• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 284 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TUSB1002: How to characterize (S parameters) TUSB1002?

traian mitrache
Prodigy 125 points
Part Number: TUSB1002

I have this part designed in on a very large board 22" x 26"

There are several USB 3.1 Gen 2 sources on the board, at different distances from the edge of the board where the connectors are. Each of them has a TUSB1002, to help drive the long traces.

The ones closer to the edge of board work, the ones further away do not.

In order to figure out the proper equalization settings for the TUSB1002, I have created a breakout board, to allow me to bring a signal source into the board to test the S-parameters and create an EYE diagram

To characterize the circuit (PCB traces + TUSB1002 + Connector) I am using a PNA: Keithley N5222B.

THE PROBLEM: when the TUSB is in the circuit, stimulated by the PNA, I am getting no signal at the output of the TUSB1002 (power is ON, and the input EN pin is left floating). When I removed the TUSB1002 from the circuit and bypassed it, I DO get signal, which proves that my cables/connections/traces are OK

QUESTION 1: Is there an issue with using a PNA to characterize the TUSB?

QUESTION 2: How did TI generate the SDD21 plots in the datasheet? Can you share with me what instrument was used for characterization? 

  • 0
    TI__Guru** 104630 points

    you need to set MODE pin to 0 to measure S-parameter

  • 0 in reply to Brian Zhou
    Prodigy 125 points

    Brian,

    Thanks for the tip. There is nothing in the datasheet to indicate this. All I see in the datasheet is that this is a "Test Mode" pin.
    It made an immediate difference: the signal is no longer at -80dB levels, it is now somewhere -35dB at 5GHz. Still not quite what I would expect.   
    I'm getting better signal if i remove the TUSB1002 and jumper across from input to output pads: i get -20dB...and this is with the most ordinary wire...
    Is there any other settings that I should tweak? Is there any app note regarding how to characterize the part ?
    Do I need to worry about the "Receiver Detect Functionality" as well?
    Do the CFG settings work in this mode?
  • 0 in reply to traian mitrache
    TI__Guru** 104630 points

    pls accept my friendship request and I will send some docu to you

  • 0
    TI__Intellectual 1740 points

    If you are using the TUSB1002A, then you will need to have MODE pin = "0". This will place TUSB1002A into basic redriver mode.

    If you are using the TUSB1002, then you will need to put device into TI Test mode (MODE = "R"). Make sure you have TX1P, TX1N, TX2P, and TX2N AC coupled (75nF to 265nF) to 50-ohms to GND.

    MODE = "R"

    CFG1 = "0"

    CFG2 = "0".

     

  • 0 in reply to Michael Campbell
    Prodigy 125 points

    But having 

    CFG1 = "0"

    CFG2 = "0".

    Would make this whole test useless, as my main goal is to figure out the proper settings in my system. If the CFG pins HAVE to be 0, I would not be able to try different CFG settings...

  • 0 in reply to traian mitrache
    TI__Intellectual 1740 points

    I understand.  You can still test the 16 different EQ levels but the linearity and DC Gain will remain fixed.  Linearity is set to 1200mV and DC Gain is at 0dB.

    If you want to test all combinations, then MODE pin must be at "0" and then you will need to bias RX1 and RX2 to 1.8V.  For TUSB1002, the internal BIAS for RX1 and RX2 is disabled when MODE = "0".

    Another option is to fool the USB state machine into thinking the link is always at U0.  This can be accomplished when greater than 500MHz signal is fed continuously thru TUSB1002 in one direction while in the other direction you measure the s-parameters.  Again make sure TX1 and TX2 is terminated.  Make sure signal is present on one of the receivers within 200ms of powering up the TUSB1002.

  • 0 in reply to Michael Campbell
    Prodigy 125 points

    The main problem is, with this configuration

    MODE = "R"

    CFG1 = "0"

    CFG2 = "0"

    I am not seeing any signal come through the TUSB1002. VNA connected on one side, getting -70dB on the other.

    I don't have a clean way of biasing the RX1/2 to 1.8V. There was nothing in the datasheet to suggest that I might need to do this, and I did not make any provisions on the board for it. Any "reworks" to add biasing resistors would significantly affect the signal integrity.