Speaker or Amplifier - which should be more powerful?

[IndianEars]

Audiophiles often feel that an under powered amp will protect their speakers. This is unfortunately not true, and a 20 watt amp driving 50 Watt speakers ( whatever that means .... but that's a different story, for another post ) can easily damage the 50 Watt speakers.

How ???????

1. Most Loudspeakers have atleast 2 drivers:
A Low Freq (LF) Woofer which handles most of the Power.
The Woofer of a 50 watt speaker will typically be able to handle 50 watts of power

And

2. High Freq ( HF ) tweeter which typically is rated to handle only about 10% of the rated power, since High Frequencies in normal music have very little power.

A 50 Watt Speaker system could therefore have a 50 Watt Woofer + a 5 Watt Tweeter ( and ofcourse a cross-over network that steers the Low Freq to the Woofer and the High Freq to the Tweeter ).

If this system is connected to a 20 Watt amplifier under heavy clipping, the clipped amplifier waveform, according to Fourier's theorem will contain a LARGE amount of Harmonics... ie High Frequencies that are not present in the original signal.

There could easily be 10 Watts of power, from the clipping amp ( @ 50% distortion) , at 10 KHz and higher frequencies.

The cross over network will steer all the 10 Watts to the Tweeter............................ The Tweeter will tweet its last ...

Corollary to Fourier's Theorem: A Lower Watt Amp can blow a Higher watt speaker.

 

[Thad E Ginathom]

Yes, just because the power of my amp is less than the "power" of my speakers does not mean I can forget that risk --- although, in practice I can, as my listening volumes would never even approach clipping.

 

[IndianEars]

A more important criteria for selecting a speaker is whether it is capable or creating enough Acoustic output, to deliver music in your room, at the levels you want.

Its like ensuring that the Airconditioner that you purchase is capable of cooling your room ...

 

[jls001]

Thanks for the knowledgeable replies, Kanwar and IndianEars.

So what I understand now is the clipped portion - the flat top and bottom - is the DC component.

Further, the clipped portion of the waveform becomes "squarish" and, by definition, a square wave is made up of an infinite number of sinusoidal harmonics (courtesy our friend Joseph Fourier whom I continue to hate for giving me a tough time in college:lol: but, to his credit, I hate Laplace also equally). So it can undesirably shift the energy to the treble where the tweeter may not be rated to handle the extra power.

So my take away from this is that we should not operate a low-powered amp at clipping or near clipping. I guess this is deducible by common sense too. By corollary, as long as we are OK with the compromised SPL and not driving an amplifier hard, it is OK to use a low powered amp to drive speakers with higher power handling (e.g. many happy users of the 5Watt Pass AmpCamp driving various speakers like the 88dB efficient GR Research X/LS Encore BSs).

Now this brings us to the second part of the original post on this thread - what is sacrosanct about the 8? (or for that matter 4?) rating of a speaker?

I understand some vintage amps have multiple taps like 16, 8 and 4? to drive the power into the correct speaker load, but 8? seem to be the most common modern impedance rating, and the maximum power handling of amplifiers are usually specified at this load impedance. Why the choice of 8?? Why not 6 or 4 or 2 or 16? Or 600?

Further, I can't reconcile the concept of having low output impedance and high input impedance (in the ratio of 5 to 10?), whereas the maximum power transfer theorem says that the source and load must be of equal impedance for maximum power transfer from source to load. Can someone explain this for me?

 

[IndianEars]

So what I understand now is the clipped portion - the flat top and bottom - is the DC component.

Sorry to pick bones.... this is not correct.

As I have mentioned :

When an amplifier is over driven into clipping, the waveform may not be symmetrical... the Top Half of the wave form may be taller than the bottom half of the waveform.

The unsymmetrical extra is the DC portion

If the amplifier clipping is completely symmetrical, there is Zero DC.

Further, the clipped portion of the waveform becomes "squarish" and, by definition, a square wave is made up of an infinite number of sinusoidal harmonics (courtesy our friend Joseph Fourier whom I continue to hate for giving me a tough time in college but, to his credit, I hate Laplace also equally). So it can undesirably shift the energy to the treble where the tweeter may not be rated to handle the extra power.

So my take away from this is that we should not operate a low-powered amp at clipping or near clipping. I guess this is deducible by common sense too. By corollary, as long as we are OK with the compromised SPL and not driving an amplifier hard, it is OK to use a low powered amp to drive speakers with higher power handling (e.g. many happy users of the 5Watt Pass AmpCamp driving various speakers like the 88dB efficient GR Research X/LS Encore BSs).

