• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 16 subscribers
  • Views 433 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

SN74LVC1G3157: application for PCIE4.0 REFCLK

user6257239
Prodigy 20 points
Part Number: SN74LVC1G3157
Other Parts Discussed in Thread: TMUX1072, TMUX136

dear TI guys：

can I use SN74LVC1G3157 as switch  for PCIE4.0 REFCLK ? which is 100Mhz and had some jitter requirement.

thanks.

  • 0
    TI__Guru 52882 points

    Hi,

    The bandwidth is not a problem as this part can typically handle 350MHz. That being said - what are the jitter requirements/budget ?  As we don't typically spec jitter as our IC's typically don't add a lot of jitter - especially in one channel devices - and random jitter is the biggest concern because our ICs typically aren't a source of deterministic jitter - but there is a small propagation delay that will effect each similar pulse similarly.

    Please let me know!

    Best,

    Parker Dodson

  • 0 in reply to Parker Dodson
    Prodigy 20 points

    dear  Parker:

    thanks for your reply. for PCIE differential refclk, its cycle-cycle  need be less than 150ps. and its RMS jitter  after filtering by  LF and HF need be less than 0.5ps.

    furthermore, it required the rise edge of refclk+ to match the fall edge of refclk-, and less than20%.

    so  this switch is available for differential refclk? did your customer have used it like this before?

    by the way, can I use  the ibis model of this switch to simulate ,and add the clock soruce ,if the reslut is ok, it means the switch can work?

    thx

  • 0 in reply to user6257239
    TI__Guru 52882 points

    Hi,

    So this part shouldn't add much jitter - as the channel is really just a RC low pass filter - so there is a prop delay but it should apply to all pulses through the channel the same way. 

    This switch is only 1 channel and I wouldn't use is as a differential switch as two IC's can have slight differences in prop-delays can cause issues. For a differential clock a 2 channel 2:1 device would be best - we have a few options that have low On resistance and On Capacitance. Some options are the TMUX136 and the TMUX1072 -> these are higher bandwidth devices so that means they have lower capacitance which will distort the signal less.

    As for the IBIS model - it models SI of a signal going through the channel in typical operating conditions. It would be a good approximation to see if this part will work under normal operating conditions. However, if the application is pushing the parts to their limits or you are looking for guaranteed performance IBIS simulations will not be the greatest tools as it is used more for system level integrity tests at nominal operating conditions.

    Please let me know if you have any other questions !

    Best,

    Parker Dodson

    The IBIS model