The signal from the pulse generator appears on channel 1.  Note that the signal is balanced with
respect to ground.  The signal is transformer coupled. The signal would be distorted if it were
not symmetrical about ground.  The output signals faithfully reflect the input signal with a bit of peaking.
The picture above is with no current in the twisted pair. The picture below has a 50 mA current in the
twisted pair.  Note that the signal quality is still good, but that the slope (sag) has nearly doubled, but
appears to be within acceptable limits.


 Click for a figure made with 100 mA into the pair. The goal is to reduce the DOM power to 2.5W or
less at approximately 100V. So, 100 mA is a worst case example.

Suppose there is a power supply failure in one of the two DOMs on a cable pair.  Then all the current in the
magnetic components flows through one half of the power splitter (a coil), with the possible consequence
of causing core saturation which would disrupt the signal waveforms.  The simulation of this error produces
a negligible degradation of the quality of the communications signal.  The current drops from 100 mA to
75 mA as expected for a fixed voltage source. The resultant waveforms are pictured below.


Signaling goes from surface to DOM, and from DOM to surface.  The next figure shows the communications
waveforms when one simulates the DOM transmitting, while observing the pair connected to the surface and
at the other DOM.  Note that the isolation of the power splitter depresses the DOM signal (Ch 2 send)  when observed
by the other DOM (Ch 3 receive)

The above waveforms are with no net current in the twisted pair.  The waveforms below are with
a 100 mA signal in the twisted pair.  Note the scale change on Ch 3.  The amplitude of Ch 3 is 20%
the amplitude of Ch 2 at twice the current anticipated for the system.

Communications thresholds in the DOMs of mid level +/- 20% to +/-30% should result in
acceptable communications performance.