Of course, jcx.
Just look at those sims from the same amp in two configurations: One current feedback, one voltage feedback.
Don't take care with the high end low and high differences, the current feedback is low pass and high pass filtered for production, while the voltage feedback is simulated at the extreme bandwidth as possible. No time to equal the two.
To Wavebourn: Do you really think that everybody apart you is stupid ?
Attachments
Last edited:
seems like just yesterday...
http://www.diyaudio.com/forums/anal...ch-preamplifier-part-ii-2916.html#post3236109
as far as CFB/VFB questions I was trying to engage in a fair evaluation in Elvee's thread http://www.diyaudio.com/forums/solid-state/218689-cfb-topology-superior-why.html - have you abandoned it?
in short, you can play with self reinforcing assumptions - CFA can be more linear than diff pair with large input error V where the diff pair is hard limited by its bias current, low open loop gain/feedback factor creates large input error V...
http://www.diyaudio.com/forums/anal...ch-preamplifier-part-ii-2916.html#post3236109
as far as CFB/VFB questions I was trying to engage in a fair evaluation in Elvee's thread http://www.diyaudio.com/forums/solid-state/218689-cfb-topology-superior-why.html - have you abandoned it?
in short, you can play with self reinforcing assumptions - CFA can be more linear than diff pair with large input error V where the diff pair is hard limited by its bias current, low open loop gain/feedback factor creates large input error V...
Last edited:
It has nothing to do with comparison, i used the file that i have. It is about looking inside closed loops.
You are free to make your own conclusions. Please, do not deviate, i'm bored with this VFB vs CFB controversies, my choice is made since decades and it was not my thread.
Last edited:
That's a pretty vicious test, Wave, especially the 60dB gain: the very best discrete amplifiers in the world would sound pretty sick if configured to do the same ...
But, agree that low level info is what it's all about; having a system with gain set such that a peak level signal will momentarily deafen you, but the whisper stuff still comes through perfectly clean -- just like the real thing ...
Frank
It wasn't much I said that it was a nice academic overview and summary of most of the previous research on the AM/PM issue including an analog technique for separating them (which he chose to try and patent). He, much to my relief, presented results for several kinds of amplifiers without having an agenda. The results were in general what I would expect op-amps with slew rates < 1.5V or so did not measure well. Several more modern higher speed (some fairly mundane) op-amps measured below the noise floor on his test while some highly regarded (by audiophiles) amplifiers measured markedly worse. This begs the questions, is this test relevant and are there audible abberations of reproduced sound that for some reason create listener preferences?
It seems nothing is more like poking a stick in a hornets nest than stating the distortion of an SET amp IS what creates the preference for the sound. It's time to question some beliefs. It has been proposed for years that the THD has nothing to do with it and there is some "unmeasured" phenomena responsible. Frequently this PIM, FIM thing has been put forth as the answer.
Wave, what happens when you use 10 of your favourite class A tube amps in that test? Have you tried this test yourself with OPAs or class A tube amps?
What do you recommend to replace evil OPAs to get 60dB gain?
Very little seems to have been said about memory effects affecting sound, and this is not talking about cap DA, but that the sound quality changes over a period of time, of the order of hours at times, due to myriads of subtle material behaviour mechanisms all doing their thing. Thermally related variations are an obvious one, but there are plenty of others too. It's a huge juggling act, keeping all of these factors from impinging too much on the sound.
To me, the SET thing overlays just the right level of syrup to mask most of these audibly disturbing characteristics, without getting too much in the way of revealing what's in the recording ...
Frank
Yes, those factors will do it nicely, if the problem areas are constrained totally by, and are within the FB loop. If however, in the real world, the PSRR and loop gain starts to collapse at the higher frequency end, and you then combine that with the increasing human sensitivity to distortion the higher the frequency it starts to look less assured ...
I've spent many years trying to sort out, bring under control these elements, and it is far from an easy task ...
Frank
I've spent many years trying to sort out, bring under control these elements, and it is far from an easy task ...
