KeithEmo
Member of the Trade: Emotiva
- Joined
- Aug 13, 2014
- Posts
- 1,698
- Likes
- 868
You seem to insist on treating certain generalizations as facts - while discounting others disproportionately.
As a few other people have noted, you can't get a true square wave out of digital audio at all - due to band limiting.
However, in order to be potentially dangerous, you don't actually need an especially good approximation.
For example, many popular speakers have a crossover between 1 kHz and 3 kHz, and second order, and even third order high-pass filters, are actually quite common.
So, if I start with a 1.5 kHz tone, as a pure tone it will be handled by the midrange, but even the first harmonic present in a vague approximation of a square wave (the third harmonic - at 4.5 kHz), will fall squarely into the tweeter.
And, since it is also relatively common for tweeters that are used above 3 kHz, with a third order filter, to have rather low power handling capabilities, this can be significant.
And, yes, you do have to clip the signal "pretty hard" to generate any significant power in the harmonics...
Because of this, it probably is far more likely that a speaker will be damaged by the high average power level of a heavily clipped signal than because of its spectral content.
The general way I've heard the assertion is "people with low powered amplifiers, who turn them up until they clip, in an attempt to get them to play loudly, are more likely to blow tweeters than people with more powerful amplifiers".
The assertion I'm familiar with is NOT that people with more powerful amplifiers are less capable of blowing tweeters if they make bad decisions - just that they are statistically less likely to do so.
And, while that may not always, or even necessarily, be the case, most people in the industry agree that it often is.
Incidentally, most companies are not able to collect data about exactly why tweeters get blown (customers tend to avoid admitting that they drove their amplifier into hard clipping, because that would count as abuse, and might void their warranty).
The reality is that people routinely harbor the belief that "if they have a 100 watt amplifier and 200 watt speakers they don't have to worry about how loud they turn up the amp because it isn't powerful enough to hurt their speakers".
And this frequently ends up being proven wrong by the practical lesson of a blown driver.
I would also disagree with your assertion about amplifier power (unless you're simply suggesting that "most people have at least a reasonable amount of power these days - which I tend to agree with).
We still routinely see AVR and smaller amplifiers in the range between 20 watts/channel and about 50 watts/channel.
The amplifiers we currently make here at Emotiva range from 50 watts/channel to 650 watts/channel into 8 Ohms... up to about 1000 watts into 4 Ohms.
In discussions like this - test data ends up being not much better than anecdotal data.
I could easily find a specific example where a 100 watt amplifier, driven into hard clipping at a certain frequency, would burn out the tweeter in a certain speaker.
While a more powerful amplifier, operating cleanly with that same test signal, at a higher power level, would not.
However, since continuous sine waves aren't especially typical of music, it would prove very little - other than that the possibility exists.
(And, if it's simply a matter of boosting the average power level, without clipping, then the danger will always be to the driver handling the primary frequency of the test signal.)
The actual operation of various software limiters varies quite widely (and depends on how you set them).
Some will allow you to clip, or "soft clip"; others will absolutely prevent clipping, but will cheerfully allow you to turn music into a solid wall of full-amplitude sine waves.
However, since, statistically, with typical music, the average power level is usually around 10% of the peak power, raising the average power level significantly above that is going to increase the risk of overheating all the drivers in a speaker.
And, while tweeters often receive a relatively small percentage of the total power, they are also often only able to tolerate a very low average power level long-term.
(So, if your music has a lot of high frequency content, raising the average power level is going to be more likely to damage the tweeter than the other drivers, independent of any spectral shift due to the clipping itself.)
In the old days, various types of deliberate limiting were also commonly found in the "protection circuitry" of amplifiers - but most modern amplifiers avoid this because it tends to compromise sound quality unacceptably.
As a few other people have noted, you can't get a true square wave out of digital audio at all - due to band limiting.
However, in order to be potentially dangerous, you don't actually need an especially good approximation.
For example, many popular speakers have a crossover between 1 kHz and 3 kHz, and second order, and even third order high-pass filters, are actually quite common.
So, if I start with a 1.5 kHz tone, as a pure tone it will be handled by the midrange, but even the first harmonic present in a vague approximation of a square wave (the third harmonic - at 4.5 kHz), will fall squarely into the tweeter.
And, since it is also relatively common for tweeters that are used above 3 kHz, with a third order filter, to have rather low power handling capabilities, this can be significant.
And, yes, you do have to clip the signal "pretty hard" to generate any significant power in the harmonics...
