Jump to content
The Klipsch Audio Community
Sign in to follow this  
Jabez Scratch

"Clipping" and tweeter damage

Recommended Posts

The following is from: http://www.audiovideo101.com/dictionary/dictionary.asp?dictionaryid=100

Clipping

Amplifier distortion occurring when a high energy wave form (a very loud sound resulting in a large output) is input into an amplifier and the amplifier is unable to fully reproduce it due to power supply limitations or amplifier design limitations resulting in the audio output waves being cut off (the rounded tops sliced off resulting in short waves with flat tops). Clipping creates audible distortion and can be damaging to speakers especially if the clipping is hard and frequent.

Clipping generally occurs when an amplifier is playing at a high level and it is asked to output a large amplitude waveform (tall wave with lots of power). The amplifier clips when it does not have the power capability to correctly create the waveform. Instead, as the wave is built it hits a ceiling essentially not allowing the wave to go any higher. Since the amplifier cannot recreate the remaining portion of the wave rising above the ceiling, the wave is cut off.

Generally, the more power an amplifier has (especially relating to the quality of the amplifiers power supply) the more immune it is to clipping. For this reason, larger amplifiers tend to provide better quality sound at loud listening levels since they clip less often (if at all) compared to similar but less powerful amplifiers.

Clipping may be heard in loudspeakers as an abnormal, non-musical sound. It is unpleasant and it may damage speakers (with tweeters being particularly susceptible). This occurs because a speaker cannot produce the flat-topped waveform sent to it by the clipping amplifier.

In order avoid clipping, do not play a sound system at excessively loud levels and make sure the amplifiers being used are large enough to recreate the sound levels in the given space. Generally, use the largest amplifiers reasonable, as they are less likely to clip and damage speakers. It is much more dangerous to clip loudspeakers by using a small amplifier lacking in power than to use a large amplifier even if that amplifiers power ratings are greater than those recommended for a given speaker (distortion, and clipping in particular, causes damage not clean power from a quality amplifier).

For more information check out these related terms:

--Amplifier

--Amplitude

--Distortion

--Power Supply

--Sound Wave

--Volume

--Watt

Share this post


Link to post
Share on other sites

An amplifier can push a lot more power than it is rated for. The specs tell you what it will reproduce cleanly without distortion. There is also some usable headroom above that, eg. a 250 W amp can produce say 290W. Above that all you have is distortion when the waves are clipped. Distortion is heat. You can break a coathanger if you bend it enough times as it heats up. The coil in the speaker is wrapped with wire not much thicker than a hair, it doesn't take long to burn and break. Just an easy way to understand what's happening to the poor tweeter when an amp is overdriven to the point where it can no longer provide a clean signal.

Share this post


Link to post
Share on other sites

Good answer Lynnm.

I was going to respond to the question when I read it eariler today, but I didn't have the time.

This is why I believe that one is better off having more power then too little.

Share this post


Link to post
Share on other sites

Everybody is spot on!

Specifically, tweeters get killed first because a flat-topped, clipped wave contains large quantities of very high frequencies, required to make it flat. On top of that, harmonic distortion consists of multiples (2x, 3x, 4x, 5x, 6x, 7x, etc.) of the original signal adding more HF energy. The voice coil wire is small making the tweeter intolerant of excess power.

If you can hear any distortion, turn it down. Mild clipping sounds hard or brittle

Share this post


Link to post
Share on other sites

I'm being picky here.

I was thinking about this the other day while looking at a technical book about the frequency components caused by clipping. Actually it was the frequency analysis of a square wave.

Certainly, there are odd order harmonics generated and these go to the tweeter, potentally causing burn out. However, even with a square wave, the harmonics are not nearly as strong as the primary signal.

Most music has high frequency components which are well below the midrange and bass.

However, if you're driving the amp into clipping the bass and midrange, you're also pushing a lot of "normal" tweeter range frequencies. And tweeters are fragile.

