How much power do I need? (How big or small an amp?)

[A.S.]

Covered in great detail here:

http://www.harbeth.co.uk/usergroup/showthread.php?t=41.

Also refer to here:
http://www.harbeth.co.uk/usergroup/s...=1920#post1920

and here:

http://www.harbeth.co.uk/usergroup/s...=2292#post2292

also here:

http://www.harbeth.co.uk/usergroup/s...=1932#poststop

Remember! You control the amplifier! It is not like a runaway train! If the amp damages the speaker that's because you didn't contol it! It does not control you!

[matthewz]

Hi all,

just want to share my experience which kind of relates to this subject.

I recently purchased a used NAD 3020 integrated amplifier ( just for fun ). But boy, let me tell you this, this amplifier is full of fun, and it's a serious piece of amp for such power output. Since I don't have its original manual, I guess it delivers 20 Watts per channel into my Compact 7.

Perhaps, as Mr. Alan Shaw mentioned somewhere in this forum before, listening to music does not necessarily mean having your neighbour knock on your door and remind you her baby's asleep. To make my point here, my NAD proves me that less that less than 20 Watts can certainly deliver the sweet melody in your listening room.

*** Merry Christmas 2007 and Happy New Year 2008 ***

Cheers,

J.L.

[Hu]

Happy New Year 2008 to all!

I havea question here, you use small amplifier to drive Harbeths and feel happy, may I ask you what sort of music you listen?

[Gan CK]

Well i have tried Matthew's combination briefly when i had the old C7 & 1st generation Nad 3020 & was rather surpirsed that they sounded pretty ok provided you don't crank up the volume too much. It was only when i switched to an Onix integrated amp then i realised that the 3020 is a little long in the tooth. The C7, like SHL-5 is extremely amp friendly but move up in the amp dept & it will reward you accordingly.

Just to share with Matthew, i got very good results from the C7 using a very inexpensive hybrid tube amp in the form of Jolida 1301. Compared to the old 3020, the Jolida was fuller, sweeter, more refined & detailed.

[David Schalkwyk]

This is, I suppose, a "technical" question for Alan, but I'd be happy if anyone else has any suggestions. I'm moving continents very shortly and will have to leave most of my hi-fi equipment behind. I do aim, however, to take over my HP3s and C7s. I'll need to get a CD player, amplifier and tuner. In the first year or so I'll be on a very restricted budget, which is why I've been attarcted by reviews of the Shanling MC-30 music system: CD player, iPod connection, tuner and single-ended class-A valve amp for about $900. Point is, the amp produces 3 watts! In a small room, listening nearfield, mainly to jazz and chamber music (with the occasional Mahler symphony for fun), will this be in any way a proposition for either the P3s or C7s. Or should I forget it?

Thanks

David

[A.S.]

...Point is, the amp produces 3 watts! In a small room, listening nearfield, mainly to jazz ...David

Well, 3W is a very tiny amount of power. And there may be a degree of 'over egging' that figure - there so often is. 3W would produce what I'd call 'conversational level' sound. But, as you say, you are listening in the nearfield, and that may (just) be sufficient.

Personally, I'd be happier to see you use 15W as a sensible minimum, especially with a mini-monitor (with necessarily low efficiency) like the P3.

[aawcy]

a lot of speaker manufacturer prefer their speaker to be driven with slighty more powerful amp....Can underpowered amp damaged the Harbeth speaker??...really caught my attention when harbeth can pair up with smaller output amp....

[A.S.]

More on amplifier power reserve here.

[A.S.]

A trawl through my archives and we can trace the history of the relationship between speaker amplifier power* and the size of the drive unit (hence its efficiency, its loudness) and the use of a rubber cone surround or not.

If we refer to the BBC LSU10, a gigantic wardrobe sized speaker in use from the late 1950s, we know that inside this powered-passive speaker was a Leak valve amplifier, rated 12W. This tiny amp produced an adequately loud sound because of three happy coincidences:

1 - The woofer was as enormous 15" unit with a powerful magnet
2 - The woofer did not use a modern rubber surround around the perimeter of the cone
3 - The 'surround' was in fact, merely an extension of the cone, corrugated to permit some (small) fore and aft motion. Unlike a heavy modern rubber surround attached to a much smaller cone, the weight of the corrugated self-surround would have been negligible and, importantly and unlike the modern rubber surround, the proportion of the diameter given over to the surround area was small. The modern surround generates negligible sound but its high mass acts as a drag - not so with these old paper units.

