Double down, good or bad?


I came across this article on Atma Sphere's website:

http://www.atma-sphere.com/papers/myth.html

In short, Atma Sphere believes having a power amp that is capable of doubling its power when impedance is half is not necessarily a good thing because speakers in general do not have a flat impedance across all freq range.

On paper, it does make sense. Though I am sure speaker designers take that into consideration and reduce/increase output where necessary to achieve the flatest freq response, that explains why most of the speakers measured by Stereophile or other magazines have near flat responses.

But what if designer use tube amps to design his speakers, mating them with solid state should yield higher bass output in general? Vice versa, tube amps yield less bass output at home?

I have always been a tube guy and learned to live with less bass weight/impact in exchange of better midrange/top end. Will one be better off buying the same exact amp the speakers were "voiced" with, not that it will guarantee good sound, at least not to everyone's ear.
semi

Showing 13 responses by atmasphere

I'll take it a step further- there aren't **any** tube amps that double power, yet the tubes vs transistors debate has been here for 40 years... what does that tell you? The tubes failed to be the 'obsolete' technology that they were supposed to be- people keep listening to them.

The simple answer is that they must be doing something right!

The ability to double power as impedance is halved has been entirely a solid state thing. The idea behind it is that the amp is considered 'load impervious' if it can make constant voltage output into all loads, IOW double power as impedance is halved (conversely halve the power as impedance is doubled), so you can get flat frequency response from a speaker that has a box resonance. But what if the impedance curve of your speaker isn't based on a box resonance, like an ESL, magnetic planar or horn system? The model starts to fall apart.

So here's where that leads:

http://www.atma-sphere.com/papers/paradigm_paper2.htm

Its not really a tube vs transistor thing after all- its a bit bigger than that, just as it is not about objectivist vs subjectivist, and we are talking about equipment matching too. Its all those things, but also about the fact that the industry has a dilemma: make it look good on paper to make lots of dollars, or make it to sound good to the ear to make lots of good sound?

Its my contention that this has been going on a long time and the industry does not like to talk about it (doing so reveals the purely monetary angle), resulting in a ton of confusion (and often bad sound) in the audio community.
Al, in the case of the B&W 802 you need to have 3 db more power into the 4 ohm woofer load as you do into the mids and highs because the woofers are 3 db less efficient. I am ignoring the effects of feedback here and doubtless there are amps that don't quite double power that can do OK on this speaker. I point it out though, because your typical tube amp is not going to work right on this speaker- it literally is designed with the constant voltage model -the Voltage Paradigm- in mind.
Unsound, I agree with you. I should point out that what you said in your last post is not anything I was suggesting. I think its nice if an amp measures well, if it also sounds good to the human ear.

What I have found is that the physics of making an amp sound good has to do first with understanding the rules of human hearing, then applying the physics to get the amp to follow those rules.

Negative feedback is a primary violator of one of the fundamental rules of human hearing (its use enhances the harmonics that the ear uses to determine volume of a sound) and a lot of amps use feedback to reduce distortion i.e. look good on the bench. Amplifiers that have good bench specs and no feedback are rare but they exist. Interestingly enough, most of those amps also have a reputation for good sound too.
Kirkus, it seems you got my point in your last paragraph. We could easily build an OTL with an output impedance of 0.1 ohms with enough feedback, but it would never sound right, and the lower output impedance would say nothing about what loads it could effectively drive.

The other aspect of the conflict of the Voltage and Power paradigms that is important here is equipment matching- any time you mix the two, you get tonal anomalies: a transistor amp driving horns for example will be shrill, a tube amp driving a set of B&W 802s will have insufficient bass.

In the case of the horns, the drivers are very reactive and the back EMF wreacks havoc in the feedback loop of the traditional transistor amp, making the presentation shrill, in the case of the B&W, the woofers are in parallel (4 ohms) and are 3 db less efficient than the 8 ohm midrange and tweeter; if a transistor amp is used it will have double the power into the 4 ohm portion of the impedance curve, thus driving the woofers with the 3 db extra needed for them to operate with the midrange and tweeter. The speaker cannot be driven properly by any amp that does not double power as impedance is halved.

