Qualities of an Exceptional Amp

[av_phile]

On the assumption you have respectable players and speakers and excellent room accoustics, what would you consider the qualities of an excellent amplifier? I'll start off with some of my thoughts:

Straight Wire with Gain: Transparency, Accuracy, Neutrality

This is perhaps the goal of every amplifier design in the quest for high fidelity. It seems to me that by reading the technical specs of many expensive amplifiers out there that, indeed, we have come a long way to reaching this level. A linear frequency response from 1hz to 100hz +or-0.5db seems to speak volumes about accuracy and neutrality. A S/N ratio of 100db seems to indicate neutrality. And a THD of 0.003% seems impressive enought for transparency.

Perhaps a definition is in order here, and you guys can add or argue on them.

Straight Wire With Gain just means what comes in an amp is what comes out in SHAPE, FORM and SUBSTANCE but with an increase in SIZE (gain in current) across ALL the shapes,forms and substances enough to drive an intended speaker load.

And the three close dovetailing attributes of this are:

Accuracy is just a part of the above defintion to mean what comes in is what comes out in SHAPE, FORM and SUBSTANCE. This relates to having a linear and extended frequency response so that the amp passes ALL the signals as sent by a source player, unimpeded.

Neutrality simply means nothing is added or subtracted by the amp circuitry so that what comes in is what comes out in SHAPE, FORM and SUBSTANCE. It is related to the above quality of an amp but with the suggested dimension of having a very low noise floor, excellent S/N ratio as well as channel separation to bring out what the source material from a player has intended. Sort of a wide doorway or highway. No "touching" of the signal.

Transparency is very much related to neutrality and is often interchanged with it. But for me it means you can hear precisely and clearly what comes in as intended by the source material because apart from being neutral, the amp adds no spurious harmonic(THD) or intermodulation distortions (IM).

Those three attributes make up essentially the definition of straight ware with gain. But they apply to other parts of the system as well: players, speakers, cables, and the room accoustics.

Applied to the system as a whole. I often wonder if the people who claim accuracy and transparency really know what they are talking about. Can they really discern how a live symphonic orchestra sounds like versus what they can hear from their system to say it is neutral, transparent and accurate? Excpet that I simply have to rely the specsheet and on the "expert" reviewers out there who have access to REFERENCE amps, speakers and systems or who are themselves conductors, musicians or recording engineers with daily access to how REAL music sounds, I cannot as a mere consumer say whether an amp, or speaker or system is indeed neutral, accurate and transparent.

Absence of Listening Fatigue

Has anyone ever wondered why conductors and musicians in a symphony or ensemble music can't seem to tire for hours on end of rehearsals? No ear fatigue. Never a listening fatigue. (Ofcourse they do tire at the end of the day, but more because of their work rather than any listening fatigue.)

If you can spend at least 4 straight hours of listening to your system at loud levels without ever craving for total SILENCE to deliver you from it, then you have a truly exceptional system. And for the most part, a truly remarkable amp and speaker combination can result in such a euphonic experience that doesn't want to see any end. In a truly exceptional set-up, there is hardly any compulsion to bring the volume down (except when the wife is calling). In fact, I sometimes get the urge to bring the volume further up, only to be held at the threshold above which the neighbors will start to call the police.

There is every reason to suspect that odd-order THD is the culprit for listening fatique. I recall reading some pscho-accoustic experiments where the subjects on headphones never became fully aware they were already receiving deadly 130db SPLs. For the simple reason that they were recieving very clean signals. But when subjected to about 0.5% THD, the subjects had to turn down the volume from around 100db. The higher the THDs, the lower the SPL tolerance. (This is often the case with solid states. I am not sure about tubes which are touted to have high THDs, apart from soft clipping at high volumes. But these are even-ordered THDs which are euphonic, so the higher THDs don't seem to induce ear fatigue.)

Many excellent and expensive SS amps out there have THDs with many zeros after the decimal place. And the more power an amp has, the less chances of really going into clipping distortions. Which leads me to the next quality.

Dynamic Power Headroom.

