[evlatests] initial cable delay prescription for EVLA antenna

[Ken Sowinski]

                   Initial Path Length Estimate
                         March 31, 2006


We now have enough examples that it is worth inventing a scheme
to generate cable delays for SYSPATHF from the OTDR measured
path lengths between control building and pad. 

Tom Baldwin has provided a compilation of measured path lengths at:
/home/evla/techdocs/fiber/Cable_Plan/Fiber_Lengths/

There have now been EVLA antennas on seven different pads and
we can compare the cable delay required to obtain fringes to the
measured lengths.  Seasonal variations should not be more than
a few nanoseconds over the short runs used so far.  

The table below presents what we know so far.  All path lengths
are in nanoseconds.  Measured lengths are converted back to time
using an assumed index of refraction of 1.466.  'Delay' is the
cable delay required to optain fringes, 'OTDR' is the measured path
length, and the last column is the amount to be added to the measured
path length to get the required delay.  Path lengths are negative
because of technique used to calculate a total delay to present to 
the correlator.

Pad    Delay     OTDR     Diff
---    -----     ----     ----
CW5    -5102    -4144    -1045
MP     -7035    -6054     -981
AW5   -41380   -40253    -1127
DN6    -5941    -5051     -890
AN2   -12105   -10939    -1176
CE7    -9836    -9262     -574
AE3   -19548   -18587     -961

Clearly CE7 is abberant, perhaps there was a transcription or
measurement error.  For all other pads an appropriate prescription
for an initial trial cable delay is to just add a microsecond to the
OTDR measurement expressed in nanoseconds and negate the sum.  This
value should be entered in SYSPATHF for the corresponding pad and
should get the delay close enough that the correct delay can easily be
found in a spectral line observation as long as the antenna is working
in all other respects.