8.4 Sensitivity
Please note that the noise floor and sensitivity
can vary widely
with spectrum
analysers, depending on frequency. They can be quite different even with only
small frequency changes of a few 100 MHz. In general, however, the following
rule applies: The higher the frequency, the higher the noise floor and thus the
worse the sensitivity. In practice this means:
You can measure significantly
weaker signals at “2GHz” than at “6GHz” with the same filter settings, etc.
8.5 Measurement inaccuracy
Aaronia states a
typical
accuracy for each SPECTRAN model. However, this
means that higher deviations are entirely possible. Especially when you
approach the so-called noise floor or the maximum sensitivity of a spectrum ana-
lyser, accuracy decreases
by principle
.
Even though SPECTRAN devices offer an impressive sensitivity considering
their price, a yet higher measurement inaccuracy than the base device itself has
is assumed in practice: The reason behind this are further factors which can
influence measurement inaccuracy: Some of them are: temperature dependen-
ce, repeatability, antenna errors, modulation dependence etc.
According to BUWAL (see also www.BUWAL.ch), one should
always
expect an
extended measurement inaccuracy of at least 35% even with absolutely profes-
sional equipment. So, expressed in dB, you should always expect a rounded
+/-
3dB
and consider this in every measurement!
Example: You measure a value of -45dBm. So, the actual result could, according
to the mentioned measurement inaccuracy, vary between -42dBm (-45dBm
+3dB) and -48dBm (-45dBm - 3dB). Thus, as you want to measure
exposure
due to radiation, you should always assume the
highest
value, namely -42dBm.
In simple words:
Always add another 3dB to the measurement inaccuracy of the SPEC-
TRAN device with all measurements.
8.0 Correct measurement
33
Содержание SPECTRAN HF-2025
Страница 72: ...23 0 For your notes 72...