[RDavidson]

As this discussion progresses I see that we are still in the low numbers for power. This post already has me thinking of how to make a good 10-20 watt Class A SS or possibly hybrid amp with a very simple circuit. What I am thinking about is single tube driving a pair of MOSFETS. I have made some of these before but had not considered making one class A at low power. At the time I was still in the 100+ watt camp with the rest of you. I'm not in that camp so much anymore.

  When should I send you a deposit?
That's great! Just earlier I was thinking to myself how amazing it'd be if you and Nelson teamed up and did something exactly as you describe.....not that you couldn't do it yourself. My thinking is that 2 great minds working together is better than one, that's all.

I think 25-30 watts might be an easier sell if you can get to that number and keep the circuitry minimal. You might attract a slightly broader audience. It's a mental thing. Maybe it's just me? Perhaps it has something to do with the old NAD 3020's 30 watts per channel still resonating in people's subconscious.......or maybe it's the Dynaco ST 70's 35 wpc? Many of Nelson's First Watt amps are 25-30 wpc also and are quite popular.

Regardless......I'm just one dude with an opinion. Whether it's 10 watts or 20, or 30, I'd give it a try.

[G Georgopoulos]

I think 25-30 watts might be an easier sell if you can get to that number and keep the circuitry minimal. You might attract a slightly broader audience. It's a mental thing. Maybe it's just me? Perhaps it has something to do with the old NAD 3020's 30 watts per channel still resonating in people's subconscious.......or maybe it's the Dynaco ST 70's 35 wpc? Many of Nelson's First Watt amps are 25-30 wpc also and are quite popular.

Regardless......I'm just one dude with an opinion. Whether it's 10 watts or 20, or 30, I'd give it a try.

Yeah that sounds very attractive despite HUGE power bypassing the spk and going to ground..
this has been done before by musical fidelity 20wpc burning idle at about 100wpc..
yeah you sould do it guys just dont blame me for the heat and electricity bill..joking

[Freo-1]

Hi Freo,

I know the company, have owned their excellent amps. I had the first one which came from the work of Matti Otala. I sought him out at CES many years ago and had a nice lunch with him. A very sharp guy who found something people were not yet thinking about. The AES paper was written for engineers.

In reading the last page of the manual I see: Due to high class A operation in all Electrocompaniet designs, it is normal for the
AW 2x120 to feel warm. Proper ventilation is important.. From this statement I infer the amp is not full class A but just high class A. How high is high. Using Nelson Pass's estimate of dissipation = 3 times total power the amp should draw 720 watts (120 per channel x 2 x 3).

I did a little figuring for a perfect lossless amp and came up with a theoretical number of dissipation = 2x idle power. But that does not include power transformer losses, rectifier losses, transistor saturation losses and the big one, how close to the rail can the out put get for any particular load. This is the biggest practical loss which brings the theoretical 2x to approach Nelson's 3x. In 1976 I designed a bridged amp to run off 13 volts in a car. I was very careful to maximize the output potential and got within about 0.2 volt of the 13 volt rail and 1 volt above ground. That amp would be like a 2.1x dissipation to output amp.

I wrote my first post here not so much to get into a discussion of who makes class A amps  but what a class A amp has to do to be a class A amp so that the reader could tell who was telling the truth . Everyone wants a class A amp. We want grade A meat, produce, milk and eggs. Manufacturers of amplifiers know we desire class A amps (frankly I  don't desire them because of their power consumption). I don't know how well the USDA is monitoring our food quality but no-one is monitoring amplifier claims to be class A. That is why I wrote that extensive post so you the buyer could determine who was telling the truth.

The OP asked "who makes amps with class A bias" I don't think we have found one yet. They do exist but you will likely find them in lower power ratings simply because of the high dissipation.

Roger, thanks for the feedback.  This is an excellent thread, lots of good information here.  I'm sure you are correct in your description of the AW2-120.  My take away is that for practical listening purposes, its running Class A at normal listening levels virtually (almost) all the time.  Again, I think how they handle TID is key to the outstanding sound.

Do not concur with Davey's assessment on Class A amps.  For example, the Pass Labs XA-30.5 is a Class A amp that shifts to AB at a certain transition point.  There is more than one method to design Class A amps.  Nelson has explained this in his white paper on Class A amps (for those that are interested).

