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.

TUSB1310A: TUSB1310A: Issue with PCLK

Part Number: TUSB1310A

Hi TI support team,

I have encountered a problem while debugging my TUSB1310A design. First of all I wanted to measure the PCLK. For that reason, I have booted up the system and then the PCLK was 250 MHz. But against to your documentation, the clock was 900 mV  /- 100 mV. Could this be correct? In the documentation I found, that the clock should meet the LVCMOS18 standard. Do you know, where the problem is?


After a few minutes, the PCLK was getting unstable in the follwing way: The clock is “stable” for 150 µs as described at first (250 MHz, 900 mV  /- 100 mV), than it is high (1.8V) for 50 µs. This is periodically. So the “stable” part for 150 µs follows the high part for 50 µs and so on:

Here is the schematic:

The DIP switches are used in the following way:

S1 – 1 -> ON

S1 – 2 -> OFF

S1 – 3 -> ON

S1 – 4 -> OFF

 

S2 – 1..8 -> OFF

S2 – 9 -> ON

S2 – 10 -> OFF

 

All signals connected to page 3 are connected to an FPGA. The signals in the FPGA are connected in the following way:

USB3_USB3_1_POWER_DOWN0  <= '0';

USB3_USB3_1_RESETN       <= ‘1’;

USB3_USB3_1_OUT_ENABLE   <= '1';

USB3_USB3_1_NEN_P5V0_OUT <= '0';

 

There are 2 of the TUSB1310A Chips on one board. Both Chips show the same behavior. 


Thanks and regards
Michael

  • Hello,
    How is the SSC_DIS pin set? If it is low, can you set it high?
    Are these two TUSB1310A connected in parallel with the same FPGA?
    Regards
  • Hi,

    SSC_DIS was low. I set it to high (pull up) and got exactly the same behavior:

    It means, that there 2 two of the TUSB1310A are on the board and connected separately to the FPGA. There are no shared signals between the two TUSB1310A.

    Maybe there is another strap pin set to the wrong value? 

    Thanks and Regards

    Michael

  • Are there any news on this case? 

    We are still stuck.

  • Hello,
    Sorry for the delay, I will talk with the Designer about this, I should have more feedback within this week.
    Regards
  • Hello,
    There is an errata talking about a relationship between SSC_DIS pin and the crystal frequency.
    According to the errata, if you enable SSC then you have to use 40MHz, and if you disable SSC then you have to use another frequency.
    I know you were using 40MHz with SSC enabled which seems correct, just to double-check, could you use 20, 25 or 30MHz and SSC disabled?
    Are you able to capture a logic capture of all the PIPE signals?
    Regards
  • We could not find anything in the device that will make PCLK behave like this. We have not seen it in the past either.

    How many devices/boards have you tried that you see this issue.

  • Hi,

     

    I am not able to change the crystal on the board. But I do not think, that the problem is here because I am able to measure the 40 MHz clock of the crystal.

     

    I am sending no data on the PIPE interface. I thought the clock is running without sending data on this interface?

     

    I have only one board here with 2 devices on it.

     

    Which signals have to be present and in which state do they have to be, that I can see the clock running?

    Thanks and regards

    Michael 

  • Hello,
    Could you elaborate more on your last question? DO you mean the state of the PIPE signals at power-up ?
    Also, the scope capture says "40.08MHz", however, if I look at the cursors and the time delta the frequency would be 25MHz, can you double check that and make sure the frequency matches the REFCLKSEL[1:0] configuration?
    Please disable the spread spectrum clocking.
    Are you able to get any PIPE communication done? If so, can you send a logic capture of the PIPE signals?
    Can you specify whether a pull-up or pull-down is installed on each configuration pin?
    Regards
  • I mean generally signals and the state which they should have to see the clock running. Which set of signals have to be present at the start to see the clock running.

     

    The scope says 40.08 MHz because the time I have captured was not enough to calculate the right frequency. The clock is 40.00 MHz. REFCLKSEL pins have pull-ups. Please refer to the schematic attached at the beginning of the thread.

     

    The spread spectrum is disabled. SSC_DIS has an external pull-up.

     

    I am not using the PIPE interface at the moment. I have no IP core to use it.

     

    I have mentioned the information about the pull-ups and downs at the beginning of the thread in the attached schematic. If pins are connected to DIP switches, I have send the information about the DIP switch settings.

    I have added pull-downs to the reset signals and delayed it that the strap pins can be sampled in the right way. But at any time I see (in comparison to my first post) no clock at any time. I have also toggled the reset signal with a frequency of round about 1 Hz. But even if the reset was set or released, there was no clock at any time.

    Do you have any new idea?

  • Sorry for the delay again. It is possible that PLL is actually not lock.

    Can you share us power up sequence and reset sequence to make sure the power up timing is correct to ensure PLL is locked.

    regards, win