On Tue, 17 Aug 2004 13:53:24 -0700, Bob Cain
wrote:



Your experiments confirm nothing. They don't eliminate
sources of false positives.



Let me explain again my processing of the pcavtech data.

I demodulated the FM (rather well, I think). This makes my data
immune to any AM sidebands. I then selected the component of this
demodulated signal at 50 Hz. This makes it immune to all the other
frequency components present. I did this both for the 4025Hz carrier
(shown) and the 1000Hz carrier (not shown at the time I posted
"bingo") both of which were frequency modulated by the large 50Hz
component of the test data. The phase deviation in radians is
numerically equal to the modulation index and in the case of 50Hz
modulating frequency you can convert this to frequency deviation by
multiplying the modulation index by 50. This was plotted as Y values
and obviously involved a lot of processing and is a very indirect way
of measuring distance. (phase angle/2*PI)*wavelength)

The X coordinates were simply calculated by taking the estimate of the
linear displacement of the speaker cone at 10 Volts RMS drive given by
Arny and assuming that the displacement was linear with Voltage.
This was converted to the fraction of the wavelength at 4025 Hz and
1000Hz this represented and converted to radians.

The X and Y values were then plotted without any further assumptions.

Both the the 4025Hz and 1000Hz data follow the same "curve" (with
some scatter) and that curve is the straight line X=Y.

Now when you can think of some distortion by the amplifier/speaker
that follows *that* maths, get back to me.