[evlatests] P-Band Tsys, for real

[Rick Perley]

    Eric has already made a change to FILLM, so I can get backend Tsys 
for all EVLA antennas without having to run FILLM twice.  So I can now 
report on the (backend) Tsys values at P-band:

    Antennas 14 and 23 give non-varying Tsys values, near 10K.  (If it 
were only true ...)
    Antenna 21 gives realistic looking value on IFs A, B, and C, but no 
values at all (not even constant ones) on IF D.  The values on A, B, and 
C are many thousands of K. 
    Antenna 17 is the only one giving really sensible data.  (And when 
IFs B and D went out halfway through my run, the Tsys values jumped to 
many thousands of K).