The problem with "know-how".

[nigel pearson]

All this is very good and it is close for me. But truth search always leads to revising of customary concepts. With a support on the already known.

Good has come from this . The class A2 amplifier might work with my speakers . I will soon be 60 years of age , I might have 20 creative years left ? If I don't do it now I never will .

I was saying yesterday the A2 amp would not drive my speakers . I was forgetting the panels crossover at 600 Hz using 1st order filters . That could be " The Priest " as the bass and A2 as the other . It even removes horrible problems . I could even have a special transform that would be better and cheaper ( - 3dB 200 Hz ? ) .

The Priest looks like a Goldmund amp circuit ( chance , not intention ) . The Priest did battle with the amp below .

SoundStage! Equipment Review - Art Audio Diavolo Amplifier (2/1999)

They are very similar sonically . Listening to a Decca recording a train enters a station . The steam escapes , doors slam . And them the most glorious sound you every heard . The whistle . The Priest equals it in every way except the sense of space is slightly reduced . Then the whistle . No longer just tonality . You are riding a rocket and the roof of the building is heard . Pre amp was Art Audio also . I have to say the Art stopped me having prejudices against valves . Tight and not obviously distorted . My amp has the the virtues so nothing to say extra about depth and space . The Priest has virtually zero distrotion . If I can ever persuade my friend to send the Audio Precision files of the Priest I will show them , she intends to sell the amp as L'Art du Son so I guess she never will . They are about 20 db better than I had guessed . The Priest can give 1000 watts into 1 ohm . No idea what she is doing as she never says . She likes my very correct current mirror she says ( 4 transistor , no big deal ) .

HH 1200 was my inspiration for the Priest . I don't use MPSA 42/92 .

One thing to make clear . Mostly I don't like valves . A bit like friends who let me down . One has to know the limits .

 

[Sergey A]

I have an apartment, the wife and the son, and also a dog . All other things too.
My data still need expanded operation with them, and can give only advantage in search time and narrowing of a zone of search.
And the customer at all isn't interested in process . For him only the good sound is important.
And I am not the hero of Vanity Fair.
Only a generator of new ideas.
In Spain was not, but I think there is quite good. No worse than Italy, Greece.
If you like process and result ,what can be better? The main thing that the desire didn't overcome mind.

 

[Pano]

Rightly or wrongly I believed what I read by Jean Hirage in translation of testing of listener preference . This was about 1980 . He said approximately that a high level of distortion which has harmonics that show something like an exponential curve getting lower as the harmonics increase may be thought to be zero distortion .

Hi Nigel. You probably read Harmonie et distorsion by Mr. Hiraga from November 1981. There is a lot in that article, but the take way is about harmonic masking. Mr. Hiraga cites work going back to the 1920s (and more recent) that shows certain harmonic structures are indistinguishable - by ear - from pure tones. Thus a signal can have relatively high harmonic distortion without us being able to hear it. That is, with a certain fall off of the harmonics.*

The trouble comes in maintaining the masking a different levels and different loads. Not all circuits do this well. Of course one can rightly claim that very low levels of harmonic distortion are better than higher, masked ones - and that's true to some extent. But the ear seems rather sensitive to harmonic imbalance even at very low levels so you have to be careful with the trade-offs. An example I find over and over is the analog output section of a DAC. It may have low levels of THD, but the higher order harmonics (>5th) dominate. That does not sound good. The suppression of the 4th harmonic also sounds strange, and that's common in a lot of circuits I've measured.

I worked on a few projects with Mr. Hiraga back in the 80s and early 90s and he continued to design by harmonics, trying to keep them inaudible across the spectrum and power scale. To my ears, his designs and builds (tube or transistor) are the most transparent I've ever heard.

*I can post those ratios if you are interested.

 

[Sergey A]

Level of intrinsic noises is very important, you are right. As well as pure supply. Irrespective of circuit engineering.

