Port measurement of the Receiver

Dear all,

I have a receiver included LNA+downconversion mixer+VCO. It just has RF input and the IF output.
Actually, how could I measure the port isolation between RF-IF and LO-IF and RF-LO?
I have two ideas to do this. First method, using the network analyzer, connect port 1 to the RF and port 2 to the IF. S11 measures the input RF impedance and S21 to measure RF-IF isolation.
Second method, using the signal generator and feed the signal to the RF input port. Using spectrum analyzer to measure the output frequency and see the different power between the input and the output to get the isolation.
Actually, which method is correct and get more accurate isolation value?
On the other hand, how could I measure the LO-IF isolation? It is because the LO is generated by the VCO directly.

Best Regards,

wccheng :roll:

1 method is wrong : you can't measure RF - IF isolation with VNA because the input and the output frequency are different.

Sorry, I got mixed up !

2 method: you will have more accuracy if you use a power meter instead of the spectrum analyzer.

LO-IF isolation - disconnect the VCO and then apply the 2 method.

Dear marcomdd,

Are you mean that I could feed the RF (1.8GHz) to my circuit's RF input port, and then use the spectrum analyzer (sorry my lab does not have power meter ) (which is connected to the IF output port) to see the power of the frequency at 1.8GHz? Then compared what is the power drop in order to find out the RF-IF isolation?

Moroever, my VCO output is connected directly in the input of the LO mixer input in the IC. How could I measure LO-IF isolation?

Best Regards,

wccheng

Sorry I got mixed up (with losses).

Use the VNA, I think it is more accurate than Generator - Spectrum Analyzer.

When you perform full two port calibration, calibrate for the isolation of the VNA also .

wccheng said:

Dear marcomdd,

Moroever, my VCO output is connected directly in the input of the LO mixer input in the IC. How could I measure LO-IF isolation?

Click to expand...

Do you know the nominal value of the lo power ?

You can do this with a VNA but it is tricky to calibrate since the VNA drive and measured frequency are the same. With a mixer in the line, the measured frequency is not the same as the drive frequency. I think for your use it is easier to use the signal generator and spectrum analyzer. Apply the RF input signal. Measure the IF output signal. Peak search the marker on this and then delta marker to the RF signal leaking through. That is your RF-IF isolation in dBc. Peak search the marker on the IF signal again and then delta marker to the LO. This is your IF to LO isolation in dBc. Remember that the IF signal is your desired output and any other signal is unwanted.

Generally, the importance of isolation measurement don't call for hi accuracy.
The VNA is accurate but it is sensitivo to its signal only.
A signal generator and a Spectrum Analyzer is everytime the best way.
Because you not only measure isolation but also you may create yoursefl an idea of harmonics products and spurious products.
To measure RF>IF path, exite RF port with all RF, Image, and IF Freq. and receive with SA at the same freq exited.

If the BW of the converter is large enough, you don't need Sintetized instruments.

sergio mariotti said:

The VNA is accurate but it is sensitivo to its signal only.

Click to expand...

What do you mean ?

The VNA (if it is a tracking generator and a Spectrum Analyzer type VNA, which is often the case) will be receiving the signal that it puts out, and no other signals (ie. a tuned detector.) So, using a tuned type VNA (or a spectrum analyzer with a tracking generator, effectively the same thing) will give you rejection of most unwanted products from your measurement. However, you do have to watch out that your input to the measuring device (whether VNA or spectrum an.) is not being overloaded because it can generate products at the same frequencies as the ones you are trying to measure, so sometimes a filter at the input of your measuring device (to select only the signals you want to measure) is a good idea.