900 mhz ltcc bandpass filter

I am trying to design a RF bandpass filter with multiple output, each output with different frequency band, Examp: one RF input and multiple output (1) 900MHz, (2)1.8GHz (3)2.4GHz (4)3.5GHz To save the space and matching problems, Any way to design it with filters in series ? (instead of a lot of parallel filters)

I would probably do it this way, using "singly terminated" lowpass and highpass filters. A standard quadraplexer design.

Rich

The standard way to make this is the multiplexer method. Since all your channels are far apart, so it is fairly easy to make this 4 channel multiplexer.

Good luck!

Multiplexers method is bulky.
I will choose biff44 method with LTCC technology.

Maybe I miss something but from the posted schematic I see only the outputs 1.8GHz and 2.4GHz having Band-Pass configuration (LP in series with HP).

900MHz output have Low-Pass, and 3.5Ghz output have High-Pass.

I think the initial request is looking for Band-Pass on all ports.

"each output with different frequency band"

Requirements met, ship it!

Hmmm. So, why don't use two combined duplexers (or diplexers), or just use a 4-way splitter.

Is that a trick quesion? Because he wants the 4 frequencies to come out of 4 different connectors.

Is not about a hidden question here, but probably the best is to wait for the initiator of this topic to clarify what exactly is looking for.

splitters are lossy. is abc123 is OK with it?

abc123 did not mentioned about the requirements on
lower cut off for 0.9GHz and
upper cut off for 3.5GHz

that probably add additional LP and HP.

only abc123 can answer for the queries.

thanks for the good help. I can figure out the design details myself. (the LPF, HPF, BPF and bandwidth of each band...) the biff44 example is a great starting point.

the only problem now is for each output port we need 50 ohms impedance, and since each filter section in series chain has 2 output load(LPF, or HPF) how to choose input/output impedance of each filter section is a question, do I need to match each sub-filter I/O to 50ohms first? then combine it together? or there are better way? since the internal node impedance is not that important to us.

Not necessary. but it helps in debug.

LTCC is my answer.

In LTCC the integration is vertical and has much advantage for space constrain designs. But it takes more time per simulation.