Speakers and power - explain this bit....

[john_the_bass]

I've just seen this in a buyer's guide on the sweetwater website:

[quote]See the Power Amplifiers section of this buying guide. Fortunately, you can determine this without having to do (much) math. JBL recommends that, in general sound reinforcement situations, you use an amp that delivers equal to or up to double the IEC power rating of the loudspeaker, i.e., a speaker rated at 300 watts capacity needs a 300- to 600-watt amp. Contrary to popular belief, you're more likely to damage your speakers with an underpowered amp than with one that has too much power, so don't scrimp here![/quote]

How does that work then? How are you likely to knacker your speakers with an underpowered amp? Is that because you're more likely to try and drive your speaker harder if you amp doesn't put out that much power? Surely though if you get a very sensitive speaker, then you can use a lower powered amp?

Answers on a postcard

[Musky]

I think you've got the answer John.

I've seen this view before for PA amps, and some advocate it for bass amps as well. I seem to remember that driving an amp into clipping can damage the speakers.

[BOD2]

Ah this old chestnut again

That oft-quoted JBL article has done more to confuse the issue than it has to explain anything.

Try searching this forum for "clipping" and/or "JBL". There have been a few threads that went into this before that you might find help.

The bottom line however, is that you can use ANY combination of amp and cab power together (i.e. big amp & small cab or small amp & big cab) provided you LISTEN and back off as soon as any unexpected distortion occurs.

Also, there was a question as to whether the JBL article in question was referring to full range hi-fi types of setups rather than purpose designed bass rigs.

Cheers

[Protium]

[quote name='Musky' post='313429' date='Oct 23 2008, 11:17 PM']I think you've got the answer John.

I've seen this view before for PA amps, and some advocate it for bass amps as well. I seem to remember that driving an amp into clipping can damage the speakers.[/quote]

Pushing the speakers past their Xlim will damage them. Clipping from an amp won't.

[Moo]

The clipping amp destroying speakers thing is a Myth. - Do guitar amp cabs blow up every time the overdrive is switched on?
Too much power destroys Cabs, either mechanically with over excursion or Thermally by overheating and melting the voice coils.
There are some circumstances where if the power of the amp EXACTLY matches the melting point of the speaker voice coils, the extra power given to them by clipping will take them over the thermal threshold and melt them, but this wont happen with a 50w Guitar amp clipping into a 100w Cab for example.

The reason that 2x amp to speaker ratio is recommended for P.A. is to avoid distortion and sound nice.
The max power rating of Amps is measured at the point where they reach 1% distortion.
This means that if you run a 300w amp into a 300w speaker at its max level you will be amplifying 1% distortion as well as the sound you want. The harder you push them, the harsher they sound.
If you want to have a nice clean sound in your P.A. then you need to run at a level which is less than the Max rated level of the amp and leave yourself some headroom.
With a 600w amp into a 300w speaker, you can run it at 3db (Half the power) below its 1% distortion threshold and keep the sound clean. It is worth noting that High power P.A. systems will also be running with crossovers and limiters that will be set up so that the Amps output voltage will not exceed the thermal rating of the speakers and melt the voice coils.
Unless you have limiters and have set the gain structure of the amps correctly, then it would be unwise to run 300w speakers with a 600w amp because if any 'helpers' who don't know your system turn up the amp and run it into the red then the speakers will be cooked.

For bass amps the requirement is slightly different.
There is only one manufacture of bass heads who routinely includes a limiter on the output of there amps (They also make P.A. amps). Unfortunately this limiter is pretty much useless for protecting the cabs, because it is not adjustable, and obviously not able to telepathically detect the thermal voltage rating of the cabs it is connected to.
Many people prefer there sound slightly distorted and obviously some like it a lot distorted. In these circumstances it is wise to have cabs that have a higher rating than the amp they are connected to.
If the amp is rated higher than the cabs, it could in theory sound very clean and 'Hi Fi', but one slip with the volume knob, or one slap too hard will result in fried cabs.

Moo.

