Quote Originally Posted by skeptic
John, I also assume that R1, R2, and R3 are really Z1,Z2, and Z3. Wouldn't Z1 and Z2 be in the megaohm range normally? What about Z3 at 3.6 and 14.4 Khz? These should be fairly low. What are typical values?
They are indeed impedances..For this analysis, I considered only resistance. And they should be in the sub ohm ranges.

A loop of conductor will develop a voltage that is proportional to the rate of change of the flux captured by the loop, dB/dt... For a sinusoidal flux change, the rate of change of flux is the derivative of the flux...as the frequency goes up, the slew rate, ie, flux rate of change, goes up, directly proportional to the frequency. So, I used the variable K as including the time derivative, to keep the equations simple.

The inclusion of f2 was not meant to represent the frequency being transferred, but a factor for the strength of the coupled signal.

Sorry for not including the simplification of terms in my explanation..

An example:

Say, a 100 mV signal input causes 100 watts at the speaker.

The supply will draw haversine current, 60 hz, 180 hz, 300 hz, 420 hz..540 hz..660 hz

Assume 2 amp line draw..60 hz component..1 amp 180 hz comp, 1/2 amp 300 hz.

Now...goto the final equation...

Assume that at 60 hz, the 2 amps will cause say, 1 millivolt of error signal..

At 180 hz, 1 amp, three times the frequency means 9 times the coupling (freq squared), so the 180 hz component will be (half the current times 9 times the coupling).

So, the 180 hz error signal will be 4.5 millivolts.

At 300 hz, 1/4 the current, but 25 times the coupling.. 6.25 millivolts.

420...49/8=6.125 millivolts..

Using these 4 harmonic terms, adding their peaks..(don't forget how a haversine is made from the frequency components), 17.75 millivolts of error signal, with a 100 millivolt signal..17% peak waveform distortion..and it's haversine derivative stuff..

Now, these numbers were pulled out of the hat, as I assumed 1 millivolt error at 60 hz..

The error at 60 could easily be just 100 microvolts...try measuring that level in an amplifier delivering 100 watts to a load....not exactly an easy task.

But, that level of coupling would add 1.7% errors, in the prev example..

Now....this has been only a haversine discussion...those nefarious loops also exist in the supply section of all amps..So what keeps the 5 Khz amp output signals from feeding into the PC loop?

And, that frequency will couple 6,800 times more than 60 hz..so even 10 mA of audio injected into the line cord will be an issue...

What would be more important IMHO, is that bleedback of that type in the audio band would not necessarily be waveform distortion per se, but phase shift...and most spectrum analyzers do not see phase shift, but measure waveform distortions.

If phase shift occurs as a result of this, then the phase shift would be amplitude dependent...and, how would we interpret that effect in a binaural soundstage??? What words would an "audiophile" use to describe that effect??? Beats me...

Mtry....skin analysis and testing: I recently moved, and all my stuff is in boxes until I can finish sheetrock, rug, and windows in the attic, then I can start to build my audio test setup in the basement. Skin effect takes a back seat to life...

And yes, I did get that document...I am amazed that even one or two people could hear 2 microsecond deviations..

Bill L: Yes, at least two people understood it..

Cheers, John