Power for its own sake is exhillarating. Ever drove a Ferrari? I don't think one can go wrong with more of it. Better to err on the side of plenty.

But whether 1 watt or 1000 watts, more than average continuous power, an amplifier better has enough dynamic power headroom to accommodate peak signal transients that often exceed twice the continuous average power at any listening level. This is often indicated in specsheets in db, say 3db dynamic power headroom. That means the amp can deliver twice the power during instantaneous peaks. As a function of current, dynamic headroom can also be expressed in terms of the amp's ability to drive low impedance loads continuously, say at 2 ohms. Thus, so-called high current amps have the following ratins:

100 wpc, 8 ohms
200 wpc, 4 ohms
400 wpc, 2 ohms

A few amps I've come across are even rated at 1 ohm.

It is a reality in struck instruments that every note delivers an Attack, a Sustain and a Decay. These three phases of a struck note extracts different power levels. And it is the Attack phase of that note that demands instantaneous power for a few milliseconds to bring out the detail of that note. Otherwise, the amp will just say "hing" when it should have said "thing." You can lose the detail that makes a musical passage truly engaging and breathtaking.

An amp can possibly get along without dynamic headroom, provided it is operated below half its maximum average or continuous power. But once you get the urge to blast the neighbors, you could end up with a laughably emasciated system that coughs at every peak signal. And you get ear fatigue as well.

Needless to say, a good amp with lots of power is heavy and hefty. I can blame physics for that. Some amps can't be lifted at all without getting herniated. Weight and heft is an indication of excellent power reserves.

 

[Rip Van Woofer]

"Accuracy", "neutrality", "transparency", and lack of "listening fatigue" are basically the same thing: vanishingly low distortion and noise with linear frequency response. All modern amps and electronics do this already. It can pretty much be taken for granted at nearly any price point. Assuming one's amp has adequate power for the speakers and room without being driven into clipping, audible distortion is a non-issue in modern solid-state electronics. Listening fatigue will come from your speakers, or from simply playing the music to dang loud!

The headroom of which you speak is the ability to deliver lots of current and voltage into any reasonable load at the maximum listening level, and then some. The ability to drive a 1 ohm load continuously is only valuable to the owners of a small number of exotic speakers. For most of us that is expensive overkill. Same goes for the ability to double the power for each halving of the load impedance: it's difficult and expensive to do, and an amp that "only" delivers, say, 180% of its 8 ohm rating into a 4 ohm load is perfectly fine as long as it is adequate for its intended use. There is no objective industry or engineering standard I know of for what constitutes a "high current" amp. It's up to the marketing department!

The ability to continuously drive 4 ohm or higher loads and to deliver short-term peak current to even lower impedance loads (since the actual impedance of a speaker can vary widely depending on frequency), as well as to loads that vary in phase from inductive to capacitive (also variable in a given speaker) is desirable and attainable at less than scary prices with careful shopping, if not necessarily at the bottom end of the spectrum. For that matter, some expensive amps aren't so great at that, either. At the current state of the art, it is the only parameter that can actually have an audible effect if one's speakers present a difficult load.

Note that tube amps don't count as "modern", whatever their vintage, and are thus exeptions to all of the above.

 

[av_phile]

Listening fatigue will come from your speakers, or from simply playing the music to dang loud!

This is partly correct. Speakers contribute in a large part to the overall THDs arriving at the listener.

But listening loud per se does not result in listening fatigue if the signals are CLEAN. That's why I prefaced my point on conductors and musicians in an extended rehearsal session who recieve anywhere from 90 to 120 db SPLs which are loud by any standard. But they don't get listening fatigue.

But if you get dirty signals, you'd get listening fatigue even at low volumes.

 

[av_phile]

"Accuracy", "neutrality", "transparency", and lack of "listening fatigue" are basically the same thing: vanishingly low distortion and noise with linear frequency response. All modern amps and electronics do this already. It can pretty much be taken for granted at nearly any price point.