Now, use that same 12W amplifier to drive a modern, 5-8" rubber-edged, plastic coned speaker woofer mounted in a small enclosure and I hope that you can see that none of the above conditions are met. A modern speaker's 8 inch cone area is about 56 square inches (pi x r 2). The 15 inch woofer has a cone are more than three times bigger! That alone means it is going to produce a lot more sound for the same electrical input. Let alone considerations of the magnet power and the low mass of paper cones.

These tiny amps are all very well, but they ignore the reality of modern speakers which have much smaller surface areas, far lower electro-acoustic efficiency and much higher moving mass which can only be overcome by one expedient: more input power. So an amp of more than flea-power is really necessary to generate anything like a satisfyingly loud sound in our listening room at home and to avoid amp strain, distortion and potential damage when driving 'modern' hi-fi speakers - those made since the mid 60s when paper started to be replaced by plastic cones. It's that simple.

* About amplifier power rating. This number defines how many watts are available to deliver to the speaker load. It does not define how many watts will be delivered to the speaker. Your adjustment of the volume control does that, until the amplifiers power reserve is fully consumed by the speakers. Then, no matter how much further you advance the volume control, no more power can be delivered under any circumstances. You have hit the end stops of the amplifier. Your foot is pushing the throttle to the floor: you cannot go faster. If you only had a bigger amp with some power in reserve you could go faster - you could play louder. And by louder, I don't just mean gross loudness, the sort that annoys the neighbours: I means transient loudness, where there is an incredible burst of energy in the music, even between phrases. To capture that with fidelity, you need amplifier power in reserve.

No flea-powered amp can deliver the required number of joules to the speaker units to accelerate them up to the necessary velocity to generate the correct, realistic sound level, note by note.

See also here.
>

[A.S.]

I have nothing against small amplifiers, although it is obvious that they hail from a time when speakers were efficient. Their application in the modern speaker era raises significant questions of appropriateness especially with digital music. That's because, no thanks to tape hiss, analogue recordings have, at the very, very best, a dynamic range of about 65-70dB. Sounds quieter than seventy decibels below the loudest sound that the analogue medium (tape, vinyl) can handle are erased by tape hiss, which is just random noise. So those microtonal details that digital reproduction can reproduce are absent from analogue because they've been smothered in hiss. And one they are below the (terribly high) background hiss level of analogue systems, they are gone forever. You cannot convert hiss back into music.

In the case of digital recording and playback, the dynamic range is about 100dB. That means the loudness range from the very loudest signal that can be reproduced to the very quietest in an all-digital system is about thirty to fifty times wider than analogue. That was an astonishing achievement for the Sony/Philips team who invented the CD. Now, it could be argued that one reason that a die-hard proportion of audiophiles like the 'analogue sound' is because they positively dislike the wide dynamics of digital sound. One sure way to convert those digital dynamics into a more analogue-like dynamics is to use a very small amplifier, feed it with a wide dynamic digital recording and connect it to modern, low efficiency/high-quality loudspeakers.

Providing that the amplifier 'runs out of steam' in a relativity benign way that does not introduce obvious chronic audible distress - i.e. the amplifiers soft-clips the music through lack of power reserve - then the small amp has compressed the wider range dynamics into something more like the analogue sound. But is that what the Sony/Philips inventors would want? Is it actually high-fidelity reproduction as we would understand it? Could the same subjective effect be achieved by passing the digital signal through a signal compressor and then feeding that to a more powerful, capable amplifier?

We can make some examples to demonstrate what happens when a small amplifier doesn't have the power reserves to reproduce the dynamics of a wider-range signal. You might be surprised by what you hear: it is counter-intuitive.

Loading the player ...

