Datasheet

Page 11
Find us at www.keysight.com
Noise multipliers
The measurement noise for the X-Series power sensor is dependent on the measurement mode and the time for the measurement.
In general, average only mode is lower noise than normal mode, and the longer a measurement takes the lower the noise is. We will
define three measurement modes and how the noise can be adjusted.
Average-only mode
The measurement noise due to the X-Series power sensor is dependent on the measurement time. In general, the longer a
measurement takes the lower the noise is. The measurement noise specification is defined for 16 averages with an aperture of 50 ms,
or a total time of 800 ms. Noise will reduce or increase with the square root ratio of the measurement time to the specification
measurement time. Thus a noise multiplier factor can be derived for any combination of averaging and aperture:
——————
0.8
N
mult
=
√
——————
N
ave
× t
a
Increasing measurement time will reduce noise at this rate until around 3 seconds. As the measurement time increases beyond 3.2
seconds the noise reduction exponent changes from 0.5 to 0.2.
N
mult
= 0.89 ×
(
1
)
0.5
, for N
ave
× t
a
≤ 3.2
N
ave
× t
a
N
mult
= 0.63 ×
(
1
)
0.2
, for N
ave
× t
a
> 3.2
N
ave
× t
a
Noise
actual
= N
mult
× Noise
spec
Where N
ave
def
number of averages and t
a
def
aperture in seconds.
Free-run normal mode
The measurement noise specification is defined for 1 average. Although the noise will reduce with increased averaging, it will not have
a significant impact on the measurement uncertainty, and the figure of 32 nW (High/Off VBW) or 10 nW (Low/Med VBW) without any
multiplier should be used in the uncertainty calculations. (Refer to the measurement noise in the noise and drift table above.)
Gated-average normal mode
The measurement noise on a time-gated average power measurement in normal mode will depend on the time gate length. 20
averages are carried out every 1 µs of gate length. The noise-per-sample contribution in this mode can be reduced by approximately
to a limit of 32 nW. (Refer to the noise and drift table above for the noise-per-sample.)