I've recently come across (and bought ) a pair of Quad 989 speakers.

Now the speaker specifications are: "Power capacity: 100W, 10V RMS, 40V peak maximum signal input."

Yet a lot of people reccommend to buy MUCH more powerful amplifiers. (http://www.stereotimes.com/amp100702.shtml says a 2x250W amplifier is the 'ideal pairing').

Can someone explain to me what the deal is? When will 2x250W give speakers more breathing room (whatever that means), and when will 2x250W burninate my poor $9000 speakers?

That is nonsense. You need an amplifier that is rated for the maximum power you need. If you used 2*250W you can expect arson attacks from your neighbours. What you need depends on how good your hearing is, what music you listen to, the sensitivity of your speakers, and your personal preference.

Not much bigger, no.

It's only that some amps, nearing its top power output, distorts.

That's what the fuss with the "breathing room". At 80% top power 99% of amps don't distort. Well, the quite good ones, that is (good <> expensive, mind you).

Of course if the amp is rated 2*250 W even 80% is big enough and, like CSMR said, will start a neighborhood riot

One thing to notice is the rating:
100W, 10V RMS
What this means is that their real power rating is only 12W continuous (assuming 8ohm, if they are 4 ohm then it's 25W), so the spec is lying to you. The next thing is that their peak voltage is rated at 40V (200W into 8ohms. which is interesting) which is obviously specified for short transients.

Amp distortion near the top power rating is mostly due to the specification being a little out of sync with reality. For example, my amp can put 162W per channel of square wave into my speakers but only 82W of unclipped sine wave.

So, if your amp vendor is honest, then a 2x100W (or even 2x50W) amp with decent full power distortion figures will be plenty (in my opinion).

what you have to take into consideration is "peak power", ie. a musical signal is never a constant level, it is more like valleys and peaks... the peaks are were for brief moments the amplifier has to generate multiples of its rated continous power to cover the dynamic range. By having a reasonably powerful amplifier you are buying yourself "headroom" in the sense that the amplifier will be capable of producing enough peak power as to not clip or distort the incoming signal. A clipping or distorting amplifier will damage your loudspeakers far quicker than having one that has enough peak power.

PS: this all assumes a reasonably loud listening level - you should be aware however, that your Quad's are not decibel monsters so don't expect to drive them to loud anyway...

Not true - I've destroyed far more speakers with low power than high power. I destroyed two pairs of tweeters with a 25W receiver that was routinely driven into clipping (thus creating a lot of relatively high voltage/power ultrasonic harmonics that kill speakers). After switching to a pair of 500W mono power amplifiers (overkill for certain) I never blew a speaker again. And the speakers were "rated" at 250W.

Remember speaker ratings are just general ballpark figures based on typical music signals in a steady state situation, not high frequency squarewaves that come from clipping amplifiers. You can usually exceed the "maximum power rating" by quite a lot for transients if the signal is clean.

You can never have too much clean amplifier power as long as you are sensitive to not overdriving the speakers. Speaker distortion rises gradually to help you know when you're pushing it - amplifier clipping just happens without much warning. So it is better to have a higher powered amplifier.

Now the Quads are electrostatics IIRC, and can be fragile relative to arcing burning the transducer, so if you're rocking out you may not have the right speakers.

Be careful out there.

I have a pair of SoundLab Pristines I drive with two 1,200 watt amps. In the '38-'39 US Worlds Fair in NYC Westinghouse found that 80 watts were needed to reproduce solo piano with the then high efficiency speakers.

If you match your speakers and amps watt for watt you will be over-driving your amps at musical peaks. This often results in clipping which translates to square waves which translates to broken speakers. My electrostatics play bass pretty well, way better than most electrostatics, and sound better than the Martin-Logans which are "mushy" to my taste. The ESL's are sweet, but could handle no power/volume in the past. They seem to have gotten past that. My speakers are rated at 200 - 500 watts, I think. The high wattage amps I have allow the amps to act as a damping agent, too, as I understand it. I do not over drive them, but they handle power well.

