John Curl's Blowtorch preamplifier part II

[scott wurcer]

Capacitor microphone? Although there the charge stays constant and the capacitance varies so maybe not.

A condenser mic works both ways, constant charge or constant voltage. The quietest mic in the world is about 98dB below 100mV in air but I wonder what the vacuum limit is in constant voltage mode.

There has been some work on single electron effects in integrated circuits.

 

[scott wurcer]

just my 2 cents regarding john´s frequency unrelated "blips" in the spectrum of a ahard driven TL741, could it be "shot noise" or "popcorn noise" wich is behind his findings?

I thought this was put to bed, fat chance, all the "blips" are accounted for by harmonics of the two input frequencies. They were listed by at least two folks months ago. The counter argument rests on trying to read something into a third generation Xerox of a printout from a strip chart recorder.

Almost forgot, the counter argument also relies on some unknown physical process that violates first principles.

 

[scott wurcer]

Well said about jfets, John! I remember how somebody here about 3 years ago criticized my condenser microphone as "Arrogant", because I used jfet there on microcurrents, unlike in datasheet examples.

That's interesting Neumann ran a 2N3819 at <400uA in one of their first JFET mics. The 48V phantom power only dropped to 47V.

 

[john curl]

First of all Scott, a chart recorder was not used. A precision HP X-Y plotter was used. It was more accurate than you presume. Second, the HP3581 used has a nominal resolution to 1 Hz, and I have measured it to be accurate to that precision. IF you don't want to believe that the 'controversial blips' were the result of FM modulation, so be it. However, it has NOT been proved that it is not so, AND the same degree of evidence (having generated a similar waveform through a different method) which I did by taking a 3.18KHz square wave and FM modulating it, has been used by JN, who consistently challenges Hawksford with his 'exotic' waveform that mimics Hawkford's mathematically derived and measured wire distortion waveform. Let's have a level playing field here.

 

[DF96]

Hawkford's mathematically derived and measured wire distortion waveform

Would that be the waveform which simply shows the effect of a linear filter? Calculated by taking account of internal inductance, but without realising/admitting that is what it is? So nothing new, and no distortion?

 

[john curl]

I think that a 3db/octave slope should not be ignored.

 

[scott wurcer]

IF you don't want to believe that the 'controversial blips' were the result of FM modulation, so be it. However, it has NOT been proved that it is not so, AND the same degree of evidence (having generated a similar waveform through a different method) which I did by taking a 3.18KHz square wave and FM modulating it, has been used by JN, who consistently challenges Hawksford

I even posted the figure from the Engineer's Handbook, FM sidebands are and can only be at the same frequencies as the AM. This whole myth is a very strange one to hang on to. If the data was kept in phase and magnitude form the AM and FM might be seperable. As it is I don't get it, some FM is no surprise and nothing new.

 

[simon7000]

Calm down, this your personal view on this topic. It is not useful to cry Ohm's Law isn't so we must question everything just because resistors have a temperature coefficient.

Scott,

You are welcome to be an omphaloskeptic.

But the rest of the world does try to use the same language. See:

http://www.lectures4you.de/pdf/chris_prot/Ohmlaw.pdf

http://www.physics.isu.edu/~hackmart/114-214ohm.pdf (page 4)

Ohm's law - Wikipedia, the free encyclopedia (see the linear approximation section)

I don't understand why you try to make this a personal issue. You are free to believe what ever you like no matter how silly.

 

[PMA]

Linear or non-linear, it is always valid at the point of interest (dynamic resistance).

 

[DF96]

Almost true, apart from discontinuities. With a relay and a resistor you can make a two-terminal circuit which is approximately linear almost everywhere, apart from a discontinuity with hysteresis. So what, you ask? Well, exactly. I can't even remember why I pointed out that Ohm's law is not a law of nature - I guess someone must have claimed that it was. Can we drop it, and get back to audio?

 

[wrinkle]

First of all Scott, a chart recorder was not used. A precision HP X-Y plotter was used. It was more accurate than you presume. Second, the HP3581 used has a nominal resolution to 1 Hz, and I have measured it to be accurate to that precision. IF you don't want to believe that the 'controversial blips' were the result of FM modulation, so be it. However, it has NOT been proved that it is not so, AND the same degree of evidence (having generated a similar waveform through a different method) which I did by taking a 3.18KHz square wave and FM modulating it, has been used by JN, who consistently challenges Hawksford with his 'exotic' waveform that mimics Hawkford's mathematically derived and measured wire distortion waveform. Let's have a level playing field here.

