Constant Power

The fatal flaw in that, if I read it right, is that speakers are (with very few exceptions) far from ideal single-impedance loads. High output impedance and significant response variation with load (and thus frequency) are hallmarks of the vast majority of tube amplifiers. I love them, but they are nowhere near as accurate as SS.

As far as distortion, we are generally more sensitive to IMD than THD, and tolerate much higher levels of either in music (or movies) than test tones.

All IMO - Don

I try to design speakers with a high impedance /impedance minimum , nominal impedance doesnt tell you much .
Ohms law U=I*R .
When you keep R (impedance ) as high as you can , you dont need much current to stay on the same voltage .
Tube amps in general are not capable of delivering a lot of current , so keeping the speakerimpedance high means you end up with a optimal dynamic loudspeaker.
If its not the case you end up with a soggy slow tubey sound iow a underpowered speaker

This high impedance (and high efficiency) also improves the sound with SS designs iow a efficient motor which doesnt need much power to get you where you want.
Put the money in expensive loudspeakers as a small/cheaper amp is enough ,lets say from 8 watts and up

andromedaaudio said:

I try to design speakers with a high impedance /impedance minimum , nominal impedance doesnt tell you much .
Ohms law U=I*R .
When you keep R (impedance ) as high as you can , you dont need much current to stay on the same voltage .
Tube amps in general are not capable of delivering a lot of current , so keeping the speakerimpedance high means you end up with a optimal dynamic loudspeaker.
If its not the case you end up with a soggy slow tubey sound iow a underpowered speaker

This high impedance (and high efficiency) also improves the sound with SS designs iow a efficient motor which doesnt need much power to get you where you want.
Put the money in expensive loudspeakers as a small/cheaper amp is enough ,lets say from 8 watts and up

Click to expand...

where [roughly] do you draw the line of efficient, high impedance? My X1s (as you may recall) are about 6 ohms across the board (except 3ohms at 17khz and 80hz) and 95db efficient.

My voyager model is +- 90.5 db and has a min impedance of 10 ohm , but i also suspect its the way the filter is designed iow there is probably more to it .
From expirience this one is about as low as i want to go with 8 watts , the new model is not ready yet but that has much higher eff.
regarding your wilson model i think the real test is to put a 8 watter in the chain and open the throttle , ive read a lot of claims which fall short , proof is in the eating in the end

mep said:

So it really boils down to who is designing their speakers to be optimally driven by tube amps and who is designing their speakers to be optimally driven by SS amps? Are any speaker companies consciously even doing that?

Click to expand...

I'm not sure about the latter, but for the former there are several custom speaker builders (mostly horns) that tailor speaker and amplifier combinations for their customer's room, listening level, etc. Though they are rarely, if not ever mentioned in the hifi magazines...

FWIW, I had a RM9 in my system for a few weeks. It's biggest flaw, almost no bass below ~40Hz, little sense of impact with my SCIVs. The rest of the spectrum sounded fairly liquid and dimensional with a nicely delineated sense of depth. No real bass, though, no thanks.

The link in post 1 did not take me to the article ... missing some characters. This is the full link.

Overall, a pretty useless article IMHO. No evidence offered to back up claims. Like this:

It is very easy to tell how an amplifier will sound using measurements based on the Power Paradigm as the measurements are made with regards to understanding and working with the rules of human hearing.

Click to expand...

Ok, what measurements, what rules of human hearing, and how do they relate?

BTW, it's not that hard to make a solid state amp that measures the same power at 2, 4, and 8 ohms. Give it a soggy power supply.

mep said:

I'm sure the "rules of human hearing" has something to do with how we don't like like odd order harmonics and we like them even less as the harmonics go higher and conversely we just love boatloads of second harmonic distortion. As for the measurements, I never heard about a different set of measurements that are used to measure a power paradigm amplifier either. I guess when your amplifier can only come close to making the advertised rated output power at 10% THD, you need some other type of measurements. One thing I do know, tube amps with output transformers basically aren't measured any differently by JA in SP than SS amplifiers, and it doesn't matter if they are single-ended or push-pull.

