• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 20 subscribers
  • Views 1061 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.

TXS0108E: eSPI translate

Patrick Chen
Mastermind 8875 points
Part Number: TXS0108E
Other Parts Discussed in Thread: TXB0108, LSF0108

Hello

        Does TXS0108E support eSPI translation and support VCCA=VCCB?  if not, has other suitable solution? Thank you. 

BR

Patrick

  • 0
    Guru 295240 points

    At 1.8 V, the TXS cannot go above 40 Mbps even with push/pull driving. And the TXS pulls all lines high, so it cannot be used for the eSPI clock.

    Why do you want to use the TXS when you are not doing level shifting?

  • 0 in reply to Clemens Ladisch
    TI__Mastermind 45045 points

    Hi Patrick,

    What are the requirements for eSPI? For example what is the needed data rate?

  • 0 in reply to Dylan Hubbard
    Mastermind 8875 points

    Hello

         eSPI operates at 1.8V only and supports one or multiple slave devices (SIO/BMC). SPI slave devices can share the same bus with eSPI slave devices. This interface is not shared and distinct from the SPI interface used for flash device. The eSPI interface on PCH supports 20MHz, 24Mhz, 30MHz, 48MHz and 60MHz. eSPI supports single, dual and quad mode. eSPI and LPC are both supported but only one of them can be implemented as they are mutually exclusive.

         Do we have other device support above requirement? CPU is 1.8V eSPI then translate 3.3V to FPGA. If do through LPC, it is needed to 3.3V to 3.3V. eSPI and LPC are the same pin that is through strap pin to decide which nterface to use. Thank you. So level shift is also need VCCA=VCCB support. Thank you. 

    BR

    Patrick

  • 0 in reply to Patrick Chen
    Guru 295240 points

    The TXB0108 might be fast enough (just barely), but its output drivers are too weak for the eSPI pull-up/pull-down impedances.

    The LSF0108 has the same problem as the TXS.

    Fast bidirectional buses are not meant to be level shifted. If anybody asked me, I'd recommend to run the FPGA I/Os at 1.8 V.

  • 0 in reply to Clemens Ladisch
    Prodigy 30 points

    @  is the problem with TXS that it is too slow?

  • 0 in reply to Andrew Colletta
    Guru 295240 points

    The common problem of the TXS and the LSF is that they use pull-ups, which is wrong for the eSPI clock.