 

[nigel pearson]

I hope to meet Jean one day . I speak French so can engage on details ( technical French better than everyday French ) . The lady I refer to is Martina Schoener . She is a friend of his . There was a fire or something and she hasn't seen him since then . I have begged to meet him . The Coral 777EX I bought having read the 1981 stuff . Best cheap MC ever ?

By accident I got involved with the medical side of hearing . Then my daughter got asked to do tests by the same people 20 year later . It has given me a different view . It is almost like Quantum mechanics where we have to describe seeming madness as a better model of reality . When we can show is to be real as they did we go forward . Personally I saw no debate on that . Do what works best for every situation . I hope I show both zero distortion and Hiraga amplifiers sound like real music . It is difficult to make Hiraga amp . With valves it is the best option ( I failed a bit ) . A low distortion valve amp seems the hardest . Croft OTL is getting somewhere I think ?

Please post the measurements .

 

[nigel pearson]

Level of intrinsic noises is very important, you are right. As well as pure supply. Irrespective of circuit engineering.

I will buy you a big drink for that . An amplifier is a musically controlled power supply . Preferably a constant voltage source in the world we live in .

1000 W 1 ohm is OK ?

Vanity fair is a great book “Never lose a chance of saying a kind word.” . The Santander area of Spain is if you like Europe's Sweet Spot . Any plant will grow there it is said . That is from Alpine to Tropical . The English love it as the can have dream gardens . The people are nice also . Hemingway writes about the area approximately . He learned Spanish and Portuguese by begging for food . In fact what he learned was dialect which was neither . The old man of the sea is a great book .

 

[gootee]

I will buy you a big drink for that . An amplifier is a musically controlled power supply . Preferably a constant voltage source in the world we live in .

1000 W 1 ohm is OK ?

That is absolutely the best way to think about it. The power supply current is the music output and almost all of it comes from reservoir and decoupling capacitors.

So those capacitances and the transformer must be sufficient, in order for the result to then depend only on the "controller". Decoupling caps having very low inductance paths through which to service the needs of the active devices is necessary for accurate transients/timing and also the upper end of the feebback processing. Accurate transient amplitude and timing up to high frequency also means phase angles of Fourier components stay accurate vs frequency, so complexly-shaped signals and sharp edges are accurately reproduced without ringing, and precise timing cues are retained, for more-coherent imaging.

If supply voltage staying constant is important, and especially at higher frequencies, then the decoupling caps' configurations are even more important, in terms of both large-enough values and small-enough distances from points of load (inductances). That implies desirability of very small caps very close to active devices, then slightly larger caps slightly farther away, and so on. It is easy to calculate both the minimum capacitance and the maximum distance (inductance), for each. Ideally, of course, they would all have zero loop length conductors to point of load and load ground.

 

[Sergey A]

Unfortunately to the 90th, the knowledge of speech English or French, was to anything, behind our iron curtain. I studied English very not bad, but absence of application always affects. French was interesting, with the same success.
Why to you 1 kW? All manage with music power less than 1 Watt, usually. Some Watt can. If it simply possibility of the amplifier, at will. I don't think that such inventory is mandatory.
Bonus ,Click on the picture
Ìóëüòôèëüì "Ñòàðèê è ìîðå" (1999 ã.)

 

[gootee]

With less than 150 watts into 4 Ohms rating available, with my Magnepan MG-3.6/R speakers, the amplifier could start clipping just before the sound is loud enough for me, which could blow the fuses of the ribbon tweeters. But with my Adcom 565 (or is it 585?), which is also rated to do 1000w into 1 Ohm, that doesn't happen.

It almost seems paradoxical at first, but a too-small amp is much more likely to blow the ribbon tweeter fuses.

So that is one possible reason for needing higher rated power.

Another good reason might be in order to have more voltage headroom.

 

[nigel pearson]

Unfortunately to the 90th, the knowledge of speech English or French, was to anything, behind our iron curtain. I studied English very not bad, but absence of application always affects. French was interesting, with the same success.
Why to you 1 kW? All manage with music power less than 1 Watt, usually. Some Watt can. If it simply possibility of the amplifier, at will. I don't think that such inventory is mandatory.
Bonus ,Click on the picture
Ìóëüòôèëüì "Ñòàðèê è ìîðå" (1999 ã.)