If we really want to give credence to voltage paradigm amplifiers and power paradigm amplifiers, I think there has to be a third category created for OTL amplifiers because IMO, OTL amplifiers don't fit in either category. OTLs don't fit the definition of a power amplifier that maintains the same power output at 4,8, or 16 ohms. The measurements show that power increases as the impedance increases which should surprise no one based on the fact an OTL amp by nature has a very high output impedance and needs to be matched to a high-impedance load. That's the price you pay for getting rid of those pesky output transformers that actually match the output to the nominal impedance you are hooking up to. Another thing that I find interesting is that there is a subset of Atma-Sphere owners who use autoformers with their amps in order to raise the impedance of their speakers in order to better match the amp to the load. You can google "autoformers for use with Atma-Sphere amps" and come up lots of articles to go read. Here is a link to a company that makes autoformers for Atma-Sphere amps: http://www.zeroimpedance.com/multiplier.html
The ironic thing is people buy a tube amp because it has no output transformer and then they hook them up to an external transformer to try and match the load to the amp.

So, if we have voltage paradigm amplifiers and power paradigm amplifiers, maybe the third category should be called high-impedance paradigm amplifiers because they don't follow the rules for the other two categories as explained in Ralph's paper.



Or run them through autoformers like McIntosh does...

Click to expand...

JA does it right, tube amps shouldn't be treated different from solid state amps. He even gives frequency response into various simulated loads which is helpful. I would like to see measurements of various orders of distortion plotted against frequency, does anyone know if AP software can do this?

mep said:

Felix-I just read a review of your speakers and the measurements taken by JA. Your speakers are efficient and present a fairly easy load to drive.

Click to expand...

How about you take several amplifiers, laden with differing topologies, and try them on the SCIV? Clearly, you have not done that, so what are we talking about? A dissection of the specs and an extrapolation of their imaginary sonic attributes? Really?

I suspect that your RM-9 used the EL-34 output tubes. If so, switching to 6550s or KT-88s probably would have made a big difference in the bass. I also didn't mean to infer that I think the RM-9 is one of the best tube amps out there in tube amp land, I just picked it because it comes damn close to matching the power paradigm as described by Ralph. The RM-9 is also recognized as being a very well engineered tube amplifier.

Click to expand...

The RM9 I used was equipped with KT88s. It did great on everything we sold, except SCIV and above. I'm saying this with the expectation of a full frequency response.

I used to sell these speakers. Many folks expected to drive them with smaller amplifiers (read, Premier 11A), because they read the naïve specs, and were not terribly happy with the results. Shocking! lol Some went for smaller YBA amps, BAT VK500 ...I can go on and on. No dice, again.The BAT was big, but sounded small and shrill. Sunjectively nothing below 30Hz. FWIW, I sold the first and secord VK500 in the country. The second had the optional BAT-PAC; same owner as the first. The guy hated it on his 801s and traded it on a Coda 20.5. That was a real amplifier. OTOH, an Aleph3 drove them very well, albeit with dynamic limitations, same goes for a MC275. To be fair, the MC275 didn't reach all that low, but it did have better control than the RM9.

Yes, the RM9 can be a fantastic amplifier in the right system. Judging system compatibility solely by the specs is...well...you know.

mep said:

Did I miss the above sentence the first time I read your post or did you subsequently add it after?? I don't know why you felt the need to get snarky with me for commenting on what the measurements indicated for the SCIV speakers. Usually if speakers are a tough load to drive, JA mentions that in his measurements. I mistakenly thought that your RM-9 used EL-34 tubes. The bottom line is that we all know every amp doesn't match every speaker and you found the RM-9 didn't provide the bass that you knew the SCIV speakers could deliver. The only reason that I even mentioned the RM-9 amp in the first place is for the reason I explained before-it is a good example of a power paradigm amp. Never did I say that the RM-9 was the ideal amp for your speakers nor did I recommend any amplifier for your speakers. If your RM-9 did use EL-34s, I thought that switching to 6550s or KT-88s would make a positive difference in your bass extension. Obviously since you were already running KT-88s, that wasn't the case. Why you needed to get in a snit is beyond me. I really didn't deserve that.

Click to expand...

Sorry, Mark. Rough morning with the wife.

MEP, you missed some things and made some incorrect assumptions in your opening post.

So I will try to clarify. Please refrain from Logical Fallacies as you have used in the past. I want to keep this as an intelligent debate. If you do not understand what Logical Fallacies are (such as the Strawman), please google the phrase.