The image that was shown to us was obviously a photocopy as even the harmonic peaks were not on an even spacing, it also had the typical photocopy blurring and spots on it, this was totally and exhaustively laid to bed some time ago. The parrot is most definitely not pining for the fiords, its mortal coil has well and truly shuffled.

Wrinkle

 

[john curl]

But NOT as Scott Wurcer said it was.

 

[Wavebourn]

John; if it is not what Scott said, it was obviously some kind of miracle. Do not matter what kind of modulation it was: amplitude, phase, or combined, frequencies of sidebands would be the same, arithmetically. The differences are only in amplitudes and/or polarities of them.

 

[Chris Hornbeck]

<Thoughtful and pertinent example snipped for bandwidth>

However, nothing in an audio circuit is isolated, as otherwise you don't have a circuit. An electron can be, in effect, halfway onto a node but still halfway off the previous node. The quantisation does not occur. Even in a valve, the electrons approaching an element affect the potential of the element before they arrive.

Would you see this quantisation anywhere? Yes, possibly. You need a combination of low capacitance and low signal level. Capacitor microphone? Although there the charge stays constant and the capacitance varies so maybe not.

Note that what matters is not the capacitance of, say, the base of a transistor but the capacitance of that entire circuit node. My guess is that if this was a problem it would be seen in integrated circuits, as they have tiny circuit nodes with very low capacitance.

The circumstance in which Dr. Hawksford raised the issue was in the base current injection of a BJT, with the implication that discrete charges exist there. Is this circumstance similar enough to charging a capacitor that the same intuition would apply? I wish I were smart enough to say yea or nyet myself, but I ain't.

Thanks very much for your comments,
Chris

 

[gorgon53]

Random speculations create confusion.

My intention was to stear into a new way of looking at things that often turn
out to be very fruitfull. Without natures random trial and error approach we would not even be here....

 

[Wavebourn]

The circumstance in which Dr. Hawksford raised the issue was in the base current injection of a BJT, with the implication that discrete charges exist there. Is this circumstance similar enough to charging a capacitor that the same intuition would apply? I wish I were smart enough to say yea or nyet myself, but I ain't.

Chris

Discrete charges indeed exist there, according to theory of physics of semiconductors: each discrete charge carrier needs certain charge to jump over potential barrier. Then it looses charge when recombination happens, emitting again one discrete photon. Similarly to a steam engine that works on heating and cooling, except in transistors heating happens electrically. We all know that steam engine produces noise, so transistor should as well, by analogy.

 

[john curl]

Chris, you are probably right about caps as well as bipolar transistors.
I just reviewed Hawksford's 'Fuzzy Distortion in Analog Amplifiers' JAES Oct. 1983.
WOW! I forgot how revolutionary this paper was. I can't find anything wrong with his reasoning, and I am glad that I reread it. Try it yourselves, anyone who wants to know and understand the essence of this thread.
No wonder bipolar IC's sound 'homogenized' compared to discrete jfet open loop circuits.
I was attributing it all to PIM, but Hawksford has opened my eyes, once again.

 

[DF96]

If this mechanism creates problems in audio ICs, then it ought to be much worse in RF ICs with their smaller geometries and potentially smaller signals. Is it seen? I have never heard of a receiver designer complaining about the discrete charge on the electron.

 

[wahab]

Not sure that RF IC have smaller signals since they are low impedance.

-120dbV on 50R is about the same as -60dBV on 47K , as in MM preamp.

 

[DF96]

For electron quantisation noise/distortion it is voltage which counts, not power. RF ICs are likely to have smaller capacitances than an audio IC. Let's guess 0.5pF. One electron then represents 0.32uV. I think we would see that, as receiver front-ends routinely work at sub-uV levels, down at the thermal noise. Some are cooled, so less thermal noise.

Many of the 'bright' ideas that people come up with to 'explain' perceived audio problems would have repercussions in other branches of electronics. These repercussions are not seen, so you then have to explain why (or hope your readers don't notice). Of course, some things do effect audio more than elsewhere: flicker noise is a good example.