• State TI Thinks Resolved
  • Locked Locked
  • Replies 23 replies
  • Answers 1 answer
  • Subscribers 34 subscribers
  • Views 1105 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.

TUSB1044: Config modes for USB3.1 Only

Luis Jimenez1
Prodigy 120 points
Part Number: TUSB1044
Other Parts Discussed in Thread: , TUSB1002, TUSB1002A

Hi,

I am trying to use TUSB1044 for USB3.1 signal conditioning with long passive USB type-C cables.

I plan to use TUSB1044 pin configured because I want to use captive USB cables and control by myself what set of pairs I am using.

My application includes also another redriver for DisplayPort, and I want to use the same device for both (with different configurations, of course)

To check the different configuration sets I have a TUSB1044RNQEVM. I am making tests with minimum cable lengths just to check the configurability of the direction of the redrivers.

I have been able to use seamlessly the ports TX1 and RX1 (DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L and L/H/L/H/H). (See page 49 of TUSB1044 datasheet)

But I would like to use the ports TX2 and RX2 (DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H or L/H/L/H/L) but the USB3 device I connect is not recognized.

TX2/RX2 ports should work exactly as TX1/RX1 ports or I am missing something?

JFI, the USB3.1 device I am using for the tests is a USB3 Loopback tester from PassMark. It is connected via a USB Type-C to USB micro B cable. I am taking into considerantion the orientation of that Type-C connection because only a set of RX/TX pairs ends in the microB end.

Regards,

Luis Jimenez

  • 0
    TI__Guru 64326 points

    Hi,

    This is correct, TX2 and RX2 should have same function. Have your tried all orientations of the EVM and Type-C to micro B cable? Maybe try flipping the EVM when DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H or L/H/L/H/L?

  • 0 in reply to Malik Barton57
    Prodigy 120 points

    Hi Malik, thanks for yor answer,

    Yes, I have tried all orientations of the EVM and Type-C to micro B cable:

    - When the EVM is connected to a Type-C port, its orientation does not affect because the PC detects it. When I connect it via a USB-A to C cable I can confirm that there is one correct orientation.

    - The orientation of the Type-C to microB cable affects too, in one orientation the device is not connected, but in the one that is supposed to be the right one (by following the schematics and pinging the cable) the device sees something of the USB3 port, in the screen of the PassMark loopback tester it says "USB3 enumerated", but the device never appears in the device manager or in USBView.

    I have made the same tests with L/L/L/H/L and L/H/L/H/H and the behaviour is always the expected according to the orientation of the cables.

    I started to think that my EVM was wrong, but I have two of them and both have the same behaviour.

    I think I will have to move to TUSB1002 for this part of the design.

  • 0 in reply to Luis Jimenez1
    TI__Guru 64326 points

    Hi,

    If you are going to move, I recommend trying TUSB1002A instead, this part has a updated state machine. This is abnormal that you are not seeing the correct orientation. Is it possible to use another device besides the Passmark  loopback tester? Try a regular PC as the host and regular Type-C of mirco-B device as see if this enumerates correctly? 

  • 0 in reply to Malik Barton57
    Prodigy 120 points

    With another USB3 device the result is the same.

    I forgot to mention that both boards I have are the ones in which JMP10 and JMP12 have the old layout. There might be other changes in the schematics/layout of the EVM?

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    Is this your setup:

    PC (Type C connector) --- TUSB1044 EVM (Type C plug) --- TUSB1044 EVM (Type C connector) --- Device?

    If you probe on TX2/RX2, are you seeing the LFPS?

    Thanks

    David  

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Hi David,

    I have tried that setup, and also with a USB-A to USB-C cable before the EMV (in this case I have to consider if the cable is flipped or not when I want to use RX1/TX1 or RX2/TX2).

    Until today I hadn't tried with the scope.

    With the config  for RX1/TX1, that works OK, (DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L)  but without a device connected, I can see in the scope a polling on LFPS in TX1, both before and after RUSB1044. I see nothing on RX1.

    With the config for RX2/TX2 ((DIR1/DIR0/CTL1/CTL0/FLIP = L/H/L/H/L or L/L/L/H/H) I cannot see LFPS

    These last tests have been made with the EVN directly connected to the USB-C port of my laptop.

