The receiver equivalent noise temperature has been determined in the
standard hot and cold loads experiment in front of the receiver. The
best receiver sensitivity has been achieved at the telescope
(2850m altitude) where the physical temperature of the mixer was 4.2 K, about
0.5 K lower than in the laboratory at sea level. At the telescope
the cold load temperature drops, with the liquid nitrogen temperature,
to 75 K.
For noise measurements,
the output receiver power is integrated over the whole IF band
(500 MHz wide).
The best measured value of the Y factor in this experiment was about 3.
It corresponds to a receiver noise temperature of 30 K (
level).
Expressing this noise (or sensitivity) in photon numbers we get down to
the four--photon level. Minimum fixed tuned receiver DSB noise
temperatures are presented in Fig. 5. In Fig. 6
the receiver
sensitivity is given in photons. The black parts of the curves
in these figures correspond to the frequencies supported by the
existing LO remote control unit (this year the band available for
the observations was 320-360 GHz). For the other frequencies the
local oscillator has been tuned manually.
An example of the receiver DSB noise behavior in the intermediate 1250-1750 MHz frequency band is given in Fig. 7 for the 325 GHz LO. This curve was measured at the radio telescope with the autocorrelator backend, and with the calibration loads directly in front of the receiver. The steps in the curve are due to slight differences in the noise behavior of the correlator units.
The mid--band DSB noise temperature is about
32 K in this measurement, and the receiver noise in the 500 MHz IF band is
rather constant (variation of 
K).
The output power stability is better than 
over timescales of 1s.
The spectrum of low frequency instability in the IF power is clean
(no periodical modulation, with the exception of 50 Hz).