This is Purrrfect :clapping:

 

[kanwar]

Take a case of LF signal such as Bass, if the amplifier is clipping heavily[6dB overdrive], the sinewave burst in bass signal will almost turn to squarish at top and the bottom, there the wave which is flat for finite interval of time [when at bottom or top]is nothing but plain Rail DC voltage of either polarity subjected i.e output stage entering saturation and the output is almost touching rails for some finite interval of time, that condition is nothing but plain DC current entering the voice coil though the duration is much much less[not a proper flat DC, its a squarewave, hence can't be termed as DC component if Fourier Analysis is taken]. But the heating of voice coil at that condition is much severe because it will mostly get converted into thermal power not any acoustic power by the voice coil.

 

[IndianEars]

Dear Kanwar,

With all due respect , I disagree when you say :

the wave which is flat for finite interval of time is nothing but plain Rail DC voltage

A Flat Wave top is NOT Pure DC.

Whether the falt wavetop is due to the Output devices saturating and in effecrt connecting the Load to the power rails is not relevant.

An unclipped, symmetrical Square wave will have ZERO DC....

DC is the difference in the waveform symmetry, so even an unsymetrical, smooth sine like wave, which has un-equal top and bottom halves, will have DC.

Most Loudspeakers have less than 5% acoustic efficiency ( some Horn Speakers excluded ) so atleast 95% of all electrical power reaching a speaker ... AC or DC gets converted to heat.

DC is a concern in the following 2 scenario :

1. A significant and permanent DC Offset in the amplifiers output will continuously Cook the speaker voice coil, without producing any sound ( eg no Hum ) sdo the listener does not even know his speaker coil is cooking ...

2. A Large amount of waveform unsymmetry will yield DC which will displace / use up a part of the speaker's excursion, so the speaker cone will not have its full mechanical swing available for the music. As a practical consequence, the speaker will distort, at relatively lower acoustic output. Ofcourse this is only temporary, and the speaker is restored when the signal ( DC + AC ) is removed.

3.Yes, permanent damage Can occur is the DC Current x DC Voltage ( DC watts) exceeds the speaker's rating for a substantial period of time. The Speaker magnet could also partly loose magnetisation in extreme cases, as mentioned by you, but this is really more applicable in Pro Audio where Killo Watt Amplification is used.

 

[reubensm]

having said that, there is another league of guys that drive highly sensitive higher wattage speakers with low power tube amps (with greater tollerance when it comes to distortion, clipping and overloading)

 

[kanwar]

IndianEars,

Let me put that in another words.

When you say a symmetric Squarewave has net Zero DC component, i agree to that, going by Fourier , yes the net DC component is Zero. But i am not talking about that DC component.

But when i say the clipped sinewave , its top and bottom chopped off, here i am not talking about net DC component , here what i am trying to say is that the effect of heating of voice coil is more due to the fact that the flat portion of the wave injects more DC like current for higher time in the voice coil, as compared to unclipped sinewave. That portion of clipped flat wave certainly represents DC like heating effect.

If you take 100V peak Sinewave and 100V Squarewave, the Squarewave will produce more heating in the resistive load. That is my point.

When i say DC rail voltage, i am not saying about net DC component, but the interval where the current is almost maximum for finite sustained amount of time [greater than unclipped sinewave]which can fry tweeter coil.

 

[IndianEars]

jls001 said:

Now this brings us to the second part of the original post on this thread - what is sacrosanct about the 8? (or for that matter 4?) rating of a speaker?

I understand some vintage amps have multiple taps like 16, 8 and 4? to drive the power into the correct speaker load, but 8? seem to be the most common modern impedance rating, and the maximum power handling of amplifiers are usually specified at this load impedance. Why the choice of 8?? Why not 6 or 4 or 2 or 16? Or 600?

Long long ago when radios and amplifiers used valves, speakers had high impedance. This is because valves operate at high voltage ( say 400 Volts) and low currents ( less than 1 Ampere). These operating points facilitate delivery of high voltage and low current to speakers. As a result, speakers used in valve radios had high impedances like 1,200 Ohms and 600 Ohms.

It was difficult to wind speaker coils with such high impedances. Very thin wires were required with a large number of turns.

It was much easier to wind 16 Ohm speakers and use a transformer between the valve and the speaker.

The transformer acts like a gear box which changes the ratio of high voltage and low current delivered by the valve to low voltage and high current required by a 16 Ohm speaker.




Solid state amplifiers operate at relatively low voltages ( usually less than 200 Volts) and deliver high currents ( Several Amperes). These are ideal for directly connecting to a loud speaker.

Initially, 16 Ohm loud speakers were adopted as a compromise between what the valve prefers, what the transistor prefers and what is easy to manufacture! As transistors/solid state amplifiers became more prevalent 8 Ohm speakers were widely adopted.

If 2 loud speakers are to be used in one enclosure, with each loud speaker equal to 8 Ohms, these could be wired either in series for 16 Ohms (good for valves) or in parallel for 4 Ohms (good for solid state).

Due to dominance of solid state impedances moved down to 4 Ohms for speakers. Also a 4 ohm speaker will accept 4 times as much power from an ideal amplifier, compared to an 8 Ohm speaker.

