Barney Oliver at HP also did this before Jensen. The Bob Cordell I met seems a lot different than the one John describes. Very pleasant, smart and willing to share his knowledge.
John seems to make everything personal - no room for simple technical disagreement, personal loyalty trumping objectively determinable fact
to even suggest an idea of his or his posse deserves closer examination, possible reinterpretation is a personal tort by his posting history
Cordell in fact did adopt (rather too high to my taste) diff pair degeneration, mentioned the need for larger than expected output current into complex impedance loudspeaker loads, and credited Otala for those correct, useful insights - just look in Cordell's 80's AES paper's references and the Audio Power Amp book
to even suggest an idea of his or his posse deserves closer examination, possible reinterpretation is a personal tort by his posting history
Cordell in fact did adopt (rather too high to my taste) diff pair degeneration, mentioned the need for larger than expected output current into complex impedance loudspeaker loads, and credited Otala for those correct, useful insights - just look in Cordell's 80's AES paper's references and the Audio Power Amp book
Last edited:
What me, take anything personal? Well yes, especially when I have put 10 years into researching something. '-) (by 1980)
Last edited:
In context your reaction makes sense, if someone has a disagreement and puts themselves out to prove their point it seems like a personal attack. It's like dredging through the entire history of IRE or IEEE papers to prove someone else did something first.
It does not matter how long you have believed something when you realize you are wrong you were always wrong.
There's also this earlier 1980 paper that was.
"CONVERSION OF AMPLITUDE NONLINEARITIES TO PHASE NONLINEARITIES IN FEEDBACK AUDIO AMPLIFIERS"
Last edited:
Add to that signal embedded noise, circuit/components intrinsic and excess noise plus 1/f noise continually changing operating points and you have a cloud of dynamically moving junk that the ear picks out and objects to.
With all these erratically dependent dependences, it is to be expected that changing the nature of system noise will appreciably change the output signal.
Open loop and CFB dynamic and intrinsic noise behavior is fundamentally differently to input diff pair NFB, and that difference is audible with OL intrinsic noise sounding the most 'natural' of the bunch.
We all agree that these dependencies should be so far 'down in the mix' as to be irrelevant, but in practice this is not so, and is large part of what separates the the A Class from the also rans for long term audio enjoyment.
Being able to instantly and at will change system intrinsic noise behaviour is a very interesting experiment, and significant.....and educational.
Dan.
care to show some examples? - mostly linear feedback circuits seem in practice to really follow superposition
several professional lifetimes of feedback texts show that Bode Sensitivity analysis gives practical results for typical analog amplifier circuits
in my experience you need to go as complicated deliberate oscillation, multipliers, large signal nonlinear cirucits like Delta-Sigma converters to start seeing significant "noise modulation" by low level signal
perhaps you are refering to circuits bathed in out of band EMI? - the defenive arts for that are known
of course then audiophiles will be twitching over ferrites in their audio path
several professional lifetimes of feedback texts show that Bode Sensitivity analysis gives practical results for typical analog amplifier circuits
in my experience you need to go as complicated deliberate oscillation, multipliers, large signal nonlinear cirucits like Delta-Sigma converters to start seeing significant "noise modulation" by low level signal
perhaps you are refering to circuits bathed in out of band EMI? - the defenive arts for that are known
of course then audiophiles will be twitching over ferrites in their audio path
I have this and remember it when first published as one of the truly important papers for audio design and construction. After I read this, I extended it to C and R leads, chassis, connectors, cables etc.
THx-RNMarsh
Last edited:
Samuel Groner and Bruno Putzeys extend this in their Vol 13 article to tests showing magnetic distortion in reed relays.
Jan
Dan, you certainly mention things that are real and manifest in several ways, but I can not swallow some of your conclusions and some of the claims involving noise removal in favor of the signal which do violate the second law (entropy reversal). So if you have a buddy versed in the physics put together a story for us to at least read.
As to these effects one is, as I mentioned earlier, a long-tailed pair disturbed from equilibrium will have a slight change in trans-conductance which will change the noise but this is a very small effect and with FET's because the the noise goes as the fourth power of current almost non-existent.
Really bad carbon comp resistors have excess noise that can be modulated by signal. I have only been able to make this easily observable by selecting particularly bad components deliberately.
Last edited:
Maybe with microvolt/milli volt signals (phono cart), but does this hold for higher levels found in A-D-A systems?
What would be the conditions your comment does Not hold?
-RM
Reed relays? The HP 339A has two reed relays and it can go down to below -100dBv. But the ShibaSoku 725 series and Panasonics VP---- and Audio Precision do not use any reed relays nor any signal parts with steel in them. Same with AP.
Linear Audio mag is one of the best audio related design magazines ever.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.