Our amps are really not spinoffs of anything. And they are tube. So I don't agree we're at the end of the line by any means...
But to make it more interesting, we've been working on a class D amp of our own for which we're also working on a patent. Not going to reveal too much, but we solved one of the major sources of distortion in class D amps. And we have proof of concept.
Now if we can do that on no budget and without any prior *recognized* expertise in the field of endeavor (people tend to think that just because we only do tubes that we can't know solid state as well, as if solid state is not taught at the University of Tubes or something...), what does that tell you? That perhaps there is still more to be done?
One area that is a problem for all amplifier designs is that most are designed to have specs that look good on paper and are not really designed to also sound good. Now this is a simple engineering problem (understanding the rules of human hearing and designing to those standards rather than the existing set of arbitrary rules); the bigger problem is tradition- the tradition of how we say what are good measurements and what are not is at the heart of the issue. How do you get the industry to move off of standards set in place 60 years ago??
Until we fix *that* problem, progress will only be had by the outliers who are willing to buck the tradition and pay the price. And they are out there.
Some years back I had some troubles when some people tried to steal my company. I remember getting a call from David Berning, who simply called to offer moral support; he told me (paraphrasing) that 'the industry needs people like you that bring diversity to the field'. I really appreciated hearing that from him and who better to say it as he is exactly one of those individuals: a brilliant designer and no-one makes amps like he does either!
There are brilliant designers in this field and there are those that recognize that if their amp is simply competent, someone will buy it even if it is a rehashed 1950s circuit. I don't see that progress has slowed down at all- if that is what Nelson (whom I see as one of the world's top designers) is saying then I disagree! I do think that we see a lot of derivative circuits but we're always going to see copycats.
|
@kosst_amojan The gold standard is class A and nobody has come close to building a
class D amp that touches the qualities of class A. Even the very best
draw very mixed opinions. Amplifiers are art. Every amplifier imparts
distortion with some character. For those of us accustomed to low order
distortion imparted by simple topologies using as few gain stages as
possible, class D isn't even on our radar. It may interest you to know that our tube amps are class A with only one stage of gain. It can't get a whole lot simpler than that. And I've yet to hear a class D amp that can keep up. But it would be a mistake to simply write off class D. Its still on the steep part of the price/performance curve. |
@inna Why would a respected tube amp designer want to play with something very different? To see what is possible. Nelson Pass sold a kit very briefly that employed an output transistor known as a Static Induction Transistor or SIT. This was and so far is the only transistor type ever made that had the linearity of triodes and even had a soft clipping characteristic- so it should be possible using them to make a solid state amp that sounds exactly like a tube amp. Sony made them back in the 1970s for their Vfet amps. Nelson made a kit using them and you can bet that I wanted to know what that sounded like so I bought a kit. Why would you not? is more the question! |
^^ I have exactly that in mind :)
|
The digital data to be converted to analog far out strips the ability of the analog section to pass along the stored image. This statement is false. |
^^The premise of the post above is false.
Try running an analog source through that output stage and you will see why.
As a bit of a tip: its probably not a good idea to talk about 'distortion-less' in the way that you do, it strains credulity.
|
the reason is because you need a method of detecting velocity first
before you can correct it - and then you have to do it in real time
(without any delay).
Although I find much of this amusing, I do agree that speed is important. Our amps don't sound like tube amps partially because they are so fast; there are very few solid state amps that are as fast as our amps. In that way our tube amps differ quite a lot from the vast majority of tube amps. As I understand it Roger, you have a means of detecting this 'velocity' but what I found peculiar last time I engaged in this topic was that you had not quantified this velocity as a specification. I'm pretty sure you're not talking about risetime/slew rate. |
So Roger, have you quantified this axis then?
That is the question I asked before and got a long answer that seemed to contain neither yes or no. Have you quantified, do you have a spec for this circuit?
|
OK- so this sounds like to me that the correct answer to my question is 'no'.
|
As I understand it Roger, you have a means of detecting this 'velocity'
but what I found peculiar last time I engaged in this topic was that you
had not quantified this velocity as a specification. I'm pretty sure
you're not talking about risetime/slew rate. The question is, have you quantified this circuit operation with a spec? Or put another way, have you quantified the 'velocity' spec? At least you confirmed one thing- that you are not talking about risetime/slew rate. |
Your analogy of analog tape does not hold water when applied to amplifiers. Amps don't cause pitch variation. Speed variation in analog recorders does.
From your description, if I had to guess its almost as if you are describing the speed of the amplifier in terms of propagation delay- that the delay from input to output must be at the speed of sound.
But I suspect that interpretation is not correct either...
But at any rate you've not answered the original question- which is: have you got a means to quantify this?
|
Alright- so several times I've asked the same question and met with obfuscation. So I have to assume that you (Roger) have been unable to quantify the effect you are talking about.
Since you claim to have a circuit to compensate for this effect, the means to quantify the effect is apparently at your fingertips and this should be obvious to the casual observer.
Therefore I can only conclude that this 'effect' is non-existent. Otherwise you would have already told me how you quantify it. Kosst-amojan stated the exact same problem in a different way in the post just prior to this.
|
Do you honestly think I'm blowing smoke or selling snake oil? Since you have consistently avoided answering a rather simple question, one that **should** have been very easy, I'm forced to conclude regardless of what I want to believe that the correct answer to the question above is 'yes'. Here's why: you claim to have a circuit to correct this 'speed' issue. At this point we don't have to know anything technical about it; its mere alleged existence points to a means of quantifying the 'speed' of an amplifier circuit, which it then corrects. This has been true apparently for some years as there is another thread on this site wherein this circuit is discussed, and that thread is several years old. But apparently despite the existence of this circuit, it has not helped in the means of measuring the speed of the amplifier circuit, which it surely must do, otherwise how could it correct; for that matter you would have to have a reference to know how to set the 'speed'. The logical issue here is obvious: obviously the 'circuit' does not exist. Otherwise you would have a specification to look for so that when you shipped a product out the door, you would know it was correct. |
Hm. This explanation is very different from the one you gave several years ago.
I'm very open to the idea that there are new vistas to be explored. That's why I have patents. But I think you need to get your story straight.
One problem with your story is that you don't have a measurement means. That's a problem that anyone with an engineering background will point to; and like me they will find it odd that you've had a circuit to compensate for this 'effect' but you've not developed a means of measurement. Anyone can see that such will lead to QC problems.
So my advice to you is to develop solid repeatable measurements. You already have the circuit (allegedly); so it itself is your means of measurement, as if that was not abundantly clear earlier. It is this particular fact that will send up red flags for anyone with a logical mind.
|
OK- so despite my asking several times it appears you did have a measurement system after all! Why didn't you just come out and say so the first time I asked?
Air is anything but a constant when dealing with the speed of sound. Humidity and air pressure both play a role. Do you have compensation for pressure and humidity?
I'm having a problem with several other comments here as well. One was the speed of the power supply and another was the speed of the amplifier. Long ago I discovered that any power supply has a timing constant and if the amplifier goes lower than that constant IM distortion will rise. So that is a rule I've been careful to follow. But from your comments it sounds as if you are talking about something else.
Also the speed of the amp is another area that is well understood. Are you referring to the speed of the amplifier as propagation delay?
|
I don't understand much, if any, of Mr. Paul's explanations for his discovery/invention I would not feel bad about that. I suspect that mach 1 has nothing to do with the inside of an amp. Given two explanations, its likely that the simpler one is the correct explanation. We have Mr. Paul's rather lengthy explanations and we also have a very simple explanation for the same thing. |
Hm. I don't suppose you see where the problems are in your response here?
I'll point them out: First, apparently you don't have a measurement system as one of your recent posts seemed to suggest. You have a simulation method. Simulations are great when the simulation system works and they really suck when they don't. The thing is, if you don't then back up the sims with actual real world data, the sim might be helpful but you can't know that its accurate. Some simulators are pretty good these days, but I've seen them be a mile off.
One example I know of resulted in an ESL to be really hard to drive. Turned out the sim was nothing like the real world measurement- it was a country mile off. Once that was pointed out to the designer and corrected, the speaker became a whole lot easier to drive and better sounding at the same time.
In your case we have several problems. First, there is nothing to suggest that an amplifier has to operate at the speed of sound. I think any designer would agree that a good amp will treat all frequencies with the same speed. Most are much faster but none run at the speed of light. So you have to make a far more convincing argument; one that is backed up by some sort of physical law. The conclusion you jump to by simply declaring that amps have to be as slow as air is not logical- but I can see it appealing to those who don't have an engineering background.
The second problem is a physical measurement of the performance of your circuit is required. Without it no-one (including you) could say if it even works at all- right now the position you seem to be in is that you have an amplifier that sounds nice but to separate it from the competition you've come up with a panacea that no-one in the industry seems to have even heard of!
The problem with that scenario is that sooner or later you will run into someone that takes it personally when they see that there is an attempt to pull the wool over their eyes. Further telling them that they are just not up to 'speed' (if you will pardon the expression) is not actually the way to solve that problem; in fact its a recipe for making it worse.
What's hurting you right now is you don't have any means of proving that what you say is true. Its not enough to ask someone to trust you. Its further not enough to say that you've been working on it for however many decades (or that you found a bug with someone else's product; I can tell you from personal experience no-one cares). People often devote their lives to ideas that are mistaken. So I would encourage you to come up with a physical test that allows you to differentiate your amplifiers in the way that you say they are different. Put it in a box and then show how other amps to measure up.
|
You were closer to understanding what I’m talking about when you refereed to propagation delay or group delay. I did ask you about that earlier but you seemed to not respond. So is propagation delay what you are talking about? IOW, if the amp has constant propagation delay over the entire band of frequencies to be amplified, would that satisfy what you are looking for? |
Didn't we find out a long time ago that some amplifiers with relatively
high THD - I.e., more than an order of magnitude higher than those amps
with vanishing low THD - actually sounded considerably better? Quite often, yes. The ear does not seem to care so much about lower ordered harmonics, but cares a lot about higher ordered harmonics (5th and above) as well as IMD and inharmonic distortions such as aliasing. The latter three can be in trace amounts that are hard to measure but are easily heard by the ear (which converts them to tonality) as brightness and harshness. Its not that hard to design an amplifier that is absent the higher ordered harmonics. Such an amp will not use any loop feedback though, as feedback will add harmonic content and IMD of its own regardless of the left/right 'speed' of the amp (as seen on an oscilloscope). See Norman Crowhurst. SETs have been taking advantage of this for some time now (part of the reason they have such an avid following in high end audio). We do as well, but get rid of considerably more distortion merely by using fully differential balanced topology and only one stage of gain. |
^^ That is correct.
The ideal for amp and loudspeaker interface has been the idea of the voltage paradigm or voltage rules, wherein the speaker is 'voltage driven' and the amp behaves as a 'voltage source' which means it can make constant voltage regardless of load.
But SETs and a good number of other amps (including ours) do not work on those rules. Fortunately there are a good number of loudspeaker manufacturers that prefer the sound of tubes, and so make speakers that do not require that the amp be a perfect voltage source- far from it, in some cases.
The idea behind the voltage rules is plug and play flat frequency response with no adjustment needed to the amp or speaker. However no speaker is entirely flat and it turns out that the human ear/brain system has tipping points where tonality created by distortion is favored over actual frequency response errors. So the benefit of having perfectly flat response is reduced in the face of the fact that to get it, many amps have to employ loop negative feedback. Such use is known to add harmonic and IM distortions of its own, so many designers choose to use no feedback at all. This means that you have to be a bit more careful about speaker choice, but the result is more musical and more neutral.
This has been going on in high end audio for decades. So the result is that you simply have to be careful to match the amp and speakers together. Its my opinion that you go with the amp first and the speaker second, because the first problem to solve is whether you prefer tubes or transistors. People that prefer tubes usually don't like it when things are bright, so if you already got the speaker first you may wind up flushing a lot of money down the loo without satisfaction.
|
I usually tell people that the room is 50% of the system. There's not a lot you can do with the very best system if the room sucks. Conversely a very modest system can do great things in a good room. If you had the perfect amplifying method - a pair of Radio Shack book shelf speakers would still give you incredible imaging.
I used to work at the Allied Radio service center in Minnesota. Radio Shack speakers imaged just fine if they were set up right, even with one of their lowly receivers like the STA-76. If you put better gear on them they did even better. One of their speakers used the Lineaum tweeter; Tandy bought out Lineaum just to get the technology for a $250.00 pair of speakers. They got good reviews in TAS on that model FWIW. |
I said it before on this thread and I very much feel this is where the big improvements will be made: One area that is a problem for all amplifier designs is that most are
designed to have specs that look good on paper and are not really
designed to also sound good. Now this is a simple engineering problem
(understanding the rules of human hearing and designing to those
standards rather than the existing set of arbitrary rules); the bigger
problem is tradition- the tradition of how we say what are good
measurements and what are not is at the heart of the issue. How do you
get the industry to move off of standards set in place 60 years ago??
Until
we fix *that* problem, progress will only be had by the outliers who
are willing to buck the tradition and pay the price. And they are out
there. IOW its simple engineering, but if we apply our engineering to making equipment that looks good on paper, but at the same time does not acknowledge how the human ear/brain system perceives sound, then we won't make any progress. We have to overcome the traditions of decades to do that- most of the specs we revere on paper were developed in the 1960s and a lot has been learned about human physiology since then! |
ok, now I’m curious - where can I read about the advances in
understanding human hearing & perception thereof (psych. not
physiol.) since the 1960s?
You gotta pound the heck out of google. Here's a fun one: https://pdfs.semanticscholar.org/2f03/7a2adec2af5953b4f84ca6b2e2b3f30a3bf3.pdfAnd its the tip of the iceberg. Note how the limbic system is involved... more: https://pdfs.semanticscholar.org/2f03/7a2adec2af5953b4f84ca6b2e2b3f30a3bf3.pdfGE did a study back in the 1960s that showed that the brain uses higher ordered harmonics to gauge sound pressure, but the study does not appear to be online. However this particular fact is easy to prove with very basic test equipment. Oddly though, the implications of just this fact are largely ignored by the audio industry. Going even further back, we've known since the 1930s that certain harmonics have certain audible effects, for example the 7th can impart a metallic quality (and in small amounts; see the Radiotron Designer's Handbook) but this fact again is largely ignored. |
No takers? Nobody can tell me what the perfect amp is like? Am I to take that as conceding that Nelson is right? No takers so far because in your prior post Anyone out there banging on that "perfect amp" drum mind explaining
exactly what the perfect amp would sound like or perform like? - you ruled out anyone credible. |
@roger_paul How do you know your circuit works??
I ask simply because in the past you've not indicated that you have any means of quantifying this timing thing you talk about. Just so you know, that's a bit of Red Flag.
|
Thanks, I appreciate your letting me know what I'm doing wrong. I don't
want anyone to think that there is actual snake oil under the hood. Thinking it and Being it are two different things. No-one wants people to think there is snake oil in their stuff. The problem you have is that you refuse to explain your timing theory in any sort of meaningful way that makes sense to anyone with an education. The fact that you have this circuit to do the job for you and at the same time you've not quantified the timing issue is a tale-tale that the circuit either a) does not exist or b) does not work as you think or c) does not work at all and is a fantasy The reason is that the measurements you need are insanely easy to produce- yet in all the years you've made these claims the proof of your claims have been conspicuously absent. If you want to not be constantly challenged on forums like this one you have to overcome those shortcomings with something real- not just endless text about nothing. Post some measurements that show how your circuit works where others fail. Its simple and easy. |
...applies some positive phase second order distortion through the NFB loop balance to warm the sound up and open up the stage.
Translation: without adding second order - the raw circuit is "cool " sounding and has a narrow or restricted sound stage. Yes?
No. |