The Power Paradigm is the method of design, test and measurement that existed before the Voltage Paradigm rules were developed in the late 1950s and into the early 1970s. In many fields of endeavor, a succeeding art will eclipse the prior art, but in the case of audio that did not happen. Instead the two methods/techniques/camps have existed side by side since the 1950s to this day. Examples of of Power Paradigm technology are a good number of tube amps, but transistor amps can be Power Paradigm technology too. In order for them to work properly, they must be mated to loudspeakers that are also Power Paradigm technology. Examples include:

all traditional horn loudspeakers, most modern horn loudspeakers, ESLs like old Quads and all Sound Labs, Audiokinesis, Wilson Audio, Coincident Technology (of Canada), the original Acoustic Research AR-1 (the world's first acoustic suspension loudspeaker, mentioned here because of its historic nature), Classic Audio Loudspeakers (a horn/bass reflex hybrid), Merlins, High Emotion Audio, all headphones and many more that I am not thinking of off the top of my head.

In short a Power Paradigm loudspeaker is any loudspeaker that can be driven with reasonably flat frequency response by an amplifier of moderate output impedance (1-20 ohms is typical; impedances above that we shall call 'high impedance', as there is also a current-drive method which has been mentioned by Nelson Pass but remains mostly a DIY matter; such amplifiers may have an output impedance as high as 80 ohms), wherein the amplifier has little or no negative feedback. There may be several means of obtaining flat frequency response; examples are port tuning (ex.: Audiokinesis) and midrange/tweeter level controls.

There are several points to keep in mind:
#1 If you mix technologies you will get a frequency response error. That is to say that a Voltage Paradigm amplifier (one that is capable, IOW, of constant voltage response with respect to load impedance) will have a frequency response error when driving a Power Paradigm loudspeaker and vice versa. This can be because the crossover design rules for a Power Paradigm loudspeaker are different. The result can be that the drivers in the speaker (in the case of a dynamic loudspeaker) may be trying to reproduce material outside of their passband because the crossover does not work correctly with the amplifier (this is why many older horns have a reputation for being 'honky').

#2 No amplifier is a perfect Power Source, just as no amplifier (in the Voltage camp) is a perfect Voltage Source. There will be an impedance, different with all amplifiers, wherein the models fall apart.

#3 to obtain flat frequency response with a Power Paradigm amplifier you do not need a speaker with a flat impedance curve (although they can work just fine with that, and as a result, there is some cross over between camps). To get flat bandwidth, the overall impedance of the speaker has to be either a) high enough overall or b) designed to expect a certain behavior out of the amp if the impedance is lower. An example of this is the Wilson Watt/Puppy, which has a 2 ohm impedance at 2KHz because of a 2KHz trap which is intended to kill excess energy. This works great with amps that have a higher output impedance such as a tube amplifier, but not so great with a SS amp. The result is that the Watt/Puppy had a reputation for being bright (see #1 above). However it had no such brightness if used with the equipment for which it was designed.

#4 The Voltage Paradigm was developed in the late 1950s and into the early 1970s. Electro-Voice and Macintosh were leaders in setting up the technique, with the goal being predictable flat frequency response.

****

In one sense this is very much about equipment matching, which is a conversation that has gone on it audio for decades, about 5 decades to be a little more precise. There is also the tubes/transistor debate and the objectivist/subjectivist debate, which all stem from the existence of the two means in the field.

Put another way, if this was talked about more openly, audiophiles would spend less money making expensive mistakes.

Now there is a different question: why is the Power Paradigm still around 40-50 years on? Why did it not become the dust of history? The answer is that it still offers to this day, a method of extremely accurate musical reproduction. The reason has to do with loop negative feedback, a design feature that has been used in amplifiers very commonly in the last 50-60 years. The problem is that all amplifiers have a measurable propagation delay (the time it takes for a signal to propagate from the input to the output). What this means is that if you apply global feedback to such an amplifier, it will at some frequency be coming back too late to reduce distortion and instead will begin to add to it; in particular to the odd-ordered content. Additionally, intermodulations at the feedback node will contribute to harmonics as high as the 81st, creating a noise floor of harmonic and inharmonic distortions (Crowhurst).

In a nutshell, this causes the amplifier to be brighter and harsher sounding than the original musical signal. In short this is a coloration, and one that our ears are very sensitive too, on account of the fact that our ear/brain system uses odd ordered harmonics to gauge how loud a sound is. IOW, such distorted waveforms sound louder than they really are and to many they can also be outright painful.

If you want to avoid that you have to avoid negative feedback in the amplifier. This means that you have to find other ways to obtain linearity in the amplifier, and other ways to get flat frequency response in the loudspeaker. Enter the Power Paradigm. It is a lot easier to build an amplifier with good linearity with no feedback if you use tubes. So quite often tube amplifiers operate with little or no feedback. This results in a higher output impedance which must be dealt with differently and properly in the loudspeaker crossover and box (if there is a box) design.

