User`s guide

18
Signal-to-Noise Ratio
The accuracy of the measurement
is directly proportional to the
signal-to-noise ratio (SNR) of
both signals. For best
measurement results, the
signals at both inputs to the
HP 71500A should be as close to
(but not above) 0 dBm as
possible. Tables 4, 5, and 6 show
the accuracy degradation due
to sub-optimum RF input signal
levels.
When using the FFT zoom
mode, choosing the modulation
index to maximize the
measured sideband levels is
also important (see step 2 on
page 14).
When measuring a DUT with a
frequency response
characteristic of a filter, the
group delay accuracy will degrade
when measuring the filter skirt,
due to a smaller output signal
level.
The bandwidth of the DUT will
also affect the measurement.
The narrower the bandwidth,
the more accurate the
measurement will be due to less
noise. Similarly, a DUT with
excess noise will degrade the
measurement accuracy. DUT
noise may be in the form of LO
phase noise or broadband
amplified noise.
Mismatch Error
Impedance mismatch in the
system between cables, the
DUT, and test equipment causes
group delay, phase and
amplitude ripples versus
frequency. Error caused by
mismatch can be difficult to
eliminate since normalization
only removes mismatch error
present without the DUT in
place. If the DUT has poor port
matches, the mismatch error
introduced when the DUT is
added will be measured as part
of the DUT’s response.
Judicious use of attenuators
and isolators will improve port
matches, insuring that the
measurement contributions of
the test setup are minimized.
Mismatch error can occur
within the DUT as well. For
this reason, a DUT consisting of
a filter and a mixer may have
significantly worse group delay
performance as compared to the
filter alone, due to the
interactions between the filter,
mixer and internal connections.
LO Stability
This measurement application
assumes that both the internal
LO and the RF source used as
the stimulus are of high quality
(i.e., typical synthesizer
stability and phase noise).
While the LO needs to be stable,
its exact frequency need not be
known initially. The Find LO
routine can be used to determine
the LO frequency with sufficient
accuracy to perform a group
delay measurement.
Group Delay vs. RF Input
Power
RF=1 GHz RF=20 GHz
0 dBm +/- 275 ps +/- 1.1 ns
-10 dBm +/- 300 ps +/- 1.5 ns
-20 dBm +/- 350 ps +/- 2.0 ns
-30 dBm +/- 500 ps +/- 2.5 ns
-40 dBm +/- 1.2 ns +/- 3 ns
-50 dBm +/- 4 ns +/- 7 ns
Table 4: Typical group delay
accuracy versus RF input power
with no averaging. Zoom size = 16.
Excludes mismatch error.
Phase and Amplitude vs.
RF Power
RF = 1 GHz
Phase Ampl.
0 dBm +/- .4° +/- .1 dB
-10 dBm +/- .5° +/- .1 dB
-20 dBm +/- .7° +/- .1 dB
-30 dBm +/- 1.0° +/- .1 dB
-40 dBm +/- 1.5° +/- .1 dB
-50 dBm +/- 10° +/- .25 dB
Table 5: Typical phase and
amplitude accuracy versus RF input
power with no averaging. Zoom
size = 16. Excludes mismatch error.
Phase and Amplitude vs.
RF Power
RF = 20 GHz
Phase Ampl.
0 dBm +/- 1.0° +/- .15 dB
-10 dBm +/- 1.5° +/- .2 dB
-20 dBm +/- 2.0° +/- .2 dB
-30 dBm +/- 2.5° +/- .2 dB
-40 dBm +/- 4° +/- .3 dB
-50 dBm +/- 10° +/- .5 dB
Table 6: Typical phase and
amplitude accuracy versus RF input
power with no averaging. Zoom
size = 16. Excludes mismatch error.