Sound Quality Vs. Measurements

[nigel pearson]

Dvv you said about Italian engineering . This example is a hydro power station ! 1906 and I think most as designed and still in use . The Hoover dam I have visited which I thought to be pristine in every detail ( it looks new ) . This is as impressive in the generator hall .

The hydroelectric power station Crespi Taccani - Trezzo d'Adda (IT) - Vale's Monkeys Photography - Valerio Li Vigni

[gootee]

Originally Posted by a.wayne

Gootee ,

Dont we have to divide by 1.4 for rms output or not ...?

I didn't see it specified and am not familiar with the iec test protocols. So I was doing peaks.

But the rail only dropped 20 volts for that so there's still plenty more where that came from.

[dvv]

Originally Posted by a.wayne

Yes , Magic ! he is getting more power out than he is putting in ....

He gets out more than anybody except you would ever need.

But seriously, I asked him about that. He said he was down to spitting blood by the time he got the transformers to perform as he wanted them. At those levels, many rules in our little audio world either don't apply any more, or are damn hard to come by.

Consequently, prices go astronomical. At least for this mere mortal.

[dvv]

Originally Posted by gootee

Sorry to interrupt the very interesting ongoing conversation to go back to this.

It's actually easy to get that power level for such a short time, and almost couldn't be otherwise. And the 2 Ohms just makes it easier.

Starting at 0 watts, the no-load voltage would apply. So initially you'd get 52v/2 ohms = 26 amps. And 26 squared x 2 = 676 Watts.

And if the rail voltage was allowed to drop by 20 volts over 20 ms, then using

I = C dv/dt = .0164 • 20/.02 = 16.4 amps gives 16.4 • 16.4 • 2 = 538 Watts.

Parasitics and other things were ignored in those calcs but they're close-enough to see how it's possible.

Completely agreed, Tom, but that was not the aim of my question.

By the time it's delivering exceptionally high power levels, I was wondering why was its power supply not sagging more? I mean, 8,200 uF is anything but exciting, in fact, it's kinda stingy for its price class in those days.

Off hand, I'd expect its power supply to sag more, a lot more.

But we can turn this around. This shows that caps are NOT as important as we seem to think, not unimportant, of course, but not the determining factor. Also, I'd like to mention that the caps it has are from a totally unknown to me South Korean manufacturer, no usual "High Class" fare from Elna, Nichicon, Sanyo, etc.

It appears that two factors seem to be much more important than the electrolytic caps: the engineering, which was not giving up nor doing anything untoward, absolutely stable, and the power transformer. This is quite in line with what DF96 was saying, it's not the size alone, it's the RIGHT size.

I would venture a guess that it's also the topology, especially (but not solely) regarding the output stage. Two pairs of 130W power devices is hardly impressive, but it does make me wonder about all those amps sporting 3 pairs of 180W devices which can't get anywhere near this, despite greater nominal power into 8 Ohms.

Brains again triumph over brawn?

[gootee]

Originally Posted by dvv

Completely agreed, Tom, but that was not the aim of my question.

By the time it's delivering exceptionally high power levels, I was wondering why was its power supply not sagging more? I mean, 8,200 uF is anything but exciting, in fact, it's kinda stingy for its price class in those days.

Off hand, I'd expect its power supply to sag more, a lot more.

But we can turn this around. This shows that caps are NOT as important as we seem to think, not unimportant, of course, but not the determining factor. Also, I'd like to mention that the caps it has are from a totally unknown to me South Korean manufacturer, no usual "High Class" fare from Elna, Nichicon, Sanyo, etc.

It appears that two factors seem to be much more important than the electrolytic caps: the engineering, which was not giving up nor doing anything untoward, absolutely stable, and the power transformer. This is quite in line with what DF96 was saying, it's not the size alone, it's the RIGHT size.

I would venture a guess that it's also the topology, especially (but not solely) regarding the output stage. Two pairs of 130W power devices is hardly impressive, but it does make me wonder about all those amps sporting 3 pairs of 180W devices which can't get anywhere near this, despite greater nominal power into 8 Ohms.

Brains again triumph over brawn?

I'd say it was almost all just the 16400 uF of caps, for that 20 ms test. And it sagged a heck of a lot.

[fas42]

Originally Posted by dvv

My most delightful surprise overall though was that it sounds exactly like all other Karan amps, power and integrated, do. That's the one thing I take my hat off to Milan, all 6'6'' and 300 lbs of him, he has managed to keep his sound more constant and consistent than most out there in the High End. My KA-i180 integrated, delivering 180/250 WRMS into 8/4 Ohms, sounds exactly the same as that beast, only that beast has more of it to flaunt.