It is obvious why speakers impedances have moved down not up.

Incidentally, all the talk about 4, 8 and 16 Ohm speakers is predominantly theoretical. When the loud speaker is inside a cabinet, its impedance fluctuates very widely with frequency. For e.g., an 8 Ohm speaker could dip to 2 Ohms at 20 Hz and rise to 15 Ohms at 20 Khz !

So don't focus too much on a single numerical value of the speaker impedance.

 

[kanwar]

Also a 4 ohm speaker will accept 4 times as much power from an ideal amplifier, compared to an 8 Ohm speaker.

How come ?

 

[IndianEars]

If you take 100V peak Sinewave and 100V Squarewave, the Squarewave will produce more heating in the resistive load. That is my point.

The Heat produced (power) is represented by the area enclosed by the curve. A square wave produces more heat ( power) because a square encloses a larger area under its curve, than a Sine wave.

The square wave will deliver more power ( and No DC if its symmetrical ) irrespective of whether the amp is clipping or not .

Any ways, lets drop it . I think the question posed by jls001 has been answered

 

[IndianEars]

Further, I can't reconcile the concept of having low output impedance and high input impedance (in the ratio of 5 to 10?), whereas the maximum power transfer theorem says that the source and load must be of equal impedance for maximum power transfer from source to load. Can someone explain this for me?

You are correct. The Maximum power transfer theorem states that FOR AN IDEAL VOLTAGE SOURCE the maximum power is delivered when the source (amplifier) impedance is equal to the external load impedance.

A solid state amplifier will typically have an output impedence of 0.1 Ohms. Yes, it would deliver maximum power if the speakers impedence was 0.1 Ohms but the amplifier is NOT AN IDEAL VOLTAGE SOURCE. It cannot provide indefinite amount of current required for its full voltage swing, if the speaker had a 0.1 ohm impedance.

 

[IndianEars]

Quote:
Originally Posted by IndianEars View Post
Also a 4 ohm speaker will accept 4 times as much power from an ideal amplifier, compared to an 8 Ohm speaker.
How come ?

OOps !

Typo !

 

[jls001]

OOps !

Typo !

I thought so too

Thanks for the excellent explanations. The difference between an attentive student (like you and Kanwar) and an inattentive dunce (like me) is the former remembers "minor" details like "FOR AN IDEAL VOLTAGE SOURCE" and understands the difference between sinusoids and square waves, whereas the latter glosses over it. Anyway, as they say, "education is what you remember after you forget what you learned":lol:

 

[keith_correa]

Ok, so let me introduce a twist to this. How much more "powerful" does an amp need to be than the speakers? Numbers please. The numbets that are typically quoted in specs. It's high time this s##t was decoded.

 

[kanwar]

1.5X times is norm in pro-audio.

 

[gilfredpaul]

In that case, if my speaker continuous power handling 100w ...then I should choose 150w amp

 

[jayants]

Reading through all the posts when I discussed with some other audiophile greats in other forums this is what they observed:

"1) Over Powering the Speakers with 1.5 times Amp. and If turned the volume all the way up to 80% you have a Toasted Speakers for Dinner

2)Hmmm so these guys say use an amp that is rated more than the handling power of the speakers! Does that make much sense to anyone? It's doesn't to me. All in the name of avoiding clipping? Hmm! Again. An amp clips when the speaker can't handle it's power and current backs up into the amp. Many high end amps have circuits to stop this overload.try connecting a 100 watt rms amp to a 50 watt rms rated speaker and turn the volumes to 60% and hear what happens. Do it at the risk of blowing your drivers.

3) Speakers always come with rms power handling ratings. It is wise to choose an amp within that range. If one uses an amp more powerful than the handling capacity, they could end up frying the drivers within a short time frame, esp. at higher volumes. The above scenario also will cause excessive amp clipping and if the amp is not adequately protected, blow a cap in the amplifier and /or overheat it to failure

4) if the rms rated power of the amp is 1.5 times that of the handling capacity of the speaker, the speaker will be doomed to failure. If anything, the reverse might be true."





So now its very confusing.It seems there are 2 schools of thought completely different from each other.

 

[koushik]

Everywhere they mentioned that when you drive it loud it will fry the driver. That is true BUT you have take care of that. You will reach the loudness you want within the 50% of the powerful amp. Let me site an example -

- If you use 100W pc amp for a speaker with max handling capacity of 120W
You would reach say x dB at 90% of volume setting and that is your desirable volume. The amp is driven at 90W pc.

- If you use 150W pc amp for the same speaker with max handling capacity of 120W
You would reach the same x dB at 60% of volume setting. The amp is driven at 90W pc.

In the first case, you are pushing the amp to the limit and hence the distortions comes in but not for the 2nd amp at only 60% of the rated power. Why would you again pump up the control when you have already reached your desired volume level! In any case you can't pump it up more than 10% in the low powered amp.

That is my 2 paisa!