John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
see my clarification. I am afraid to me, if a power amplifier has 'character' something is wrong with it or the listener has a bias. I am however willing to be proved wrong.

As you point out competent amplifiers have a tiny fraction of the distortion of speakers. I am confused as to how 2 amplifiers of a 100ppm or less can have any sort of character. Toole and Geddes and Linkwitz I believe are of similar views.
Oh, please, apologize, I'm confused, I did not knew you were totally deaf.
 
I have noticed the same, and even for MMs where the generator impedance is low. For example Grado Platinum into as low as 2K - and quite a bit better than 47K.

.

Lucky has already posted this is nonsense, the electro-mechanical reciprocity of a MM (Grado is MI BTW) is infinitesimal. For the Grado especially there is neither a coil or magnet attached to the cantilever just the flux coupling disk. It is observable under a microscope you can't move the tip with current in the coil.
 
diyAudio Member RIP
Joined 2005
Global comment about cartridge loading, loudspeakers, amplifiers: remember that frequency response changes (some call them linear distortions) are the easiest things to hear.

Nonlinear distortion is more difficult to hear. For harmonic distortion the coincidence of distortion products with the overtone series makes monotone-input distortion of low order (second, third, fourth) essentially indistinguishable from changes in frequency response. But the nonlinearities that effect the production of these have different "curves of growth", typically. So for ~slowly-varying amplitude one can possibly perceive that the system is doing something to the program material. The "problem" is real instruments and voices also change timbre with level, so this can be ambiguous. And we're usually listening to music, not seeing if we can discern distortion.

Now look at two-tone intermodulation. Then we are in a new territory, particularly when dealing with material based on equal temperament, where only octaves are based on powers of two and hence coincident with specific sum and difference frequencies. We (western ears) have learned not to be bothered by this very much, although explanations for why live string ensembles sound good, due to adjustments made by players in long-held notes to come closer to just intonation, seem not to hold up most of the time.

One could go on.
 
diyAudio Member RIP
Joined 2005
And they are technical arguments i can oppose to BJT for power stages, like secondary breakdown or need of temp compensations.

MOSFETS operated in "linear" modes do suffer from localized heating and potential thermal runaway. It is only in hard conduction that the positive tempco of the silicon enforces some current sharing both on chip and among multiple devices.
 
Not exactly, SY. I was wondering who could, today, print the distortion specifications of an amp with the same numbers than the ones of a speaker...
With other words, if you ask-me what are the threshold distortion numbers for a transparent reproduction, i will answer: "Excuse-me i have an urgent appointment..." ;-)
On the thread "Sound quality VS measurements", the are talking cars and motors...

Christophe, these are subjet people eher definitely do not want to talk about. I hve asked several times what threshold of THD and IM should we take as borderline between shall we say acceptable and not acceptable and nobody took it up with even a presonal view (which is what I asked for). If nothing else, Otala & Lohstroh at leasted hinted at it by saying that 0.1% THD was acceptable, while in case of IM our hearing was more sensitive. Agree or not, but they did say it. There was ne fellow who claimed he could hear THD down to levels of 0.01%, which I strongly doubt.

Look at the CFA vs VFA state of affairs in the world today. All of sudden, there are myriads of people, almost all laymen, who "know" CFA yields better results. How? Well, many will say thet if mass produced items are adopting CFA, then it must b better, why else would they do it? How about marketing? How about they ran out of "exotic" standard topologies and it was time they made a new "discovery", which fits nicely into their ad campaign. I have heard only a fw amps using the CFA and in all honesty, they sounded exactly like well put together VFA amps. Possibly their lab data is better, but if I don't hear it, that means nothing to me.
 
I hve asked several times what threshold of THD and IM should we take as borderline between shall we say acceptable and not acceptable and nobody took it up with even a presonal view (which is what I asked for).

It was pointed out why the question was meaningless, since you're basing things on marketing specs rather than engineering. You chose not to address that. If you want to actually understand the issue, read the papers on the Geddes Metric.
 
diyAudio Member RIP
Joined 2005
damping beyond a short

They make it up and you have to prove otherwise. For one thing here the concept of impedance matching is silly, max damping is a short.
Agree about impedance matching. Actually one could go further with negative resistance loading, but as mentioned there is a problem with realizations in terms of nonlinear distortions, as the negative R is concocted using positive feedback, and not to mention potential instabilities. I wouldn't advocate it, but it came to mind after reading Bob Cordell's partial realization of the RIAA characteristic and applying it to low inductance MC cartridges.
 
