It's causal and stable and there is a unique output for any input, I don't see it. Causality is the key, one could use a long pipe with a speaker at one end and a mic at the other and create a real memory loop, this would be a crude acoustic delay line memory. Echos and reflections are the most common non-minimum phase events. John knows Dick Heyser covered all this.
Oh come on. Is it even possible that the old pioneers had to evaluate with tube amps, transformers, and tape and LP, tone arms and cartridges and speakers of those days, that had high distortion and noise? -- Compared to today's sources and amps and speakers. And, that they masked the effects we are now able to more clearly hear and address? Maybe?
The good old days of the pioneers - sound-wise- werent so good, as I recall. [And, I had the old Quads and still have a pair of the new ones.]
There are plenty of current listening exercises that do not support any more than the grossest examples and differences.
Maybe you are right, Scott. Then, why do we make IC amps that have .0000001% thd+n? Its an intellectual excercise? Well, for me it is. Reminds me of when Brian Elliot (PHD) engineer at IBM and later HP was studying the limits and causes of distortion.... down to the elements we got. There we ended at the valance level shift. But who knows.... I think there are people who do and those that dont hear a differnce.... like many people said you couldnt hear a fraction of a DB level change in the frequency response but F.Toole and others showed it was detectable with certain Q. So a zillion people do and a zillion people dont. Lets move on and I give back my time to the man -- JC.
I have no stake in it one way or the other. So lets call it a draw for now.
I have no stake in it one way or the other. So lets call it a draw for now.
A lot of research never gets to mainstream. I had UC Berkeley TID do a search for me on causes of distortion in resistors. One paper recieved was an unpublished PHD thesis at UK university. He showed there was evidence of two second harmonics but each was out of phase 180 degrees. (opposing diodes???) The usual thd tests would show no 2nd. So we test for 3rd... I owned a Quantec transistor and same brand resistor noise/distortion analyser at the time. This an other such research work has led me to be suspect of sine wave testing. We should be doing more with asymetrical tests IMO. No IMHO. 
Welcome to MY world, Dick. I can't even talk about preferring servos, without a battle. Weren't you one of the first users of servos in audio?
We haven't gotten as far as differences in film capacitors, either, like your invention, in making a 'better' cap. Maybe, we will make progress, over time.
We haven't gotten as far as differences in film capacitors, either, like your invention, in making a 'better' cap. Maybe, we will make progress, over time.
What are you going to prove John, that it is 'ignorant' comment, or that adding active components to passive components decrease noises and distortions? If the first, my silly comment was regarding your silly comment. We are even.
I can prove you in turn that all depends on topology. Topology where capacitors "do not sound" can't be improved by servo, if servo involves additional active components. However, if you open my Pyramid amp driver's schematic, you can see a servo, in form of R-C network from the tail of LTP driver to the screen grid of the first pentode. But it is a different case: the servo there does not include additional active components. And direct connections between stages were used not because "capacitors sound bad", but because I wanted stability on low end, since the amp uses feedbacks, one of which is over all stages.
Though, more linear than real active component.
Last edited:
Right. Like, red dog steals hens because it is red, or because it is actually a fox?
Similarly, some capacitors with high DA distort not because of DA, but because they are also non-linear. Also, caps can be innocent, when they couple stages with non-linear input an output resistances. That's why I prefer to couple driver to the output tube's grid directly instead of using mega-expensive boutique capacitor: voltage swing on control grids of output tubes is enough for caps to feel non-linear input resistance. Especially on peaks that always present when listening to decent records.
Yet there's threads here about how speakers were better in 1928, vinyl and tape are "better" than digital, and that the recordings from the '50s waste anything made now. So...
To me, those are about likes and dislikes.... not about measured data being better... like thd and IM and s/n and Freq response.
Last edited:
Yes. i am the intellectual inventor of the dc servo concept given to Walt Jung in a telephone conversation as a solution to my thinking to go direct coupled without offset and drift being an issue. Walt was the one who put my idea into practice, first. And, the rest is history.
Last edited:
This problems were considered as problems long time ago, today it is easy to measure all of them and make them more than satisfactory on steady measurements. What is most overlooked, is dynamics. The shorted is time interval between "measurements", the more precise is "measurement" by perception. While our electronic tools prefer longer intervals. Also, what for tools is close to threshold of recognition, for perception is most significant.
Audiophiles don't complain about THD on forte-fortissimo; they complain about lack of details. Like air, walls, lips, nails and fingers on strings...
You probably forgot about the other side of the fence where servos were used in analog control systems because digital ones were not so good.
Last edited:
Hmmm... yes, lets call it a draw or merely a flesh wound. This audio stuff is almost as bad as that pesky Higgs boson.
Well, on to servos, if we can. The reason for servos is mainly that 'really good' capacitors are usually large and very expensive. To get the same 'performance' a servo can be added, that will replace the need for coupling caps. It is not absolutely necessary to make a complicated servo, I hope to show a few examples in future.
- Status
- Not open for further replies.