Tubes can be made to more closely follow human hearing rules (IOW, they make less odd ordered harmonics and triodes otherwise are some of the most linear amplification devices known to man) and this is why tubes are still in production a good 50 years after becoming 'obsolete'. IOW the Power Paradigm is still around as the use of vacuum-tubes is still the most common means of reducing the distortions that the ear finds most objectionable. Is it perfect? No- nothing is. But it does work, hence an entire industry is still alive to service it.

mep said:

So it really boils down to who is designing their speakers to be optimally driven by tube amps and who is designing their speakers to be optimally driven by SS amps? Are any speaker companies consciously even doing that?

Click to expand...

Oh yes, there are many speaker manufacturers who design with tube amps in mind.

One way to spot a loudspeaker that will work well with power paradigm amplifiers is to look at the impedance curve. If it is substantially flat, and not too low, it will work fine. Coincident Speaker Technology is one that comes to mind. Incidentally, the flat impedance curve of the Conicidents means that they will work well with voltage paradigm amps too.

Or you may find a speaker whose impedance curve "zigs" where its frequency response curve "zags" (based on measurements made with voltage paradigm amps). Silverline Audio comes to mind, I forget which model, but I'm sure there are many others that I just haven't taken note of.

I am a dealer for several amplifier lines, both tube and solid state, including Atma-Sphere. So I design my home audio speakers to work well with a wide variety of amps by keeping the impedance curve fairly high and substantially smooth (perhaps varying between 7 and 12 ohms above the bass peaks for my 8 ohm models, and scaled up accordingly for my 16 ohm models). That being said, there is a "free lunch" available with high output impedance amplifiers: Increased bass extension. One way to take advantage of this is to provide a user-adjustable bass tuning system. Basically you want to be able to tune the cabinet a lot lower than would be "optimum" for a solid state amp (there's a little more to it but that's a good start). With this technique, it is quite possible to get an additional 1/4 to 1/3 octave of bass extension.

I'd like to comment on a couple of other things you said. My own measurements indicate that Atma-Sphere's posted output impedance specs are correct for the S-30 (haven't measured their other amps, but the S-30 should have the highest output impedance of all). This is based on frequency response measurements into a known variable impedance speaker. I found this out the first time I set out to design a crossover for a speaker that would be used with both Atma-Sphere and Nuforce amplifiers (representing pretty much opposite ends of the amplifier spectrum).

Regarding the high distortion measurements that were reported for Atma-Sphere amps driving a low impedance load, 10% THD is a non-issue from a perceptual standpoint if the distortion is second harmonic. Forty percent second harmonic distortion has been shown to be statistically imperceptible. On the other hand, very small amounts of high order, odd order harmonic distortion are both audible and objectionable. Crossover distortion (in amplifers) is also far more audible and objectionable than one would expect, and is usually ignored as being buried in the noise floor. This has been documented by Earl Geddes and Lydia Lee in a pair of papers on distortion perception published in the Journal of the Audio Engineering Society a few years ago. Earl told me in a private conversation that his data shows an inverse relationship between amplifier THD specs and subjective preference - in other words, higher distortion amplifiers tend to sound better. The reason is, the techniques used to lower the THD spec tend to actually increase the audibility of the distortion products. So the amplifier world is using the wrong yardstick. Total harmonic distortion numbers do not begin to correlate with subjective preference. The reason has to do with a property of the ear/brain system called "masking", an explanation of which is beyond the scope of this post.

Here are those papers. The above synopsis is based on conversations with Earl.

http://www.gedlee.com/downloads/Distortion_AES_I.pdf

http://www.gedlee.com/downloads/Distortion_AES_II.pdf

Anyone attending the Rocky Mountain Audio Fest is invited to see what a high-THD power paradigm amp sounds like by visiting either the Classic Audio Reproductions room or my room. If THD is an accurate yardstick of amplifier sound quality, you should easily be able to hear the inferiority of amps that have hundreds or thousands of times higher THD than the competition. I won't even ask you to bring an open mind - bring your honest skepticism, just give a fair listen.

^^ FWIW dept.: between the time of the Soundstage! review 12 years ago and now we have found ways to reduce the HD distortion by about 90%. The amplifier of that review was discontinued in 2002.