A joke or are you being rude? No smiley so have to assume latter. My question was serious. Surely if an amplifier has a noticable 'character' it is an effects box and not 'high fidelity'?

"Noticeable character", as I understand the term, is a specific character the amp will produce when using several highly linar transducers, i.e. it is repeatable and even foreseeable.

In such a case, I agree with you.

However, a more intersting case is when two amps consistently deliver the same or similar feeling that the one which sounds better actually measures worse than the other. Notwithstaning the easily measurebale effect of well knonw phenomena, such as THD, IN, TID, Slew Rate, damping factor nd so forth, any one of which, or a combination of some could be the reason, I believe one should look at how the better spec was achieved.

In essence, most of use about the same or similar amounts of NFB, the key difference being whether you go for little local and much global NFB, or more local and less global NFB. I sincerely believe, on basis of my own experience, that amps using more locl and less global NFB tend (but with no hard guarantees) to sound more relaxed and easy going. However, that does not mean that another amp, using almost exclusivey global NFB cannot sound good. In my own case, I've made it clear that I believe Karan Acoustics' KA-i180 integrated amp is I believe an excllent amp however you look at it, yet it uses almost exclusively global NFB (it has an op amp input stage). By comparison, my H/K PA2400 power amp, with almost identical specs which uses merely 12 dB of global NFB, but predictably quite a lot more local NFB, sound almost the same, and if anything, has a little better 3D sound "picture", especially in terms of front-back rendition.

Just my 2 cents' worth.
 
Darn, I keep hearing sonic signatures in even my own amp designs. But that should be impossible! '-) I guess I should change the face plate of the amp. to improve its sonic character. Of course, I already make amps that are already direct coupled, all push pull, very fast, and have low harmonic distortion, so it must be my imagination. '-)
 
diyAudio Member RIP
Joined 2005
How? Well, many will say thet if mass produced items are adopting CFA, then it must b better, why else would they do it?

Although it is an asymmetrical-input design, the power amp section of the ancient NAD 3020 is a CFA. It looks pretty crude on paper but works very well. The intrinsic slew rates are also somewhat polarity-sensitive but fast enough (therefore symmetrical enough for bandlimited signals) to satisfy Keith Johnson, who said upon acquiring one from a friend while he waited for time to repair his own amp, "I didn't have to do anything to it to make it sound right".

And Erickson did this with 2N3055 and 2N2955 as the output devices.
 
And on amplifiers: two things. (1) How do they overload? (2) Are they stable with various loads and with temperature, both ambient, and self-heated from the signal history?


It's been my experience, at least, that more often than not "audiophile" listening sessions end up straining most amplifiers out of their comfort zones. I remember a pair of poor Quicksilver mono-blocks playing huge bass towers at 11 until they smoked.
 
MOSFETS operated in "linear" modes do suffer from localized heating and potential thermal runaway. It is only in hard conduction that the positive tempco of the silicon enforces some current sharing both on chip and among multiple devices.
My remark was based on practical experience with huge PA systems. With enough safety margin, lateral Mosfets based amps are a lot, a lot safer. They are flat on tempco around 150mA... very good for class A operation range at normal levels, and... they like to be run hot.
In case of paralleling devices, they tend to equalize themselves, too, because the hottests will be in their negative tempo area.
 
Last edited:
Well, I keep working at 'improving' my amps and preamps. I do hope there will be a difference once I get the latest efforts out. It certainly won't be because of 'measurable' distortion. My best stuff, already does 'distortion' pretty well. Of course, my competition is not SY or SE, but the likes of Nelson Pass and Charles Hansen. I will look at their efforts to keep me on my toes.
Personally, I'm kind of lazy. I don't really want to make a newer design of power amp, but I do want to match or beat the competition.
I am now an old man, facing a finite future, and frankly, 'I don't give a damn' about most things anymore, but I am not so foolish to think that my amps are virtually perfect. I can do better, even though I'm not sure what will make a real difference. I will just 'tilt at windmills', including improved power supply regulation and still lower harmonic distortion, and hope for the best.
Richard Heyser blamed the use of negative feedback in amps and preamps, for why they can 'measure' better than speakers, yet still make a sonic difference. I am pretty sure he was right, but I will keep on using negative feedback when I need good specs., which is most of the time, as always. Oh, the compromises we have to make in order to make commercial products. '-)
 
Status
Not open for further replies.