    Regards,

    Luis

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    1. Please have the Type-C device connected directly to the USB-C port of the laptop, both the flip and non-flip orientation

    Flip Orientation:

    Non-Flip Orientation:

    2. Please set TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L, and have the Type-C device plugged into the TUSB1044EVM Type-C receptacle. While keep the orientation of the Type-C device the same, change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    Flip Orientation:

    Non-Flip Orientation:

    3. Please change TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H, and have Type-C device plugged into the TUSB1044EVM Type-C receptacle in the opposite orientation of experiment #2. Now change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    Flip Orientation:

    Non-Flip Orientation:

    Thanks

    David

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Hi David, here are the resutls:

    1. Please have the Type-C device connected directly to the USB-C port of the laptop, both the flip and non-flip orientation

    Flip Orientation: Status is "Speed 5Gb/s", device OK

    Non-Flip Orientation: Status is "Speed 5Gb/s", device OK

    2. Please set TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L, and have the Type-C device plugged into the TUSB1044EVM Type-C receptacle. While keep the orientation of the Type-C device the same, change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected non-flipped into the TUSB1044EVM

    Flip Orientation: Status is "Speed 5Gb/s", device OK

    Non-Flip Orientation: Status is "Speed 5Gb/s", device OK

    If the device is connected flipped into TUSB1044EVM (USB3 shuld not work this way) the status is "Not connected". OK.

    3. Please change TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H, and have Type-C device plugged into the TUSB1044EVM Type-C receptacle in the opposite orientation of experiment #2. Now change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected flipped into the TUSB1044EVM

    Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView.

    Non-Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView.

    If the device is connected non-flipped into TUSB1044EVM (USB3 shuld not work this way) the status is "Speed 480Mb/s". OK, USB3 is not connected but USB2 is.

    Regards,

    Luis

  • 0 in reply to Luis Jimenez1
    Prodigy 120 points

    I have repeated the same with a different USB3 device and the results are the same.

    Luis

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    For experiment #2 and #3, would you please measure the voltage on CC1 and CC2 for each case?

    2. Please set TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L, and have the Type-C device plugged into the TUSB1044EVM Type-C receptacle. While keep the orientation of the Type-C device the same, change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected non-flipped into the TUSB1044EVM

    Flip Orientation: Status is "Speed 5Gb/s", device OK

    Non-Flip Orientation: Status is "Speed 5Gb/s", device OK

    If the device is connected flipped into TUSB1044EVM (USB3 shuld not work this way) the status is "Not connected". OK.

    3. Please change TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H, and have Type-C device plugged into the TUSB1044EVM Type-C receptacle in the opposite orientation of experiment #2. Now change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected flipped into the TUSB1044EVM

    Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView.

    Non-Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView.

    Thanks

    David

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Here are the voltage measurements on CC1/CC2

    2. Please set TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/L, and have the Type-C device plugged into the TUSB1044EVM Type-C receptacle. While keep the orientation of the Type-C device the same, change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected non-flipped into the TUSB1044EVM:

    • Flip Orientation: Status is "Speed 5Gb/s", device OK. CC1/CC2 = 0.923/0.467V

    • Non-Flip Orientation: Status is "Speed 5Gb/s", device OK. CC1/CC2 = 0.916/0.468V

    If the device is connected flipped into TUSB1044EVM (USB3 shuld not work this way) the status is "Not connected". OK. In both cases CC1/CC2 = 0.45/0.92V

    3. Please change TUSB1044 EVM to the setting DIR1/DIR0/CTL1/CTL0/FLIP = L/L/L/H/H, and have Type-C device plugged into the TUSB1044EVM Type-C receptacle in the opposite orientation of experiment #2. Now change the orientation of the TUSB1044EVM Type-C plug to both the flip and non-flip when inserted into the USB-C port of the laptop.

    The device is connected flipped into the TUSB1044EVM:

    • Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView. CC1/CC2 = 0.453/0.916V

    • Non-Flip Orientation: Device says "USB3 Enumerated", but it does not appear in Windows or in USBView. CC1/CC2 = 0.453/0.924V

    If the device is connected non-flipped into TUSB1044EVM (USB3 shuld not work this way) the status is "Speed 480Mb/s". OK, USB3 is not connected but USB2 is. In both cases CC1/CC2 = 0.92/0.47V

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    Do you have a regular device that can be plugged directly into the Type-C receptacle and repeat the same experiment? 

    And where do you measure the CC1 and CC2 voltage?

    Thanks

    David 

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Hi David,

    I have repeated the measurements with another USB3 device (it has a micro B USB3 connector, so I am using a type-c to microB cable) and the measurements are the same.

    I haven't used a pure type-c device, because the only one I have right now is my phone, and it exeedes the power of the USB port without power negotiation.

    I am measuring CC1 and CC2 voltages at pins 12 and 11 of U3.

    Regards and thanks,

    Luis

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    With Type-C to Micro-B cable, only A2/3 and B10/11 of the cable are connected, B2/3 and A10/11 are NC. So you will have the orientation which USB3 will not work.

    Thanks

    David

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Hi David,

    No, that is not the problem.

    I have identified the orientation of the cable, even pinging it. I am sure that the SSRX/TX of the device are connected to the pins in the EVM that match the amplifiers of the redriver I am configuring.

