Interference Problem in BDA

[Squirtle]

Hello guys,

I am trying to design a bidirectional amplifier system (BDA). In the figure below i drew the block diagram of the system. I used 2 duplexer with 4 cavity each. 4 cavities for Tx (2 at the i/p of LNA and 2 at the o/p of LNA) and 4 cavities for Rx overall. 2 cavities have at least 60 dB isolation. LNA's have 24 dB gain and they work in linear region, no P1dB or OIP3. The problem is when the system is connected just like shown in block diagram, we obtain nonsensical graph because i think interference occur between Tx and Rx. However when i disconnect one of the antennas node and measure Tx and Rx separately, they work just fine. I cannot figure out how to solve the problem. Any recommendation would be very welcomed.

 

[BigBoss]

You have to switch off one of the amplifier depending on TX or RX.
Otherwise there will always be an feedback loop and it will probably gets into oscillation regime.
OR
You can also TX/RX switches at both Input and Output.

 

[Squirtle]

Isn't it the main feature of the BDA to work bidirectionally? We use a full duplex system. If we have to switch off one of the amplifier depending on Tx or Rx, it means that we don't have to use duplexer or any filter.

 

[FvM]

Similar topologies are e.g. used for GSM repeaters. It can work due to frequency separation between up- and downlink.

I don't know what you mean with "nonsensical graph". Presuming suitable filter design, insufficient antenna isolation could be the problem.