That's certainly fine if you cannot hear the distortion of a dozen op amps in series.
A few are used full range all going through the initial high pass if memory serves. Nelson Pass analyzed the design here a while back. View post 96.
So others, including Nelson Pass and Sy share my opinion.
Agree entirely. The vast majority of such tests are nothing more than parlor tricks. Here's another really funny one from Brad Meyer and David Moron.
There was extensive discussion about this over at AA back in 2007 found here.
Last edited:
Some of his writings can be found here. The "Tire Testing Without a Pressure Gauge" article was interesting (to me) and seems to have been written by someone with his head screwed on the right way.
Maybe. Or maybe he's well aware of that. However...
That's a very questionable assumption. With a lot of amps, badly distorted output voltage will result in badly distorted input current. That in turn will result in distorted input voltage unless the source has a very low output impedance.
This isn't hard to understand. One of the effects of NFB is to dramatically raise the inherent input impedance. The flip side of that is that removing the feedback will dramatically reduce the inherent input impedance.
Let's take a look at a worked example for a high-feedback amplifier:
Say the input stage is a BJT long-tailed pair with a tail current of 5mA and no emitter degeneration. Assuming the transistor's current gain is 250, the input impedance without feedback will be about 5K. When negative feedback is applied, the input impedance is raised by a factor of hundreds if not thousands. The nominal input of the amplifier is then set to say 10K, by essentially putting a 10K resistor in parallel with the input.
So far, so good. The input impedance of the amp will indeed be about 10K, and nice and linear, as long as the neagative feedback is doing it's job.
Now what happens during clipping? At the onset of clipping, there's suddenly no negative feedback and the inherent input impedance drops back down to it's open-loop value i.e. 5K. In parallel with the 10K input resistor, that means the incremental input impedance at the onset of clipping drops almost instantly from 10K down to about 3.3K. As the input voltage increases further, the transconductance of the input stage drops, and the incremental input impedance rises back up to 10K. Similar story as it comes out of clipping.
Long story short, during clipping, the incremental input impedance drops from 10K to 3.3K for parts of the input waveform (four times per cycle, FWIW).
If driven from a source with significant output impedance, this will result in easily measurable (and quite likely audible) distortion of the input voltage.
Bottom line: If the inputs of two amplifiers are both connected to a source with significant output impedance, then gross misbehaviour of one amp (e.g. clipping, current limiting or slew rate limiting) can distort the input voltage of both amps.
Maybe he has never ran into proof of otherwise? That's a real possibility. How can he know his system is really bad if nobody showed it to him?
Having heard his home system on several occasions, I too failed to point out to him how bad it really sounded
Jan
Godfrey, yes I fully agree with you.
But I would submit that if one of the DUTs is clipping or otherwise grossly misbehaving, the whole exercise of trying to hear small differences between them becomes rather moot. Or rather easy, depending on what you are looking for.
Jan
Hi i think is this one ... very very interesting
If
compared to a direct connection i would say ... may i have the schematic and/or the part no. ?
It must be a very "transparent" buffer indeed ... or the case can be extended to the average buffer op-amp ?
But anyway very very telling.
I assume that the writer had a very resolving system ... and ears.
Thanks again. Kind regards, gino
When you're a hammer, every problem looks like a nail to you.
When you're a speaker designer every problem looks like an acoustical one.
Clearly, he is indifferent about amplifier quality. Those ATIs and Samson amps he recommends really are great aren't they? I had to smile when I viewed his speakers at this show.
Estat,
Those were not novel viewpoints of mine; they were factual findings supported by measured data by a multitude of researchers and their journals. Anyone can test and acquire those results and enlighten themselves. Without feedback, overall performance and the subsequent waveform accuracy really does worsen. Feedback received a bad name in audio from the erroneous use of such with BJTs and the original faulted Transient Intermodulation Distortion papers, and oddly enough, some people even today haven't got the memo that TIM does not exist in the absence of slew rate limiting. Back on harmonics, don't know of a single person who would honestly be able to tolerate even second order harmonics at -20 to -30dB from an amplifier or preamp for extended periods with all audio material. In instruments it is fine, and it works well with jazz, but painting the entire reproduction envelope with the same orders of distortion that are not necessarily present in things like recording venue acoustics or voice can vastly degrade the sonic presentation becuase the auditory system may identify the problem. For some, it sounds euphoric, and of course some music and tastes will enjoy it. Sometimes even a little bit adds character, but the sheer distortion levels that some of the audiophile press advertised gear reach look more like an excuse to create the guise of credibility where they don't know who to reduce the distortion. High-end rhetoric is almost pseudo scientific, and believers who do not have the experience to both listen and measure it in completeness jump on it as though it were truth. For other listeners, they hear it for what it is, unnatural alteration of the material. When harmonics (and other distortions) are added to sounds that the auditory system knows are not natural components, that person will hear it as distortion. These are two very different listener groups and much of the discontent between the low and high distortions proponents may be caused by their own preference.