Lastly, check out a company called Valvet.   They make SE SS amps:

http://valvet.de/?p=2783

[Carl V]

The buying public should realize that they have a powerful effect on what we as designers do. When asked "why do you make a 100 watt/ch amplifier I say: I don't think anyone needs one but they think they do. A low power amp is hard to sell." Fortunately low power is getting some attention. There is a low power circle here. http://www.audiocircle.com/index.php?board=195.0

I will say it again. A 100 watt amplifier with 99 watts of unused power does not control a speaker or make any difference. What controls a speaker is high damping and available current. What we need is an amplifier that sounds the same right up to clipping. The whole concept of headroom needs to be looked at. Shall we start a post on that?

Being merely a hobbyist...
it was my understanding that full spectrum musical reproduction
can be challenging with ambitious Recorded Crest factors etc.,
That making use of Powered Sub-woofers or Woofers can
relive some of the heavy lifting burden of both Amps & monitors.  Vandersteen has
gone down this path amongst many others....not too mention the DOLBY THX specification.

I had a conversation many years ago with  Tim de Paravicini of EAR. He had some strong
opinions about Transformers & circuits etc.,  one take away was that many amps/transformers
would suffer from some saturation losses when subjected to full spectrum higher dB reproduction.


So, is conventional wisdom correct...does making use of a powered woofer/subwoofer benefit us?
Bi-amplification....active/passive or a hybrid of these?
Can an RM-10 driving a typically modern speaker with 4ohm/85dB at reasonable 75-95 dB listening
levels sitting 9' (3m) away cut the mustard so to speak?

Nostalgia and rose colored glasses being what they are....I have always regretted selling my Wolcott
and Sound-lab A3 system to buy books & tuition.

[G Georgopoulos]

One thing i discovered in class-b is all the power goes into the power device from the lowest signal levels,this gives high quality damping from these low levels,something class-ab and class-a cant do,another thing is at idle class-b is dissipating less power than either class-ab and class-a,another thing is distortion class-b amplifiers had been known to have high distortion something that's not anymore with the work of Douglas Self who proved according to his meachanisms he can reduce distortion to vanishing levels...I have come across similar low distortion operation as well,i know this thread is about class-a,i merely compare different amplifiers to see the pros or cons they may have...hope you guys you'll forgive me..cheers...

[RDavidson]

Being merely a hobbyist...
it was my understanding that full spectrum musical reproduction
can be challenging with ambitious Recorded Crest factors etc.,
That making use of Powered Sub-woofers or Woofers can
relive some of the heavy lifting burden of both Amps & monitors.  Vandersteen has
gone down this path amongst many others....not too mention the DOLBY THX specification.

I had a conversation many years ago with  Tim de Paravicini of EAR. He had some strong
opinions about Transformers & circuits etc.,  one take away was that many amps/transformers
would suffer from some saturation losses when subjected to full spectrum higher dB reproduction.


So, is conventional wisdom correct...does making use of a powered woofer/subwoofer benefit us?
Bi-amplification....active/passive or a hybrid of these?
Can an RM-10 driving a typically modern speaker with 4ohm/85dB at reasonable 75-95 dB listening
levels sitting 9' (3m) away cut the mustard so to speak?

Nostalgia and rose colored glasses being what they are....I have always regretted selling my Wolcott
and Sound-lab A3 system to buy books & tuition.

From my understanding, a subwoofer is meant to augment speakers that aren't full range, not to make up for lack of an amp's ability to "cut the mustard." So, if my understanding is wrong or only 1/2 the equation, I'd appreciate enlightenment.

Your question about power (and headroom), speaker sensitivity / efficiency, listening distance, room size, typical levels, etc. has already been discussed quite a bit. You should read through the thread and I think you'll find your answer.

[RDavidson]

One thing i discovered in class-b is all the power goes into the power device from the lowest signal levels,this gives high quality damping from these low levels,something class-ab and class-a cant do,another thing is at idle class-b is dissipating less power than either class-ab and class-a,another thing is distortion class-b amplifiers had been known to have high distortion something that's not anymore with the work of Douglas Self who proved according to his meachanisms he can reduce distortion to vanishing levels...I have come across similar low distortion operation as well,i know this thread is about class-a,i merely compare different amplifiers to see the pros or cons they may have...hope you guys you'll forgive me..cheers...

George, I'm sorry, but you're REALLY beating the efficiency horse to death in the wrong thread. You should start a new thread about the benefits of non Class-A amps. Nearly everything you have said in this thread has little to do with Class A amps except for their poor efficiency.........which we (people who like Class A amps) are all very aware of, but chose to ignore. This has been covered by several in this thread already. Please stop.