There are many other examples!
Magfan, voltage source / current source are Voltage Paradigm terms. A Current Source amplifier is one with a high output impedance. The Power paradigm term for this is 'power amplifier' since such amplifiers attempt to make constant power into all loads rather than constant voltage.

The two approaches require different speaker design considerations! Neither requires a flat impedance curve, but it you want to compare the two types of amplifiers, you need a speaker with a flat impedance curve and an adequate minimum impedance (8 ohms as opposed to 4 for example).
Kirkus, you are right in most of your posts and apparently either did not read something somewhere in mine or misinterpreted it. First, I'm not talking about clipping here- at all. I am well aware of the significance of the woofer in the box! The very point of a woofer in the box is that there is a resonance. Voltage paradigm amplifiers ('voltage source' amps) will throttle back their power as they encounter the resonant peak of the box, and so not exacerbate that resonance. That is the basis of the theory of the idea of making an amp that is a 'voltage source'.

Maybe you missed the link, so here it is again:
http://www.atma-sphere.com/papers/paradigm_paper2.html

Before the Voltage paradigm was proposed (MacIntosh and EV were two proponents in the 50s and 60s) the Power paradigm was the only game in town. I have had to create the terms 'Voltage Paradigm' and 'Power Paradigm' simply because the industry is mum on this subject in general- its inconvenient.

Yet the web audio forums are full of people that are constantly harping about aspects the the conflict between these two paradigms- tubes vs transistors is one of those debates that seems to arise from this conflict. But if you think my posts are *about* tubes vs transistors you have missed the point entirely!

The Voltage Paradigm uses a sort of short hand nomenclature that leads to confusion and I think this is happening for you in this case. Of course all amplifiers are power amplifiers, what the Voltage Paradigm seeks to do is call an amplifier a 'voltage source' when it refers to an amplifier that makes constant voltage into any load (i.e. doubles its power as the load impedance is cut in half). Amplifiers that have a high output impedance and make (or attempt to make) constant power are termed 'current source' amplifiers.

What is important to note here is the word paradigm. If you are operating solely within a paradigm, anything outside that paradigm can be construed as blasphemy.

The Power Paradigm amplifier is a 'power source', i.e. it will make constant power into any load. That is the voltage and current will both vary. I don't know of an amp that does this but that is the ideal, just as there are no true 'constant voltage' amplifiers out there either- that is the ideal. Does this clarify things?
Unsound, we got a Golden Ear Award for our MA-1 driving a set of Apogees. This is a zero-feedback tube amp BTW.

Kirkus, its clear to me that your perspective is that of the Voltage Paradigm. As to papers on the subject, one was written by one of the designers at EV. The way tube amp manufacturers were getting their 'output impedance' down was by adding feedback- to the detriment of the resulting sound quality. The HK Citation 2 is a good example- a fair amount of feedback, used to reduce the distortion imposed by the AB pentode-based transformer-coupled output section. Despite that though, the amplifier fails to double power as the load impedance is halved. And we are not talking about clipping!!

clipping power is precisely what this thread, and "doubling-down" is all about. What I'm confused about is why you seem to be discussing clipping-power specifications and output-impedance specifications as if they're interchangable . . . or at least a common debate. They're not.

I am not discussing clipping at all, nor do I see clipping power as interchangeable with output impedance- no idea where you got that.

FWIW the 'Grand Conspiracy' thing you mention seems to me an example of adding meaning where none existed prior. All I am pointing out is what is causing the tube/transistor debate, the objective/subjective debate and the equipment matching conversation- they are all the same thing. There is a secondary conversation regarding the rules of human hearing/perception, wherein I contend that it is important to understand those rules and adhere to them as design principles. FWIW, that, for the most part, is not happening in audio.
Unsound, I've heard a set of our MA-1s driving the old Full Ranges, which were one ohm (we used a set of autoformers that had one ohm taps). There is an old TAS review which mentions that from the early 90s as the speakers were owned by a TAS reveiwer.

