Could anyone please help me explain this RF circuit

Hi,

I am using Qualcomm Atheros's AR9331 to design one router.

But about the reference circuit provided by Atheros, I am confused with the RF front end part design of AR9331, which is shown

in the attached picture.

The following is my understanding about the design.

As shown in the attached picture, RFOUTP/RFOUTN are from the output of the AR9331, while RFIN1P/RFIN1N are the input signal

to AR9331. Part A circuits are the design of the Balun. Part B is the matching circuit between PA of AR9331 and balun while Part C

is the matching circuit between LNA input and balun.


Then why Part B and C are connected together? And how are they getting to work?

Looks like you are dealing with two frequencies here. Path B is via a balanced HPF while path C is with a balanced LPF. Part A works like a diplexer and steering the signal (incoming or outgoing through the correct path).

Hi Joshuacp,

I have doubt on the design because Tx and Rx are directly to each other.

Part A : a balun and change from differential to single end input or output

Part C : matching for rf input

Part B :

C203 : Resonant with RFOUTP, RFOUTN pins and bond wire to remove reactance so that only resistance left behind
C181, C182 : matching + dc bias for PA
C66, C67 : coupling + dc blocK

Instead, the circuit should including a T/R swtich to separate Tx and Rx so that they will isolate from one another. A typical pin diode switch can give an isolation of 12 ~ 13 dB and it is cheap. The loading effect between Tx and Rx can significantly improve.

Hi,rogerynt,
Thank you very much. Your explaination has helped me understand the circuit clearly, especially for the part B.

For my understanding, Part A is a balun+diplexer. It steers Tx signal and Rx signal. Normally, Part A should be replaced by a T/R switch.
This circuit is to save the cost and just use the part A to replace a T/R switch.

I also warry about the isolation between Tx and Rx signal. But as the reference design of AR9331 is a very mature design, maybe the design works well.

rogerynt said:

Hi Joshuacp,

I have doubt on the design because Tx and Rx are directly to each other.

Part A : a balun and change from differential to single end input or output

Part C : matching for rf input

Part B :

C203 : Resonant with RFOUTP, RFOUTN pins and bond wire to remove reactance so that only resistance left behind
C181, C182 : matching + dc bias for PA
C66, C67 : coupling + dc blocK

Instead, the circuit should including a T/R swtich to separate Tx and Rx so that they will isolate from one another. A typical pin diode switch can give an isolation of 12 ~ 13 dB and it is cheap. The loading effect between Tx and Rx can significantly improve.
View attachment 114082

Click to expand...

- - - Updated - - -

hi,E-design,
Thank you very much for your help.
Yes, Part A works as a Balun+diplexer. It steers Tx and Rx signal.
Normally, Part A is replaced by a T/R switch. But here T/R switch is replaced by part A. Maybe this is to help to save cost.

E-design said:

Looks like you are dealing with two frequencies here. Path B is via a balanced HPF while path C is with a balanced LPF. Part A works like a diplexer and steering the signal (incoming or outgoing through the correct path).

Click to expand...

- - - Updated - - -

Hi, rogerynt,
Because AR9331 is a wifi Router design. As wifi is a time-division system.
So maybe part A could replace a T/R switch here.

rogerynt said:

Hi Joshuacp,

I have doubt on the design because Tx and Rx are directly to each other.

Part A : a balun and change from differential to single end input or output

Part C : matching for rf input

Part B :

C203 : Resonant with RFOUTP, RFOUTN pins and bond wire to remove reactance so that only resistance left behind
C181, C182 : matching + dc bias for PA
C66, C67 : coupling + dc blocK

Instead, the circuit should including a T/R swtich to separate Tx and Rx so that they will isolate from one another. A typical pin diode switch can give an isolation of 12 ~ 13 dB and it is cheap. The loading effect between Tx and Rx can significantly improve.
View attachment 114082

Click to expand...

- - - Updated - - -

Hi, rogerynt,
Because AR9331 is a wifi Router design. As wifi is a time-division system.
So maybe part A could replace a T/R switch here.

rogerynt said:

Hi Joshuacp,

I have doubt on the design because Tx and Rx are directly to each other.

Part A : a balun and change from differential to single end input or output

Part C : matching for rf input

Part B :

C203 : Resonant with RFOUTP, RFOUTN pins and bond wire to remove reactance so that only resistance left behind
C181, C182 : matching + dc bias for PA
C66, C67 : coupling + dc blocK

Instead, the circuit should including a T/R swtich to separate Tx and Rx so that they will isolate from one another. A typical pin diode switch can give an isolation of 12 ~ 13 dB and it is cheap. The loading effect between Tx and Rx can significantly improve.
View attachment 114082

Click to expand...