damage will not occur...regardless of whether or not the signal is being clipped.

DC does not inherently damage drivers, nor does clipping. Damage occurs when power exceeds the capabilities of the driver.

That said, clipping can certainly be a predecessor to over-powering a driver, and can increase the likelihood of doing so (as a function of higher continuous power in the frequencies being clipped).

In all cases, clipping is not good. At best, it distorts the signal. At worst, it increases the likelihood of damage...to not only the speakers but also the amp (clipping can throw an amp into oscillation, which in turn can cause a variety of nasty problems).

Last thought...while some members have jumped on me for saying so without strong empirical evidence, I believe that clipping is much more common than most audio enthusiasts would expect. For instance, consider the large number of members using amps in the 25-40 watt range, running speakers rated at 86dB or thereabouts, and playing them LOUD.

 

[fill35u]

Last thought...while some members have jumped on me for saying so without strong empirical evidence, I believe that clipping is much more common than most audio enthusiasts would expect. For instance, consider the large number of members using amps in the 25-40 watt range, running speakers rated at 86dB or thereabouts, and playing them LOUD.

Is that scenario really that common? I can't imagine wanting more than 85dB average (with +15dB peaks) in my living room during a party. If somebody really thinks they need more than that, they should rent a DJ. Alcohol and temporary hearing loss will ensure that after the first half hour, sound quality won't matter anyhow...

I was actually about to comment how silly this whole discussion seems: if somebody is clipping a 30W amp that's prone to oscillation and has no protection against it, into an insensitive dome tweeter with no protection - it's hard to feel pity for them. Bottom line, they abused their equipment.

The scenario of "drunken party" is inexcusable, of course. And no matter how much protection built into the amp and speaker circuits, overturning a beer or piece of gear is just as destructive.

It's no mystery how to assemble a system that will play loud and not self-destruct while doing it. For those who would rather abuse their beautiful vintage gear, I *hope* they burn out their tweeters: at least the smoke might stop the party and prevent further damage, and they'll upgrade their unobtainable drivers and crossovers to Econowaves (which are better suited for the job).

 

[spark1]

There was a thread a while back asking the level at which members listen. Responses of 90dB or more (when wanting to listen "loud") were plentiful.

When sitting in front of my speakers, I sometimes listen near that level (continuous/average), about 9 feet away...depending upon mood and music choice. Rough math with my speakers says that's 6-8 watts, assuming 7dB loss due to distance, and 3dB gained back from 2nd speaker. Not much left for clean peaks with a 45 watt amp, assuming a high quality recording.

 

[spark1]

It's no mystery how to assemble a system that will play loud and not self-destruct while doing it.

Given the number of times I (and many others) have had to explain it, even to long-time enthusiasts, I would suggest that this knowledge is not as common as you might think.

Heck, I would say that it was not until 10 years after I bought my first decent system that I finally got a good grasp on the variables. Of course, that was a long time ago...before the Web, when you often had to learn the facts and then put them together yourself.

Another reference point is the experience I had when invited to a gathering of 15-20 local enthusiasts about 10 years ago. The host asked me to give a "presentation" explaining the relationship of power, loudness, speaker sensitivity, speaker impedance, etc after reading an email of mine. It became clear early on in the presentation that only a few of those in attendance - both young and older - had a good understanding of the variables and how they inter-relate.

 

[damacman]

Hypothesis - Driving an amplifier into clipping generates DC at its output.

Who would now like to use science to prove or disprove this?

It's just that easy.

Or, I suppose we could continue passing the wrong answer around until enough folks have heard it that it's just taken as gospel.

 

[pustelniakr]

Hypothesis - Driving an amplifier into clipping generates DC at its output.

Who would now like to use science to prove or disprove this?

It's just that easy.

Or, I suppose we could continue passing the wrong answer around until enough folks have heard it that it's just taken as gospel.

That issue was put to bed in this thread. There is no need to rehash it here.

https://audiokarma.org/forums/index.php?threads/clipping-amps-and-blowing-tweeters.570917/​

The topic of THIS thread has been stated as: can a brief burst of excessive amplitude from a low wattage tube amp damage the OP's speakers that are rated to handle much more power than the amp can produce?

Rich P

 

[damacman]

That issue was put to bed in this thread. There is no need to rehash it here.

https://audiokarma.org/forums/index.php?threads/clipping-amps-and-blowing-tweeters.570917/​

The topic of THIS thread has been stated as: can a brief burst of excessive amplitude from a low wattage tube amp damage the OP's speakers that are rated to handle much more power than the amp can produce?

Rich P

That is quite a thread with some excellent and thorough work by several. This should be required reading.

 

[tubed]

Yes, that was quite a thread.
I came away with the notion that basically overpowering a tweeter beyond its rated power handling was damaging, and clipping merely coincidental depending
on the power rating of the amplifier.

 

[elitopus]

^Yes, good thread.