I really wish that were true for ALL modern amps. When I see a modern Sherbourn receiver power rated only across 40Hz up to 20Hz, I wonder what happens between 20Hz and 39Hz. When I see a modern Pioneer receiver power rated at 1% THD, I must admire at the candidness of the marketing guys. And when I listen to some modern recievers and integrateds , I have to wonder why i get listening fatigue even at a mere 10 o'clock volume position.

 

[hopjohn]

Component of the month club

Room acoustics will play a larger role in accuracy than any one component. I would think listening fatigue would occur most when your brain is being pounded by 2nd, 3rd, 4th etc order reflections that so many enthusiasts ignore in favor of becoming full fledged members of the "component of the month" club.

 

[av_phile]

Rooms accoustics do play a substantial role in what we hear. No argument there. I probably should have worded my question to include, "with excellent room accoustics." In fact I am adding that to my first post after this.

I agree that a very live reverberant room can induce listening fatigue even with clean signals. And there's really nothing an excellent amp, player or speaker can do to correct this. Not even an equalizer. Only the right accoustic treatments will do. And the listening fatigue I got when listening to some modern gears are in a reasonably treated room with adequate decay at the listening position.

 

[jeffsg4mac]

High headroom was a marketing tool that NAD used to promote their PowerTracking amps. I have one that needs repair, a Monitor series 2600. The Adcoms, and HK's of the same period claimed High Current designs. They all had balls because they were quality amps compared to the cheap receivers at the time. Receivers have caught up though, as my Denon has more balls than my Adcom GFA 535II, which is a high current-dual mono design. The Mark Levinsons of the same time had enormous amounts headroom, but they never really advertised that aspect of their amps, in one review I remember reading, the guy said he could arc-weld with a levinson amp it had so much current? Not sure if that means arc-welding power supplies make good amps for home audio, Hmmm?? In agreement with Rip, a well made 100watt amp will be more than enough for the average room and speakers. Only when you start trying drive planar speakers such as MagnaPans, Logans, Quads, ect. Do you really need more power. I am not against having an amp with a ton of headroom, but it seems no one is making them like that anymore, not even NAD has one that has the headroom rating of the one I have sitting in my workroom waiting for repair. So what does that tell you? More important than a lot of headroom is the damping factor. An amp with an adequate damping factor will have better control over the cone movement than an amp with poor damping. This you will notice more than headroom every time.

 

[Rip Van Woofer]

Hmmm...if you can arc weld with a Levinson amp, I wonder if you can drive speakers with a Lincoln arc welder??

Double blind (ABX) testing routinely fails to reveal audible differences between el cheapo mass-market recievers and high-end amps. I know that listening fatigue can be due to lots of other factors besides your gear or room (amen there, Hopjohn). Some days, or at different times of day, I can listen fatigue-free at window-rattling levels. Other times I have to turn it down. So unless you A/B'd the amps at matched levels under identical conditions I must raise my eyebrows in your general direction!

The one factor you might hear between amps is, again, if they're driving difficult speaker loads. Leaving aside the exotics, even many "conventional" speakers will have high impedance swings and/or wide swings in phase from inductive to capacitive (we're talking swings from inductive to capacitive here, not phase coherency between drivers) The lesser amps will cry uncle and I grant that disortion might then become audible. That's the one thing that separates the men from the boys in amps.

 

[Dan Banquer]

Amplifiers

It's threads like this, which are regurgitated consumer advertising/marketing material that make me want to scream.

To Rip; This is from the last section of my FAQ's:
" Dave: Your equipment is so quiet Dan. I'm sure you would always hear the difference in a head to head test. Have you ever done a double-blind test with it?

Dan: I have a great story about that. One Saturday morning I went to this person's studio, who happened to be a very well-known audiophile in the Boston area. I was there to do a field test with a prototype pre-amp and D/A Converter that I had designed.

First, we played some music and listened to his system. You know... the usual. Then I hooked up my two units in his system. Wow! The noise floor fell so low that both of our jaws dropped. It was that blatantly obvious to the both of us. You could hear so much more of the music. Now mind you, this guy was a die-hard double-blind ABX adherent who kept saying that he heard no differences between records and CD's, and he was out to prove it. He invited me to bring back my D/A Converter for the ABX test but not the pre-amp. I'm sure the reason is because if we used my pre-amp, you would have clearly heard the differences between records and CD's. He also made sure to tell me as I was leaving, that I was not to tell anyone that his system was so noisy! Can you imagine what his pre-amp sounded like?