Because of this, it probably is far more likely that a speaker will be damaged by the high average power level of a heavily clipped signal than because of its spectral content.
The general way I've heard the assertion is "people with low powered amplifiers, who turn them up until they clip, in an attempt to get them to play loudly, are more likely to blow tweeters than people with more powerful amplifiers".
The assertion I'm familiar with is NOT that people with more powerful amplifiers are less capable of blowing tweeters if they make bad decisions - just that they are statistically less likely to do so.
And, while that may not always, or even necessarily, be the case, most people in the industry agree that it often is.
Incidentally, most companies are not able to collect data about exactly why tweeters get blown (customers tend to avoid admitting that they drove their amplifier into hard clipping, because that would count as abuse, and might void their warranty).
The reality is that people routinely harbor the belief that "if they have a 100 watt amplifier and 200 watt speakers they don't have to worry about how loud they turn up the amp because it isn't powerful enough to hurt their speakers".
And this frequently ends up being proven wrong by the practical lesson of a blown driver.
I would also disagree with your assertion about amplifier power (unless you're simply suggesting that "most people have at least a reasonable amount of power these days - which I tend to agree with).
We still routinely see AVR and smaller amplifiers in the range between 20 watts/channel and about 50 watts/channel.
The amplifiers we currently make here at Emotiva range from 50 watts/channel to 650 watts/channel into 8 Ohms... up to about 1000 watts into 4 Ohms.
In discussions like this - test data ends up being not much better than anecdotal data.
I could easily find a specific example where a 100 watt amplifier, driven into hard clipping at a certain frequency, would burn out the tweeter in a certain speaker.
While a more powerful amplifier, operating cleanly with that same test signal, at a higher power level, would not.
However, since continuous sine waves aren't especially typical of music, it would prove very little - other than that the possibility exists.
(And, if it's simply a matter of boosting the average power level, without clipping, then the danger will always be to the driver handling the primary frequency of the test signal.)
The actual operation of various software limiters varies quite widely (and depends on how you set them).
Some will allow you to clip, or "soft clip"; others will absolutely prevent clipping, but will cheerfully allow you to turn music into a solid wall of full-amplitude sine waves.
However, since, statistically, with typical music, the average power level is usually around 10% of the peak power, raising the average power level significantly above that is going to increase the risk of overheating all the drivers in a speaker.
And, while tweeters often receive a relatively small percentage of the total power, they are also often only able to tolerate a very low average power level long-term.
(So, if your music has a lot of high frequency content, raising the average power level is going to be more likely to damage the tweeter than the other drivers, independent of any spectral shift due to the clipping itself.)
In the old days, various types of deliberate limiting were also commonly found in the "protection circuitry" of amplifiers - but most modern amplifiers avoid this because it tends to compromise sound quality unacceptably.
I welcome you to show, with data, how the actual measurements I posted do not encompass common conditions. Just saying they don't doesn't count.
Hmmm.....perhaps you need to check up on what a square wave actually is. I've already shown what clipped audio looks like, and already given you actual levels of clipping at which you start to get pseudo-square waves with continuous sine waves as input, but you can't get a square wave at all out of actual music audio, nor does clipped audio approximate the spectrum of a square wave. I've already shown proof of this.
But the claim of the myth is that if you changed to an amp that does not clip but provides adequate drive, you won't damage your tweeters. To do that you need a higher power amp with similar gain. Then, if you want to turn that amp up, you can. But for comparison, we have to establish common factors, so with gain held constant we can then compare the effects of clipping on output power. I've shown that in the initial graph where you can see without doubt that a higher power non-clipping amp provides more RMS power to the load. Always.
Agreed.
No, I don't have that data at all. What I have is a lot of assumption based on unconfirmed data collected anecdotally. Do you, as a speaker manufacturer, have actual test data that confirms anything?
What I'm seeing these days is amplifiers have leveled at somewhere between 80 and 120 wpc, which lands pretty much every amp within a 1.8dB range of power output. It's actually hard to buy an underpowered amp given today's typical speaker efficiency. I would think a speaker/amp manufacturer would recognize this. Before you go to the high end exceptions, note that in the above you're talking about "the sort of people who buy under-powered amplifiers".
PLEASE DO THAT! You may learn a few things rather quickly.
Absolutely NOT. PLEASE TRY IT!
Nor have I.
OMG. Would you PLEASE read my other posts and look at the graphs? What, exactly, do you think I've been doing and posting????