Anyway, the point is that more speakers are destroyed by underpowered amps than overpowered. Old ESL's would arc when driven too hard! Yeah, the old 63's. I just preached on this in another post: check JBL's pro web site for info on speakers blown from too little power. Having huge amps does not mean you will be playing the speakers that loud. You control the volume with that little round knob. I think the rule of physics is to double the volume you need square the power, so do not worry.




They will work just fine out of the box. If you feel better about it, give them a week. Most hi-end audio lore is BS.

Maybe I missed something, but isn't 200 Watts quite a lot for home use? A friend once showed me the volume of a pair of speakers running with 1 Watt (measured at the speaker cable), and it was already too loud for normal listening imo. I know that double power is not double volume, but still, it seems a bit much for me. My speakers are rated for 15 Watts each, and I never used them anywhere near the point where they would start to distort (in fact, when I tried to find this point it was very uncomfortably loud). I do get them to distort at lower perceived volume when I turn up the bass all the way, but that's not how I would normally listen.

Maybe I should add that the most expensive piece of audio equipment I own are my Sennheiser PX100s, so I have no experience with good speakers, amps etc.

The efficiency of the speakers comes into play quite heavily here. This can range over about 100x (20dB), so judging one set by another is not really that educational.

I have linked to Rod Elliot's article on why small amps kill tweeters many times. It's really worth a read for those of you who haven't read it yet.

Boojum: What is the real power of those 1200W amps? If they can really put 1200W of sin wave into an 8ohm load then I am seriously impressed. I have a 6kW test load that I use as a footrest at work (don't use it much) which could do with some warming

Note also that twice the power is only +3db an increase in volume, assuming the speaker has a relatively linear sound output increase vs. the electrical input increase. 1dB is considered the smallest perceivable change in volume that an average human can notice in tests (this may be the definition of the dB).

I've read that increasing +6db to +10db is the range people generally count as "twice as loud". If we work with the upper end (perhaps a good idea since the speakers might be losing efficiency as the power increases), 1W to 100W is generally going to be a change in perception of "four times as loud".

Essentially, you're not supposed to be buying the 200W vs. 100W amp for volume purposes. You're buying it for greater rail-to-rail headroom before the amp starts to clip and sends a frequency distribution that the speaker drivers are not designed to handle.

-brendan

I second that! Volume is not a purpose in itself. What is most important is that your amp should have more than enough power to handle all the dynamics in the audio signal before it will even start producing audible harmonic distortion. And it should absolutely be capable of handling these dynamics without clipping.

Now how many watts is 'enough' ?
The amplifier volume you need depends on the sound pressure level (SPL) you want to achieve in YOUR listening environment with your preferred kind of music.

To determine the SPL you want there's a very nice article on this subject here.

Especially for puntloos being a Dutchman here is a Dutch site that gives some info too.

Cheers!

Obviously you were using an amplifier which is not adequate for your power need, hence the clipping. Nothing wrong with an amplifier being higher power, it just is not useful if you don't need the power, so you restrict your choices and likely pay more than necessary.

Another one of Rod Elliot's articles is somewhat relevent to this discussion. It was something I found very interesting the first time I read it - Power Amplifier Clipping. It's really worth a read if you are interested in amplifier clipping.

I don't really buy into this "your amp needs to be much more powerfull than your speakers". Just as speakers can take transients with much more power than their RMS rating, an amplifier can deliver transients much stronger than its rated RMS power without much distortion. There's nothing wrong with some headroom to steer clear of clipping but you don't have to overdo it.

This depends on the limits of the amplifier - if it's power is limited by clipping (most designs behave this way) then they can't deliver transients above rated power without distortion. For example, an amp with 35V rails can only deliver 153W into 8ohms. Granted, such an amp could only deliver about 80W of unclipped sine, but that extra 73W only gives you 3dB of headroom. If the amp is power limited by dissapation in the output devices (heat) then you will likely have more headroom - but results do vary.

Another thing is that some amps (badly designed ones, granted, but this is not uncommon) behave nastily for a little while after they clip. Bad behaviour can include oscillation, sticking to rails, spurious low frequency signals and other ugliness. Speakers, on the other hand, generally behave decently for sharp transients (within reason). The original poster's set, for example, can only handle 10V RMS but 40V transients.