Those were not novel viewpoints of mine; they were factual findings supported by measured data by a multitude of researchers and their journals. Anyone can test and acquire those results and enlighten themselves. Without feedback, overall performance and the subsequent waveform accuracy really does worsen. Feedback received a bad name in audio from the erroneous use of such with BJTs and the original faulted Transient Intermodulation Distortion papers, and oddly enough, some people even today haven't got the memo that TIM does not exist in the absence of slew rate limiting. Back on harmonics, don't know of a single person who would honestly be able to tolerate even second order harmonics at -20 to -30dB from an amplifier or preamp for extended periods with all audio material. In instruments it is fine, and it works well with jazz, but painting the entire reproduction envelope with the same orders of distortion that are not necessarily present in things like recording venue acoustics or voice can vastly degrade the sonic presentation becuase the auditory system may identify the problem. For some, it sounds euphoric, and of course some music and tastes will enjoy it. Sometimes even a little bit adds character, but the sheer distortion levels that some of the audiophile press advertised gear reach look more like an excuse to create the guise of credibility where they don't know who to reduce the distortion. High-end rhetoric is almost pseudo scientific, and believers who do not have the experience to both listen and measure it in completeness jump on it as though it were truth. For other listeners, they hear it for what it is, unnatural alteration of the material. When harmonics (and other distortions) are added to sounds that the auditory system knows are not natural components, that person will hear it as distortion. These are two very different listener groups and much of the discontent between the low and high distortions proponents may be caused by their own preference.
Last edited:
Some of us make judgements on audio quality using our ears rather than our eyes. You're welcome to your opinion while my experience follows that of many companies like Ayre (Charles Hansen), Pass Labs (Nelson Pass), Audio Research (Rick Larsen), AtmaSphere (Ralph Karsten) and a host of others who make very nice sounding amplifiers/preamps using zero corrective feedback.
The devices they use are linear enough. To each his own.
The problem is that they are using corrective feedback. Rod Elliot and a couple others have highlighted a variety of problems with Pass' designs and improved on them greatly. Zero-feedback claims are outright false. Any amplifier with a non-zero emitter ground impedance is using feedback by definition, unless audio designers are now attempting rewrite the terms for their own fame and financial gain. All single ended triode amplifiers also have to use feedback, ie a cathode resistance or paralleled capacitive impedance, to set the bias Q point half way along the load line. This, in itself, is feedback. I am not trying to argue with you, and I think that you know what you like in music. You've found what really makes your music sing and that is everyone's goal here. It's just that there are also engineering facts behind the design principles used to reach those goals and others, and they have to be shared. People have been misled for too long by the audio press and certain seemingly-intuitive, yet incorrect ideals.
Last edited:
When you're a hammer, every problem looks like a nail to you.
When you're a speaker designer every problem looks like an acoustical one.
Clearly, he is indifferent about amplifier quality. Those ATIs and Samson amps he recommends really are great aren't they? I had to smile when I viewed his speakers at this show.
Well I actually listened to the system. Of course, it is possible that I am utterly deaf. Or maybe you have no clue. No way for me to know!
Jan
I don't think that's correct, but it's moot: the assumption of the Y-connector is untrue, at least for the ABX brand box and the systems used by people like Lipshitz and Vanderkooy. The red herring about pops leads me to believe that the poster has never actually had his hands and ears on an ABX box nor actually participated (much less set up) in a serious double-blind, ears-only test. It's trivial to prevent pops in switching circuits unless there's DC coupling and severe offsets, and the engineers who designed the boxes were full aware of them. Even a dumb chemist like me could design to avoid them and document how it's done.
Yes, it is and can be; it's not difficult to design a perfectly transparent buffer as long as that transparency is tested ears-only.
The writer's system is a pretty good one and his hearing isn't terrible.
He spends much time with live music and has bothered to investigate sonic issues himself rather than parrot the nonsense rampant in audio magazines, non-technical internet forums, and promotional pieces by people trying to sell something.Yes, it is and can be; it's not difficult to design a perfectly transparent buffer as long as that transparency is tested ears-only.
The writer's system is a pretty good one and his hearing isn't terrible.
Hi and thank you very much indeed for the very valuable advice
I must have been a little (or quite a lot) brain washed also myself
I will try to stop reading such magazines
I hope my rational side will prevail for my own sanity
And i will give opamps the attention they deserve
Thanks a lot again
Regards, gino
Yes, it is and can be; it's not difficult to design a perfectly transparent buffer as long as that transparency is tested ears-only.
The writer's system is a pretty good one and his hearing isn't terrible.
Sounds like a nice guy to meet
Sounds like a nice guy to meet
You wouldn't like him; he's offensive, violent, and an inveterate pervert. Be glad that all you have to do is edit and publish his stuff.
What's a spike to one person may merely be a pulse to someone else ... I'm referring to the "pulses" of current that flow when the rectifiers in conventional supplies switch on - when the amp's working hard the smoothing caps are signficantly depeleted between each charging period, so high current draw on the transformer during the recharge period. This means pulses of current in the primary winding, which translates to voltage noise in the mains from the resistance in the house wiring. Using a light dimmer is a notorious way of causing audible problems in audio systems, and this is just a variation of the same 'issue'.
Switching supplies largely alleviate this but of course introduce other aspects that have to dealt with.
The "more linear magnet structures" may be part of the scenario, but I have not had problems driving conventional drivers to levels normally associated with compression drivers, if the electronics were working correctly. The "working correctly" is helped enormously if, yes, the voltage rails are not modulating severely from the load requirements.
A single counter-example "proves" the case: I heard a Dynaudio midrange unit, the very type of thing gino was "complaining" about, being driven far beyond what compression drivers in audiophile setups are normally asked of, in an audio show - and doing a superb job of it. Very high PA sound levels, without sounding "PA" - far superior to 99% of normally heard hifi sound.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.