Edited October 24, 2008 by Moo

[alexclaber]

[quote name='Moo' post='313497' date='Oct 24 2008, 05:03 AM']If the amp is rated higher than the cabs, it could in theory sound very clean and 'Hi Fi', but one slip with the volume knob, or one slap too hard will result in fried cabs.[/quote]

Nope. Thermal damage requires enough continous power to overheat the voice coil. Bass playing has loud and quiet bits - the only time you might be able to sustain full power for more than a brief moment is when using overdriven feedback or an all-out fuzz pedal.

Using an amp with much greater power than your speaker cab is often futile due to the miserable low frequency (excursion limited) power handling of most bass cabs. Having enough power that your amp never clips will reduce the risk of damage to tweeters and midranges. Woofers are not bothered by clipping, only by too much continuous power (thermal damage) or massive spikes of low frequency power (mechanical damage).

The vast majority of damage to bass guitar speakers is mechanical, not thermal and can always be avoided by turning down or reducing the bass EQ on your amp in response to farting noises from your cab.

Alex

[Adrenochrome]

JBL want you to spend more money on their bigger amps.

[Thunderhead]

Speakers are in fact only ever blown by too MUCH power - you cannot blow one with too LITTLE - but the mechanism by which too much power reaches the driver is not quite as straightforward as you might think, and it IS possible for an underpowered amp to blow a higher-rated speaker.

[quote name='Moo' post='313497' date='Oct 24 2008, 05:03 AM']The clipping amp destroying speakers thing is a Myth. - Do guitar amp cabs blow up every time the overdrive is switched on?[/quote]No, but that's because guitar cabs are very crude with only "low/mid" drivers and no crossovers, and the valve amps normally used to drive them have a limited frequency range too, due to their output transformers which block very low and very high frequencies almost completely. (Even solid-state guitar amps are designed to have this sort of frequency response, precisely to make them sound more like valve amps.)

But in a cab with a crossover driven by a solid-state amp with a full-range frequency response, it's very different. A clipped signal starts to approximate a square wave, which is effectively a series of sections of DC separated by sharp transients - ie entirely composed of extreme low and extreme high frequencies - and a solid-state power section is quite capable of delivering this since there is no output transformer. If the crossover is really efficient, it will then send the full power of the amp alternately to the bass driver and the tweeter, instead of evenly distributing a range of frequencies... which may fry both units, particularly the tweeter (since in normal program music the power in the highs is substantially lower than that in the lows, so the tweeter does not normally have to handle as much), and might even damage the crossover since it's not designed to handle that sort of signal either.

[quote]Too much power destroys Cabs, either mechanically with over excursion or Thermally by overheating and melting the voice coils.
There are some circumstances where if the power of the amp EXACTLY matches the melting point of the speaker voice coils, the extra power given to them by clipping will take them over the thermal threshold and melt them, but this wont happen with a 50w Guitar amp clipping into a 100w Cab for example.[/quote]Exactly, but it CAN happen with a 500W PA amp clipping into a 1000W cab, because a clipping 500W amp can be putting out up to 1000W (the energy contained in a square wave is double that in a sine wave of the same peak voltage, and even if it's not a full square wave the energy rises drastically as soon as clipping is reached), and the individual drivers in the PA cab won't be able to handle this - especially the tweeter, which will more likely be rated around 300W. The bass driver may not like being hit with bursts of DC at that sort of power level either... DC is very bad for speakers because it does not allow the voice coil to cool itself properly as it is not actually moving while it's absorbing the power.

