|
Comment on "Crosstalk
(Noise) in Digital Systems"
I made an advance in this
subject today.
In 1980 I wrote;
Space is
the ability to accommodate energy.
(From Electromagnetic
Theeory vol. 2 , p236 , p238 )
The first thing is that
energy (current) can enter a region of space if its aspect ratio is
acceptable – 377. That is, the ratio between the electric dimension and the
magnetic dimension of the energy needs to be 377. If the region contains a
dielectric with permeability of permittivity which is different from that
of vacuum, the energy current has to have a different aspect ratio in order
to enter it. If the aspect ratio is wrong, then some of the energy current
will reflect at the entry point to that region of space.
We know something about
the collision of two energy currents at 180 degrees. Superposition applies,
except that while they overlap, there is a brief lateral force between the
guiding conductors. This is discussed starting at page 258
and 260
of my book
.
At page 13 of my other book , there is
discussion of what happens if two pulses collide at a four way junction. If
they are of the same polarity – called “dissimilar pulses”, they help each
other across the gap. If they are of opposite polarity, “similar pulses”,
they are deflected through 90 degrees.
The advance I made today
relates to what I have found out from my 1967 paper
"Crosstalk (Noise) in
Digital Systems" , pub. IEEE Trans. Comput.,
vol. EC-16, no. 16, December 1967, pages 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 31 , some of which is in two
of my books. The argument starts at page 30 of one book , and at page 4 of the other book
, continuing on page
55 . Here in
figure 9.2 we see “a very narrow pulse introduced at the front end of the
active line. If there were no parallel passive line nearby, this pulse
would travel down the active line (at the speed of light for the
dielectric) more or less unchanged,” in a TEM mode. “However, as the other
two traces show, the presence of the passive line caused the original
narrow pulse to break up into two similar pulses.” In figure 9.3 we see
that the two pulses are of opposite polarity on the passive line, and
initially were superposed. Thus, two energy currents of opposite polarity
were travelling together in the same direction at the same point, close to
the passive line. Had there been uniform dielectric throughout, with conductors
buried between voltage planes, these two opposite energy currents in the
vicinity of the passive line would continue to travel together. This leads
to my new insight gained today, which is that a region of space can
accommodate two energy currents of opposite polarity at the same point in
space.
If we attach a 500 ohm
oscilloscope probe to the inner and outer of a coaxial cable, it has a
series 450 ohms at the probe point followed by a 50 ohm coaxial cable
terminated by 50 ohms at the other, oscilloscope end. If a 10 volt pulse
travels from left to right down the coax cable, a 1 volt pulse will travel
up the probe cable to the oscilloscope, where it will be represented as 10
volts. Now if the 10 volt pulse came from the right, an identical 1 volt pulse
will be picked off and travel to the oscilloscope. At the probe point,
information as to the direction of the 10 volt pulse is lost. If two
pulses, one from the left and one from the right, overlap, the oscilloscope
will receive a 2 volt pulse, and register 20 volts. If however one of the
two pulses had opposite polarity, the oscilloscope will register zero
volts.
Neither the 20 volts nor
the zero volts represent the real situation at the probe point. The reality
is that at that point, one 10 volt signal is travelling to the right, and
another to the left. The oscilloscope’s 20 volts or zero volts are
creations of the probing system, which neglects direction. The reality is
that when the two pulses overlap, they remain two 10 volt pulses, and
should not, and cannot, be mixed together. There is no 20 volts and there is no zero volts.
After centuries of
brainwashing we are trained to think that there exists a single potential
difference between the inner and outer of a coaxial cable. The truth is
that there are two voltages, one associated with a TEM wave travelling to
the right, and a totally independent one associated with a TEM signal
travelling to the left. Also, they have a magnetic field just as much as
they have an electric field. We ignore the magnetic field value because it
is much more difficult to measure, but this should not have been allowed to
make us neglect it.
Ivor
Catt March 2010
|
