Input matching of a 3dB power splitter

Hi everyone,

I am wondering if the input matching of a 3dB splitter will be degraded if two filters with different passband frequencies are inserted at the two outputs of the 3dB splitter?

You get mismatch at the out-of-band of your filters. But in a 3 way Power Splitter, mismatch at splitted ports is not critical if main port is properly matched.
Take a look to this document, page 5:
http://www.minicircuits.com/appnote/pwr2-4.pdf

Thank you for the link. What if the power splitter is now preceded by a wideband amplifier? Should I be concerned with the power outside of the filters passband (-> frequencies at which the input of the power splitter is not matched) beeing reflected at the input of the splitter?

Thanks for your help

If you replace the BPF with a wide-band amplifier could be better, because usually the wide-band amplifiers has a wider 50 ohms input match.
One concern specific for Wilkinson power splitters is the internal resistor that can be damaged for high power and port mismatch.

Hi,

concept of two filters connested to the outputs of power spliter (a hybrid or Willkinson divider) is working, you can get diplexing performance of whole network without any problem. Isolation resistor is taking care of isolation, so out of passband reflection is not doing to much damage.

However, thic concept suffers from not being optimum one for the job! Spliting/combining with a hybrid or a spliter/combiner has inherently 3dB insertion loss due to power splitting + any termal loss involved in spliter + insertion loss of filter. Depending on a frequency this could add up to 3.5-5.0 dB in total!

There is not a successful replacement for a true diplexer- you will have to pay the price!

flyhigh

hi all,

I wander how to modify the isolation resistor
in HFSS(how to draw the current flow).

I am wondering why you start with the 3dB coupler.
You may combine the two filters to make a duplexer,
and you can avoid the 3dB loss.
What is the signal level?
You may have different choices depending upon power
levels.

If you hook up two different bandpass filters to the outputs of a splitter, you should expect a lot of passband ripple! That is a poor way to do a diplexer filter.