[quote]The reason that 2x amp to speaker ratio is recommended for P.A. is to avoid distortion and sound nice.
The max power rating of Amps is measured at the point where they reach 1% distortion.
This means that if you run a 300w amp into a 300w speaker at its max level you will be amplifying 1% distortion as well as the sound you want. The harder you push them, the harsher they sound.
If you want to have a nice clean sound in your P.A. then you need to run at a level which is less than the Max rated level of the amp and leave yourself some headroom.
With a 600w amp into a 300w speaker, you can run it at 3db (Half the power) below its 1% distortion threshold and keep the sound clean. It is worth noting that High power P.A. systems will also be running with crossovers and limiters that will be set up so that the Amps output voltage will not exceed the thermal rating of the speakers and melt the voice coils.
Unless you have limiters and have set the gain structure of the amps correctly, then it would be unwise to run 300w speakers with a 600w amp because if any 'helpers' who don't know your system turn up the amp and run it into the red then the speakers will be cooked.[/quote]Absolutely! The 'over powered amp' theory only applies if the user does not allow distortion to occur - if it does, the speakers will be toast immediately since the power applied to them is then even further over their rating.

[quote]For bass amps the requirement is slightly different.
There is only one manufacture of bass heads who routinely includes a limiter on the output of there amps (They also make P.A. amps). Unfortunately this limiter is pretty much useless for protecting the cabs, because it is not adjustable, and obviously not able to telepathically detect the thermal voltage rating of the cabs it is connected to.
Many people prefer there sound slightly distorted and obviously some like it a lot distorted. In these circumstances it is wise to have cabs that have a higher rating than the amp they are connected to.
If the amp is rated higher than the cabs, it could in theory sound very clean and 'Hi Fi', but one slip with the volume knob, or one slap too hard will result in fried cabs.[/quote]I agree - I prefer to treat bass systems more like guitar ones. I do use fuzz bass myself (which makes it essential) but even for clean sounds it's safer to go with more power handling for the speakers.

In the end, there is only one truly safe method - make sure the speakers (including all the individual elements inside the cabs, for multiple-frequency cabs) can handle the full distorted power of the amp. This does mean using much heavier and more expensive drivers than you really need for a clean sound though, so it isn't a great way to build a PA system usually... although I have done this when working as a repair engineer for rehearsal studios (ie PA systems being run unattended by musicians!), which I mentioned in the thread about protecting speakers.

Normally for single-range guitar or bass speakers, double the clean power rating of the amp is enough - an amp can't produce more than double its clean power even fully distorted (although you can still run into the problem that the speaker power rating doesn't necessarily apply at near-DC frequencies, and mechanical cone-excursion damage can also occur even when electrical damage doesn't). For PA, you may need three or four times the amp rating to be really sure you won't blow the tweeters, which obviously is going to require otherwise unnecessarily expensive and heavy cabs, so it's not the right solution for most users, when simply avoiding distortion will do the job.

Edited October 24, 2008 by Thunderhead

[alexclaber]

[quote name='Thunderhead' post='313633' date='Oct 24 2008, 10:49 AM']But in a cab with a crossover driven by a solid-state amp with a full-range frequency response, it's very different. A clipped signal starts to approximate a square wave, which is effectively a series of sections of DC separated by sharp transients - ie entirely composed of extreme low and extreme high frequencies - and a solid-state power section is quite capable of delivering this since there is no output transformer. If the crossover is really efficient, it will then send the full power of the amp alternately to the bass driver and the tweeter, instead of evenly distributing a range of frequencies...[/quote]

No. A square wave is not sections of DC separated by sharp transients, it is an AC square wave. The crossover will definitely not send full power alternately to the bass driver and tweeter, it will continue to filter the output as described by its highpass and lowpass functions.

This shows the harmonic composition of a perfect square wave (first 20 harmonics shown):



Note that the voltage drops with frequency so the tweeter will never receive as much power as the woofer. However tweeters always have much lower thermal power handling than woofers and the increased HF content in a distorted signal can cause overheating.

[quote name='Thunderhead' post='313633' date='Oct 24 2008, 10:49 AM']Exactly, but it CAN happen with a 500W PA amp clipping into a 1000W cab, because a clipping 500W amp can be putting out up to 1000W (the energy contained in a square wave is double that in a sine wave of the same peak voltage[/quote]

This is the risk with a clipped amp - increased total power output. It isn't as bad as the simple analysis suggests though because of the lowpass function of the woofers' inductance reducing the current flowing at higher frequencies.