Therefore, harmonics may be the straw which breaks the little camel's back (the little camel being the tweeter). None the less, most of the problem may well come from an extraordinary amount of normal treble.

Amps with a lot of headroom may well avoid harmonics from clipping. None the less, they may still be more than capable of distroying a tweeter.

Best,

Gil

Share this post


Link to post
Share on other sites

OK Gil, let's get picky, especially if it will save someone from damaging their system.

----------------

On 1/3/2004 1:33:00 PM lynnm wrote:

The following is from: http://www.audiovideo101.com/dictionary/dictionary.asp?dictionaryid=100

...It is much more dangerous to clip loudspeakers by using a small amplifier lacking in power than to use a large amplifier even if that amplifier’s power ratings are greater than those recommended for a given speaker (distortion, and clipping in particular, causes damage not clean power from a quality amplifier)...

lynnm, I think this is a very important note. I would explain it different though, you are more likely to clip a small wattage amplifier than a large wattage amplifier trying to put out high sound pressure levels. It is safer to drive a speaker over its suggested power handling capabilities with clean/undistorted watts from a larger amp than trying to drive it louder with a smaller amp that is distorted. Pushing any size amp into extreme clipping will inevitably result in blown speakers.

John and Gil have started explaining the nasties that make up clipping distortion.

Bottom line I guess is if it's not loud enough with your chosen amp and speakers then you had probably better look for bigger and more expensive equipment or keep the volume down enough to avoid burning/destroying/wrecking your sound system.

Oops, blew another speaker. It does happen.

1.gif2.gif

Share this post


Link to post
Share on other sites

Also the peaks of a clipped waveform are DC. Large amounts of DC current, unblocked by a capacitor ( as in woofers) will overheat and distort the voice coils. The first signs one often hears of this is when the woofer' coil is permanently distorted and sounds raspy when moved by hand. not to mention it will sound bad.

Rick

Share this post


Link to post
Share on other sites

The urban legend about 'clipping distortion' damaging your speakers is just that, a legend.

As with all legends there is some basis in fact.

Over 25 years ago I contributed to a JBL 'tech note' titled: Danger, low power. It was written to try and slow the rash of tweeter failures we were seeing (I ran a service center).

The only portion of that note that I consider valid today is the caution about the use of tape lifters in RWD and FF with the volume up.

Speakers die from two root causes:

heat

mechanical damage

With the advent of what PWK called 'the stove' (an amplifier of 100W or more power), the K77s started to die.

The type A network became the AA with the 18dB slope and the zener diode protectors.

The main form of death was mechanical failure, the tweeter lead out wire broke.

The zener diodes limited the clean signal to 2W and turned everthing into a 4W square wave at high power.

That's right, the protection ciruit deliberately converted clean signal into square waves before feeding the tweeter.

In 1982 the elliptical filter was added, and the zeners dropped (they were actually quite expensive). The elliptical filter was basically the 18dB Chebychev with a zero added at the resonant frequency of the tweeter.

The interesting thing was that EV up-graded the voice-coil assembly during this time period and the network changes may not have been appropriate (the new network sounded sibilant on female vocals and had time/energy storage issues, see the Heyser review of the Klipschorn in Audio).

EV had changed the lead-out wire to a flat BeCu lead, the same as their top-of-the-line DH1 compression driver.

The lead-out wire no longer broke from flexing, but the coil came apart.

EV had changed from a long-cure thermo-set adhesive to a quick-cure UV-set adhesive.

The UV-set adhesive became soft under heat and the coil windings came apart.

The newest coils have the flat lead-out wire and the green thermo-set adhesive. These are very reliable.

Woofers die from clipping too, usually mechanical damage.

It is an amplifier design issue, and caused by the feedback loop being clipped. It can cause rail-to-rail pulses in the 1hz~4hz region. This in turn causes so much cone motion that someting has to give, the coil hits the magnet, the spider hits the top plate, the spider or cone tear, etc.

