Amplifier Power: How Much is Enough? Page 2

Thus far, we have fairly well established the power that we must have in order to avoid outright overload when reproducing original orchestral level through a speaker of known efficiency. But it is not all the power we should have on hand, because there's more to fidelity than just reproducing sound at the proper volume.

Anyone who has perused an amplifier's power-vs-distortion curve will have noticed that distortion rises gradually with output until just below the overload point, beyond which the distortion skyrockets. This is one reason why a high-powered amplifier is likely to sound better than a low-powered one even at every low power levels. They may both be operating at well below their overload point, but the fact that the high-powered one is running at 1/10 of full power when the other is at 8/10 of full power will mean that the former is contributing less distortion at all times (fig.1), and this will generally show up as cleaner, more "comfortable" sound.

Fig.1 Distortion curves for two typical amplifiers rated at 25 and 44 watts.

There's a second reason why high-powered amplifiers should outperform low-powered ones, even at low output levels. It is customary to equip an amplifier with an output transformer that is no larger than it has to be in order to yield full rated power in the middle range. The British are still making low-powered amplifiers with substantial output transformers, but the prevailing attitude in the US seems to be that the low-powered amplifier is sort of a stopgap component, to tide the buyer over until he can afford to purchase something good. There is rarely any attempt to design a really good low-powered amplifier. As a result, the typical 10-watter, even though it may well meet its rated power at 1kHz, is severely limited in power capability at both ends of the spectrum. The power loss is usually most severe at the low end, where there is often a great deal of energy in the audio signal, so the unit may only be able to deliver half, or less, of its rated power before the program material overloads it (fig.2).

Fig.2 Test setup for observing amplifier overload from program material.

Even the biggest, costliest amplifiers exhibit this power loss at the frequency extremes, but in these, the losses don't usually start until well beyond the audible range.

Let's assume now that we have access to an amplifier's power response curve, and can see that it will deliver its full rated power to 20Hz. Is this any guarantee that it will sound the same, at low levels, as a high-powered unit? It is not.

Power response curves show the power levels at which different frequencies will generate the same 2% distortion at which the midband power is usually rated (fig.3). What they fail to show is distortion at less-than- maximum power levels. An amplifier that yields 2% distortion at full rated output may yield 0.2% at half power, or its distortion may never drop below 1% regardless of how little power we drive from it. And since we do most of our listening at power levels far below overload, the amplifier's minimum distortion, or "residual" distortion, is of considerable interest to us. Here, again, is where the typical low-powered job falls far short of its heftier ilk.

Fig.3 Output versus power at 2% distortion, in an inexpensive 12-watt amplifier.

The light output transformer in most low-powered amplifiers is susceptible to core saturation at low frequencies, and even though this may be held low enough to meet overload limits down to, say, 20Hz, it nonetheless imposes a severe limit on the amplifier's low-frequency residual. Thus, typically, the low end will exhibit increasing distortion with decreasing frequency, even at the very lowest output power levels. At 1 watt, where the mid-band is contributing only 0.3% or so distortion, there may be 1% distortion at 30Hz.

Actually, it is a rare low-powered amplifier that will produce as little as 0.3% distortion at low levels, even through the midband. Most of them, sloppily designed as they are, have enough distortion in their earlier stages to hold their residual at about 0.75% no matter how good their output stage may be, so they can never sound as good as the more carefully designed high- powered units. The few exceptions to this rule are so costly that one might just as well buy a higher-powered unit and be done with it.

There are extenuating circumstances occasionally, though. Loudspeakers and amplifiers that ate designed specifically for one another should be used together regardless of the amplifier's power rating. Some speakers are fragile, and will burn out if hard-hit by a hefty amplifier. Fusing helps, but the series resistance in the line reduces the electrical damping applied to the speaker, inhibiting the amplifier's ability to prevent spurious cone vibrations. Consequently, if you must use such a speaker, it's advisable to bypass its fuse, and couple the speaker directly to an amplifier that won't be able to damage it.

High-power advocates have always claimed that one reason a high-powered amplifier sounds better than a top-notch low-powered job, even at low levels, is because the big one's reserve power gives it better control of the speaker's voice-coil. It was reasoned that a large reserve of power, operating through a tight negative-feedback system, could bring more power to bear more rapidly for suppressing spurious vibrations of the speaker cone. This sounded plausible, until the first of the all-transistor amplifiers came along and befogged the issue.

Transistors just do not behave like tubes. Transistor amplifiers whose measured distortion is higher than that of the cheapest "hi-fi" amplifiers somehow manage to sound much better than they should, and the absence of an output transformer from most transistor amplifiers (the low-impedance transistors connect directly to the speaker) eliminates most of the annoyance value of marginal overload on peak passages. As a result, a transistor amplifier seems to produce far more clean power than a tube amplifier of the same rated output.