[quote name='Thunderhead' post='313633' date='Oct 24 2008, 10:49 AM']and even if it's not a full square wave the energy rises drastically as soon as clipping is reached), and the individual drivers in the PA cab won't be able to handle this - especially the tweeter, which will more likely be rated around 300W.[/quote]

The energy seen by the tweeter (and/or midrange speaker) increases dramatically at clipping. The energy going into the woofer does not.

[quote name='Thunderhead' post='313633' date='Oct 24 2008, 10:49 AM']The bass driver may not like being hit with bursts of DC at that sort of power level either... DC is very bad for speakers because it does not allow the voice coil to cool itself properly as it is not actually moving while it's absorbing the power..[/quote]

This is completely untrue. In an amp that is not broken there should be no DC output. A square wave is not DC. Even if a woofer had a square wave applied to it the inductance of the woofer would make the wave less square. The woofer does move during the flat sections of the AC square wave - woofer position does not match the wave shape, it matches the double integral of the current wave (which itself is derived from the square voltage wave applied to the impedance function of the speaker).

I agree that if you're into dirty bass sounds and/or use lots of FX pedals it is wise to use speakers with more rather than less power handling. If you prefer clean bass sounds then you are extremely unlikely to ever run into thermal problems. Distorted signals are prone to overstressing high frequency components but a good speaker cab should be designed to handle these signals and if need be should incorporate protection circuitry.

Alex

[Thunderhead]

Alex, a square wave IS sections of DC - each the length of half the complete waveform, and of alternating polarity - separated by vertical transients. Just look at it on a scope. Yes, it's [i]also[/i] an AC square wave.

A clipped solid-state power section is quite capable of producing this (subject to the slew rate of the output devices, which makes the transients not quite vertical). Again, look at it on a scope.

To the crossover, the sections of DC are 0Hz and so will be sent exclusively to the bass driver.

I know the bass driver cone does not perfectly follow this - it can't due to its mechanical inertia - but it will try, and the lower the frequency the closer it will get. Connect a signal generator to a speaker and watch what happens to the cone as you lower the frequency with a square wave input.

[alexclaber]

[quote name='Thunderhead' post='313658' date='Oct 24 2008, 11:27 AM']Alex, a square wave IS sections of DC - each the length of half the complete waveform, and of alternating polarity - separated by vertical transients. Just look at it on a scope. Yes, it's [i]also[/i] an AC square wave.[/quote]

DC never alternates. That's it, end of argument.

[quote name='Thunderhead' post='313658' date='Oct 24 2008, 11:27 AM']A clipped solid-state power section is quite capable of producing this (subject to the slew rate of the output devices, which makes the transients not quite vertical). Again, look at it on a scope.[/quote]

Yes it can, where did I say that it can't? I just pointed out that your incorrect definition of there being DC involved is very misleading and incorrect.

[quote name='Thunderhead' post='313658' date='Oct 24 2008, 11:27 AM']To the crossover, the sections of DC are 0Hz and so will be sent exclusively to the bass driver.[/quote]

The flat top sections of a square are not zero Hz, they have the same fundamental frequency as the unclipped note. See the FFT I posted previously.

[quote name='Thunderhead' post='313658' date='Oct 24 2008, 11:27 AM']I know the bass driver cone does not perfectly follow this - it can't due to its mechanical inertia - but it will try, and the lower the frequency the closer it will get. Connect a signal generator to a speaker and watch what happens to the cone as you lower the frequency with a square wave input.[/quote]

You may understand the workings of an amp but you clearly do not understand how a speaker works! The voltage applied to the voice coil causes current flow. That current flow generates force in the motor which causes cone acceleration. That acceleration causes velocity. That velocity causes a displacement. Therefore the displacement can be calculated by integrating the current (i.e. force i.e. acceleration) to get velocity and integrating the velocity to get displacement, all the while bearing in mind that voltage and current are related by the variable impedance curve including the reactive elements (the speaker is not a simple resistive nominal load).

Integrate a square wave and you get a triangular wave. Integrate a triangular wave and you get a sine wave.