This is mainly a solid state problem, tube amps do not have the extended power bandwidth (for the most part) to cause the same damage, and the feedback loops are different.

Raw power (volume) seldom damages speakers, I have seen a Klipschorn driven to 250W peaks with a 1A inline fast-blow fuse.

Another interesting (and valid) viewpoint on this subject:

http://www.rane.com/pdf/note128.pdf

Share this post


Link to post
Share on other sites

I should probably add that the primary failure mechanism in the newer Klipsch speakers is:

heat

These are two-way systems with lower crossover points and more energy reaching the tweeter.

The tweeters are also less efficent than a K77 and therefore are driven harder (attenuated less).

The Rane note discusses these type of failures.

Share this post


Link to post
Share on other sites

Heat is right.

Forget we are dealing with sound for a min.

think voltage and resistance.

sound is AC alternating curret

the voltage changes from pos to neg in the wave form , but it is all ways changing.

when you clip a wave (as stated) it has a flat responce = no change

NO Change = DC direct current

when DC is input into a voice coil, the coil becomes a heater.

Then sometime later, it becomes a fuse..a broken fuse.

PWK has a note about this in his Dope From Hope writings.

I am sure someone can post that.

Share this post


Link to post
Share on other sites

I like the Rane white paper very much. This is mostly because I agrees with what I was saying. Smile.

He is talking about compression. That term is used to describe a lot of electrical signal effects which are different from each other.

Here though, let's assume that normal tweeter components of the music, the average high frequencies are about 6 or 9 dB down from the average midrange. This means that if you drive the midrange into 6 or 9 dB of clipping (the peaks would be that much higher) the tweeter high frequencies are not clipped, but are at full amp power. That could fry things without harmonics.

That is his story.

I was also thinking about whether a K-77 could be blown out by just (only) the harmonics of a square wave (hard clipping). He shows that the lowest numbered harmonic generated is the 3rd. It is almost 10 dB down from the fundamental.

So let's assume a 2000 Hz fundamental driven to clipping at 50 watts to make a square wave. (This is really hard clipping.) The 3rd harmonic is at 6000 Hz and 10 dB down is 5 watts. That could fry a K-77 which PWK said is good for 5 watts continuous (note, continuous).

So, I tell myself, that proves clipping at 50 watts can fry a K-77. But wait a minute. If the speaker as a whole is rated for 104 dB at one meter, that is. . . 10 watts brings output to 114 dB. 20 watts up to 117 dB, 40 watts to 121 dB. And we'll not worry about the rest.

121 dB is horribly, horribly loud. I don't think anyone could stand to be in the next room, let alone the listening room.

So, it does make sense that harmonics alone are not at fault, unless you're doing something very radical. And if you're doing something that bad, the normal signal can fry it sooner.

- - - - -

I must take some issue with my buddy Trey and others to the extent they argue that a square wave (going plus and minus) has a d.c. component. It doesn't by the Fourier math. Also by the Fourier math, it doesn't have any frequency component below the fundamental.

It is correct that there is a flat top, and d.c. is flat. And,if you just hooked up a d.c. signal source (long term) to a K-Horn, it could fry the woofer.

Here I'm assuming the simple inductor crossover having a 6 dB per octave roll off at 400 Hz. Or even the more recent ones with a cap added to give a 12 dB per octave roll off at 400 Hz.

The following is to say, the effect of the short term d.c. plus and minus, do not equate in an LC circuit to d.c. being passed to the woofer. If it can't get through a low pass filter, it ain't d.c.

If you hook up the battery, LONG TERM, the inductor is going form a magnetic field at a constant strenght and the cap is going to charge up an electric field to a constant strenght. Then, and only then, will d.c. get to the woofer voice coil. All power is being sucked off to build these fields first.

