No announcement yet.

impedace control on flex rigid

  • Filter
  • Time
  • Show
Clear All
new posts

  • impedace control on flex rigid

    hi i finish the high speed design coruse and i tring to design flex rigid board for our connector (Fischer_MR11WL08_0420_BK1_E1AP to LSH-030-01-G-D-A) ,
    i have eternet pairs on my deisgn and im tring to calculate the width for 50 ohm single ended and 100 ohm diff parirs,
    acoording to this stackup
    Click image for larger version  Name:	cap1.png Views:	0 Size:	41.0 KB ID:	16510
    the numbers dont add up,
    Click image for larger version

Name:	cap2.png
Views:	68
Size:	150.6 KB
ID:	16512
    what do you think?
    Last edited by Kfir2020; 02-01-2021, 02:11 AM.

  • #2
    Better to ask the PCB Manufacturer with whom you got this stack-up.


    • #3
      I agree with Lakshmi, if it is important I would try to ask the PCB manufacturer. But it looks like, you are using PCBWAY, so there may not be many options.

      The problem is, that 2mil (50um) between copper layers is really small (it means, you would need to use very narrow tracks to get 50 OHMS) and I double checked PCBWAY website and it looks like they only have that kind of stackup:

      What do do? I am not sure .... maybe risk impedance mistmatch? I have seen many Ethernet connector breakouts connected by ribbon cables working just fine (not exact impedance matching, I found that ribbon cable impedance is like 110-130ohms diff (?)). Maybe it may work also with flex just fine ???


      • #4
        Hi Kfir2020

        I've experienced a very minor amount of discrepancy between Altium, Saturn PCB Design and my fabricator. But I'm talking ±1Ω for both single ended and differential signals. Even coplanar edge coupled is within that range.

        One thing to keep in mind is that the Er listed in the dielectric materials datasheet is almost never the value that you use because it changes by thickness.
        Your fabricator should be able to provide you with the active Er (for lack of a better term).
        They may also provide you with the Er of the adhesive and the mask (or bikini coverlay).
        The equivalent Er of the active part of the stack is what you use.
        e.g. if I have a bunch of cores at 100µm thickness and 3.31 Er, and prepregs at 66.7µm and 3.51 Er, then the equivalent Er between the 1 or 2 reference planes might be 3.4 or something.
        I always get this value from my manufacturer because I have no idea how to calculate it for each construction thickness and gum percentage. I would love to know but haven't been able to find out.

        The next thing I would look at is the plating. In my experience, if you don't specify how you want the flex to be plated, the manufacturer will probably use button plating (or whatever different manufacturers call it). Button plating is where they only surface plate exposed pads, untented vias, and other sections of unmasked copper. This improves the board flexibility, but also throws off your calcs because you've set Saturn PCB to use a plating thickness of 0.5oz when your manufacturer has probably removed it.

        Saturn PCB is also missing the ability to consider solder mask (flex coverlay). Altium and Polar can do this. Although the effect on the impedance is minor in comparison to my previous points, it will still contribute to the error. I honestly don't understand why it has an effect, but Altium seems to think it does. My understanding isn't enough to say otherwise.

        Kind regards,
        Last edited by DanR; 03-29-2021, 05:19 AM.