Clip 1, our reference. I have normalised the audio so that the peaks hit the maximum possible loudest signal that a digital system is capable of. We call that 'Zero dB' or 0dB. Normalised to 0dB, the original dynamics of the recording are preserved; that is, the ratio between the loud, medium and quiet parts of the music are preserved in a locked relationship to each other. To play louder, you have to generate more sound in your room, which means using an amplifier that also preserves this relationship and does not alter the dynamics of the signal.

Loading the player ...

Clip 2. Here I have synthesised the effect of taking the normalised, fully modulated digital signal in clip 1 and passing it through a small amplifier that does not have enough power reserve to drive the speakers. I've actually increased the signal by 6dB (that is, doubled it) going into the amp. It does its best to follow the dynamics of the input signal, but just doesn't have enough joules available in its power supply to pump out into the speaker drive units.

Loading the player ...

Clip 3. I've taken Clip 1 and applied 10dB of level boost - about three times the level of Clip 1. The amp is completely out of steam.

Now what can we hear if we carefully listen to the clips? First, like it or not, Clip 1 is what the performers, recording engineer and producer laid down on the CD. That is, and has to be, what we should be trying to reproduce in our home audio system if we accept that it is not our role in the replay process to significantly alter the dynamics of the recording. Are we observers of other's art, or artists ourselves?

Second: and this may be counter-intuitive. I think this is the core of the misconceptions concerning small amplifiers. Do we agree that Clips 2 and 3 seem to be significantly louder than the reference Clip 1? How can that be when we know that in 2 & 3 the small amplifier is significantly out of puff and is working at and beyond clipping point? The reason is that the relationship between the quiet (or in this clip, relatively quieter) moments and the medium loudness and peak loudness phrases has been altered in the amp. The amp did have enough power to elevate the quieter and middling loudness elements but it did not have the power to lift the peaks at all: those are clipped-off. But the subjective impression by the listener who doesn't recognise the sound of an amplifier that's run out of power reserve (in its PSU) is of greater 'warmth, involvement, resolution, detail, engagement, fuller, bigger sound' etc.. Listen for yourself and see how 'breathy' Clip 3 is: you could forgive the inexperienced listener for believing that Clip 3 is more 'analogue' sounding. That is, until you recognise that the peaks are squashed and completely unnatural. Those cymbals are too soft sounding. Their dynamics have been crushed along with all other high-level notes. Does this make sense?

Loading the player ...

Clip 4. The last step then is to take Clip 3, where we drove the amp into severe clipping by lifting the input by 10dB, and then reduce the level of that output by 10dB. This synthesises an approximation* of returning the level to that of the original Clip 1, but retaining the irrevocable dynamic range compression that took place during clipping in step 3. Could we agree that we've converted the rather "cold" hard-on-the-ears digital sound in Clip 1 into a more comfortable "warm" easy-on-the-ears analogue sound of Clip 4? Clip 4 would be ideal for inoffensive background music.

Many audiophiles prefer the 'analogue' sound and we can hear why. But if the reproduction of the dynamic range of real life is our true goal, we have to accept that loud sounds can, will and do have a bite to them which tickles our fight or flight mechanism. If we manipulate those sounds to be more 'comfortable' and hence less of a threat, can we truly say that our objective is high fidelity at home? I don't think that we can. We have become artists ourselves, reinterpreting the sound recording.

There is no substitute for joules (aka watts) if we attempt to reproduce the huge dynamic range of live sound in our homes.

[Pluto]

Dynamic compression. The very foundation on which the modern record business is built.

Here are pictures of units designed in the fifties that exploit, in a controllable way, the phenomena described above. It's all based on the fact that when a tube is driven hard, its transfer characteristic is such that the louder sound gets 'squeezed' and a consequence of this is that quiet detail becomes relatively louder - hence the often stated belief that valve amplifiers reveal more musical 'detail' than their vastly superior solid state brethren.

EMI limiter.jpg Fairchild limiter.jpg Teletronix.jpg

[A.S.]

Dynamic compression. The very foundation on which the modern record business is built....