But LONG TERM is not what is going on with a square wave. Consider a square wave with a fundamental of 1000 Hz. It goes "high" for 1/2000 of a second. Then it goes "negative high" for 1/2000 of a second. Essentially, the first duration is too short for the magnetic and electical fields to reach a steady state d.c. value. And following, there is the negative voltage-current trying to reverse everything.

It is true that the woofer crossover is a "low pass". That also means it is a "no high pass". In other words, it is big, fat, dumb and happy (these are large value component), and does not allow quickly alternating voltage-currents to get through to the voice coil of the woofer.

The very reason why the low pass network works when fed with a square wave (or a sine) above the crossover point is that the components draw off the energy, and then give it back when the signal changes sign. The driver absorbs little of it. It is all sloshing around, back and forth in the cap and inductor.

- - - -

Too complicated? I'll have to work on it.

Gil

Share this post


Link to post
Share on other sites

I replace diaphragms in these tweeters for a lot of customers. I too have tried to understand the falure modes for these. I can tell you what I can see when replacing them. Usually I can see nothing wrong. Just that by measurement the VC is open somewhere in its 78 inches of 40 AWG length, (read that in a spec somewhere). I think those likely failed due to old age for lack of a better term. The other case that I see fairly often but not as often as the above mentioned is a mechanically damaged coil. I can almost predict these from what the customer tells me. It usually goes something like this. "We got a bit loud at the club last night and suddenly one (or both) tweeters just quit." Those voice coils typically are loose and look a bit like a slinky when I remove them. I call this one a mechanical failure. I think is likely due to over excursion. The coil was just asked to move further than it could move so it came apart.

Here is a personal example of the mechanical damage case. I was a few days ago playing around with a K-77M with a new diaphargm just installed. I was looking at sine wave frequency response using a spectrum analyzer and calibrated mic. After looking at a few points high up in the frequency range using an input of 2 volts RMS, I moved the frequency down to 3 k to look around at the low frequency cut-off point. The tweeter produced that frequency for about 3 seconds, sounded fuzzy for about 2 more seconds and then quit. The voice coil came apart in the manner I described above. I still find this hard to believe that happened at considerably less than one watt and would like to duplicate that test to see if it was just a fluke or a genuine reproducable event. The cost of diaphragms has so far made me resist the desire to duplicate that. By the way, this is, I think, very unlikely to happen to a tweeter connected to a crossover. The type A crossover would have had that 3k signal knocked down by about 10 db. I was not using a crossover and was protecting the tweeter with a 1 amp fast blow fuse. The fuse survived the episode in fine condition.

Bob

Share this post


Link to post
Share on other sites

I guess I was depending too much on the reader deducing that a clipped signal that generated lots of HF energy, added to the original signal, would overpower (overheat) the tweeter resulting in its failure.

I'm convinced they fail mostly due to heat. Perhaps some of Bob's failures are heat related as well as mechanical. The only drivers I've ever damaged were my sub drivers. All 4 went in about a minute one day when I was home alone rocking the house. After about 15 minutes of high power, hard rock (SPL Meter in use), the drivers signed off one at a time. By the time I realized it all 4 were damaged, but 2 were still making "noise". All 4 voice coils had unwound and were mostly in the bottom of the magnet structure. Two were open, as well.

Share this post


Link to post
Share on other sites

John,

I don't mean to say that I see all cases or even that the ones I see are representative of the typical failure modes. The only ones I see are those tweeters whose owners for one reason or another send their tweeters to me for repair rather than doing it themselves. Far more choose to do the work themselves. The only one that I have personally seen that gave visual evidence of having been hot was one that also exhibited a distorted voice coil which had been "rubbing". That one looked and smelled as if it had been hot and showed abrasions at the point where it had rubbed the side of the magnet gap. The evidence of being hot in this case could have been from mechanical rubbing creating heat or electrical heat could have caused distortion of the coil leading to the rubbing. Can't say which is cause and which is effect.