Even more significant, however, is the "transistor sound" at low output levels. Even the feeblest transistor amplifiers we have heard (a 3-watter, for instance) sound like high-powered amplifiers when operating at low levels. They are transparent, crisp,and have the same kind of bass solidity that high- power advocates have always attributed to the monster amplifier's reserve of speaker-controlling watts. So the superiority of the high-powered tube amplifier is not just a matter of reserve power.

Just what it is a matter of is still open to question, but we may be in a better position to answer this when we get the opportunity of comparing high-powered transistor amplifiers with their betubed competitors. Tube amplifiers have fouled up the power question for years, because the low-powered ones so often suffered from shortcomings that had nothing to do with the simple fact that they were 10- or 12- or 15-watt amplifiers. Transistors may change the picture.

So, where do we stand? For the nonce, let us say that computed power may be taken as the power we should have on hand if we use a transistor amplifier or a high-powered tube amplifier. In the lower-power categories, tube amplifiers in general will not produce the best sound that the average speaker can furnish. They may be adequate, and can nonetheless provide enjoyable listening, but they do leave room for improvement. Whether or not the improvement is worth an additional outlay of money to you is up to you. But it's there for the buying.—J. Gordon Holt


HammerSandwich's picture

A 10dB increase in level represents a 100-fold increase in power

Actually, 10dB represents 10x power, so 100x is 20dB.

Odd that JGH missed this after correctly writing that doubling power is 3dB. I suppose even he could make a potential mistake.

John Atkinson's picture
HammerSandwich's picture

Your keyboard seems to have moved while you were correcting this. :)

audiolab's picture

This was written the year I was born (53 years ago), and just like me things have changed alot. To many out of date misconceptions & assumptions to be of any real value today. However, it makes interesting reading to see where we have come from. I have recently finished reading two books by Briggs on loudspeakers and it was all somewhat a familiar experience.

DaveinSM's picture

Actually, What struck me is how much of this article is still relevant today. It reads like it was written last week. Too many people underestimate the impact of listening habits in terms of volume level, room size, etc., on power requirements.

Bkhuna's picture

You can never have too much overhead.

dalethorn's picture

Seems like, a company that makes premium loudspeakers, should be able to tell you more than "Use an amp with at least 50 watts of power", or, "We like the ABC amp best, but the XYZ amp is good also". Seems like they should have a booklet or set of papers that describe their experiences more fully.

Allen Fant's picture

JGH was the master! We audiophiles could never have enough high-power nor high-current, IMO.

Ronald Koh - SG's picture

Just in hope that we it'll be clearer in simple terms for those who are not familiar enough with technical terms too. As, we need to know in simple terms that:

1. Sound levels in decibels and light in lumen too follow the inverse square law with change in distance. Meaning, twice the distance from sound or light source, 4x power is need. Inverse that distance to 1/2 and the power requirement becomes 1/4.

So SPL, simply put is 1 divided by the change in D sqd. Thus:

2x the distance is 1/2 sqd = 1/4 of the original SPL, means 4x SPL is required to maintain the same SPL at the original distance from the sound source. Thus, the original SPL at 1x is only 1/4 of the the new 2x distance from 1x.

2. In the earlier JGH article on "Too Good to Be True' of 1 Jan, 1963 at:

I pointed out about source to load matching impedance-modulus to the technically knowledgeable about this. So it will suffice to note that the loudspeaker impedance curves and voltage to current lead-lag phase angles comments by JA in his reviews of speakers are additionally also helpful in determining/choosing power amplifiers' RMS & dynamic/transient power capabilities.

And important is also to note JA points of speaker sensitivities are usually measured at the standard distance (or other by him accordingly) at 1kHz etc per his expertise qualified methods in typical listening rooms. So they are not Anechoic Chamber measurements which are also without typical room gain in the bass regions.

And since we are about bass, in the JGH era, it was told that even 20% speakers distortion is not audible? Would JA like to give us a brief on this if you can find time?

3. Just sharing for how anyone may view it from any perspective per their experience.

I use Infinity IRS dipole ribbons with integral servo unit drive between preamp & power amp. Twin 12" bass units wired in reverse in each bass bin, are directly driven. Their foam surrounds are modified & coated for long term durability to suit our very humid tropical climate for longevity. So the electronic servo unit was modified to offset the modified surrounds tension & motion sensors' sensing behind the dust cap in one bass unit in each speaker. Also, channel tuning are individualised for fine tuning to integrate L & R performance as seamlessly as possible acoustically.

They were made in 1987, and I had them for as modified for some 15 years without needing to change surrounds every 10 months. Yes! Their bass rendition change that fast due to foam deterioration! And both don't deteriorate equally too! As replacement costs increased with foam surround quality & replacements skills become questionable to now a dying trade. Was why I did it as a do or die effort attempt.
Positioning them quite an effort too for best results. I read Stereophile's 3 reviews long ago of a higher 4-units infinity model with separate bass towers. Can't remember the model now.