[Roger A. Modjeski]

Being merely a hobbyist...
it was my understanding that full spectrum musical reproduction
can be challenging with ambitious Recorded Crest factors etc.,
That making use of Powered Sub-woofers or Woofers can
relive some of the heavy lifting burden of both Amps & monitors.  Vandersteen has
gone down this path amongst many others....not too mention the DOLBY THX specification.

I had a conversation many years ago with  Tim de Paravicini of EAR. He had some strong
opinions about Transformers & circuits etc.,  one take away was that many amps/transformers
would suffer from some saturation losses when subjected to full spectrum higher dB reproduction.


So, is conventional wisdom correct...does making use of a powered woofer/subwoofer benefit us?
Bi-amplification....active/passive or a hybrid of these?
Can an RM-10 driving a typically modern speaker with 4ohm/85dB at reasonable 75-95 dB listening
levels sitting 9' (3m) away cut the mustard so to speak?

Carl,

I agree with what you have stated above though I see others have disagreed.

Low frequencies at high levels do drive output transformers near saturation.  Tim, as usual is correct. Vandersteen is also as to speakers. These are two of the smarter guys in the industry.

Using a Sub-woofer to relieve both the amp and speaker of information below 100 Hz is a very good idea. Anyone who has looked at music on an oscilloscope or spectrum analyzer (even those with rows of LEDs in octave bands) knows this. However many sub-woofer users make a big mistake by sending the full range signal to the power amp and speaker rather than crossing over before the amps. If the full range signal reaches the main amp you will still have the effects of saturation and IM distortion that you are trying to avoid. Even if you use the speaker level passive crossover in the sub (usually some cheap electrolytic caps that go in series with the main speaker) you are still taxing the amp with the full range signal.  I think most people run subs this way because they fear going through the electronic crossover in the sub will hurt their signal. It also requires long cable runs to and from the sub. So don't go that way and get a good crossover like an RM-3.

We make a great sub system with an external amp and crossover. The woofer box is only 1/3 cubic ft, very small. We use two to spread the launch point and reduce peaks and nulls in the room. The woofer is not mass loaded like other woofers so it is very fast. The crossover under sits under or next to your preamp. It is 1 unit rack mount size and deep enough to form a base for the preamp. I find it very handy to have the sub-woofer level control right there by the preamp so I can adjust is at will for different music. Because the sub level is right there I trim it for every record according to my desire for bass. If I had to reach behind a sub I wouldn't be trimming the bass so often. I can hardly express how this simple control  has increased my enjoyment of my system. This is entirely different from a bass control on a preamp. It does not muddy up the mid bass as a regular bass control as it only changes the level below 100 Hz.

I use an RM-10 to drive a ESL that is around 80 dB efficient to 95 dB peak at 2.5 meters. I am using a sub below 100 Hz, crossed-over as described above. My speaker is a line source, a point source requires more power especially at that distance as SPL falls off more rapidly. If you are talking about 95 dB average level you might be pushing it. An RM-200 would be more appropriate in that case.  If we have the ability to measure peak levels we should be doing it because peak to average is so different with different music.

[Carl V]

I will locate my Volt meter and run some numbers.

DONE> (from post #76 )

around 2.1 volts @ 85~ish dB
2.7 volts @ peak at 95dB.



I have had had  Snell active X-overs, marchand Bassis
and Feedback destroyers, Ranes etc., today, handing off to two Tumults.
And have also utilzed Powered bass cabinets...e.g., two Aurasound
10" drivers with Hypex plate amps with a passive High Pass for the MTM.


A colleague has hinted at selling me his Apogee Stages & Llano A100 amps.
Actually his wife has done most of the encouragement. I will need to reconsider
my next direction in Hi-Fi. As stated previously, I miss those Soundlabs.
I have followed Richard Vandersteen's lead with a powered Woofer/subs.
and decent quality 6dB passive filters.