Let me recap all this a little bit.

- Amplifiers that start to clip can easily damage speakers with the dirty output
- (attempting) playback at higher volume than the speaker can do is Bad™ in any case.

So the key here is:
- Buying an amplifier that can give clean output up until the levels that will cause the speaker to clip
- And try to never reach that speaker-clipping point.
- if you do try to find the absolute max, reach it carefully, and once determined never reach it again with ample safety margin.

Therefore buying an overdimensioned amplifier depends a little bit on your assessment on how badly the amplifier manifacturer lies about its amps capability to deliver clean output. If they say it can do 400watt, does that mean 300watt clean and 100 watt dirty?

All in all would you guys agree that (also given that double wattage only increases volume (i.e. wave amplitude) 3db, if a speaker would be linear) an amplifier of 2-2.5x the rated speaker power is ample, and not too much, if you act sensibly with your volume control?

With that theory the evo 4 would (@360Watt rated) would be a decent pairing for my 100-140Watt rated speakers, a bit on the hefty side, but not insanely so..

These are mono-bridged halfer 500 watt amps. They were measured floating from 1,100 to 1,300 watts each into 8 ohms steady output. My dynamic speakers are KEF 104/2's which are 4 ohm, yielding roughly twice the power. Yet the KEF's will cause the amps to overheat and shutdown to a cool-down mode as they can absorb ~4,500 watts each! Yes, check the specs. Funkytown, Saint-Saens' 3rd (Organ Symphony) pedal notes and jets taking off will rock you.

They are also excellent for playing at room levels objectionable to no one. Plenty of reserve or "headroom." I always wanted a stereo which could play louder than I could stand it, and I got it. Now I just play music with it. Good for quartets and trios, too.

Considering that this topic keeps coming up - maybe this thread (or one like it) should be linked to in the FAQ. There is some good information here that people might find helpful.

According to this link:

http://www.tnt-audio.com/casse/quad-esl989-2_e.html

it's an electrostatic transducer designed for a tube amplifier, so most of the power handling considerations usually regarding dynamic speakers may not be helpful. The first goal has to be stability, as resisitive load is
only a small part of the impedance.
I suppose it has a transformer with an 8 or 16 Ohms input which in combination with the capacitive load
has the mentioned voltage limitation that must not be exceeded. Some audio freaks put a resistor in
series to avoid stability problems with transistor amplifiers. Additionally, some put a bipolar capacitor of
~600 µF in series to avoid saturation effects with low frequency burst signals in that transformer.
Anyway, I would cautiously try to test it for stabiliy first slowly increasing level without entering the
saturation region.

Good luck!

yes

It may be ample. An amplifier of 1/100 the rated power may be ample also. And it may not be ample if it doesn't give enough volume for you, but then you would have to change your speakers as well as your amp.

No, the correct answer is 6dB.


Exactly. You want your amp to be ample

Twice the power is 3db I believe. Twice the wave amplitude is 6db. Twice the amplitude is four times the power.

Yes, twice the power is 3 dB gain in volume.

This is correct, I didn't read carefully enough earlier.

The problem here is that amplifier ratings are relatively (!) accurate vs. speaker power ratings. Speaker power ratings are a general guideline - many tweeters can only take 25W, but because the majority of the music spectrum is below about 3kHz, in a multi-way system (midrange, woofer, etc.) you could likely get away with 100W no problem on typical music. Unless you send a 100W sine wave at 20kHz which will pass through the crossover a toast the tweeter pretty quickly.

Speaker power rating is a very ambiguous, general guideline whereas an amplifier output can be unambiguously measured and verified. Speaker power handling will depend on the spectrum of the signal, duration of the high amplitudes , crossover design and potentially how heated the electromagnetics are.

I would never fear hooking a 1,200 watt amplifier to any speaker because I could easily tell when the speaker was being overdriven. My 25W receiver example (from the poor school days) would make me leary of underpowering.

