Hi Scott, Can you share any of the studies you've performed to characterize the dozens of materials? I'm especially interested in clever techniques you might have used to ensure you measured the maximum for each material, given the statistical nature of the problem (whenever I think of performing such a study, I am put off by "how do I ensure my data includes some representative 'worst case'?"). To all, I believe there is extreme ambiguity in the terms used for spread glass, I would be wary of any particular terminology (is "ultra" better than "hyper"?). And, even if you do standardize the definition of "spread", the net effect is going to vary significantly for the various glass styles (1080 will probably always be more problematic than 2116, for instance). I don't know of any method of easily quantifying the electrical properties of any particular spread glass. There is work on-going for a visible light test, but as far as I know those results have not yet been correlated to Vp differences. BTW, this might be a nice opportunity for a clever person to come up with an easy (cheap, HVM-compatible) method of shining a light through glass weaves and quantifying the difference between the brightest and darkest areas - ideal "spread glass" would have very little difference, corresponding to very little difference in Vp across the weave. And finally, I don't know of any silicon solutions, ours or otherwise, used explicitly to solve the fiberweave issue. Adaptive equalization may make it less of an issue than it was previously, however. Jeff Loyer -----Original Message----- From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of Scott McMorrow Sent: Friday, December 05, 2014 5:42 AM To: Havermann, Gert Cc: billh@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx Subject: [SI-LIST] Re: AW: Glass-Weave Skew / Fiber-Weave Effect Gert, Even flat weaves have skew, as easily evidenced by a picture of the weave that shows gaps in the overlap of fiber bundles in one direction. I've found that the killer problem is skew on line cards causing diff to common mode conversion that bursts as crosstalk within connector PTH via fields and connectors themselves, which are not designed to control common modes. This was seen specifically in a significant loss of NEXT margin at receivers from Tx or Rx card skew causing excessive crosstalk in the connector pin fields for a well-known standard. This required two solutions, one was to guardband margin to allow for the additional skew caused NEXT. The other was to use some skew abatement methods. Doing 1000's of sensitivity runs I've found that this skew sensitivity is much worse on cards than it is on backplanes. That is, the card skew allowance has much less tolerance than does the backplane or total end-to-end skew. Where the skew is located is much more important than how much skew there is. To put some numbers to this phenomena, a 10G link had 12 ps of skew, which translates to around -0.7 dB of additional insertion loss. That is generally not a serious issue for 99.9% of all designs. However when the effective loss due to crosstalk was factored in, due to diff to common mode and common mode to diff conversion, total additional eye loss was -3.9 dB. which is a huge impairment. In the end, the only solution for skew, especially at 28G rates and above, is to utilize special techniques to mitigate skew on backplanes, and skew-free materials on cards. I've measured and characterized dozens of materials and can say that I can easily demonstrate the potential for any woven PCB laminate to have significant skew, even those with spread weaves, except for those laminates that are specifically engineered for zero-skew with matched glass and resin dielectric constants. Regards, Scott Scott McMorrow Teraspeed® Consulting - A Division of Samtec 16 Stormy Brook Rd Falmouth, ME 04105 (401) 284-1827 Business http://www.teraspeed.com On Fri, Dec 5, 2014 at 3:29 AM, Havermann, Gert <Gert.Havermann@xxxxxxxxxxx> wrote: > Hi Bill, > I've seen it quite often, but I never seen it killing System > performance as we work with good margin, and the Silicon accepts even > more skew that expected. > One Design was not usable due to weave effect skew. It was a TRL > Cal.Stucture where the shortest line hat skew exceeding the Phase > difference. After that I redesigned with different Material and flat > weave Glass, and that worked great. > > BR > Gert > > > ---------------------------------------- > Absender ist HARTING Electronics GmbH, Marienwerderstraße 3, D-32339 > Espelkamp; Registergericht: Amtsgericht Bad Oeynhausen; Register-Nr.: > HRB 8808; Vertretungsberechtigte Geschäftsführer: Dipl.-Kfm. > Edgar-Peter Düning, Dipl.-Ing. Torsten Ratzmann, Dipl.-Wirtschaftsing. > Ralf Martin Klein > > -----Ursprüngliche Nachricht----- > Von: si-list-bounce@xxxxxxxxxxxxx > [mailto:si-list-bounce@xxxxxxxxxxxxx] > Im Auftrag von Bill Hargin (Nan Ya, USA) > Gesendet: Freitag, 5. Dezember 2014 06:04 > An: si-list@xxxxxxxxxxxxx > Betreff: [SI-LIST] Glass-Weave Skew / Fiber-Weave Effect > > Hi Folks … > > I’m doing a bit of research on glass-weave skew / the fiber-weave effect. > I’ve read the articles/presentations, and understand the nature of the > beast, but I’m interested in getting data from the design trenches. > > Are you (or do you believe you’ve) seen it in your designs? What > happened, and how did you resolve it? (E.g., angled routing, low-Dk > glass, homogeneous resin/glass, etc.) What were the characteristics > of the material and signals? > > I have no idea what I’m going to hear in response, but if you respond > offline, I’ll hold the info you provide in confidence. If you reply > publicly – I promise not to share your secrets beyond the SI-List … > > Bill Hargin > Director of North American Sales / Marketing Nan Ya Copper-Clad Laminates > billh@xxxxxxxxxxxxx ▪ 425-301-4425 ▪ Skype: bill.hargin > > > > > > > > ------------------------------------------------------------------ > To unsubscribe from si-list: > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field > > or to administer your membership from a web page, go to: > //www.freelists.org/webpage/si-list > > For help: > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field > > > List forum is accessible at: > http://tech.groups.yahoo.com/group/si-list > > List archives are viewable at: > //www.freelists.org/archives/si-list > > Old (prior to June 6, 2001) list archives are viewable at: > http://www.qsl.net/wb6tpu > > > ------------------------------------------------------------------ > To unsubscribe from si-list: > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field > > or to administer your membership from a web page, go to: > //www.freelists.org/webpage/si-list > > For help: > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field > > > List forum is accessible at: > http://tech.groups.yahoo.com/group/si-list > > List archives are viewable at: > //www.freelists.org/archives/si-list > > Old (prior to June 6, 2001) list archives are viewable at: > http://www.qsl.net/wb6tpu > > > ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List forum is accessible at: http://tech.groups.yahoo.com/group/si-list List archives are viewable at: //www.freelists.org/archives/si-list Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List forum is accessible at: http://tech.groups.yahoo.com/group/si-list List archives are viewable at: //www.freelists.org/archives/si-list Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu