Specifications

12
The cascade equation shows how F12 is very sensitive
to uncertainty margins in the second stage term
[(F2 - 1) / G1]. (To see how F12 would vary with marginal
changes in F2 or G1, see reference 6.) If the DUT has
insertion loss (e.g. a mixer, attenuator, etc.), use a low
noise pre-amplifier before the instrument to reduce
the uncertainty margin. Choose a pre-amp with the
lowest noise figure and a gain of more than 100 (20 dB +).
(See Agilent Technologies Application Note 57-2 for more
information on selecting the pre-amp.) Make the noise
figure of the second stage as low as practical and the
uncertainty of F12 (and hence F1), as low as possible.
Adding a pre-amp also gives the measurement some
resilience against noise figure variations versus frequency
of the second stage.
In the case of a high-gain DUT, there may not be a need
for a pre-amp. In order to make that decision, place the
DUT’s linear gain and the measurement system’s linear
noise figure (i.e. noise factor) into the Cascade Equation.
Notice the noise figure of the cascade will converge to the
noise figure of the DUT.
Figure 7-1.
● HINT 7:
Use proper measurement correction
Take the following steps to ensure the measurement
system itself does not add error to the measurement.
• Remove the noise figure of the measurement
system with regular user calibration.
• Avoid exceeding the maximum input power of
the measurement instrument. Modern instruments
can handle around 65 dB of device gain for narrow
band devices. For wider band devices with high gains
it is likely that an attenuator will be required after the
device to keep the overall power within the
instrument’s range. Use the analyzer’s compensation
feature to account for the losses of the attenuator.
Use appropriate filters/isolators/circulator to suppress
out-of-band responses that would otherwise
contribute noise power at the high gain level and
overpower the instrument’s input.
A Y-Factor noise figure analyzer measures the noise figure
of the measurement system and the DUT combined (see
Figure 7-1). Below is F12 in the cascade equation:
F12 = F1 + [(F2 - 1 ) / G1] (7-1)
F1 and F2 are linear noise figure values for the DUT and
the measurement system, respectively, and G1 is the
gain of the DUT. User calibration (termed “second stage
correction”) determines F2; measurement determines
F12 and G1. The analyzer calculates F1 from the cascade
equation.
Perform user-calibration prior to the measurement to
remove the second stage contribution. Calibrate out the
second stage contribution at regular intervals depending
on how sensitive the noise figure and gain are to
temperature drifts.
Noise
source
Measuring
system
DUT
Gain, G
Calibration path
Measurement
path
F2 = noise figure of the measurement system
F1 = noise figure of the device under test (DUT)