Bob

Share this post


Link to post
Share on other sites

Gil,

You confined your treatice to mid and high freq's. Since most like a little slam in their music, and low freq's because of there sustained notes demand more power from the amp, what happens when the amp clips on a sustained bass riff at 30-100HZ?

Rick

Share this post


Link to post
Share on other sites

I think the Rane article answers most of the questions. There are harmonics and they are in decreasing levels at 3, 5, and 7, etc. time the fundamental.

I was looking at this another way. Let's compare the amount of power in a square wave and a sine wave of equal peaks.

The square goes between plus and minus 1 volt. So it has the power of a 1 volt d.c. signal.

The sine wave goes between plus and minus 1 volt. It has the effective power of a 0.707 volt signal. There is a lot of math behind that fact. However this is "RMS" (root mean square) voltage of a sine wave.

Power delivered into a resistor is proportional to voltage squared. So in the square wave, we multiply 1 x 1 (this is squaring) and get 1. In the sine wave we multiply 0.707 x 0.707 and we get 0.5.

The result is that the square wave (which is a worst case clipping condition) puts out twice the power of the sine wave. That extra power goes into harmonics which are always higher in frequency than the fundamental. But it takes going out to infinity of frequency to get that all to add up.

Now, you may say "horrors". Obviously doubling power when the bass is thumping and forcing that energy up into the midrange and treble is going to have an adverse effect. But doubling power is fairly insignificant given (1) the range of dynamics of music and (2) what you can do with an amp.

The classic logic started with a 5 watt amp. By the above, it could put out 10 watts when driven to clipping. The cure was sometimes a 100 watt amp which wouldn't clip and we didn't have that extra 5 watts going places we might not have wanted them (harmonics to the mid and tweeter).

BUT, loud music being more appealing, people would use those 95 watts. One rule of thumb is that, not withstanding power and 3 dB, it takes 10 dB to make music sound twice as loud. It is easy to do with the volume control.

- - - -

My guess is that most voice coil failures are indeed mostly mechanical. This assumes music with dynamic peaks is being used. OTOH, if you hit the system with a long term sine wave or even a square wave, it could heat up the voice coil, the glue, the former, etc., while still being within mechanical limits.

The actual cause of any failure is a bit speculative unless we can take the voice coil apart and do a post mortem. Even then. Do we blame the harmonics or just raw power when there is a train wreck? From the above, raw power is much more easy to find in abundance as a driving force.

On the receiving end: If both the heat dissapation limits and the mechanical limits are challenged at the same time, who is to say which is the "cause".

- - - -

I had some adventures with a sine wave generator. If you hit a horn loaded bass system with even a moderate amount of power at resonance, it is not difficult to bottom out the driver. That is to say, the voice coil mechnism hits up against the metal and you get an odd clanking sound. My guess is that the crackling sound I hear when the Telarc disks are putting out infra sound is actually that same thing.

Either that, or the amp is going nuts.

- - - -

The bottom line is that the clipping / harmonics can only double power in the most theoretical sense. The cure of going to a bigger amp creates other risks which may be more deadly.

Gil

- - - -

It is interesting to read the theory that some amps when overdriven go haywire and may put some gross

Share this post


Link to post
Share on other sites

3d-zapper,

It is my guess that, in general, woofers are more apt to fail for exceeding excursion limits in a "normal" baffled-type direct radiator enclosure where the cone is free from restriction. In the case of a Khorn, its probably a combination from being prevented from excessive forward excursion by hitting the motor board on the front (I've heard this called "slapping") and over-excursion towards the basket in the other direction; the result is that heat build up in the voice coil will eventually fry it unless cone tearing occurs first. Whether the failure is to the cone, spider, or over-heating voice coil or all of the above is up for grabs.

I have temporarily clipped 300 watts into my Khorn K33E's although they are only rated at 150 watts without any apparent damage (luckily). Tweets and mids are also fine.

I wouldn't want to repeat it. If it doesn't damage the drivers, it certainly will damage your ears.

DM

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...