They are so transparent & musical that they are literally almost like computers - software rubbish in rubbish music out and vice-versa! They can be Heaven or Hell literally, depending on recordings. No back wall too with some ~40ft in front & ~10 ft from the front wall with windows I can open.

Just wonder what designer Arnie Nudell will say if he gets to hear about this "idiotic" effort of mine. He has retired from Genesis too years ago. Am 70+ with still unbelievably good balance ears able to hear tinkling highs beyond 13kHz. Can't complain at my age!

Thanks for your Merkin Hall CD in 2008 dear JA. Cheers! :)

Roger Paul's picture

Mr. Holt,

I am the designer of H-CAT and recently had a major breakthrough. This has a direct impact on your assessment of how much power you need. The significant difference is in the distortion present at various SPL. As you know most if not all amplifiers will begin to strain or breakdown in purity as you drive them harder. You were correct in assuming that a higher powered amp will at least take you farther. What happens if there is no distortion at low levels or high levels? What do you get?

The breakthrough at H-CAT is that you can now amplify without distortion. (Zero) As incredible as it sounds - its true. (It only took me 20 years) So what does it sound like? It sounds un-amplified, 100% effortless and live. Since there is no indicator that alerts you that you are running out of power, you have no idea if you are using a 50 watt amp or a 500 watt amp. In addition to having no distortion this new amplifier method includes the "wave" as captured at the original venue and its velocity is synced to the real world at Mach One. Your central nervous system accepts the sound event as happening live. No other amplifier can perform this function.

I am not trying to "pitch" my products - I am simply making you aware that sound reproduction as we know it has just had a complete game changer. The movie industry especially the 3-D film companies are the first to jump on this. I have been in touch with 4 of the top 10 film makers already regarding things like compatibility with Dolby Atmos and THX certification.

The amplifier method I'm using is 100% pure analog. You simply replace the analog amps in the theaters with "Live amplifiers" and the audience gets a bonus experience they never had. They get to hear actual live sound from the block buster movies. You can see my press release via my website.


corrective_unconscious's picture

While claiming zero distortion.... And you are not pitching, you are spamming.

Roger Paul's picture

I do apologize for addressing Mr Holt on this site as I was not aware of his passing. Perhaps it is a testimony to my absolute disconnection with the audio industry in general. As you can see from my web site I have currently one review and one press release. I did not want to make an amplifier that sits on a shelf with every other amp. I did not want to spend money by sprucing up the aesthetics of a product with solid gold handles and force customers to hire piano movers to bring them their new mono blocks. I would rather keep my day job.

Instead I have been laser focused on one single issue that has grabbed my attention since I built my first amplifier in 1969 - Distortion. My curse, if you will, has been the puzzling nature of electronic amplification and why it does not just work.

I think outside the box and it is the only reason that I have been able to conquer distortion. I would have to include my faith as well because I do believe I have been given a gift of insight into the inner workings of electronic circuitry. Many people who know me privately will tell you it is true. At the risk of ruining my own reputation, I will gladly tell you why I have been able to accomplish a nearly impossible feat. I can look at a schematic diagram of a complete circuit a see it operate with all of the devices appearing as simultaneous moving parts. Some engineers would say that's nothing and they can also envision a working circuit having been schooled in the field of electronics. The difference is that they can't see or failed to notice something that has been hidden in analog circuitry since the grid was added to a rectifier to make a triode.

Almost like the glass "half full or half empty" it is all in the viewpoint. Take two approaches to making a statue of Abraham Lincoln.
One artist may start by continuously adding clay from the bottom up and shaping the arms,legs and head etc. Another artist may start with a solid block of granite knowing that Abraham Lincoln is already in the center and all you need to do is chip away everything that is not him.

This is how I built the first true 3-D playback system. I already knew it was possible to project sound into mid air. The trick was finding out why it wasn't working. This took decades. No other designer or giant multi million dollar corporation bothered to pursue this goal because they thought that amplifiers have come a long way and distortion has already been reduced to "low enough" levels to where it "doesn't matter". Wrong.

There are many audiophiles reading this posting that I'm sure are yelling "what about the distortion in loudspeakers being higher than the amplifier so you would not be able to hear a perfect amp if there was one?". The answer is simple. Whatever the impurities are in the transducer, one thing is still certain - however they alter the sound "image" it is constant and more important it is stable. Analog amplifiers are unstable and do most of the damage to the sound of a system. By locking the delivery speed to the speed of sound, an amplifier can appear as a parallel path to sound waves. In other words, the amplifier output can produce a "wave" of sound that would be in sync with a real sound wave traveling outside the amplifier. Conventional amplifiers have a tendency to accelerate or de-accelerate the flow of information headed to your speakers. Even though this instability in speed is tiny, your brain knows instantly it is fake and any hopes of hearing live sound are lost.
Real live sound is naturally stable because the medium of air is stable.

