Perhaps someone could do a test board with two sets of signals, one "twisted" as described and the other done in the more traditional way. Make measurements and post your results :-) --- Matthew Lamkin <[log in to unmask]> wrote: > I'd suggest that UNLESS the engineer can come up > with hard evidence that > this is good practise then you should stick to your > guns and refuse to do > it. As evidence to the contrary is available to > prevent you wasting your > time & the cost involved in up-issuing & retesting > the boards. > > > > -----Original Message----- > From: Dennis Ward > [mailto:[log in to unmask]] > Sent: Wednesday, November 19, 2003 1:36 PM > To: [log in to unmask] > Subject: Re: [DC] Twisting Differential pairs > > > Thank you for all your replies. I'm glad I'm not the > only one who thought > this was > not good design practice. > > Thanks for the links, I've been to these sites, > checked the archives on the > IPC site, > did a google search on "differential pair routing", > "differential pair", > etc... and quite > honestly cannot find an article on this twist. Not > to single you out Denis, > but an reference > would be great to look at. The engineer said the > same thing, he saw an > article recommending > this but can't find the article right now. > > If the engineer comes up with any hard info on this > I will send it out under > this subject heading > so it can be archived. > > Dennis > > -----Original Message----- > From: DesignerCouncil > [mailto:[log in to unmask]]On Behalf Of Denis > Lefebvre > Sent: Tuesday, November 18, 2003 5:53 PM > To: [log in to unmask] > Subject: Re: [DC] Twisting Differential pairs > > > Don't freak out, Patrick et al. > I have seen some models of this and articles as > well. (sorry, can't > quote right now - I just remember seeing them - it > may have been the > CID+ material) My take is that its one of those > things that looks good > on paper - but show me some emperical data. Before > commiting to it, I'd > ask the engineer to do a "test coupon" to verify if > it'll work, and if > so, how well. > Just my $.02 > > Denis Lefebvre, CID+ > Senoir PCB Designer > Finisar Corp. > > -----Original Message----- > From: Patrick Jabbaz [mailto:[log in to unmask]] > Sent: Tuesday, November 18, 2003 2:47 PM > To: [log in to unmask] > Subject: Re: [DC] Twisting Differential pairs > > > Dennis > You must be kidding! > It sounds that the Engineer is thinking in terms of > cabling no external > reference. Differential signaling on a twisted > cable, you get excellent > coupling, however on a PCB it is pretty lousy, even > if you spaced them > very close together. You must also consider that > Broadside differential > routing, example L2 AND L3. The layers must be cores > in order to > maintain good registration layer to layer. The best > choice that I > recommend is to use edge couple side by side with a > good reference and > match the skew within the pair since so will not get > very good coupling > on the diff signals + > - > > There are several good articles on differential > routing go to > > http://www.sigcon.com/pubsChron.htm > > > Inkra Networks > Patrick Jabbaz CID > Sr Board Layout Eng. > 40971 Encyclopedia Circle > Fremont, CA 94538 > Work (510) 249-4835 > Mobile (408) 621-6533 > [log in to unmask] > > > -----Original Message----- > From: Dennis Ward > [mailto:[log in to unmask]] > Sent: Tuesday, November 18, 2003 2:36 PM > To: [log in to unmask] > Subject: [DC] Twisting Differential pairs > > