John Curl's Blowtorch preamplifier part III

[Jakob2]

@ PMA,

i am still interested in your answers/more specific points, as written in:
John Curl's Blowtorch preamplifier part III

You surely wouldn´t post something like "you posted no facts" without having specific (and correct) complaints, would you?

@mmerrill99,

In a certain sense, I empathize with PMA as what he (& many here) want is a definitive listening test which truly can be trusted to audibly differentiate between devices/files.
I understand the frustration in not having such a listening test & the defensiveness in trying to defend one's chosen test methodology as being somehow the answer.

When people point out the various flaws in such chosen methodology, it's natural to be defensive & to accuse the bringer of bad news with not having anything definitive to offer.

I think, in some way, this is what PMA is leveling at you, Jakob2 - not offering the final solution?

Could be, but usually i prefer to assume that people are interested in a honest discussion and up to now i had no reason to think differently about PMA.

The solutions about doing good experiments are simply not that easy, as there are a lot of variables/confounders to consider, but the underlying principles are the same; test have to be objective,reliable and valid.

So in fact it doesn´t matter what specific protocol/conditions someone is using,as the task is always the same, be able to show that it is objective, reliable and valid.

So far the existing evidence provides corrobation for the hypothesis (long favoured in the audio field) that the ABX protocol provides more difficulties for the participants. Does that mean nobody could use it? Of course not, as i´ve written before, people like Putzeys or Frindle reported quite impressive results from ABX tests with music.

But as an experimenter you have to ensure that the participants are really used to the specific conditions and you have to provide positive and negative controls as well.
These demands are basically the same regardless of the specific protocol used.
For good reasons it is accordingly mentioned in the ITU-Recommendations.

 

[TNT]

The modern Porsche and Ferrari design processes are the exact opposite of "high-end" audio. They are data-driven and rely on extensive FEM simulation of every component. They have objective targets. Ferrari has multiple of their own state-of-the-art datacenters just for their F1 program.

Yes it is really a scientific process steered by real-life KPI like lap-times on the Ring etc. Due to these KPIs, all tweakers and boutique components are since long gone. Our "game" don't have the corresponding undeniable KPIs so wodo may still reign. From this, one realises that it is not the audio components that is in dear need to be progressed but the KPIs - would we want to advance the field.

//

 

[john curl]

I really don't think than most here understand serious audio design at all. It isn't just computer simulations, although it might help. It isn't just putting something together and listening, although that might be part of it, and it cannot not be just a specific set of measurements, like THD that look good in print, either. It is a combination, of good initial concept (topology) coupled with proper parts selection, and finally tuning the combination with both listening tests and measurements. Usually something compromises the design effort. It might be cost, it might be heatsink limitations, it might be size, it might be bull-headedness of one of the principals as to what is important or not important. Usually there is no perfect execution of any design, like an amp, and that is what makes it a challenge to try to do one better.

 

[scottjoplin]

Amps are boring, they all sound the same, well know fact!

 

[scottjoplin]

Accurate reproduction of the recorded sound or of the actual event?

That old meaningless chestnut

 

[billshurv]

important. Usually there is no perfect execution of any design, like an amp, and that is what makes it a challenge to try to do one better.

It's called Engineering. Making something that meets spec (including all the bits they forgot) for minimal ex-factory cost with optimal reliability. No one can ever keep iterating to get the perfect design or the company will go bust. (Cern etc just build the best they can then ask for more cubic dollars to build better).



High end audio just has a 'flooby' spec to meet.



Odd that no one has thrown their hands up in horror about an uber amp with not only tone controls, but also feedforward error correction. Should have the low feedback crowd in palpitations

 

[Zung]

... It isn't just computer simulations, although it might help...

Sez the guy who taught me Microcap

 

[Max Headroom]

I really don't think than most here understand serious audio design at all. It isn't just computer simulations, although it might help. It isn't just putting something together and listening, although that might be part of it, and it cannot not be just a specific set of measurements, like THD that look good in print, either. It is a combination, of good initial concept (topology) coupled with proper parts selection, and finally tuning the combination with both listening tests and measurements. Usually something compromises the design effort. It might be cost, it might be heatsink limitations, it might be size, it might be bull-headedness of one of the principals as to what is important or not important. Usually there is no perfect execution of any design, like an amp, and that is what makes it a challenge to try to do one better.

Hi John.
What do you mean by 'tuning' in this context ?.

Dan.

 

[TNT]

If you by "most" mean the forum members in general - probably so, but the readers in this thread - I think they do actually.

//

I really don't think than most here understand serious audio design at all. It isn't just computer simulations, although it might help. It isn't just putting something together and listening, although that might be part of it, and it cannot not be just a specific set of measurements, like THD that look good in print, either. It is a combination, of good initial concept (topology) coupled with proper parts selection, and finally tuning the combination with both listening tests and measurements. Usually something compromises the design effort. It might be cost, it might be heatsink limitations, it might be size, it might be bull-headedness of one of the principals as to what is important or not important. Usually there is no perfect execution of any design, like an amp, and that is what makes it a challenge to try to do one better.

 

[hhoyt]

The factor also applying to hearing which cannot be factored out:
Why a Placebo is Real Medicine:

So let's talk about circuits.
Howie

 

[scottjoplin]

Could this be why Nelson's amps sound so good? They have a beautiful simplicity of design and function and he's a nice guy who everyone loves and trusts.

YouTube

 

[DF96]

If it sounds good, it's high fidelity.

Of course, we have to allow creative people the freedom to empty words of their meanings. That in turn gives us (who are trying to use words for meaningful communication) the freedom to ignore what the creative types say.

Of course I understand the difference between the creation of music and the reproduction of music.

Then why pretend that you do not?

Is this the 5 minute or the 30 minute argument?

I think the real issue is that your father never gave enough of his time to be with you as a child. Time that you needed, and without realising it, projected it onto me. You maybe in denial of this, but this is what upset me. How dare you? Would you say this to my face? I very much doubt it. Believe me when I say that amongst my peers, I am also one of the good guys and would do anything to help other photographers, including your father if I was ever lucky enough to have met him.

So now you are a world expert on cod psychology too?

So we must say that HI FI REPRODUCTION is something different than guitar amp amplification, so let's keep the two separate.

I am glad that from time to time JC says something that I can fully agree with.

 

[dadod]

I really don't think than most here understand serious audio design at all. It isn't just computer simulations, although it might help. It isn't just putting something together and listening, although that might be part of it, and it cannot not be just a specific set of measurements, like THD that look good in print, either. It is a combination, of good initial concept (topology) coupled with proper parts selection, and finally tuning the combination with both listening tests and measurements. Usually something compromises the design effort. It might be cost, it might be heatsink limitations, it might be size, it might be bull-headedness of one of the principals as to what is important or not important. Usually there is no perfect execution of any design, like an amp, and that is what makes it a challenge to try to do one better.

John, I think that you mystify audio amp design to much, just to justify an amp (power or pre) cost in tens thousands of dollars. It should contains some “mystic components” and some “tuning” and that to be secret of chosen ones.
A good amp (excellent one too) is just good engineering with good proven components. Design now days start with simulator and that is most useful and important tools which gives the result, if one knows electronic theory, very close to final amp. Testing and measuring are next steps of course, the PCB layout is equally if not even more than the simulation.
Listening is at the end, very important, but that one is contaminated with audio magazines and golden ears.
Best wishes, Damir

 

[scott wurcer]

if one knows electronic theory

That does help, I forgot one yesterday The Art of Electronics, by Paul Horowitz and Winfield Hill. If you can find a reference here to "voicing" ,etc. please point it out.

 

[billshurv]

Quite. And I am sure you are considered well versed in the art!

Thinking about it, the problem is that 'art' appears to have shifted in meaning from back when the only degree you could get was an arts one. D.Self for example is MA!

 

[PMA]

PMA, i beg your pardon, but you have to be more specific. For what did i nor present any "fact" ?

Wrt the preamplifier test i wrote a description (inluding the measured numbers of both variants) of the test approach overall several times and reported the results.

So, you must mean something different, but i again i don´t really understand what missing "test description" or missing "test files" you are referring to.

I apologize if I missed something in this extremely long thread. It might have happened easily, as I am often occupied with my job and thus I do not read everything in the forum.
Is there a test protocol or a written description of your preamplifier test, if yes, I would be interested to see it. Or an external link to your website or to the test description and its results? I assume it was a serious, professional test.

 

[PMA]

The modern Porsche and Ferrari design processes are the exact opposite of "high-end" audio. They are data-driven and rely on extensive FEM simulation of every component. They have objective targets. Ferrari has multiple of their own state-of-the-art datacenters just for their F1 program.

That was exactly my point and that was why I have shown that vintage 1936 Mercedes-Benz 500K car, as a parallel to audio high end. Nice, glossy, shiny, heavy, big fuel consumption. However, nice to look at and I like it . Below one near Prague.

 

[scottjoplin]

I'd hope it came with a chauffeur

 

[Max Headroom]

.
Listening is at the end, very important, but that one is contaminated with audio magazines and golden ears.
Best wishes, Damir

Hi Damir.
Agreed, simulation allows 'virtual' optimisation (according to models) that can get incredibly close to the target depending on the application.
In the case of audio this still applies, but the models are ultimately too course to predict the very fine effects of parasitic characteristics in components active and passive, and the consequential subjective effects presented to the listener.
One can read infinite reviews but the final choice is in the ear of the listener.....some gear sounds good or 'right', and plenty of other stufff does not, hence the eternal audiophile 'quest'.
It seems that you have invented and proven many designs (we are interested to to hear the final results of your RNM collaboration I am sure) and these designs of yours have have evolved....why ?.
'Why ?,' as in what have you found wanting in your incremental designs that has inspired you to chase 'advancement'....in what areas...max power, THD, THD Spectrum, IMD, IMD Spectrum, N, N Spectrum, BW, DF, DF/Spectrum, etc, etc.
Have you practiced subjective 'voicing' by for example swapping parts that SIM (models) says should be indistinguishable or moving NFB take off point and/or NFB series/shunt components, or different pcb substrate etc etc.......?.
I applaud your efforts, and RNM's initial appraisal of your mutual design is by all accounts is first class, I am interested to know your design process ?.


Dan.

 

[john curl]

As I said, guys.
Now when it comes to 'tuning', it is just a word I chose 'on the fly' to describe adjusting and looking at distortion 'quality' and getting it optimized. Just today I have a fairly large power amp waiting in my lab to be tested. It already works, but I had to change the value of the output emitter resistors from what the factory put in. I hope that I calculated the right value.