• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 703 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AWR1243: How to design antenna feed line length and LO length in AWR1243 Cascade usecase

Yudai
Prodigy 100 points
Part Number: AWR1243
Other Parts Discussed in Thread: TIDEP-01012,

Hi

We're designing AWR1243 cascade prototype like TIDEP-01012.

Please let me ask you how to design antenna line length and LO length for 4 AWR1243 cascade usecase.

①Could you tell me TIDEP-01012 design policy of antenna feed line length and LO length? I'd like to know which line should be isometric wiring or not.

We're considering 4 mode for AWR1243 cascade prototipe.

　3Tx(Array mode) x 16Rx by IC-A

    2Tx(MIMO) x 16Rx by IC-B, IC-C

    3Tx(Array mode) x 16Rx by IC-B, IC-C

　4Tx(MIMO) x 16Rx. by IC-C, IC-D

②16 Rx will be placed at lambda/2 interval. All of antenna feed line length should be isometric wiring? Or it should be isometric wiring for every 4 ICs?

③ When operating as an array antenna with 3Tx of the same IC-A, antenna feed line length of 3 Tx should be isometric wiring? 

④ When operating as 2Tx MIMO with 2Tx of the IC-Ｂ and 1Tx of the IC-C,  antenna feed line length of 1Tx+1Tx should be isometric wiring?

⑤ When operating as an array antenna with 2Tx of the IC-B and 1Tx of the IC-C, antenna feed line length of 3 Tx should be isometric wiring? And LO line length of IC-B and IC-C should be isometric wiring? 

⑥ When operating as 4Tx MIMO with 3Tx of the IC-C and 1Tx of the IC-D, antenna feed line length of 3Tx+1Tx should be isometric wiring? 

I'd like to know from the view point of LO wiring length and antenna feed line length.

Thanks

Yudai

  • 0
    TI__Guru 58035 points

    Hello Yudai,

    The LO routing to all the 4 devices should be matched in delay such that they are phase aligned on the PCB. The trace loss from SYNC OUT/ CLKOUT to SYNC In should be kept <6.5 dB.

    Ideally the trace routing of all the 12 Tx should be matched with each other and routing of all 16 Rx should be matched with each other so that there is minimal phase mismatch between the channels.

    Regards,

    vivek 

  • 0 in reply to Vivek Dham
    Prodigy 100 points

    Hi Vivek

    Thank you very for your kindly answer.

    I understood it from an ideal point of view.

    I'm afraid but I'd like advice from you from a realistic point of view.

    About 16 Rx, I found it difficult to be matched with each other because long trace causes bigger loss.

    Realistically, we set the RX trace length for each IC and minimize trace length differences between ICs.

    Finally, can the phase mismatch be corrected by antenna calibration?

    About 12 Tx,  I think 12 Tx is the same situation as 16 Rx.

    I think Tx should be matched with each operation  mode and array antenna operation, realistically. 

    Could you give me advice and your perspective from a realistic design point of view for me?

    Thanks

    Yudai

  • 0 in reply to Yudai
    TI__Mastermind 18500 points

    Hi Yudai, 

    There is no way to loosen the constraints on the 20 GHz LO distribution matching. I think from your reply that you understand that. 

    Regarding TX channels, for a single TX system, say for TDMA-MIMO, there is no real requirement. However, for any system requiring multiple TX like, SIMO or TX-Beamforming, if phase matching is not achieved then the system cannot operate optimally. 

    In the TI EVM designs, we have opted to try to create TX antenna paths which require minimum phase calibration. Normal TX phase path calibration does make use of these TX phase shifters, but usually only a few taps. So maybe for TX channels there is some additional margin there as long as the delay is less than a single wavelength, so the phase shifters can still compensate it out. I will not recommend doing this without researching more. 

    Regarding RX channels, any delay between channels is normally calibrated out with a known target in any case. I cannot see any fundamental downside to introducing larger static delays in the PCB channels. I still will not recommend doing this without researching more. 

    I do recommend experimentation with such a system to see what effects are seen. 

    Thank you,

    -Randy

  • 0 in reply to Randy Rosales
    Prodigy 100 points

    Hi Randy

    Thank you very for your detailed explanation.

    I understand TI EVM design policy and your perspective.

    Regarding LO and Tx channel , I will design LO length and Tx antenna feed line in same mode with equal length.

    Regarding RX channel, You and I understand that it is difficult to design all RX antenna with equal length. So I  will design feed line of Rx antenna to minimize PCB RF loss and the length difference. I expect the length difference of Rx channel to be corrected by Antenna Calibration.

    Thank you again

    Yudai