Frank
Thermal de-biasing could happen but these effects are outside of this discussion tending to be non-minimum phase and not yielding to a straight forward analytical analysis.
On the other hand, they keep some advantages in regard to industrial processes, like the ability to pair parts and fine tune... like replacing adjustable resistance by adapted fixed ones...when time is a pleasure more than money.
This test needs to be applied to everything brought to the next Burning Amp event. Anything that fails goes into the bonfire.
Bob Carver [*] won his Stereophile Challenge where he said he could emulate the sound of ANY Golden Pinnae amp by modding the response, overload & THD behaviour of one of his evil M400 amps. The results were assessed in a simple Blind Listening Test by Stereophile.
For me, the surprising thing was that Stereophile (pre John Atkinson) had Golden Pinnae who weren't deaf and could return reliable results in a Blind Listening Test.
I would love to conduct a Blind Listening ABC test with the presentations as follows
- undoctored signal
- improved by SET
- improved by eg a good chip Power Amp
Sssh! Don't tell them.[*] IMHO, Bob was probably the most truly innovative PA designer of the previous century. Always thinking outside the box and asking what needs to be done for good sound rather than what Golden Pinnae bits or topology needs to go in .. or even which specs to go for. Pity so many of his amps had a tendency to burst into flames.
It remains to be seen if the new generation of Class D gurus can provide as good a sound.
Last edited:
Which is the big problem in this audio game, and why so many chase the pixy dust solutions. Being hard shouldn't be an excuse not to try and achieve an over-arching solution ...
Frank
The best performances always need to conclude with fireworks ...Pity so many of his amps had a tendency to burst into flames.
Frank
My favorite class A tube amps have waists in different places than my favorite opamps. And they need different tests to show how sick they are: they need forte-fortissimo of a nice symphonic orchestra, and a pad on it's output. "Whisper-test" they will pass perfectly. But not "screaming" one. However, on high SPL your brain may fail this test: it is hard to tell what causes such intermodulations, the amp or the ears.
And yes, stretching and interpolating tests are my favorite tests. I do them always in order to see the problem under the microscope. Why? Because it's not enough to test for conscious recognition of errors, when deal with high-end. The point is to fool subconscious perception tat has quite different thresholds than conscious recognition.
I would use 2 or 3 OPAs instead of one for such gain.
Last edited:
The latest switchmode offerings from Mark Levinson and Anthem fared quite badly in the recent Stereophile (December 2012). JA makes an attempt to correlate some spuriae with Fremer's dissatisfactions with the ML and also reports finding the "overall sonic picture ... flat and uninvolving". For the Anthem, JA complains of higher-order distortion appearing at higher power levels, and Kalman Rubinson declines to recommend the product.
A tough business.
The point being, with real sounds your brain doesn't fail the test: high intensity sound from actual instruments doesn't faze the hearing system; otherwise musicians in an orchestra would go into mental overload quick smart. The trick is to achieve the same intensities, with the accompanying cleanness that those players experience, on the playback of recordings. You know you're there when the hearing system perceives the sound of the combined instruments as a rich, enveloping texture ...
Frank
Presumably you've carried out your "Whisper-test" on these class A amps and are reporting the results of Listening Tests. I'm sure no one will mind if you use 30-40dB gain instead of 60dB.
Which OPAs did you do your "whisper-test" on?
But that means you are passing your signal through 2-3x more evil OPAs
Even I wouldn't do that.BTW, taking a music signal down to and below the noise floor is a very good test of an analogue or even an evil digital signal chain. I've done this for many different things but not on OPAs.
It was the performance of the Sony PCM-F1 on this test that convinced me that good digital was possible. You can take a piano signal 20dB below its noise floor and still hear a clean if noisy piano.
Not many analogue chains can boast this.
With bad digital (or analogue) chains, the signal goes nasty well before you get to the noise level.
And also pretty savage in Home Theatre on the Anthem: Anthem Statement M1 Amplifier | Home Theater. Since we're talking about time dependent altering of quality, a quote from the article is quite enlightening:
Frank
- Status
- Not open for further replies.