If I have learned anything over the years since I have been in the audio industry, it's that a great majority of business is usually based upon a personal or political agenda which, as far as I can tell, has nothing to do with proper engineering or the accurate reproduction of music. I really detest getting into the politics that surround the audio industry. As far as I am concerned, it has no place in audio engineering."

O.K. Rip, whenever you want to take the politics out of ABX testing please let me know.
I can't wait for Gene to get some available time so we can write a little truth about amplifiers and mayber reorient some of the consumer regurgitated material on this thread to some common sense. Maybe one of you will get a scope and connect it to your speakers and get some idea of how much power you are really using. After you do that maybe some of you will start to rethink what you have been regurgitating here.

If anyone of you are looking for a little common sense about your systems go here:
http://www.audioasylum.com/forums/prophead/messages/8150.html
Hopefully this will give you some real food for thought.

Dan Banquer
R.E. Designs

 

[Rip Van Woofer]

I wondered when you'd jump in here, Dan! But I thought the whole point of ABX testing -- indeed, double-blind testing in general -- was to remove all the "nuisance variables" like experimenter bias, placebo effect, and yes, politics. Dan, I respect your knowledge and your firmly held beliefs about amplifier design, but your anecdote is just that: an anecdote. And to any scientist, anecdotal evidence is considered weak.

And I have to wonder about a die-hard ABX advocate who would think there is any doubt about the audible difference between records and CD's (not to open the jumbo-sized can of worms about which is "better"!). ABX testing is hardly needed to demonstrate that (but it could surely confirm the obvious differences to the skeptical).

Anyway, what I tried but failed to get across in my original reply was that subjective terms such as "transparency", etc., are of little use in describing the performance of an amp or other electronic component: better to use generally accepted and specific parameters such as THD, noise, stability into varying loads, and frequency response. These are easily comprehended by a reasonably intelligent layperson; you don't need to be an E.E. And they pretty much tell the whole story as far as amps go. There is no "X-factor" that can cause unexplainable audible differences. I'm sure there is no disagreement between us there! I know I tend to flog the ABX and "all modern components sound alike" thing a bit too enthusiastically sometimes. Mea culpa.

I have no political or financial agenda myself, BTW. I'm just an average non-technical guy who's trying to get to the truth of the matter in audio. I belive that since audio is primarily an engineering discipline, the scientific method is the most appropriate way to reveal actual differences vs. imaginary ones, and that those differences are most usefully discussed in purely engineering-related terms as I said above, as much as one is able to do so. I grant that I will never be able to understand things on the level of an actual scientist or E.E., but I'm doing the best I can with a degree in theater arts!

By the way: somebody, someday, is bound to ask: "OK, Rip, since you're such a hardcore ABX advocate, have you done it yourself?" Answer: no, for the same reason I haven't done my own testing of Einstein's Theory of Relativity, or of the efficacy of antibiotics: I lack the facilities and training to do so properly, and enough better minds than mine have already done it and re-done it and peer-reviewed it to the point where it seems it's been proven well enough for any reasonable person.

Thanks for the link on chassis grounding: I've only skimmed it for now, but as I'm considering doing a DIY amp in the near future it looks valuable. I'm just about to finish a DIY lab power supply.

 

[Dan Banquer]

Amplifiers

Rip:
It was so blatantly obvious that the noise floor was drastically lowered that the person in question had no questions about it either. That's why the test never happened. The conclusion was foregone.
Cheer up; I'm confident that engineering will finally win out over politics, it just takes a few decades for the political idiots to figure it out.
d.b.

 

[jeffsg4mac]

Dan, I am confused I thought with the first sentence of my post I was arguing against the marketing hype that AV had put down in the first place, I also think Rip was doing the same thing. It was the headroom thing that stuck out at me because NAD used to tout this all the time in their marketing. What exactly is your stance on all this amp stuff?

