SL2015
Measurement of Phase Noise.
This is best measured by looking at the IF output of the
SL2015. The IF should be fed to a spectrum analyser, where
it can be interpreted.
There are two common methods of doing this:
a) using phase noise analysis software (such a HP85671A
phase noise program)
b) direct measurement of the noise floor at the chosen offset
frequency, and conversion to a dBc/Hz figure.
Since method a) will depend on the software used, a
description of method b) will be given only.
To measure phase noise at 10kHz offset:
a) tune the centre frequency of the spectrum analyer to the
IF - e.g. 479.5MHz
b) Set the span initially wide (10MHz or greater). Gradually
reduce this until it is set to 50kHz or less, taking care to ensure
that the centre frequency of the display matches the IF peak.
c) perform a peak search
d) set marker delta to 10kHz
e) set video averaging ON to ensure that a representative
measurement of the noise floor at the chosen offset frequency
is made.
f) record the level of noise at the 10kHz offset compared to
the peak IF level (in dBc).
Care must be taken to ensure that the LO is stable, since any
instability will reduce the averaged peak LO value, thus giving
a falsely low phase noise reading.
g) convert the measured reading to a 1Hz bandwidth.
e.g.
A measured phase noise of -50dBc/1kHz bandwidth
(RBW of 1kHz) corresponds to -80dBc/Hz.
Since noise floor must be reduced by the ratio of the two
bandwidths
i.e. 10 log 1kHz/1Hz = 30dB.
Measurement of Conversion Gain
(from a 50
Ω
source)
a) Connect an RF signal generator to the RF input to the
SL2015.
b) Connect an IF output to a spectrum analyser.
c) Feed the SL2015 with the appropriate signal level,
depending on AGC setting, required output, etc.
d) Note the relative difference in the input and IF level in dB.
This is the conversion gain of the device.
For increased accuracy, the input signal level should also
be checked with a spectrum analyser, since any level
measurement errors that exist within the analyser will then be
relative, rather than literal.
The AGC voltage may be varied and conversion gain
measured at different AGC voltages.
9
MITEL [ MITEL NETWORKS CORPORATION ]
