Printer FriendlyEmail Article Link

Why are my results for WhiteNoise @-140dBm/Hz not as expected, I see a measurement of -138.6dBm/Hz to 139.6 dBm/Hz?


Measuring white noise might not be as easy as we thought.

If the measured white noise was -140 dBm/Hz to 139.6 .. 138.9 dBm/Hz, the measurement might be reasonable.
Lets assume that the user was loading white noise with 32 MHz sampling frequency.
The Settings for the Spectrum Analyzer in this case we typically use a Agilent 4395A
Analyzer is set to measure noise
Start frequency 10 KHz
Stop Frequency 16 MHz
Video BW 10KHz
Background noise of the analyzer: -146 dBM/Hz
Averaging factor 100
We could see that from the low frequency to 10 MHz area, the measurement might be at -139 to -139.5 dBm, after that it could be -140 dBm/Hz or lower.
When we see the result on the analyzer say -139.0 dBm/Hz, that was the sum of BGA (back ground noise of the analyzer) and the white noise.
Then the Sum of the WhiteNoise and BGA=10xlog10(10^(-140/10)+10^(-146/10))=-139.03 dBm/Hz
therefore, if we got the measured result of -139.0 dBm/Hz, we can say the result was good.
In the high frequency part, we might get -139.5 dBm/Hz or lower, the loss might be from the transformer, the loss from the transformer could be from 0.1 to 1 dB from low frequency to high frequency.
This is dependent on the type of balun used for the measurement.
The loss from Spirents Noise Generator like the DLS5500/DLS5405 or the DLS5800/DLS5410DC systems,  from the low frequency to high frequency will be less than 0.5 dB.

Product : Injector,Noise Generator,Noise Files