Customer has done ETSI testing for both TX channels, and with TX2 everything looks good, but the same test script with TX1 do not pass. Please see below.
They have issues with TX out-band emission and got the output emission issue at one TX(1) of two TRX path. TX(2) is operated normally as shown in attchment. They have also separated the PA portion circuit to see the signal generated by AD9361. The out-band noise is high and the total 125KHz BW in-band power is correct and the power level can be controlled. These two TRX sharing the same LDO and TCXO clock and they have set the same RF parameter to TX script such as AD9361 ATTEN, modulation, GAIN..etc. They're trying to troubleshoot why there're two different TX signals.
The frequency range of operation is from 902-928MHz and 836.105-869.89MHz. The atten is set to zero for initial setting. They have narrowed down the noise source coming from the AD9361.
They are operating the IQ bus as "Dual Port Full Duplex" mode. It seems they might actually be using the LVDS diff mode(have yet to confirm)
The specification of the mode they are using to drive the ADI digital interface for 2R2T mode is as below(FDD, Dual port).
Extracted the relevant images for RX and TX timing
1) IF they are indeed using the LVDS mode, can they use the Tx Channel Swap function. i.e., swap the Tx1 I/Q with Tx2 I/Q ordering. This way they can determine if an IQ bus timing issue can cause the Tx1 portion to be corrupted due to marginal timing.
2) Any possibility that analog noise can be introduced onto Tx1, but not Tx2?
3) Any possibility that TCXO phase noise can impact Tx1, but not Tx2?
4) Anyway to swap Tx1 and Tx2 paths internal to the AD9361?
5) Any possibility that IQ bus timing can effect Tx1, but not Tx2?
- Identical test scripts and test data is used to test Tx1 and Tx2.
6) Any suggestions on how to debug this?