• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 101 subscribers
  • Views 670 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.

C6657 power supply design considerations for PCIe with a PC environment

Daisuke Maeda
Guru 19835 points
Other Parts Discussed in Thread: DS80PCI102

Our customer designs the custom C6657 boards which can use with a PC environment via PCIe.

The power supply of the C6657 board is not supplied from a PC module. Separate power supplies are used for the C6657 board and the PC module each.
Only one side of those power supplies can be supplied, and then only one side of a PCIe lane can be driven. Those conditions may give the damage to the devices.

If the C6657 board is powered-up and operates before the PC module is powered-up, can it avoid the damage to the devices that the power domain of C6657 PCIe is disabled by PSC?

Can the PC module be powered-up and operate before the C6657 board is powered-up?

Best regards,

Daisuke

 

  • 0
    TI__Mastermind 42545 points

    Hi Daisuke,

    The PCIE interface for the C6657 is not designed for hot-swap operation where the signal is driven from one side before the power is present to the SOC. We have seen boards designed for the lab environment that are connected with SMAs where one side is powered while the other is not but this is not recommended for production systems. You should design your power supply to begin sequencing once the PC power supply is present or you should add a component which will prevent the PCIE from driving the receiver before the C6657 power is present.

    Regards, Bill

  • 0 in reply to Bill Taboada
    Guru 19835 points

    Hi Bill,

    Thank you for your reply.

    When the C6657 power is not present and the PC power is present, does this condition damage the C6657 side because the PC side drives a PCIe lane?

    Best regards,

    Daisuke

     

  • 0 in reply to Bill Taboada
    Guru 19835 points

    Hi Bill,

    is it really possible the power supply is designed to begin sequencing once the PC power supply is present?

    I check the PCI express specification. The PCIe transmitter holds a constant DC common mode voltage when Fundamental Reset (cold reset) is asserted.

    It can be momentary that the C6657 power is not present and the PC power is present, does this condition not damage the devices?

    C6657 SerDes input voltage range: -0.3 V to CVDD1+0.3 V

    Best regards,

    Daisuke

     

  • 0 in reply to Daisuke Maeda
    TI__Mastermind 42545 points

    Hi Daisuke,

    The IO cells in the C6657, including the PCIe interface, were not designed to be driven if the supply voltage for the cell was not present. I can't tell you whether a short exposure will result in damage because we don't test for that condition. 

    Regards, Bill

  • 0 in reply to Bill Taboada
    Guru 19835 points

    Hi Bill,

    Thank you for your reply.

    Does TI have a component which will prevent the PCIE from driving the receiver before the C6657 power is present?

    For a system using separate power supplies such as this case, is it needed that the PCIe on PC supports the hot-swap operation?

    Best regards,

    Daisuke

     

  • 0 in reply to Daisuke Maeda
    TI__Mastermind 42545 points

    Hi Daisuke,

    TI makes a family of redriver/repeater parts that offer the hot-plug capability. A component like the DS80PCI102 is designed specifically for PCIe. This part has a control signal, PRSNT#, which will disable the output drivers which would be connected to the C6657.

    Regards, Bill

  • 0 in reply to Bill Taboada
    Guru 19835 points

    Hi Bill,

    Thank you for your reply. Sorry for my late reply.

    Best regards,

    Daisuke