Don't blame the speakers - blame the electronics. I have had customers that were baffled by their own speakers saying "I had no idea my speakers could do this". What they thought were speaker problems were in fact amplifier problems.


corrective_unconscious's picture

Accordingly, I am unable to gather anything from your website as I have not visited it.

How are the efforts at perpetual motion devices coming along?

Roger Paul's picture

This is why I have enjoyed being disconnected from the audio circus.
Since you don't use your real name its fun to poke fun at something that you don't understand.

BTW Its 2016 Isn't it about time somebody got rid of distortion?

I don't blame most audiophiles for being skeptical because it is quite a profound claim. But that is not my problem. The open minded ones will pursue an opportunity to listen for themselves. Everyone who has listened to this technology needs only about 10 seconds to determine that this really works and are totally baffled about how it is possible. After hearing the process they are spoiled and every other system now sounds distorted to them. You won't have to go to my web site to learn about it - just keep watching the evening news. You will still become familiar with the H-CAT logo. It will be in the credits at the end of your favorite movies. (This process has already started)

If you happen to be a designer of amplifiers yourself - well I know how you must feel. That's not my problem either. I did my homework.

This will be my last post to you. I will not engage in a discussion with someone who has predetermined that they know better even though they did not give this the benefit of the doubt until you hear it for yourself.

I wish you luck and hope you find your ultimate sound system someday.


corrective_unconscious's picture

While ruminating (spamming) about zero error....

On the plus side, there can be big money in communicating with those who have passed over to the other side. And by "the other side" I do not mean "digital."

Ronald Koh - SG's picture

Hi Roger Paul,

Thank you for sharing of your new technological breakthrough in audio analogue power amplification.

Please don't get me wrong. Your H-Cat analogue design at your do indeed sound amazing. And so at this juncture, more technological information at point must indeed be proprietary.

So, without prejudice, I'll must you the benefit of the doubt from our understanding of present non-H-Cat technology. It is only sensible as subjectivity is an intensely complex thing that can lead to fierce controversies versus objective technicalities too. This understanding is why I only read Stereophile with JA's technical measurement with his methods clearly explained. And also co-relating them with his own and/or his reviewers' review comments.

1. So I shall begin with other than power amplifiers and distortions in loudspeakers and our ears perception of distortions without factoring ear deficiencies & ageing defects to not complicate and culminate into long drawn technicalities & subjective controversies.

2. Being electro-mechanical, the inherent distortions in loudspeakers are invariably much higher and also different from electronic power amplifiers. But that is not to say that their lower distortions can't or won't be heard through higher distortions loudspeakers for obvious reasons. Truth is, their type of distortions that our log scale loudness sensing ears can still differentiate them to electronic distortions. Such as in power amplifiers of the push-pull varieties from high-efficiency & distortion Class B to > Classes AB1 > AB2 > Class A. Each have their pros & cons like push-pull crossover, phase, intermodulation and harmonic even & odd distortions. And like Class A that most ardent tube and vinyl analogue followers favour, my experience it that Class A lacks that dynamic peak of rise to fall speed "slam' that has been mentioned by Stereophile reviewers. Thus I use MosFET output power amps for my ribbons for their tube-like warmth and speed of transistors. And normal silicon Class AB1 with higher dynamic & damping factor power amps for the servo driven bass drivers. Just what I consider as a "best" of both worlds compromise that's all.

And thus, last but not least is also how our ears perceive them all distortion types as irritations or not. Along with its ageing & personal understanding of musicality by not focusing on sound-bites instead! Many audiophiles do per my earlier comments in the earlier JGH article on "Too Good to Be True" of 1 Jan, 1963.

3. Thus I'll end here in the hope of someone else with good technical & subjective respectable knowledge to join in and enlighten us all on an even keel, so that we shall all not end up none the wiser.

Thank you Mr. Roger Paul. I shall be looking forward to know more of your new H-Cat design for consideration.

Roger Paul's picture

Thank you Mr Koh for sharing your thoughts. Yes it is controversial and I have had my share of skeptics. If you look at the above posting - I do touch on the reasons that speaker distortion has less effect on the overall system distortion. The secret to an amplifier used for audio frequencies is that it must include a speed governor. Amplifying sound waves have to include the wave itself. The flow of sound as a wave is the phenomenon that has to be passed through the amplifier along with the electrical signal that represents the exact changes in air pressure. Transferring changes in air pressure without maintaining a constant speed will cause the holographic nature of the presentation to collapse.


corrective_unconscious's picture

The article is from about half a century ago?