Single Rx and Tx filters instead of duplexer

In PCS-CDMA normally a duplexer (SAW, FBAR, ceramic) is used in mobile phones. But for example if you separate Tx and Rx antennas (or feed points) or even you have an additional Rx diversity path, single filters are required which take over the Tx-Rx-isolation functionality. But I heard that if you take the Rx filter part of the duplexer as stand-alone filter it has less isolation, then if it is built with the Tx-filter part together to built a duplexer. Can someone explain me this?

I have never heard of this. What may be referenced is what happens if the two halves are not retuned after separation. The isolation between the separated antennas will add at least 20 dB to the total depending upon how far apart they are.

Duplexers and Diplexers (note they are not the same thing) are electrically simple filters (combination of BPF, LPF or HPF).
While duplexers use BPF, a diplexer is a simple LPF and HPF connected together.
They allow transmitting and receiving on the same antenna at the same time, rejecting unwanted signal.
CDMA is a full-duplex system and is transmitting and receiving in the same time. You can use simple filters, to reject the unwanted (TX signal in RX band) but the duplexer is simplifying the situation. You have only one filter with 3 ports (antenna in the middle) instead of two filters.
A standard rejection of the CDMA duplexer (TX in RX band) is 50dB (this is the system requirement in CDMA).

Hi vfone,

first of all thanks for your reply.
Yes, all what you say is clear for me.
But my problem is that one guy have told me, that for example if we take the Rx part of a duplexer. it rejects around 55 dB your own Tx signal. But if you take this part and used it as a "stand-alone" Rx filter ( without connection to Tx path) you only get around 35 dB rejection in Tx band. As stand-alone it means, you have separated Tx and Rx antennas, so no directly duplex function in filters. His explanation was that in a duplexer, part of the Tx signal goes back in the Tx path, so that the Rx filter part only needs to reject a lower Tx signal.
I do not answer immeditaly to this guy, but at my home I reflect this issue and it couldn′t understand it.

Regards,
zampa

As long as you terminate the filter appropriately, you shouldn't see a difference. When you measure a duplexer on a network analyzer, you measure the 45 or 50 dB or rejection that the manufacturer speciifies. I'd have to say that your friend is wrong in this.

I was refering to a ceramic block or SAW filter, where you cannot physically seperate the two filters. If you have a lumped circuit implementation, than you will probably see different results if you do physically seperate the two filters.

zampanampa,
Sorry, but your friend is wrong in the theory of reflected TX signal.
One thing is true. If you have a Duplexer (or Diplexer) and take separate the filters, they will not work as they work together.
The insertion loss and the rejection are optimized when the internal filters works together.

regards

Duplexers made in lumped ellement techique are not simple connection of two bandpass filters, they have compensation network where they are connected together. If you use only one filter, you must correctly terminate another filter. If you separate one filter to be used as standallone you must retune it or even make some modifications.

I agree Borber's opinion,you must use filter

Everything stated by everyone all seems to have merrit, I guess the trick is to see
how well it works.

Yes, you will have to change the terminations at the adjoining ends but the issue
was, will you get betteratttenuation from the filters together than seperated.

I could see this as a possibility, sending the rejection of ther signal to the other side could effectively bleed the signal.

Does anyone know much about splitters and combinners, would this have a similar effect?

Ignoring the fact that a single filter taken out of a diplexer would be mismatched all by itself, you would lose a theoretical 6 dB of isolation also.