Seems to me Paul Bolin used Duetta Signatures.
Magfan, I also find the Magnaplanars easy to drive.

Unsound, the TAS review using the Duettas did not use the autoformer, but in the case of the Full Ranges, without an output transformer of some sort, driving 1 ohm is difficult, even for transistors. Its not 'cheating'- its simply a tube amp that was able to do the job. The speaker otherwise is easy to drive.

Kirkus, I see now why we were not on the same page- I was referring only to my posts. You are right- the article is over-simplified. So I think we are really on the same page here. With regards to 'belief systems', what is happening in high end audio is tubes are effectively competing with transistors, and there are speakers (like the Wilson Watt/Puppy) that are more effectively driven by tubes than transistors and other speakers (like the B&W 802) that are more effectively driven by transistors.

People are constantly trying to mix the two technologies and in the process are flushing a lot of money down the loo.

The Paradigm paper seeks to expose why, and in the process also expose something that I regard as the Emperor's New Clothes- that aspect wherein products are made to look good on paper and ignore human hearing/perceptual rules; that whole thing that all of us are familiar with where the specs tell you nothing about how it sounds. IMO/IME we as an industry are measuring the wrong things, and there is little or no discussion about that!

I know of a neuro-scientist that has done studies of human hearing- he has found that when the audio playback system violates our perceptual rules, the processing of the sounds in the brain is actually handled in a completely different area!
Peterayer, The drivability has to do with whether the driving amplifier is based on Power theory or Voltage Theory, and which approach the speaker designer used.

Power Theory (or Power Paradigm) is where the amp seeks to make constant power into all loads. It will not succeed, but that is the goal. The Dynaco ST-70 is a good example, 4,8,16 ohms its 35 watts. Our own MA-2 is another, 4,8,16 ohm 220 watts. Some transistor amps fall into this category.

Voltage Theory (or Voltage Paradigm) is where the amp seeks to make constant voltage into all loads. It will not succeed, but that is the goal. Such an amplifier doubles power as the load is cut in half. Lots of transistor amplifiers do this but only a handful of tube amps do.

The speaker designer may have used an amplifier for reference. If so the speaker obeys the same rules that the amp does. So if the designer favors a tube amp, his speaker will likely follow the Power Theory. An example is Wilson Audio. If the designer used a speaker design program or favors a transistor amplifier, the speaker will be a Voltage Paradigm device. An example is the mbl 101E.

Anytime one kind of amp is used with the other kind of speaker, a tonal anomaly will result due to inappropriate power response. Here are some examples:
*Transistor amp on ESL =>bright high end, weak bass
*transistor amp on horn =>shrill high end
*tube amp with mbl101 => harsh midrange
*tube amp with B&W 802 =>weak bass

I have used the term 'Paradigm' in the past as those who operate within a paradigm seek to describe that which is outside that paradigm as wrong, on account of a restructuring of that individual's world view that the individual is understandably reluctant to do. It is easier for such individuals to see the alternative as wrong, but in so doing does not change the nature of reality.

This reluctance to accept reality has resulted in several debates in the audio world that are on-going and have been so for decades: tube vs transistor, subjectivist vs objectivist and the equipment matching conversation which you have asked about.

To deny the existence of the Power Paradigm denies the validity of ESLs, horns, magnetic planars, full-range drivers and a variety of acoustic suspension and bass reflex designs, as well as many tube amplifiers. The Power Paradigm places a high value on design responsiveness to human hearing/perceptual issues.

To deny the existence of the Voltage Paradigm is nearly impossible- that is what is seen in bench measurements from all magazines. Many acoustic suspension and bass reflex designs use the Voltage Paradigm rule. The Voltage Paradigm places a high value on design responsiveness to bench specification, particularly low distortion.

