Dual frequency microstrip power combiner

[ingv]

microstrip power combiners wiki

Hi everyone,

I'm interested in designing a microstrip power combiner to connect two modules to one antenna (only one module will be connected at a time). The idea is to have just one antenna connector on the PCB regardless the module I'm using at the moment. I've been looking Wilkinson's power divider and T power divider, the thing is that the modules operate at different frequencies.

Is it possible to make some kind of dual frequency Wilkinson or any other kind of power combiner? Is there any book or paper you recommend? I'm thinking in making this because I believe it could be a problem if I have a microstrip ramification even if it has nothing connected to the other end, am I right?

Thanks in advance for any help you can provide me.

Regards,
Victor

 

[RF-OM]

If you connect antenna to the splitter you will have two antenna connectors at the splitter's outputs anyway. Design of Wilkinson or any other splitter/combiner will depends on your requirements. What is the frequencies for each signals? Please keep in mind that splitter will have at least 3.5 dB of loss and you need to terminate unused splitter's port.

 

[ingv]

Hi RF-OM, thanks for your answer.

What I want to do is to make one PCB which sometimes will have a Zigbee (2.4 GHz) module and sometimes a GPS (1.575 Ghz) module, but I would like to have just one SMA, so the device case won't vary.

Maybe is not a power splitter what I'm looking for, I was thinking in something like the image attached, would it work fine? How can I terminate the unused splitter's ports, a 50Ω resistor would do it? λ2 corresponds to Zigbee's frequency and λ1 to GPS' frequency.

Regards,
Victor

 

[RF-OM]

Your picture looks like Wilkinson combiner, but in this case two lines to outputs must be 70.7Ohm and be quoter wave by length. I think it is possible to design such Wilkinson combiner-splitter but you will lost about 3.5 or 4 dB. I would rather use diplexer with outputs for 2.4CHz and 1.575GHz. In this case you will lose only about 1 or 1.5 dB. Let me know if you need some paper on diplexer or Wilkinson design. In any case unused port must be terminated. Resistor on the board should be okay, but two in parallel but looking in opposite directions is better.

 

[ingv]

Thanks again.

My idea setting it to λ/2 was to try to make it look like an open circuit for each other frequencies, isn't that possible? I think I'll follow your advice with the diplexer so I'll appreciatte if you can send me the papers or the links to the papers.

Why is better two resistors in parallel but in opposite directions?

Thanks for your help,
Victor

 

[RF-OM]

I prepared pdf file for you with title pages and chapters for diplexer and splitter design. This is good for start and often enough to do the whole design.

Two resistors as termination is better because in this case load is somewhat distributed, not concentrated in one point. Having two parallel return paths also reduce they inductance practically twice because there is no common flux. You can check it with VNA and see what will be the difference.