This is why you can't beat premade systems. You don't have to worry about what speakers to by with what amp. Makes life easier

Maybe I'm being naive here, but the thing that kills the speaker is the total power absorbed, regardless of frequency, right? As in if you put X amount of power in for Y seconds, it burns up or blows out or whatever?

If so, couldn't you do a frequency sweep at constant voltage, measure current and have a very good idea of the safe limits at each frequency?

No, that wouldn't tell you much because impedance varies highly with frequency.

The power limits of the typical cone/dome speaker are usually the thermal dissipation in the voice coil or the physical excursion extremes of the voice coil in the magnet structure. Speaker maximum power will be highly frequency dependent because of the different properties of drivers and the absorption of power in the crossover. A speaker might absorb 100W all day long at 100Hz and only 15W at 15kHz for a few seconds.

For example, I've seen burnt out voice coils (wire melted) in tweeters and bent voice coils on woofers due to the voice coil hitting the magnet during extreme signals.

Close, but you have to deal with the speaker at the component level: the thing that kills a *driver* is overheating (which should be somewhat proportional to the energy of the signal) as well as a driver's over-excursion beyond the in/out limit set by the designer. Most audiophile speakers, however, use two or more drivers and, as has been explained above, each driver is rated for a certain amount of energy *within a certain frequency range*, which is handled by the cross-over network and the impedences of the speakers at various frequencies.

If your amp is putting out 100W at 12,000KHz, it'll probably overdrive the tweeters on your 200W-rated speakers and cause them to blow (either overheat and die or over-excurt and fall apart), since the tweeter itself generally isn't rated to handle the majority of the sound energy.

The speakers in front of me have five drivers. And, they also have fuses in back, which probably wouldn't help in this situation.

EDIT: yet again, my post is redundant. I really should learn to finish reading a thread before responding. Thanks DigitalMan.

-brendan

On that 'overloading the speakers' theory, perhaps the sane thing to do is:

1/ Buy (loan?) the amp
2/ Hook it up.
3/ Get a good mic
4/ Feed Mic and one channel (perhaps right after the preamp) into an oscilloscope
5/ Play
6/ Increase volume until Input != Output
7/ tape off everything above current output.

Or am I missing something?

Yes. For a start, all speakers introduce a certain amount of distortion and, unlike amps, they don't suddenly increase in distortion towards their top power rating. Your procedure wouldn't really be able to measure the maximum safe amount of power to give the speakers. It would, however, be able to tell when the amplifier is clipping, but you could tell that by plugging a scope straight into the amp outputs.

This isn't the sort of thing that can be measured easily. Speakers (especially ones with more than a single driver) tend to defy easy specification. For example, my speakers will play a 50W 3kHz sine happily - but the power at 15kHz would smoke the tweeters instantly. I think the best way to do this is to apply some common sense. Don't use your good HiFi for parties (get some second hand PA gear, for example) and don't abuse it too much.

It's a bit like driving a car - it doesn't say in the manual what the max rated cornering speed is, but you can generally tell when you are going too fast.

Yes and by keeping a constant voltage, sweeping frequency and measuring current you would have the impedance at all possible frequencies




So you're saying the safe power limit is itself a function of frequency, independant of the impedance at a given frequency? That certainly complicates things.

As I understand it, one of the main reasons why speakers can die when fed clipped signals is because of a lack of cooling for the voice coil, resulting from insufficient movement of the cone. This site explains it way better than I can, though.

It does complicate things a whole lot. Speakers are designed this way because most music has fairly little energy in the higher frequency bands. For example, the maximum safe power with white noise will likely be less than that with a real musical signal. Measuring the safe maximum power of a speaker is very complex. It's not surprising that manufacturers do it in a way which makes them look good.

An interesting read, which I can reccommend.

I do wonder how much of it (especially the 'cooling' part) applies to electrostatic speakers though.

Also, I'm slightly puzzled by the claims that there is no air flow if the speaker coil is not moving (in the flat area of a clip). While this sounds like common sense in general, the reality (I think!) is that you can view the speaker as a piston. If it moves 'slowly' (non-clipping), then the air will be pulled in nice and neat. If however it clips, it will draw vacuum (or much closer to that state), drawing in air while the vacuum equalizes itself even though the actual speaker 'piston' doesnt currently move.