Alex

[johnnylager]

Bear in mind I know nothing about this topic mind...

From another place... [attachment=15143:clipping_and_you.pdf]

[url="http://en.wikipedia.org/wiki/Crest_factor"]Crest factors[/url]

Edited October 24, 2008 by johnnylager

[Thunderhead]

Alex, just forget the theory of the whole waveform for a minute and think about what happens [i]within each half of the cycle[/i].

To quote the article linked above:

"One of the most notorious myths regarding clipping is that it produces DC. This
assumption is made because a square wave is seen to have flat tops and bottoms. In fact,
it's incorrect to think of a squarewave as being made up of positive and negative DC
components. The only way for it to be DC would be if there was a non-zero average value
over long periods of time.[b] If the polarity changes at all within the time frame that you are
looking at[/b], it is simply not DC."

If the time frame you are looking at is less than half the frequency, the polarity does not change. For the duration of half the fundamental frequency (and remember that clipping the output stage effectively reduces the frequency to that because it obliterates all the higher components) what you have is DC. I agree that what you have is a square AC waveform [i]overall[/i], but it [i]can also be[/i] described as sections of DC separated by transients. The fact that they do alternate in polarity is irrelevant if you consider what is happening [i]in the duration of each section[/i]. Both ways of looking at it fit the waveform but the 'alternating DC' description explains more usefully why a clipped waveform can be so damaging to speakers, especially cabinets with crossovers. It isn't [i]just[/i] to do with the increased power when clipping occurs (although that is significant too).

Sometimes it can be useful to throw away your full mathematical analysis and just look at things in a more simplistic way, especially in cases where a sudden change of operating conditions occurs (which it does when an output stage clips).

[bass_ferret]

Not again! We need a smiley for yawn!

[alexclaber]

[quote name='Thunderhead' post='313729' date='Oct 24 2008, 12:49 PM']If the time frame you are looking at is less than half the frequency, the polarity does not change. For the duration of half the fundamental frequency (and remember that clipping the output stage effectively reduces the frequency to that because it obliterates all the higher components) what you have is DC. I agree that what you have is a square AC waveform [i]overall[/i], but it [i]can also be[/i] described as sections of DC separated by transients. The fact that they do alternate in polarity is irrelevant if you consider what is happening [i]in the duration of each section[/i]. Both ways of looking at it fit the waveform but the 'alternating DC' description explains more usefully why a clipped waveform can be so damaging to speakers, especially cabinets with crossovers. It isn't [i]just[/i] to do with the increased power when clipping occurs (although that is significant too).[/quote]

You may enjoy throwing away the science to prove your point but it doesn't make you right. If you pass a square AC waveform at 220Hz (so a fully clipped amp with a bassist running extreme feedback on a sustained 14th fret G-string A) through a 2nd order lowpass filter set at ~500Hz the filter will only pass the fundamental and the 2nd harmonic fully, some of the 3rd harmonic, hardly any of the 4th harmonic and so on. It will not pass your claimed 0Hz DC signal because it isn't zero Hz DC. Any crossover (or any reactive component) has a time element in how it works with the input signal and claiming that a snapshot of the signal during a fraction of the waveform is representative of real-world usage is ridiculous.

Your 'alternating DC' description of why a clipped waveform can be damaging is totally incorrect and to be perfectly frank I am getting very tired of arguing this with every new forumite who thinks he understands the subject but doesn't. If you pop along to talkbass and do some searches there you will see that I am not the only person telling you that this DC rubbish is wrong.

Alex

P.S. If you wish to be pedantic there is a minor issue of DC offset affecting cooling but it is nothing to do with your claims of 'alternating DC' and with high Xmax drivers it makes little difference when there is always a large portion of poorly cooled voice coil outside the magnetic gap.

Edited October 24, 2008 by alexclaber

[The Funk]

[quote name='bass_ferret' post='313736' date='Oct 24 2008, 12:56 PM']Not again! We need a smiley for yawn![/quote]

A yawnie?

