Power Divider Performance Testing

[luckyali]

Hallo every one,
I have designed a power splitter for 20 Ghz, the design has been fabricated and fitted into its housing.
In order to compair its performance results from simulation with actual, I have to test this device. Can any one guide me how to do this testing with Lab equipments to get real results...

many thanks in Advance
wishes for a nice day

 

[johnjoe]

Hi,

it depends on what you have, i mean what kind of measurement equipment. Tell us! Than it's easy to answer.

 

[volker_muehlhaus]

If you have a two port network analyzer, you can measure these paths:
- input to output 1 (with port 2 connected to 50 ohm load)
- input to output 2 (with port 1 connected to 50 ohm load)
- output 1 to output 2 (check "isolation", input connected to 50 ohm load)

 

[luckyali]

yes i do have access to a network analyzer, can i get a tutorial for how to make measurements with network analyzer...

 

[eraste]

Hi, try this :
https://www.radio-electronics.com/info/t_and_m/rf-network-analyzer/analyser-basics-tutorial.php

 

[volker_muehlhaus]

can i get a tutorial for how to make measurements with network analyzer...

No manuals?

The details depends on the VNA manufacturer and model.
Try to find someone in your lab who can show this to you, for your instrument.
Don't forget to talk about VNA calibration also, because this VERY important to get accurate results.

 

[luckyali]

I tried to Use VNA from Agilent Company, and I got some Results, I saved them as *.csa format.
In order to redraw the results image, How Can I read this file on my pc...

 

[volker_muehlhaus]

Wrong file - the *.csa is the instrument calibration and instrument state.
You need to save measurement results in Touchstone (s2p) file format.

 

[KRISz]

You save file s2p from VNA and can open file form ADS simulation.

 

[luckyali]

If you have a two port network analyzer, you can measure these paths:
- input to output 1 (with port 2 connected to 50 ohm load)
- input to output 2 (with port 1 connected to 50 ohm load)
- output 1 to output 2 (check "isolation", input connected to 50 ohm load)

I have measured the divider circuit and saved the network analyser results.. here you can check them in a pdf fomate, all of the parameters are inside S11, (S12/S21), (S13/S31), S22, (S32/S23),S33.
View attachment S-Parameters.pdf
What can you comment about these results...

 

[volker_muehlhaus]

What can you comment about these results...

It looks like you need to re-design the divider ...

 

[luckyali]

oh no, can you give some suggestions for improvement bcoz i am not an expert in this thing, I just read from book and tried to design it...

 

[eraste]

If i suppose that 1 is the input, 2 and 3 the outputs of your splitter.
For a splitter, transfert from input to ouputs must be -3dB. It's not your case. Return loss seems to be to high (about -12 dB and - 8 dB). Isolation between out port seems to be low. Effectively, you have to re-design it. What technology do you use ? Microstrip ? cable ?

 

[luckyali]

yes I am using Microstrip. it was designed on Roger, RO4003C, 8 mil thickness and 17.5 um copper cladding... here is the picture to structure... input is at 3 where 2 and 1 are ouputs...
Since I was using a 2 port VNA so i renamed the pics saved from VNA according to my design...**broken link removed**