[Freo-1]

Regarding the Electrocompaniet design approach to their Class A amps, here is a technical description of their design concept:


"Traditionally, designers increased feedback to make a larger portion of the output signal control the amplifier’s response. Our listening tests showed us that simply applying more feedback was not the answer. In fact, as one kind of distortion went down, other parameters would be adversely affected, leading to an overall degradation of sound quality. We knew that the other conventional design approach of eliminating feedback completely was not the answer either, because this would cause high distortion levels, and as a result would produce a “woolly” sound.
The answer to the dilemma was found in a novel approach to feedback theory. We developed a feedback concept that allowed local feedback to be applied around individual stages of the amplifier circuit. This approach allowed us to avoid the sonic disadvantages of overall feedback from output to input. The concept was further developed to reduce phase and interphase distortion between stages of the amplifier as well. We were able to concentrate the loop feedback on the stages of the amplifier where it resulted in audible improvement.
Stability margins were also expanded because feedback no longer affected the frequency response. The use of this concept of individual gain blocks—complex in design but simple in function—allowed us to reduce distortion to minute values in all the products. The amplifier is divided into two separate sections or gain blocks. The input block is a transconductance amplifier without overall feedback. This avoids large output current being fed back to the input, and mixed with the minute input signal. The output block is a transresistance amplifier with parallel feedback. This is done to prevent higher frequencies than the feedback loop can handle from entering the loop. An approach like this will prevent Transient Intermodulation Distortion (TIM) and Slewing Induced Distortion (SID), eliminating the need for an extremely wide bandwidth.”

The items about the sub/crossover has me intrigued.  Currently using a custom DIY tube setup (6AH4 Preamp based on a Thomas Meyer design), and a pair of quad monoblock 1625 power amps.  The speakers are ATC SCM-19's, and a C1 subwoofer.  The concept of crossing over at 80 Hz with a RM3 sounds very interesting! 

[Roger A. Modjeski]

I will locate my Volt meter and run some numbers.

DONE> (from post #76 )

around 2.1 volts @ 85~ish dB
2.7 volts @ peak at 95dB.



I have had had  Snell active X-overs, marchand Bassis
and Feedback destroyers, Ranes etc., today, handing off to two Tumults.
And have also utilzed Powered bass cabinets...e.g., two Aurasound
10" drivers with Hypex plate amps with a passive High Pass for the MTM.


A colleague has hinted at selling me his Apogee Stages & Llano A100 amps.
Actually his wife has done most of the encouragement. I will need to reconsider
my next direction in Hi-Fi. As stated previously, I miss those Soundlabs.
I have followed Richard Vandersteen's lead with a powered Woofer/subs.
and decent quality 6dB passive filters.

Hi Carl,

The voltage numbers  you measured are certainly low and come out to about a watt, so the RM-10, either version would be just fine. However the 95 db number should be 3 times the 85 db number if both are measuring peak.  Peak volts should be measured against peak SPL so we don't mix apples and oranges. In any case the numbers are low and as long as they represent your peak listening level we really don't care so much about how many dB it is.

For those of you still interested in these measurements the SPL readings are only useful in conversation between us to describe how loud we listen. In your own home we just want to know how many volts you are working at to get the level you like. I can't begin to tell you how frustrating it is for us and Music Reference to have a potential customer give us all the information about his speaker, how far he sits, what kind of music he likes and then not get a measure of how loud he listens. Without a number how can I make a recommendation?  The number we really want is the VOLTAGE. Hardly anyone gives that number though it is very easy to measure. If we do get an accurate SPL number we can get pretty close to the power rating. However someone saying ... well I don't listen loud does us no good at all. Carl, thank you for your voltage reading.

I see you have a crossover and sub. Snell made good stuff. Given their cost of repair and need for power I would not recommend the Apogees. Why not get a pair of reasonable electrostats. Since you like ESLs the Quad 57 with a sub is very good. I have had several pair through here recently for service (which was minimal, no panels had to be rebuilt) and I really enjoyed them in my listening room for over a month. I have one pair of my ESL available. The information is on the sticky list.

[AJinFLA]

I invite anyone to give me a cogent argument why one needs a 100 watt amp to play a few watts.

Hi Roger. A (very?) few still listen to music with little processing/wide dynamic range, not elevator pop drone music. Thus 250 unclipped peaks, vs "100 watt" clipped is still appreciated http://www.cordellaudio.com/he2007/show_report.shtml  (scroll down to The Peak Power Demands of Well-recorded Music).

