[evlatests] 4/P Converter Noise Figure

[Paul Harden]

There are three newly built T301 4/P Converters ready for EVLA 
installation.  I performed noise figure (NF) measurements on these three 
to determine the average NF and the spread in NF between units.

RESULTS:

Average T301 NF, 74MHz   10.4 dB;  best 10.2, worst 10.6
Average T301 NF, P-band  10.7 dB;  best 10.5, worst 10.8
This includes both LCP and RCP.

The NF of these 3 T301s are very consistent with eachother, varying only 
a fraction of a dB, suggesting the nominal 10.5dB NF is typical of all 
units.

BANDWIDTHS:

There is a bandpass filter on the output of the 4/P Converter, from 
about 1050-1450 MHz IF (about 70-420MHz RF input), encompassing the 
composite 74MHz, 196MHz and P-band signals.  Much beyond the 3dB limits 
of these filters, the NF quickly exceeds 25dB.

The L-band IF output of the 4/P Converters is upconverted to X-band IF 
by the T302 LSC Converter.  There are no band limiting filters in the 
T302 except the 8-12 GHz system IF filters prior to the downconverters. 
  The LSC Converter is a 1-8GHz wideband converter, regardless of the 
band selected. Thus, the 10dB NF region for 4/P band is a narrow portion 
of the 8-12GHz IF.  I don't know how the 1050-1450 MHz segment is 
selectively chosen to maintain this noise figure.

SPECIFICATIONS:

The EVLA Project Book specifies the NF for the 4/P Converter to be 9dB. 
  This figure was established during the PDR.  Since then, an additional 
25dB of gain has been added to the 4/P Converter to compensate for the 
lower than anticipated powers from the 4 and P-band receivers, no doubt 
contributing to the additional 1.5dB of NF.

I hope this might answer a question or two regarding the sensitivity and 
noise figure issues with the EVLA 4/P system.  If not, let me know.

Paul Harden