The depressing thing is that all of this stuff is known on the inside of audio industry and has been for decades. What I'm exposing is not new - I don't have an original thought in my head. Professional sound people have been manipulating sound (using tubes, for example) to achive the sound that sells, for decades. But the artificial walls that surround audiophilia have been constructed to be so tall and impenetrable that none of the facts about valves, compressors, dynamic range, loudness and the ear, cables, connectors, room treatment and all the rest that we on the inside understand cross the boundary.

It is truly amazing that the audiophile allows himself to be marched up the isle and into shotgun marriages with unsuitable equipment not once or twice in a buying career, but constantly! What a business this is.

[Pharos]

How true this criticism is; I used to mix radio programmes and frequently used compressors to raise low level signals this often having the effect of improving intelligibility by increasing the prominence of low level detail.

In the case of my own improving system, and I have just listened to "Pick of the pops" on R2, and I regualarly hear as I did this afternoon, the use of compressors; backing sounds dropping in level when the lead singer comes in again for example.

This is all of course is revealing of the technical nature of the manipulation of sound for an 'artistic' purpose as well as for reasons of the limitations of mediium.

Does an increasingly real understnding undo, or enhance the enjoyment of music?

[A.S.]

How true this criticism is... Does an increasingly real understanding undo, or enhance the enjoyment of music?

That very much depends on your perspective. I had lunch last week with an old chum from my school days. We both - and curiously several others of our classmates - took a career path into audio; some stayed with audio, others moved on to different fields. In his case, he has been working in psychotherapy for many years and has much experience with trauma management, with military personnel for example. We talked about perceptions and faiths and the audiophile's belief system. My friend cautioned me based upon his wider observation of human behaviour, and it did make me stop and think, that as a rationalist, I could not imagine how rationalism was not a universally held view. He explained that if someone wished to believe that a certain cable was capable of contributing some magical qualities to the sound that passes along it, or that is was directional, or that the conductor material had a particular sonic signature, those views would not, could not, change. If someone believes in fairies at the bottom of the garden they are real.

This is all very well. We live in a democracy and people can buy whatever they like when they like with whatever instrument of payment they like. I have no issues with any of that as I've said before. But, as a rationalist who definitely does not believe in fairies, I see the misery that chasing the audio dream has brought to so many hardened audiophiles. The wives driven to despair because husband has spent the family weekly shopping allowance on some new gadget. Family holiday funds raided and blown on audio junk. People will do whatever they want to achieve personal satisfaction. I've done it myself.

The consequence of all this misplaced consumption which doesn't actually deliver real, tangible, long-lasting benefit is of an industry that is technologically stagnant. There is not a grain of motivation to raise audio standards (quite the opposite). There is no incentive for the Board of any audio manufacturing corporation to invest one cent in risky, uncertain, blue-sky R&D when for every cent invested in BS marketing they can hook plenty of gullible consumers. What would you do with shareholder funds?

If we want technological progress - and as a rationalist I do - we have to turn our backs on audio BS and expose audio myths, of which there is a limitless supply it would seem. Only then, when marketing BS doesn't work as a sales engine, will corporations be forced to reinvest in real R&D; or die. And then we can really get on with improving music reproduction at home as a fully absorbing experience. So the bottom line is this: the consumer, by his behaviour and $$$ spend, shapes the industry and the type of products on offer. If he puts his hands firmly in his pockets and shows a disincentive to be carried along on a sea of twaddle and consequently worthless consumption, then top management will have to re-think their entire marketing strategy - for the consumer's ultimate benefit. That means for your and my benefit.

Of course, nothing will change. The forces that promote fairies are overwhelmingly strong. Rationalism cannot overcome magic.

[EricW]

...nothing will change. The forces that promote fairies are overwhelmingly strong. Rationalism cannot overcome magic.

I don't know - that view of things seems unduly pessimistic to me. There may be, as your psychotherapist friend says, people who wish to hold a certain belief and who will not be swayed from it. But I think there are also many who buy into the BS not out of emotional need, but because that's the message being delivered, and any counter-message is hard to find, so by default the BS is taken as true. The latter are open to persuasion - someone just has to deliver the goods. And that is what this forum does.