And of course that's exactly how it should be. A 2,000W amp should sound exactly like a 2W amp, if both are running into high efficiency speakers, with peak level output of 1W. Assuming the manufacturers can be believed, and neither produce audible levels of distortion ...

And yes, all his amps have that elusive exquisite feeling of limitless power, as if you couldn't overload it no matter what you do. It's an illusion, of course, but one I would gladly go for. Then again, given that the entry level with Milan is 22 dBW or 180 WRMS, perhaps I shouldn't be surprised. You DO expect 180 Watts to sound easy

Again, this is all about power supplies. A Krell with a transplant from a Japanese receiver of nominally equivalent power, of the latter's power supply would sound like crap ...

I've been friends with Milan for some 14 or 15 years, and I've never stopped marvelling when with him. He's a giant, literally, yet I don't get to meet such gentle people as he is. I'm passionate about music, but he's way ahead of me. You should see his lab playground, it's a sight to behold, lots of projects later rejected for some reason, yet fully functional. His home loudspeakers are based on JBL professional components and weigh in at around 110 kilos (250 lbs) each. Efficiency of some 96 or 97 dB/2.83V/1m, driven by his behemoth mono blocs, capable of 33 dBW, theoretically capable of music peaks of 129 dB. I mean, like, WOW!

I visited the home of Peter Stein, the man behind ME, years ago; he had the most extreme room treatment I have ever seen. Quite an intense individual, extremely focused; he's very motivated to keep his old products running, I don't think longevity would ever be a problem with them ...

If you're into (potential!) volume you should check out www.Basspig.com The Bass Pig's Lair. Just a nickname, this is the real thing, fully pro setup easily able to maintain 135dB, and he regularly listens at these levels! Not a silly car audio type loudness setup, but for playing orchestral, and other proper recordings ...

Frank

[dvv]

Originally Posted by gootee

I'd say it was almost all just the 16400 uF of caps, for that 20 ms test. And it sagged a heck of a lot.

You are most probably right, but then how come other amps, with more capacitance relative to their nominal power, full of output devices, cannot match this situation on their level?

You know, nominally say 150 WRMS, so proprtionally it should deliver at least say 650 WRMS, or so?

Because I think you are right, and since its time for a full refresh is coming up anyway (after 15 years, I'd say it's time), I am going to stick bigger caps inside. There isn't much free space, a bit crammed there, but 10,000 uF will fit, and I'm looking at bigger values if possible. An extra 1,200 uF (or +22%) per cap isn't much, but hey, we take what we can.

[dvv]

Originally Posted by fas42

And of course that's exactly how it should be. A 2,000W amp should sound exactly like a 2W amp, if both are running into high efficiency speakers, with peak level output of 1W. Assuming the manufacturers can be believed, and neither produce audible levels of distortion ...

Ah, but comparisons were made with several speakers, some of which were less efficient at say 88 dB/1W/1m, and at several volume levels, the highest of which was (to me) ear splitting.

Nada, no way, they always came out practically the same - actually, 100% the same to me, but I have to allow for some possible differences which I may have missed.

I suppose under more extreme conditions some differences would have shown up eventually, but this was a living room test, and there's just so much volume you can take before it becomes unpleasantly loud.

Again, this is all about power supplies. A Krell with a transplant from a Japanese receiver of nominally equivalent power, of the latter's power supply would sound like crap ...

Agreed. My 180 Watt integrated amp uses two 10,000 uF/63V Fisher & Tausche German made caps per suplly line, i.e. 40,000 uF per channel, 80,000 uF for the whole amp. He gets away with 63V caps simply because all his amps run in fully balanced mode. Each half uses 2 Sanken RET 180W (I think?) in a SEPP configuration.

His biggies use as many as 40 Sanken RETs per channel. He's definitely a Sanken guy, something I like to tease him about, as I am definitely a Toshiba man (if Japanese, but Motorola/ON Semi in general).

I visited the home of Peter Stein, the man behind ME, years ago; he had the most extreme room treatment I have ever seen. Quite an intense individual, extremely focused; he's very motivated to keep his old products running, I don't think longevity would ever be a problem with them ...

If you're into (potential!) volume you should check out www.Basspig.com The Bass Pig's Lair. Just a nickname, this is the real thing, fully pro setup easily able to maintain 135dB, and he regularly listens at these levels! Not a silly car audio type loudness setup, but for playing orchestral, and other proper recordings ...