    Regards,

    Luis

  • 0 in reply to Luis Jimenez1
    TI__Guru**** 156261 points

    Luis

    I am looking at the PassMark website: https://www.passmark.com/support/usb_port_security_audit.php.

    For the message "USB3 enumerated," USB plug detected a data connection attempt from a USB3.0 port. The USB3 enumerated message means that data lines are connected, but the plug is not setup by the operating system, likely due to the lack of valid device drivers.

    From this message, it looks like the physical connection has been made through the switch.

    Do you have a way to capture the full USB3 enumeration?

    Thanks

    David

  • 0 in reply to David (ASIC) Liu
    Prodigy 120 points

    Hi David,

    Sorry for the delay.

    I don't have a way to capture the full USB enumeration, can I get it only via software or this need dedicated hardware in the middle?

    I have tried using Free Device Monitoring Studio. I get some packets when I connect the EVM in the modes that work ok, but in the config modes in which I get the message of "USB3 enumerated" I cannot capture anything.

    Regards,

    Luis

  • 0 in reply to Luis Jimenez1
    TI__Guru 64326 points

    Hi,

    Dedicated hardware would be needed in the middle. From Davids description it seems that the connection is made through TUSB1044. Would you be able to probe with a high speed scope to see if the USB signals traverse TUSB1044 in the config mode? 

  • 0 in reply to Malik Barton57
    Prodigy 120 points

    I'm affraid I don't have access to a high speed scope. I can try to use a 500MHz with a differential probe, but I guess it will ruin the signal.

  • 0 in reply to Luis Jimenez1
    TI__Guru** 122810 points

    Luis:

       do you have any update?

  • 0 in reply to Brian Zhou
    Prodigy 120 points

    Hi Brian,

    I have made some measurements with the scope, trying to capture LFPS at both sides of TUSB1044 if the different configuration modes.

    I cannot connect the EVM directly into USB-C port, so I am using a USB-A to USB-C cable and a Female-to-Female USB-C adaptor. All has been checked so that when connected the RX and TX differential pairs are routed to the desired channel.

    I am using a 500 MHz bandwith scope with 1.5 GHz differential probe. It is enough for LFPS capture.

    For all tests, the EVM is configured as (EN=H, SWAP=L, SLP_S0#=H) DIR1, DIR0, CTL1, CTL0, FLIP as described below.

    Without device connected:

    • Using TX1/RX1 channels: A_to_C cable (flipped) --- F2F adapter --- TUSB1044EVM (L L L H L):

    • LFPS measured in line TX1 both before TUSB1044 (C5, C6) and after 1044 (C13, C14)

    • Using TX2/RX2 channels: A_to_C cable (non-flipped) --- F2F adapter --- TUSB1044EVM (L L L H H):

    • No LFPS measured in line TX2 either before TUSB1044 (C3, C4) or after 1044 (C11, C12)
    • A pulse is measured on both TX2p and TX2n (not visible in differential mode)
    • It is a pulse of 0.5V 38us repeated every 14 ms

    With a device connected:

    • Using TX1/RX1 channels: A_to_C cable (flipped) --- F2F adapter --- TUSB1044EVM (L L L H L)  ---  C_to_uB (non-flipped) & PassMark:

    • LFPS measured in line TX1 both before TUSB1044 (C5, C6) and after 1044 (C13, C14)
    • LFPS measured in line RX1 both before TUSB1044 (C15, C16) and after 1044 (C7, C8)

    • Using TX2/RX2 channels: A_to_C cable (non-flipped) --- F2F adapter --- TUSB1044EVM (L L L H H)  ---  C_to_uB (flipped) & PassMark:

    • No LFPS measured in line TX2 either before 1044 (C3, C4) or after 1044 (C11, C12)
    • A pulse is measured on both TX2p and TX2n (not visible in differential mode)
    • It is a pulse of 0.5V 38us repeated every 14 ms.
    • LFPS measured on RX2 only at device side (C9, C10). It does not cross the TUSB1044.

    It looks like if LFPS does not cross the TUS1044 when I configure TX2/RX2.

    But even the LFPS is not sent from the computer when in this mode. Might it be due to some channel termination detection? Could it be related with the 38us puses?

    It looks like the device is detecting the termination and is sending the LFPS but the host seems to be sending detection pulses (they are common mode, not differential) through TX pins waiting to see something on RX.

  • 0 in reply to Luis Jimenez1
    TI__Guru 64326 points

    HI Luiz,

    I am sorry for the extremely long delay. Are you still having issues with TUSB1044? Would you be able to provide the waveform you mentioned in the last post (0.5V 38us repeated every 14 ms)? You are correct, TUSB1044 may not see the Host termination, did you also flip the F2F adapter during your second test? LFPS period (edge to edge) is typically in the 30 ns range, LFPS packets length are different depending on the state of the USB3 bus. LFPS packets are spaced with electrical idle by either 10 us or ~200 ms.