[Pharos]

I could not agree more, and the irony is that the BS extends to many, if not most, other areas of consumption. The shareholders who have made money from audio BS, probably are spending it on nutritional BS or fitness BS.

I did however have in mind, although not stating so, knowledge by the consumer of the engineering and control of the sound as in compression; does the spectacle of seeing the sound engineer altering and adjusting things at a mixing desk perhaps result in a listener feeling manipulated, and does that perhaps 'undo the magic' for him?

What you say about belief reminds me of 'truism': As the late Kevin Ayers said on "Whatevershebringswesing"; "Everything is true to you, as long as you believe it"

[Regalins]

I don't know - that view of things seems unduly pessimistic to me.

I agree that Alan's view is too pessimistic. I think there are a lot of people who want to listen to a rational argument from someone knowledgeable. When I decided to upgrade my modest hifi system (which I bought new as a teenager) to a "music studio quality" system, I had to educate myself as to where and how my limited funds should be spent. I had almost no technical knowledge of speakers, amps, cables etc, but I wanted to learn, so naturally I turned to hi-fi forums, thinking I could distil something useful from all the "information". But the widespread mix of religion and relativism (and lack of facts) on hi-fi forums actually made it harder than if i had stayed away in the first place - an absolute waste of time.

I discovered the HUG late in that process, so I initially ended up with stuff that has now been sold. The HUG is probably the best knowledge resource for hi-fi on the web, and the marketing power of this myth-busting forum is hopefully one of the reasons why Harbeth is doing so well.

[Cyreg]

I do like reading the HUG, but don't take it all too seriously; especially threads/post where the specifications rule and not the ears (skipping most of those threads/posts)

I do experience a lot of differences in soundbalance and/or -quality with almost all changes in my setup at home. And I choose the best for me without paying silly money.

I certainly do not agree with a lot of statements here on the HUG about there being no(or very small) differences between amps, cables and stands!

I am also not looking for anyone to agree with me, I make my own decision in this matter.

So cannot agree with the former post about the HUG being a mythbuster for me!

[EricW]

...

I certainly do not agree with a lot of statements here on the HUG about there being no(or very small) differences between amps, cables and stands!

I am also not looking for anyone to agree with me, I make my own decision in this matter.

So cannot agree with the former post about the HUG being a mythbuster for me!

Hi Cyreg:

I think you're right inasmuch as one shouldn't escape from one orthodoxy just to fly into the arms of another. I see the dominant views on the HUG (especially Alan's) not as positions one should uncritically believe in, but rather as views to be taken very seriously, especially as they're (1) informed by a lifetime of thought and practical work; (2) have a scientific basis; and (3) appear to have resulted in, or are at least not inconsistent with, the creation of some very fine products.

I hear differences too, but what's changed for me is that I feel like a have a larger vocabulary in terms of thinking about what might be causing what I hear, and instead of reaching for what might be fanciful audiophile explanations, I'm more apt to ask things like how the room might be affecting what I hear, whether different frequency response curves are playing a part, whether there are resonances somewhere in the chain, and so on. More physical causes, fewer mystical ones. And of course understanding the importance of playback levels, and not taking any subjective impressions too seriously unless I know levels are matched. All of that I attribute to the HUG.

[A.S.]

Hi Cyreg:

I think you're right inasmuch as one shouldn't escape from one orthodoxy just to fly into the arms of another. I see the dominant views on the HUG (especially Alan's) not as positions one should uncritically believe in, but rather as views to be taken very seriously, especially as they're (1) informed by a lifetime of thought and practical work; (2) have a scientific basis; and (3) appear to have resulted in, or are at least not inconsistent with, the creation of some very fine products....

And there is one more, perhaps the most, important consideration, as I too am on a journey. I have created hundreds of audio (and some visual) examples here to support my mere words with content that you can listen to, think through and identify with. My willingness to empower you, the reader, to judge for yourself by listening to the clips, is core to the way I want to open the mysteries of audio for the consumer's benefit. I have nothing to gain - I know precisely where I stand, and what I believe in: I don't need to create examples for my own benefit. This all takes time. More than you would believe.

