2

According to Microchip's document Implementation Guidelines for Microchip’s USB 2.0 and USB 3.1 Gen 1 and Gen 2 Hub and Hub-Combo Devices, the Superspeed differential pairs should be length matched while considering phase. I.E. the length-matching squiggles are inserted as close as possible to the location of the phase error:

Length matching considering phase.

The squiggles are inserted directly into the corners.

Questions:

  1. Does this type of length matching have a name?
  2. Does Altium have a way to generate this geometry? If so, how?
  3. If not, how can one create this geometry correctly?
Rocketmagnet
  • 26,933
  • 17
  • 92
  • 177
  • 1
    Serpentine traces is a term I (and others) use to describe those little squiggles. – SteveSh Mar 22 '21 at 12:06
  • @SteveSh - Does that refer specifically to the squiggles inserted into the corners? Or does it refer generally to squiggles anywhere along the trace? An Image Search for [serpentine traces](https://www.ecosia.org/images?q=Serpentine%20traces) seems to bring up no examples of corner squiggles. – Rocketmagnet Mar 22 '21 at 12:12
  • Usually anywhere along a trace, at least in the way I use the term. Looks like what they may be doing is changing the geometry slightly to compensate for slight capacitive changes at the corners. – SteveSh Mar 22 '21 at 12:26
  • 1
    One other thing. I see where the artwork has a comment that "all five USB pairs are perfectly matched end to end". That sound like marketing BS to me. Nothing is perfect. A good engineers would say something like "all five USB pairs are matched to withing 5 deg of each other at 2 GHz". – SteveSh Mar 22 '21 at 12:29
  • @SteveSh - Good point. I assumed (being charitable) that they meant simply that the phase matching was much better than the usual case of adding squiggles close to the corners. Knowing exactly the phase error in practice would require measurement in practice, or simulation beyond the scope of the track editor. – Rocketmagnet Mar 22 '21 at 13:45

0 Answers0