 

[Dan Banquer]

Amplifiers

Fair enough: I admit to having a bad hair day.
I do have some time this afternoon as things are a touch slow today so I'm going to write a bit about some of the amplifier testing mythology that has been posted on this thread.
I'm in the mood to cover the instaneous power nonsense. This all started when loudspeaker designers back in the 70's and into the 80's had some really low impedances due to (typically) crossover induced nulls. The result was the speaker had the average 4 or 8 ohm impedance but would drop to 1 ohm or so for a small part of the band. This started the instaneous power and instanteous high current trend in audio power amplifiers.
The premise was that in this low impedance section of the loudspeaker would only see a brief amount of energy. The premise was fallacious to begin with because music program material is notorious for throwing curve balls especially if you are a classical music lover.
Lets say that low impedance section of the crossover was in the bass part of the loudspeaker. For pop music the we could well see only brief amounts of energy in that region but classical music could well put a long tone there that could easily last for 3 to 5 seconds or even more.
Given the above; the amplifier boasting of high instantaneous power runs out off steam after the first few milliseconds or so and then proceeds to clip.
The moral of the story here is that the real measure of power is watts rms into a given load. PERIOD!
Let's take a look at a few other applications. Some of you are old enough to remember the famous audio classic loudspeaker known as the LS35A. The BBC used to use this as a monitor. The impedance of the loudspeaker went from nominal 12 ohms up to 20 ohms. No low impedance current demanding load here is there? If I were looking for an amplifer for this loudspeaker I would want to know how much power it puts into 16 ohms. A high current amp is not needed for this application.
Let's go the opposite direction and say I have a set of Klipschorn loudspeakers, the one's with the SPL rating of 100 db or better. With this kind of efficiency in the average living room a well designed 20 watt amp could make your ears bleed and the neighbors calling the cops. If any of you remember Paul Klipsch he was often quoted saying: what the world needs is a good 20 watt amp, not a 100 watt amp.
There's one very important thing to remember here: this all about application and load.
Dan Banquer
R.E. Designs
For anyone intersted in reading more about the myth of high instaneous current I recommend an article by Doug Self titled "Loudspeaker Undercurrents."

 

[jeffsg4mac]

Bravo, Dan! I agree with everything you stated. I have heard a set of CornerHorns with a 25 watt HK intergated amp, and Paul Klipsch was correct. I knew the Headroom thing was hooey when NAD was about the only ones who advertised it. If you looked at all the other companies who even bothered to list it, they were all about the same as far as dynamic power, only NAD claimed 400 watts from a 100 watt amp. I don't doubt they could do it, but so could an Adcom or Parasound or HK for those short durations.

 

[Dan Banquer]

Amplifer Power

I don't even give an instantaneous power spec for my stuff. Just watts rms; and that's without a variac holding the AC line at 120 volts.
d.b.

 

[Rip Van Woofer]

Agreed: RMS wattage into a given load is where it's at. And I think I might have been sloppy in my earlier post about short-term power into difficult loads. I wasn't talking about delivering gobs of fake "peak" power, just being able to still deliver the goods when the impedance or phase gets crazy instead of pooping out (technical term!).

Good to hear Paul Klipsch mentioned. I used to be in awe of those huge Klipschorns. Remember when they came out with a sealed box speaker way back when and called it the "Heresy"?

 

[zipper]

Speaking of Klipsch,did they ever make a speaker called a "Cornwall"? I remember going to a dudes place once about 25 years ago & he had four of these. They were about the size of a washing machine. He about busted my ears by cranking them with the 1st minute of Loverboys' "Turn me loose".
I thought he said they were called "cornwalls".

 

[av_phile]

P.M.P.O is marketing hype. Dynamic headroom is not. NAD is not the only one to talk about dynamic heardroom. Check out Musical Fidelity. They state their short-term dynamic abilities not in watts or DB but in volts. And there are some amps out there that do indicate a 3db or 4 db dynamic heardroom.

