LNA gain measurement

[huytergan]

I know when I have two port LNA, I can use VNA to get its gain i.e S21.
However, how can I measure its S21 when I terminated its input? For example, if we have a board that includes a patch antenna at the input and its ground plane soldered to the PCB ground and fed with via to PCB top. So there is nothing to be connectorized at the input, only one output connectorized at the board, but board includes chip LNAs on the path. So I have only one out and zero input for VNA. How the LNA gain is going to be measured?

 

[FvM]

If you consider the design an active antenna and don't care for LNA properties as such, you can measure sensitive in anechoic chamber compared to a reference antenna with known gain, e.g a dipole.

If you want to know LNA parameters whithout patch antenna, you'll disconnect the antenna from LNA input and connect a feed cable. This gives you the opportunity to measure complete LNA s-parameters.

 

[huytergan]

If you consider the design an active antenna and don't care for LNA properties as such, you can measure sensitive in anechoic chamber compared to a reference antenna with known gain, e.g a dipole.

If you want to know LNA parameters whithout patch antenna, you'll disconnect the antenna from LNA input and connect a feed cable. This gives you the opportunity to measure complete LNA s-parameters.

Wanna know LNA parameters. Aint there another way to measure LNA S-parameters? For exp; some guys advises that “terminate LNA input and measure the output” How is then possible? we have 1 port left again for VNA? Because assuming the antenna is not connectorized to the input of the board. lets say It’s soldered to the point that feeds the antenna, is the start point of the receiver signal trace.

 

[FvM]

some guys advises that “terminate LNA input and measure the output”

Gives s22 but nothing else. Need a coaxial cable that's small enough to be soldered to LNA input and perform VNA calibration to appropriate reference plane, e.g. cable end.

 

[BigBoss]

LNA will have following specifications.
-Input and Output Reflection Coefficients (S11 and S22) that will be measured by 50 Ohm VNA
-Power gain that is equal to S21 of whole LNA.
-Noise Figure that must be measured by Noise Figure Meter that has 50 Ohm Output Impedance
So either the antenna is connected OR disconnected. You have to measure this LNA under 50 Ohm Measurement Environment.
There is not other option to be considered.
If that LNA is connected to a Receiver part with demodulation feature, you can measure by inserting a modulated signal the other metrics such as BER, EVM.

 

[huytergan]

LNA will have following specifications.
-Input and Output Reflection Coefficients (S11 and S22) that will be measured by 50 Ohm VNA
-Power gain that is equal to S21 of whole LNA.
-Noise Figure that must be measured by Noise Figure Meter that has 50 Ohm Output Impedance
So either the antenna is connected OR disconnected. You have to measure this LNA under 50 Ohm Measurement Environment.
There is not other option to be considered.
If that LNA is connected to a Receiver part with demodulation feature, you can measure by inserting a modulated signal the other metrics such as BER, EVM.

But the problem I ask is not about 50 ohm measurement. how 1-port LNA’s gain measured If the antenna input is soldered to the signal path without any connector. I think, in such a case, I should desolder the antenna, and solder a coax cable which has connector at its other end, then I will have 2 ports now I can measure the gain like FvM said in previous answer.

 

[BigBoss]

But the problem I ask is not about 50 ohm measurement. how 1-port LNA’s gain measured If the antenna input is soldered to the signal path without any connector. I think, in such a case, I should desolder the antenna, and solder a coax cable which has connector at its other end, then I will have 2 ports now I can measure the gain like FvM said in previous answer.

Absolutely,
You have to obtain 2-ports to measure s-parameters and Noise Figure. It's obvious.