Measuring the Noise Factor of an Amplifier
Next to the gain, OIP3 and 1dBc the Noise Factor (NF) is an important measure of an amplifier so we like to know the NF of the amplifier we are using.
The NF is the degradation in dB of the SNR after the amplifier compared to before the amplifier. This implies we have to measure noise levels.
For this measurement we need some attenuators and two amplifiers, the LNA and the amplifier to be tested (DUT). The LNA is needed because the noise factor of the tinySA is too large to directly observe the noise contribution of a single amplifier.
The first step is to normalize the tinySA to the level of the test signal we want to use to measure the gain of the two amplifiers cascaded. We do this by connecting some attenuators with total attenuation approximately (+/- 10dB) equal to the total amplification of the two cascaded amplifiers to the cal output. We set the cal output to 30MHz (or any other frequency we want to use) and we observe the cal output level attenuated through the attenuators. Make sure the tinySA internal attenuation is 0dB. Then we set the EXTERNAL AMP to the same value as the observed signal level. After that the signal should read as 0dBm.
The measurement setup is like this
And the measured value after normalization looks like this (its +0.3dBm i.s.o. 0dBm)
Now we insert the amplifiers between the attenuators and the input of the tinySA as can be seen here
And we can directly read the amplification of the cascaded amplifiers under the marker (36.8dB)
Now we make some changes
- set the EXTERNAL AMP to the measure amplification (36.8dB)
- decouple the the attenuators from the input and terminate the first amplifier input with 50ohm (you can also decouple the cal output from the attenuators and use the attenuators as 50ohm resistor)
- set the noise attribute of the marker so its value is displayed in /Hz and independent of the active RBW.
- set DISPLAY/CALC/AVER 16 to get maximum averaging
The noise measurement setup looks like this
and we can read the vale of the noise marker (-168.5dBm/Hz)
To calculate the noise factor of the first amplifier we simply subtract -174dBm/Hz (the noise power per Hertz of a 50 ohm resistor at room temperature) resulting in a noise factor of 5.5dB. The spec of the first amplifier gives a NF of 4.5dB so we are close.
This measurement method assumes the gain of the first amplifier is sufficient to mask the noise of the second amplifier and the two amplifiers will mask the internal noise of the tinySA so we can safely ignore the noise of the second amplifier and the tinySA. To ensure this masking happens the total amplification of the cascaded amplifiers has to be above 30dB.
We can now also measure the noise factor of the LNA by interchanging the DUT and the LNA, put the LNA first and connects its output to the DUT. As the amplification stay the same we can directly measure the noise level with the noise marker. The measured noise factor of the LNA is between 0.5dB and 1dB which is consistent with its specification.