Because I speak French badly I understand the frustration of speaking to another human . I have difficulty in English . An Englishman finds it difficult to be understood in the USA . We understand them better as we always had US TV .

I built the Priest as I wanted to know if a public address amplifier design could be made into hi fi . I am sorry to say Hitachi did most of the work . I have no idea if my friend built the full version ? Hitachi would never give graphs from 0 to 2 watts . I had doubts . It turned out to be OK . Goldmund must have seen the same . Originally the amp was to drive a motor .

 

[nigel pearson]

With less than 150 watts into 4 Ohms rating available, with my Magnepan MG-3.6/R speakers, the amplifier could start clipping just before the sound is loud enough for me, which could blow the fuses of the ribbon tweeters. But with my Adcom 565 (or is it 585?), which is also rated to do 1000w into 1 Ohm, that doesn't happen.

It almost seems paradoxical at first, but a too-small amp is much more likely to blow the ribbon tweeter fuses.

So that is one possible reason for needing higher rated power.

Another good reason might be in order to have more voltage headroom.

Exactly .

 

[nigel pearson]

The HH is something like where I started . If an engine this is the side valve Ford V8 . Thing is that Ford V8 is still a thrilling engine . It drinks fuel . With imagination it is the Renault F1 engine with air desmodromics .

Car talk is not allowed so forgive .

From what I know the Ford could drink as much fuel as the F1 engine if pushed . We all know why . The F1 is about 5 MPG UK if I am correct at speeds one can do on a race track .

This is the translation from Russian of the link Sergey gave us . Google Translate !

The novel by E. Hemingway
Award "Oscar" for best animated film of 2000.
The story of friendship between a small country boy and an old fisherman. Santiago, strong and proud man, he can not come to terms with the inexorable passage of time-consuming physical strength.
For many weeks, he returned to the sea without a catch. Finally, he makes a decision - to leave their prey far away in the sea and not to return without a catch. Just as an old fisherman can regain the confidence and self-esteem.
Early in the morning he says goodbye to his young friend, a native shore who sees perhaps the last time, and dissolves into the dark sea water.
Still, good luck comes to the fisherman. In his tackle catches a giant fish.
Two days at sea continued their duel, but the fish did not give up and pulls Santiago farther and farther out to sea. But the old fisherman had convinced himself that the tenacity and strength of spirit - that's what will bring him victory ...

I love Cuba .

 

[nigel pearson]

Unfortunately I can't publish my tube amp as it belongs to the gentleman I built it for . This one although different was the one that inspired me . For safety I was more conservative . I got slightly better distortion . I don't use plate to plate ( drain ) feedback . I would imagine this amp to be highly satisfying . I would except it to be slightly euphonic as it crosses the 1% maximum I beleive to be the limit for the illusion . Speakers usually have more so where is the problem ? His graph says better than my predictions ! Wish there was more info , am I blind is it 1 watt ? I doubt 5 watts ?

Audio Asylum Thread Printer

 

[Sergey A]

Unfortunately diagram is available for the engine may simply APEX?
If Google is better, I'll use it.
Translated from Russian and back, looks good.

 

[Sergey A]

Audio Asylum Thread Printer[/url]

feedback through the 100k and R ds is not common?
then, can be considered mixed.
http://olc67.narod.ru/pockemon.doc
A similar project, the sound of the amateur

 

[nigel pearson]

It is a little bit like the Alex Kitic RH amplifier mk1 . I made the decision to replace the FET with a big pentode . When I built the Kitic amplifier it was more distorted than people say .It also had poor sensitivity . If you look back to Koster the harmonics are almost the Hiraga ideal . The exponential distortion harmonics curve ( Hiraga hypothesized circa 1981 , he said either exponential or straight , his instinct exponential )

RH 807 - Tube Audio ...... RH DESIGN

One thing I like about the Kitic amp is it tries to make a super triode . I spent months looking at how it worked . The transformer being outside a normal feedback loop is nice ( fast ) . I think it is no less a fake triode than connecting g2 to anode . Both are fake triodes . I will say before anyone does I understand it is not triode .