[Balcro]

[quote name='bass_ferret' post='313736' date='Oct 24 2008, 12:56 PM']Not again! We need a smiley for yawn![/quote]


That's a good idea.

A brilliant exposition, gentlemen and I wouldn't dare to intervene. There are still 2 rounds to go and by my reckoning and a slight "non-scientific feeling" Thunderhead is probably ahead by 15 points to 14.

My knowledge on the subject has increased by 2% - the rest has gone completely overthe top of my head.

I think BOD2 summed it up nicely and a nice entry from johnnylager. Personally, as I have a hi-fi cortex and a bass cerebellum I'm quite happy to match anything with anything. Never had a problem. Keep it clean; of you want overdrive or distortion - add a pedal.
Why has no one ever devised simple lockable gain/volume controls? I can password lock my suitcase for a £1.

Right bass_ferret, shall we go down the pub.

Balcro.

[Bill Fitzmaurice]

This is the source of the controversy:
[url="http://www.jblpro.com/pub/technote/lowpower.pdf"]http://www.jblpro.com/pub/technote/lowpower.pdf[/url]


Those who actually comprehend the Queen's English will note that the underpowering caveat applies only to 'High Frequency Components', ie., tweeters.

Those who do not have completely subverted the intent of the document, arriving at the conclusion posted on the Sweetwater site.

Edited October 24, 2008 by Bill Fitzmaurice

[alexclaber]

[quote name='Balcro' post='313802' date='Oct 24 2008, 01:59 PM']A brilliant exposition, gentlemen and I wouldn't dare to intervene. There are still 2 rounds to go and by my reckoning and a slight "non-scientific feeling" Thunderhead is probably ahead by 15 points to 14.[/quote]

Actually by repeatedly repeating the myth of clipping=DC Thunderhead has unwittingly opened himself up to a knockout blow. It's rather disappointing to have to get involved in these arguments but if I don't then everyone who reads this thread will continue repeating the same new bs. Ten years ago I never heard this - just goes to show why a little knowledge is such a dangerous thing.

Alex

[Thunderhead]

Whatever. Next time you blow a PA cab, don't worry, it couldn't have happened.

Edited October 24, 2008 by Thunderhead

[Protium]

Handbags!

[alexclaber]

[quote name='Thunderhead' post='313948' date='Oct 24 2008, 04:41 PM']Whatever. Next time you blow a PA cab, don't worry, it couldn't have happened.[/quote]

Wow, does my assertion that a clipping amp does not produce DC now mean that no speaker cab can be blown because only DC can blow speakers?!

If you've observed my posting elsewhere you will note that I am generally open-minded. The exception is when I am having to defend the TRUTH and FACT from those that insist on spreading inaccuracy and myth.

You make some good points and you clearly know your stuff on many technical subjects but that doesn't mean you can't be wrong on other subjects.

Alex

P.S. I shall now await a post from someone who got told that clipping causes DC which kills speakers by a bloke in a guitar shop...

P.P.S. And then later someone who's just discovered a 21" speaker cab which goes to 20Hz...

[Merton]

[quote name='alexclaber' post='313997' date='Oct 24 2008, 05:15 PM']P.P.S. And then later someone who's just discovered a 21" speaker cab which goes to 20Hz...[/quote]
I have a friend who says his cousin's girlfriend's dad has one in his garage, so it must be true.





















Sorry, couldn't resist.

Edited October 24, 2008 by Merton

[lowhand_mike]

tbh all this is of little relevance to me, while i totally respect alex and BFM and their knowledge in my real terms it means this

turn it up

if it sounds sh*t

turn it down

adjust eq

turn it up

does it sounds sh*t?

repeat if necessary.

[bass_ferret]

As I have shamelessly robbed from Alex many times:
[size=5][b]You can use any power ouput amp with any power handling cab. If any of these combinations makes bad sounds then turn down and/or stop cranking the bass EQ excessively or damage may occur[/b][/size]

It does not need any more thought than that.

The 21inch woofer is not a myth. It was an architectural hifi driver made for the US market by Electrovoice.