I guess the 'smooth' way of things is preferrable though.

I think maybe what he means is that for the periods of time when the signal is clipped, the cone moves out to its maximum excursion and just sits there and doesn't move because for those fractions of a second there's no change in the audio signal. Also, all of the electrical power from the amplifier is going through the speaker, heating it up and damaging it. But I could be wrong, I'm no expert.

Expert or not, you've got it right..

BUT

My point is that the cooling should continue. Imagine pulling all air out of a room, but leaving a crevice open. The air will not 'teleport' in, but flow in. Violently at first, lightly later, until it's equal again. And we're speaking about fractions of seconds with clipped audio. I could imagine that a tightly designed speaker will still have air flowing, even in clipped mode.. perhaps not enough though.

Anyway, I've actually gave the techs at Bel Canto a call, very nice people. Same with the Quad people, mailed them, got an answer.

Both seem to agree on a few key points: 400Watts@8ohm is not necessarily overpowering an 140W rated speaker. The speaker ratings audio manufacturers list sometimes even include some buffer for people with bad amps probably less common with high-end speakers.

The guy at Bel Canto even said he wouldn't be surprised if even at 'full blast' the amp would not break/kill/overload the speaker.

The same story came from a reviewer (on which I partially based my purchase choice): http://www.stereotimes.com/amp100702.shtml

Anyway, I've bought the amp, and we shall see(hear) what happens soon enough.

Soon enough = now - the amp arrived, hooked it up, and well, the short version is:

Selecting a volume of '80%' on my preamp, hooked up to the 400W power amp will not overload the speakers. My conclusion is that they can indeed handle at least 300Watt. I don't dare (plus I don't like to, my ears start to hurt) to turn it up higher. Im a happy camper

So is there a difference between blowing your speakers with an over-powered amp not clipping or blowing them with an underpowered amp overdriven into clipping? Either has the same end result. Surely the better option is an underpowered amp where the distortion can be heard and interpreted as a warning to reduce the volume?

I too have questions for Boojum: what do you power these amps from? 1200W output implies an input in excess of 1800W (at best, for an efficient Class B amp anyway), times two for a stereo pair is approaching the limits for a (UK) power outlet! How far do you retreat at these levels? ;-)

Amp clipping can occur without smooth onset... suddenly you end up in square wave territory and your tweeter is gone. I think distortion caused by over-driving your speakers has a smoother onset.

I think the assumption that you'll hear the clipping from an under-powered amp before the damage is done isn't correct.

-brendan

Absolutely, especially after a couple of beers, or during a party. The only time i've ever killed a tweeter (luckily the speakers were old and damaged anyway) was while I was mowing the lawn and cranked up the volume. You aren't always aware of distortion before it's too late

Would it help if instead of mere HPF on crossover we put another LPF just before the tweeter, effectively making a BPF into the tweeter?

Or even build your customized crossover with LPF into the woofer, BPF into the midrange, and another BPF into the tweeter?

Am I the only one who pictured you being drunk and running over a small bird with your lawnmower? *TWEET*

Just the thought of trying to listen to music while lawnmowing had me rofl-ing already

ANY clipping is the equivalent to some square-wave content, so by definition you're correct. However, the power levels will be small and the frequency highest at the onset of clipping. I've never tried detecting clipping audibly to see at what level it becomes noticeable, so I can't claim to know if it's a valid method. You have intrigued me enough to do some testing though, if I can figure a way to do it sensibly.

As an aside, I did do some square-wave testing on my own amp (Technics SU-V3) a few years ago, loaded by my speakers, not resistors. A 'scope connected to the output couldn't discern any rounding or ringing until over 10kHz, something I was seriously impressed with. I don't suppose I had the volume too high though, but it sure never blew the tweeters! The amp is rated at 40+40W and the speakers are, nominally, 75+75W so I'd be in dangerous territory according to the theories. Rough calculations suggest that the amp has just over 4dB headroom before clipping (allowing for some "sag" in the power supply).