by Pablo Bellinghausen –
Broken speakers are sadly a common sight in PA systems around the world, and many people can be at a loss as to why. There are many possible reasons (from feed wires getting unplugged and glue disintegrating, to someone poking the cardboard of the cone), but by far, the most common causes can all be boiled down to a single one:
Too much power will break a speaker.
The rate at which this can happen will of course depend on how badly the speaker is overdriven, and it can take anything from years to a few seconds (by the way, when a speaker coil burns, sometimes it literally catches fire!)
In most cases, overdriving a speaker will cause one of two types of failure: thermal or mechanical.
As we mentioned in our previous instalment, heat is the biggest killer of audio components, and the most common cause of overheating is the use of speakers that are too small for their volume requirements, usually due to cost or physical size. It is important to know that sound volume works in a logarithmic way, which means going louder will demand a lot of power very quickly; an easy rule of thumb is that “twice as loud” is about a 10 dB difference, which is just over 3 times the voltage and 10 times the average power; therefore, a 200 W system will only feel twice as loud as a 20 W system!
Many sound engineers will say that “underpowering”, or using amplifiers that are too small for the speakers, is worse than “overpowering” them. Although they have good reasons to say such a thing and in practice speakers do tend to break mostly when the amp is being driven beyond its operating range, the truth is a bit less straightforward than that; basically, overdriving a small amp can actually release as much power to the speaker as a big amp, and it isn’t the distortion per se that breaks the speakers but rather the increased power of the signal at different frequencies.
Here is a mono mixer feed from a live rock performance, with a little bit of compression but not much else; let’s imagine the maximum value is the amp’s peak power. The signal is clipping just a tad, but it’s pretty much unnoticeable:
The signal from the amp going into the speakers has very little info below 20 Hz and almost nothing after 20 kHz, as per the recording. The amp is outputting mostly clean, clear sound, although it’s running at the peak of its optimal range. Because of the dynamic nature of the audio signal (the peak power being 18.6 times the average power), if the rating of this amp were 1000 W peak, then the maximum average power at which it could cleanly reproduce this rock performance would only be about 54 W.
Let’s imagine the amp has no built-in limiter, and that the sound engineer wants the sound to be “twice as loud”. He cranks up the volume out from his mixer and into the amps’ input, and distorts the audio until the signal looks like this:
Besides the horribly distorted sound, amps are actually fine delivering this amount of clipping. But we end up with a signal that is almost ten times more powerful than the correct operating range, with a very high power (the above signal averages about 430 W) that doesn’t get lower in the quiet parts, which would normally allow the devices to cool down. As can be seen on the spectrum analyser, clipping also creates harmonics both in the audible treble (4–20 kHz) and in the inaudible range (over 20 kHz).
When clipping an amp in this way, because the treble increases faster than the bass, and because any frequencies too high to reproduce are converted into heat by the tweeter, this is often the part that will break first.
Many amplifiers nowadays have security features like limiters and a protect mode (in which they actually shut down instead of clipping) which makes our lives easier, but of course it’s still possible to break a speaker by using too much power; compression and limiting will decrease distortion but still increase the average power.
A speaker that is driven too hard will essentially distort by itself. Usually the sound of an overdriven speaker is nowhere near as bad as a clipping amplifier, but the audio will quickly start to feel edgy and piercing. If the amp is very powerful and the speakers sound shrill and unpleasant at higher volumes, then they’re possibly distorting and already overheating.
In many cases, engineers will make things worse by compensating for this with an EQ, which usually means removing mid frequencies and accentuating the bass. Since woofers are much less efficient at low frequencies (similarly to tweeters at very high frequencies), boosting the bass will make them produce even more heat, eventually melting or charring the voice coil.
It is important to understand that even though clipped signals are a bit worse than clean ones due to the extra harmonics, the speaker will start to overheat at roughly the same perceived volume no matter the amp; and since amp distortion is worse than speaker distortion, the former will mask the latter and make it hard to know whether the speaker is running too loudly. In any case, someone with a burnt speaker should always try to upgrade to a more powerful model; a bigger amp will clean up the sound if the smaller one is distorting, which will make the tweeter last slightly longer, but the speaker will still eventually fail at a similar loudness.
Mechanical failure is more straightforward; basically, a signal with a voltage that is too high will cause the speaker diaphragm to move beyond what it’s designed to. The speaker cone’s back-and-forth movement is called excursion, which is bigger the lower the frequency. If a voice coil moves too far (over-excursion), the suspension can stretch out of shape and lose flexibility, or the coil can hit the magnet at the back. These things will crease or deform the coil and permanently distort the sound, making it fizzy or boomy, or in a worst-case scenario rip the cone. On the other hand, if the voice coil moves too far to the front, it can get dislodged from the magnetic gap in which it’s supposed to sit, and the speaker will no longer produce any sound at all.
Mechanical failure is unusual but can happen to both tweeters and woofers, and will usually be caused by too-loud music with a lot of transients (like percussions or drums) or with a lot of very deep bass (like dance or electronic music), and by big “pops” caused by malfunctioning equipment (for example, turntable needles being dropped, products being turned on or off, mics being hit, condenser mics being unplugged without taking phantom power off, crackling leads, and poor or faulty cabling in general).
Voice coil ripped due to over-excursion
The only way to absolutely ensure that a speaker will not break in a setup due to too much power is to use an amp that is a lot smaller than the speaker rating (basically an amp peak power of half the speaker’s RMS rating) in which case the maximum clean volume of the system will be far lower than it could be. If the system is underpowered the best way to squeeze volume out and still prevent broken equipment is to use a limiter instead, and set it so that the maximum possible power ends up being around the RMS rating of the speaker.
However, the ideal way to spec up a system from the get-go is to choose speakers that will comfortably play back the audio at the required level without distorting (this is the critical bit), and then choosing a power amp that will provide that power cleanly for the best sound quality.
It’s usually a good call to err on the side of caution by choosing a bigger amplifier than strictly required; it will happily output the correct power cleanly, and if it eventually runs loud enough to break the speaker, that’s a good sign that the speaker needs to be replaced with something bigger. A smaller amp would have been clipping at that point, and most likely would have eventually broken the speaker anyway.
In our next and last instalment for PA power we will look at easy and practical ways to choose and match the right equipment for your needs according to power ratings and impedance, whether it is a simple public address in a small room, or a sound installation in a live music venue.