The one debate no one seems to want to get involved in is ultra linear feedback as Quad called it the same as ultra linear feedback of Hafler (43 % in some designs to g2 ) ? Hafler and Quad seem to have agreed co-invention . The Quad is to the cathode of a KT 66 via an extra winding . If true a triode could have Quad UL . I can not see the Quad as true UL . Doubtless it is great . Would it be that Quad UL helps a pentode more than a triode ? I see no reason why it should .

I suppose if calling it distributed load there is agreement . UL must be Hafler alone . DL for triodes anyone ? Less distortion , less gain? I could possible sell it to my friend as no loop feedback ? Nick of Trilogy amplifiers insisted Quad DL had a 2 to 1 distortion to gain advantage over all other systems . I laughed at the time, maybe Nick was right ? Trilogy were very fine amplifiers . I said to Nick I bet you sell 20 a year in the UK . He said in fact 6 ! No wonder he gave up for a while . A chance meeting as I was at a concert hall next the Festival hall . Nick was friend of my freind Amir . You don't expect to discuss UL by pure chance . It was exactly where Mr Handle and Mr Hendrix lived in Soho .

If you can do the reading .

http://www.oestex.com/tubes/ulo.html

 

[Sergey A]

At Quad feedback directly to the output stage. Here she captures the input. Fear it reduces the rate of rise in the output stage. Fans will be, and for that and this.
Now I'm just general questions for the correct sound,and cause distortion
http://www.oestex.com/tubes/ulo.html
In Russian it looks, as here
http://musicangel.ru/mess010.htm
Your authors carefully considered everything. Everything is good as far as possible.There is a lot of methods

 

[Pano]

Nigel. I just looked at the 1981 article and basically Hiraga says that harmonics are masked if:

• H2 -22dB
• H3 -26dB
• H4 -30dB
• H5 -37dB

From work in the 1960s
and:

• H2 -15dB
• H3 -18dB
• H4 -21dB
• H5 -26dB

from Wegel & Lane in 1924
Higher harmonics are not specified.

I don't know why the difference, but it may be related to SPL and/or frequency. The idea being that a regular diminution of the harmonics is masked by the ear, thus inaudible. Other distributions of the harmonics will result in unmasking of some and thus audible distortion. I'm trying to find the original papers to learn more.

 

[Sergey A]

And what can be said in terms of digital filtering, if that decimation of the odd harmonics of the push-pull amplifier? Maybe there is a lowering of superimposing of ranges of harmonicas?

 

[Kazinator]

I am not aware of any technique to 'optimize them for gain of 1', whatever that means. I do know that some opamps are decompensated for gains above unity, to increase the available loop gain and hence the available feedback at higher gains.

jan

Nigel made an interesting observation there and I will take it as the truth and try to answer how a listener/tester could come to making such a report. How is it that the op-amp could be great as a buffer, then suck at intermediate gains from 2 to 10 and then make a comeback at 50?

The main problem is probably that to have anything close to a level-matched comparison (absolutely necessary for subjective tests to have a shred of validity), when you vary the gain of that op-amp from 1 to 50 and beyond, you have to inversely vary some other gain or attenuation elsewhere in the signal chain.

The gain structure of the signal path changes. Whatever you're hearing cannot be simply attributed to the change of gain in the op-amp.

A possible explanation of the situation like this. At a moderate gain of that op amp, there is a balanced gain structure in the signal path because the other compensating gains are also moderate. The signal path sounds its best: it has low distortion. Low distortion, of course, sounds bad: boring, flat sterile.

The good sounding situations are deviations from this local minimum. When the op-amp is at unity, something else is cranked up and distorts, making a nice sound.

And likewise when the op-amp is cranked to a gain of 50, it distorts more, making a nice sound.