[evlatests] Dipole Loss -- a summary

[Rick Perley]

    I reviewed more carefully today the loss in correlation coefficient 
due to the 4-band dipoles.  A memo with the data will be issued 
shortly.  I give here the essential results:

    A) Polarization. 

    The dipoles introduce a ~1% variation in the cross-polarization at 
L-band.  This means that instead of the ~0.1% stability in the amplitude 
of the X-polarization leakage,  we are seeing temporal variations of 
order 1%.    The variation timescale is or order 1 hour. 
    There is no effect of the dipoles on C or X band polarization. 

    B) Antenna G/T. 

    a)  LBand:  The antenna G/T is reduced by 5% at elevation 55 
degrees, increasing to 10% at the zenith.  This is not a system 
temperature effect (as these were identical on the two days), so must be 
due to loss of forward gain ('efficiency'). 

    b) CBand:  The antenna G/T is reduced by 8%, with no discernible 
elevation dependence. 

    c)  XBand:  The antenna G/T is reduced by 6%, with no discernible 
elevation dependence. 

    In hindsight, I should have measured the correlation coefficients at 
K and Q bands as well ...

    C) A global change in G/T. 

    There were two EVLA antennas  -- #4 and 24 -- available the 2nd day 
which did not have the dipoles on them.  It was found, at all three 
bands, that the correlation coefficients for both antennas decreased by 
3 to 4%, compared to the first day.  This decrease was *not* seen on 
*any* VLA antennas, and indeed, no change at all in correlation 
coefficient was noted on any VLA antenna between the two days at any band. 
    I am at a loss to explain this (and only fall back on the weak 
position of a 'small sample of two'). 
    The dipole-induced losses noted in (B) above have been corrected for 
the apparent 'global' EVLA loss.