Discrete balun design for CC2640

I am sorry, but I do not fully understand your question as I cannot see the attachment. The CC26xx has the option between several front-end configurations, internal or external bias, differential or single-ended configuration. On our BLE wiki page, you will find a presentation describing the difference in performance between the set-ups, see processors.wiki.ti.com/.../Category:BluetoothLE and processors.wiki.ti.com/.../CC26xx_HW_training_RF_Frontends_and_Antennas.pdf

Hi, please see the ppt ile I uploaded. You will see the swaping connection for discrete balun circuit. I want to know whether thest two designs are similar or not??

The performance between the two solutions you sketch are the same yes, actually, your solution is similar to what is used on the the CC2650EM-7ID reference design. I recommend that you use this design for layout guideline.

From my understanding, differential impedance of discrete balun is designed to consider RX mode and TX mode. The final design is comprising solution. It is difficult to understand to have same performance for these two designs. Could you please explain again? And which design is preferred?

From my understanding of your attachment, you are considering a differential solution with internal bias (thus comp in red are removed from your drawing of our ref design). So the difference in the two pictures are simply that the components in the balun is swapped between the RF_N and RF_P pins, so far we agree, right? So this discussion is similar to this post: e2e.ti.com/.../446085

A balun is a network that transforms from a balanced to an unbalance signal, hence the name balun. The balun has a ±90 degrees phase shift implemented by using a low pass filter and a high pass filter. The important part is to keep the balun as symmetrical as possible, the pins can be reversed. That is why we recommend following the reference design (both schematic and layout).