Isn't anyone building a sensible 10-20 watt Class A transistor amp these days?

http://www.lindellaudio.se/index.php/ampx/
"Sensible" enough for me.

cheers,

AJ

[Roger A. Modjeski]

Hi Roger. A (very?) few still listen to music with little processing/wide dynamic range, not elevator pop drone music. Thus 250 unclipped peaks, vs "100 watt" clipped is still appreciated http://www.cordellaudio.com/he2007/show_report.shtml  (scroll down to The Peak Power Demands of Well-recorded Music).
http://www.lindellaudio.se/index.php/ampx/
"Sensible" enough for me.

cheers,

AJ

Thanks for the link. I studied the Cordell article, downloaded the RLJ song (Ghetto of my Mind) and indeed measured 26 db peaks above the average. For comparison I measured some rather dynamic Lyle Lovett songs and they measured 16 dB peak to average. My assistant played something popular by JayZ as an example of  something very compressed which measured 9 db peaks on our set-up.  The JayZ looked hard limited (clipped) on the scope with many clips lasting several milliseconds. By contrast the peaks on the RLJ songs were less than 0.5 milliseconds so my expensive peak reading Fluke 87 could not record their full level even on the 1 mS setting but the scope showed them.

Now we come to the question of what we hear. First I will say the RLJ cut had the highest peaks I have seen on a CD. Consequently it played rather low in level requiring the volume to be turned up more than an average CD.  The engineer on that song had to record it low to leave room for the rim shot peaks.

Now we come to the question: How audible will the rim-shot peak be affected if clipped say 6 or 10 dB. This peak is very brief. Ben who is a recording engineer thought the snare was rather loud. The high peak was of course entirely the choice of the engineer or producer of the cut. He might just have well turned down the mic on the snare. It is interesting because most producers want a loud CD and this one is quite the opposite.  I have not yet tried paying the cut on a limited power system to see at what point the clipping becomes obvious. In previous studies, as I have stated before, I can see clipping on a scope before I can hear it, especially if it is brief.

When I designed the RM-9 and subsequent amps I payed a lot of attention to how the amp recovers from a large overload. I studied traditional amps like the Dyanco ST-70 which does not recover well. I fount the highly prized Citation tube power amps to be the worst at recovery due to their large coupling caps.  I wrote a paper on it.  In my experience if an amp clips occasionally for less than a millisecond and recovers well some clipping is not noticeable. Do we really know how the ear responds to a 0.5 mSec peak 26 dB (400 times) above average.  Maybe the ear doesn't care if the peak is 26 dB or reduced to 20 dB or 16 dB which I feel is typical headroom in a well recorded CD. That would be a good study.

Now there are people out there (guys who buy the big amps) who, for some reason, would be embarrassed if their amp clipped for 1 mS on just one CD on just one song. So they will pay any amount of money to never, never, ever have this happen. They have to suffer the consequences of the big amp burden which is cost, weight, power consumption and some other psychological traumas I cannot imagine.

A more reasonable person would rather have a simple, more economical, less power hungry, perhaps more detailed amplifier that might clip for 1 millisecond once a month.. Heavens forgive!

Keep in mind that at his given average power and speaker sensitivity they were listening at 90 dB average with a peak of 116 dB. One might ask how well the speaker reproduced that peak.  Many or us do listen quite a bit lower. If one listened to the same cut at 80 dB the 250 watt power requirement would be reduced to 25 watts.

All in all it was a nice way for Ben and myself to spend a few hours looking at something that is true but perhaps matters not so much.

[G Georgopoulos]

George, I'm sorry, but you're REALLY beating the efficiency horse to death in the wrong thread. You should start a new thread about the benefits of non Class-A amps. Nearly everything you have said in this thread has little to do with Class A amps except for their poor efficiency.........which we (people who like Class A amps) are all very aware of, but chose to ignore. This has been covered by several in this thread already. Please stop.

At idle class-b is more efficient even than class-d !!! 

[Freo-1]

Roger, I listen primarily to classical, which is known to have peaks up to 40 db or more.  I can't tell you how many times I noticed over the years an occasional slight compression (clipping) when listening to symphonic music.  To mitigate this, one either has to have a very powerful amp that drives full range speakers, or, cross over the bass at 80 Hz of so to a powerful subwoofer to relive the main speakers and amp of this issue.

The ATC speakers can handle very high spl's for dynamic speakers without breakup, which helps mitigate this issue some.  Would like to investigate a active crossover for my tube setup to relieve the power amps and speakers below 80 Hz, and send it all to the sub.   