Anyone who aims to be a communicator both objective and reasonably impartial in a subject like audio, where, thanks to modern PCs, they have the opportunity to use the operating system's tools to create original audio examples, should do so. If they don't make the effort but fall back on regurgitated myths, second-hand beliefs and prose and don't introduce any new thinking via audio examples, are they worthy of our serious attention? To talk credibly about audio you have to be able to demonstrate audio. Words are not enough. Those that critique photographic equipment go to lengths to make visual examples of the optical capabilities of the camera. We can and should do that with audio. We should set aside emotion and try our best to be objective or we will end up buying our 'audio camera' not because it has the highest resolution lens, but because we like the colour of the case. The cost of demonstrating audio is almost zero; it just needs a willingness to set aside time to conceive of a fresh demonstrable insight, then create it on the computer, package it and upload it. It's not difficult.

Talking of which, I am reminded of research that Dudley Harwood (our founder) undertook thirty or so years ago concerning the audibility of 'clipping', specifically amplifier clipping, where the amp just doesn't have enough power reserve* to replicate the full dynamics of the input signal, multiplied up by a gain factor to a level suitable to drive the loudspeakers. So, for example, we know that most CD players generate an audio output of 2.0v (rms) at their phono output terminals (enough to illuminate an LED!) for a fully modulated disc that cannot hold any louder signal. That 2v cannot drive a loudspeaker with sufficient current, and that's why we connect it to a pre/power amp, and then to the speakers. In round numbers (a simplification) if we want a reasonably loud sound from our speakers, we may need to have a voltage delivered to both of them of, say, 20v. That means the pre/power amp is going to have to multiply that incoming signal by ten times (2v > 20v). But what if the amp doesn't draw enough power from the house mains supply into its PSU capacitor reservoir to actually deliver the 20v on the audio peaks? That would be the case when the amp was deliberately designed to have a small power capability. What happens if it is one of those small <25W amps?

Well, if the amp is commanded by your setting of the volume control to deliver peaks of 20v to the speakers but due to its tiny capacity it can only deliver 9v, 15v or 17v peaks when working flat-out, both channels driven, the relationship between the sonic loudness peaks reaching your ears via the speakers and the voltage peaks measured at the output of the CD player has been compressed. The small amp has acted as an (expensive) signal compressor, of just the sort that a pop engineer would use on his mixing console to squash the dynamics of a recording. Why would anyone intentionally want to compress their audio at home - the inevitable by-product of using a very small amplifier? There is no logical or engineering answer to that. Only an emotional one.

Harwood was working on the design of the BBC LS5/8, a large two-way active speaker. A decision had to be made about how powerful the internal amplifiers should be to drive woofer and tweeter (woofer, one amp; tweeter, another amp). It was easy to determine how much oomph the bass unit needed - you could hear when the amp was running out of power, but deciding how big the tweeter amp should be was surprisingly difficult and non-intuitive. The QUAD 405 (100W+ as I recall) was selected to feed the bass driver. Experiments were conducted with various amplifiers (actually, I think one amplifier was modified so that its power potential could be adjusted) and a conclusion was reached. If the bass unit needed 100W for clean reproduction of loud bass notes, how much (or little) power did the tweeter amp have to have available do you think? Obviously, after some feedback, I can create some audio example clips.

This is not a trick question, but it does lead on to a better understanding of mythology surrounding amplifiers.

* Remember: the power amp does not push power into the speakers. The speakers suck power from the amp according to how you set the volume control. Whatever you set the volume to, if the speaker cannot suck enough power from the amp for that volume level, the amp has compressed the dynamic range of the music recording. If you push the throttle peddle to the floor you will go faster but if the fuel pipe from the tank is too narrow, a point will be reached where you cannot suck enough fuel into the engine to sustain the speed and your velocity will plateau. A small amp is like a constricted fuel pipe: no problem at low speeds but with increasing speed (louder music) its sound loudness output will plateau. At that point, we cannot describe the audio system as 'high fidelity' because the relationship between the recorded dynamic range and the reproduced one has broken down. But would that power-plateau point be audible?

More here.