And I've seen amps rated for 8 ohms, 4 ohms and 2 ohms CONTINUOUSLY that indicates their ability to drive difficult loads with sufficient power reserves without any fear of going into thermal runaway or exceeding UL safety limits for heat dissipation.

Between an amp that says so and another that confines itself to just a continuous rating on only one load, I'd choose the former. Between an amp that indicates dynamic headroom versus one that is silent on this, I'd choose the former.

I agree that dynamic headroom is more applicable with classical music. Most pop, rock and jazz music are too dynamically compressed to benefit from this. But whether you like it not, an amp with dynamic headroom sounds more detailed and less congested at loud levels than one with a constricted power envelope.

I agree that different amps when level matched in an ABX can sound identical. But that's not the way I listen to music. I listen over extended periods of time. I listen until I get ear fatigue. I listen until I can get all the details I can hear. I listen until a musical piece becomes nearly a part of my subconcious.

Then when I shift to another amplifier of the same conservative wattage rating using the same preamp at the same level position. I begin to notice a more controlled and musical bass (greater damping factor), a wider stereo soundstage and a more precise center image (better channel separation), a more detailed instrumentation that was only hinted before but now comes strongly(better S/N ratio, lower noise floor) and a more airy feel that didn't sound congested at loud passages (greater dynamic power headroom) on that same musical piece. So if for you guys think there are no sonic differences between power amps, my experience points otherwise. Perhaps if we are talking about Aragons, Theta Dreadnaughts, Brystons, Perrauxs and Anthems, since these are essentially, neutral, transparent and accurate amps, the difference could very well be nil. But among lowly pedestrian amps, there are.

 

[Dan Banquer]

Dynamic Headroom

Dynamic headroom: the other great myth of audio amplifiers.
In order to explain why the dynamic headroom spec is nearly useless I need to explain some of the basics of linear unregulated power supply, which is the most commonly used type of power supply for audio power amplifiers, and how it applies to dynamic headroom. So please bear with me for a paragraph or so.
In or typical audio power amp we have a transformer that steps down the AC line voltage and then is rectified by diodes to start the conversion from AC to the needed DC to power the amplifier. The diodes takes one half of the AC sine wave coming out of the transformer and flips it so we have 2 half sine waves, (either all positve or all negative). These half sine waves are then fed to a large capacitor which smoothes out our 2 half sine waves and makes it look a lot more like a DC voltage and current. There are some key things to remember here. Our rectified sine wave doubles our 60 Hz AC line frequency to 120 Hz charging up the capacitors.
Here comes the real important part: Only half the time of our 120 Hz rectified sine wave is charging up the capacitor, the other half isn't. So if we do a little math here we are charging up our capacitors for about 4 miliseconds out every 8 milliseconds. Remeber this, because we need to understand what our power supply is doing when we think about dynamic headroom.
Now that we understand what the power supply is doing we understand that in order to get the full "dynamic power" the signal going to the speakers must be timed exactly to the charge cycle of the power supply. Well; getting that to happen on a consistent basis with music is similar to the odds of winning the lottery.
Let's take this a step further; the test signal is only a peak signal and has nothing else. Music will have other signals along with it which will require power from our power supply so this further lessens our maximum amplitude.
Now let's take it an additional step; what happens to our peak amplitude when that signal has a peak longer than our charge cycle time. It's just like the example I presented in the earlier post on instantaneous current. We simply run out of charge.
There is still only way to judge power out of an amplifier, which is watts rms. That is a repeatable and relatively stable measurement. In addition watts rms also tells us a repeatable peak power, because if we multiply our watts rms number by two we get peak power. Ex. If we have an amplifier rated at 50 watts rms into 8 ohms then we know it's peak power is 100 watts into 8 ohms.
All for now;
Dan Banquer
R.E. Designs

 

[jeffsg4mac]

Dan, I have a question looking at the specs of my receiver it says dynamic power is 120 watts into 8ohms but it is rated at 90 per channel. How come it is not 180 watts? Here are the listed specs.

All Channels 90 w @8 ohms 125 w @6 ohms
Dynamic power 120 w @8 ohms 170 w @4 ohms 200 w @2 ohms

Does this make sense?