Frank

Wow, that's some serious SPL.

I'm not that lucky. Since it provides 22 dBW, and since my speakers deliver 92 dB at 1W, theoretically I am capable of achieving (92+22) 114 dB SPL/1m. In practice, that's about 2 dB less as speakers are anything but linear power users, but still way more than most.

Anyway, that's more than enough for me, way before that the room loudness becomes intolerably high.

[fas42]

Originally Posted by dvv

Ah, but comparisons were made with several speakers, some of which were less efficient at say 88 dB/1W/1m, and at several volume levels, the highest of which was (to me) ear splitting.

Nada, no way, they always came out practically the same - actually, 100% the same to me, but I have to allow for some possible differences which I may have missed.

I suppose under more extreme conditions some differences would have shown up eventually, but this was a living room test, and there's just so much volume you can take before it becomes unpleasantly loud.

And that brings up an interesting point. I personally believe that an amplifier that can clip cleanly - how's that for an audio oxymoron?! - will be almost impossible to pick. What I mean by that is the situation is not that the power supply is sagging, rather that the combination of gain and voltage rail is such that the amplifier must clip, to be faithful to the volume required.

How can this occur, be so? Well, it you actually look at the waveform of many, highly respectable digital recordings you will see real clipping at a number, or even many points. This is not compression, this is pure clipping, the peak of the waveform has been chopped off as if by a knife. Yet, when you listen to these recordings at normal volumes you do not hear this "highly distorted" moment; it passes in an instant and the ear/brain does not register it.

And that's how a well behaved amp can work: it will clip in real terms maybe many times while playing at high volumes, and this will be completely unnoticeable, just like the "distorted" recording being played at normal volume. Key to this is that the split-second after the clip moment the amp is totally in control again, the fall waveform of that clipped peak is reproduced perfectly.

Which means, for example, that the amp can play, say, 6dB louder for all intents and purposes than it theoretically can, and easily keep up with a much more powerful amplifier, subjectively

I'm not that lucky. Since it provides 22 dBW, and since my speakers deliver 92 dB at 1W, theoretically I am capable of achieving (92+22) 114 dB SPL/1m. In practice, that's about 2 dB less as speakers are anything but linear power users, but still way more than most.

Anyway, that's more than enough for me, way before that the room loudness becomes intolerably high.

114dB is absolutely plenty, my setup I estimate peaks at 106 or so dB, but it has no problem, when running at maximum volume, of being deafening with the right recording. My ears tell me enough's enough, and I go off and do other things ...

Frank

[dvv]

I believe you are taking about dynamic headroom, Frank.

If so, let me refresh your memory - the ancient German DIN 45500 standards recognized this phenomenon since the ate 60oes. This is what they used to call "music power" - German colleagues present here, please correct me if get something wrong.

If memory serves, that was the measure of power an amp could do on a 1:16 basis, when it was driven by pulses 15 times its nominal power, and once as far as it would go for 3% of THD, no clipping. If you look over their older gear, you will notice that they usually managed 20% or more peak power.

When you think about it, there was much sense in that logic, especially for the informed. I've seen amps rated at 40W nominal, but hitting 70W in peaks, which allows me to conclude that this particular device was very unstressed under normal conditions, and that it had a rather good power supply. Remember, DIN standards are all related to 4 Ohms rather than 8, and the vast majority of devices were built to be able to drive two pairs of speakers, which means 2 Ohms.

So when you used them with nominally 8 Ohm speakers, they were usually meatier and more substantial than their Japanese counterparts, despite their much higher nominal outputs into 8 Ohms.

Again, if memory serves, IHF picked up on music/impulse power much later on, something like 1977 or 1978.

Over the years, I have heard many engineers expounding that all this was ridiculous, that it made no sense, and that nominal outpts were all that mattered. Obviously, I disagree, and would suggest to the unbelievers that they pick up a German made device of reasonable quality, refresh it with new capacitors and try again.

As for me, I'm still on the hunt for two German integrated amps I can never forget because of their sound quality as I remember it; one is a Grundig (the German mass production company) V5000, and the other is ASC LV5000 (second generation, mostly the same as the first, but with RCA Cinch plugs rather than DIN 5 pole plugs). And, I must add with some pride, my memory hasn't failed me yet - all my vintage gear performs just as I remeber it did way back then, although RELIABLE audio memory lasts around 15 minutes, so all I had were my generally formulated conclusions from way back then.