• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 35 subscribers
  • Views 547 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.

Original question:

DS90LV804: how to consider about Device to Device skew

DS90LV804: Part to Part Skew

Pero Krivic
Prodigy 40 points
Part Number: DS90LV804


Tool/software:

Hello, 

in the datasheet, the part to part skew is specified at 1.1ns. Is that the maximum allowed skew between any two parts or is it possible that, in case of multiple buffers on board, the skew accumulates?

Example: 3 buffers on a board. There is 1.1ns skew between Buffer 2 and Buffer 1. There is also 1.1ns skrew between Buffer 3 and Buffer 2. Now, is it allowed that buffer between Buffer3 and Buffer1 gets as high as 2.2ns?

In my project there will be 16 Buffers alltogheter, so I am worried that the skew explodes.

Best regards

  • 0
    Guru 318050 points

    The skew is between any two chips that work in parallel, i.e., with exactly the same input signals.

    If you have a chain of buffers that are connected in series, then the skew can accumulate.

  • 0 in reply to Clemens Ladisch
    Prodigy 40 points

    Hello Clemens, thanks for taking the time to answer my question. My buffers are not connected in series, but distributed to buffer 64 LVDS signals in parallel from the same IC. 

    So, in other words, would you say that among these 16 distributed buffers, the maximum skew would still be 1.1ns?

    Best regards, 

    Pero

  • 0
    TI__Genius 15321 points

    Hi Pero,

    The part to part skew metric (tSK(pp)) is a metric that was defined to characterize manufacturing process variability only. As footnote (4) indicates, it is not production-tested, but rather determined by statistical analysis on a sample basis at the time of characterization. It is the propagation delay time of separately packaged line circuits, and the metric is guaranteed as long as the supply voltage is the same and the temperature is within the recommended operating range. As Clemens mentions, the metric is based on devices in parallel, not series. Determining a "skew budget" for your application can also be helpful when estimating the amount of skew your system can tolerate.

    Regards, Amy

  • 0 in reply to Amy
    Prodigy 40 points

    As Clemens mentions, the metric is based on devices in parallel, not series.

    In my application, the devices will be "in parallel". So, can I expect more than 1.1ns of relative skew among them?

  • +1 in reply to Pero Krivic
    Guru 318050 points

    In your application, the skew indeed is < 1.1 ns.