[AJinFLA]

Hi Roger,

Hope you had a good 4th. Well, you Gished me there, so I'll parse and say, no offense to those that do, but I don't listen to JayZ. I did add the caveat about the few in my post. Classical, Jazz, Big Band etc., have plenty of examples of wide dynamic range (For another specific one, try Flim and the BBs "tricycle", uncompressed of course).
I agree that some instantaneous clipping may very well be inaudible, but I would hesitate to not pay heed or not rely on controlled conditions for any such JND thresholds.

Now there are people out there (guys who buy the big amps) who, for some reason, would be embarrassed if their amp clipped for 1 mS on just one CD on just one song. So they will pay any amount of money to never, never, ever have this happen. They have to suffer the consequences of the big amp burden which is cost, weight, power consumption and some other psychological traumas I cannot imagine.

Excluded argument there. Plenty of more reasonable cost/power amps out there that don't have to fall into mega-sized audiophile status symbol category. Not to mention Class D (from a sound perspective, not a "sound" perspective). Or simply wiser, technically informed loudspeaker choices when system building.

All in all it was a nice way for Ben and myself to spend a few hours looking at something that is true but perhaps matters not so much.

Well, that depends on who you ask.

Personally, I find Class A wasteful (yes, even the one I linked that you didn't comment on). But I don't live up north and I'm no audiophile, I concern myself with only music and sound (without the quotes).

cheers,

AJ

[WGH]

We make a great sub system with an external amp and crossover..... I find it very handy to have the sub-woofer level control right there by the preamp so I can adjust is at will for different music. Because the sub level is right there I trim it for every record according to my desire for bass...... I can hardly express how this simple control  has increased my enjoyment of my system.

A friend has Magnapan 3.7's and he has the sub with the remote control puck, it was great fun trimming bass for each song. The crossover was usually set at 40 Hz for well recorded music, any higher would muddy up the sound, and up to 120 Hz for early rock like the Rolling Stones because the low end on those albums doesn't exist. The bass volume was also adjusted with the remote, the beauty of the setup was the music could be fine tuned from the sweet spot, no up and downs and no risk of spilling our scotch.

Wayne

[DaveC113]

.....................

[Roger A. Modjeski]

Hi Roger,

Hope you had a good 4th. Well, you Gished me there, so I'll parse and say, no offense to those that do, but I don't listen to JayZ. I did add the caveat about the few in my post. Classical, Jazz, Big Band etc., have plenty of examples of wide dynamic range (For another specific one, try Flim and the BBs "tricycle", uncompressed of course).
I agree that some instantaneous clipping may very well be inaudible, but I would hesitate to not pay heed or not rely on controlled conditions for any such JND thresholds. Excluded argument there. Plenty of more reasonable cost/power amps out there that don't have to fall into mega-sized audiophile status symbol category. Not to mention Class D (from a sound perspective, not a "sound" perspective). Or simply wiser, technically informed loudspeaker choices when system building.Well, that depends on who you ask.

Personally, I find Class A wasteful (yes, even the one I linked that you didn't comment on). But I don't live up north and I'm no audiophile, I concern myself with only music and sound (without the quotes).

cheers,

AJ

Gished?  Whats that? I dont listen to JayZ either but it was an example given to me by a younger person. It's what the next generation is listening to.

I agree class A is wasteful and not always the best sound. In some of my bench testing I find best bias is often well below (sometimes 1/4 of) class A bias. I believe the OP of this thread and others think class A is better for reasons I put in my first post. In short A is better than B than C in most things we grade in society, including our report cards from our earliest days in school.

I meant to comment on the amp you mentioned. I looked at their website and found the idle power of 130 watts and output power of 40 watts but no statement if that's 40 per channel or 40 total watts. These numbers are one way of telling if the manufacturer is telling the truth about class A. We know the Chinese and others throw numbers around like noodles.

[Roger A. Modjeski]

Roger, I listen primarily to classical, which is known to have peaks up to 40 db or more.  I can't tell you how many times I noticed over the years an occasional slight compression (clipping) when listening to symphonic music.  To mitigate this, one either has to have a very powerful amp that drives full range speakers, or, cross over the bass at 80 Hz of so to a powerful subwoofer to relive the main speakers and amp of this issue.

The ATC speakers can handle very high spl's for dynamic speakers without breakup, which helps mitigate this issue some.  Would like to investigate a active crossover for my tube setup to relieve the power amps and speakers below 80 Hz, and send it all to the sub.   

If the peak is 40 dB above the average then what level is your average SPL? If your average SPL is 80 dB then your peak is 120 dB SPL. What does that sound like?