I side with the idea that audio equipment should obey human hearing/perceptual rules, as I maintain that human hearing is the most important aspect of audio. I get very little argument for the latter, but a lot of resistance to the former!
Peterayer, you are right, the Eggleston is designed around Voltage Paradigm rules, but some of the Pass amplifiers are not. You can find quite a bit of information on Nelson Pass' website regarding that. BTW his papers on distortion are excellent.

Bob_reynolds, Sure:
A transistor amp will make too much power at high frequencies on an ESL because ESLs have a low impedance at high frequencies, often the difference between the bass and highs can be as much as 8X difference in power response in an amplifier that can double power. Gregadd, this answers your question as well.

In the case of horns, the drivers are highly reactive due to tight gaps in the voice coils, and produce a lot of back EMF. When presented to the feedback network of an amplifier, this back EMF causes the input of the amplifier to misbehave- the voltages at that time do not represent the actual correction that the amp needs. So it tends to make excess harmonics and so sounds shrill.

In the case the the mbl, it is an easy 4 ohm load with an 8 ohm peak on the midrange driver. The designer expects the amp to throttle back its power, but an amp that makes constant power will not do so and so will have excessive upper midrange.

In the case of the B&W, the woofers are in parallel and are 3 db less efficient than the mids and highs. An amplifier with constant power response will drive the woofers with 3db less power than intended.

In all cases it is inappropriate power response. This is the major reason why you get mismatch between amps and speakers, although I would be remiss if I did not point out that distortion in both transistor and tube amplifiers also plays a significant role in what we hear as tonal aberration.
Bob_reynolds, Matin Logan has worked for years to try to make their speakers work with transistors. But ESLs in general do not behave quite the way you describe. Most of them are looking for the same amount of power to make a given amount of output, regardless of frequency or impedance. This has largely to do with the fact that their impedance curve is based loosely on a capacitor and has nothing to do with resonance. This is why amps with constant power response will be flatter in frequency response on Quads and Sound Labs, where a transistor amp will be making too much energy at high frequencies, and not enough in the bass. BTW this is a problem with Martin Logans too, despite their years of work to prevent it.

The Voltage Paradigm works best (insofar as frequency response is concerned) when dealing with speakers that have a box/driver resonance of some sort like found in an acoustic suspension speaker (although oddly enough, the very first acoustic suspension speaker, the Acoustic Research AR-1 was a Power Paradigm device and was built for amps that had a high output impedance).

Contrary to your assertion, a Power Paradigm device like an ESL, horn, or some types of box speakers (like the AR-1, but also Coincidents, Merlins and Wilsons) do **not** have to have a flat impedance curve to get flat uncolored response! All that has to happen is that the speaker be tailored in some way to deal with the constant power response of the amp. In the old days when horns were one of the few games in town, the speaker was equipped with controls to vary the output of the drivers, not to deal with room issues, but to assist the amplifier/speaker interface in getting tonally-flat response. Often all that is needed is that the impedance be *high* enough, beyond that the devil may care about the actual value.

In modern Power Paradigm designs, for example a bass reflex, using Theil/Small parameters the designer of the enclosure would set the port at a lower frequency than a designer of the same size box and driver would if in the Voltage Paradigm. Using Power Paradigm rules, the port would be set at some point wherein the amplifier will drive that set of frequencies a little harder (simultaneously avoiding the box resonance and extending the low frequency bandwidth to a lower point than possible using Voltage Paradigm rules), where if using Voltage Paradigm rules the port would be asking for considerably less power from the amp. But in either case the impedance curve would be anything but flat!
Audioquest4life, first, phase angle and the like is really not part of the discussion- so far. As far as the 802s go, sure, you can make a tube amp drive them and have them sound fine. We have had customers do so in the past.

However, that is not the same as saying that they are sounding their best! If you were to put the speaker/amp combo in an anechoic chamber, you would easily measure the effect of the reduced woofer output (which is not the same as reduced low frequency **extension**).

The problem is that the speaker is intended for use by a voltage source, but I find that I prefer how it sounds with tube amplifiers as well, despite the reduced woofer output.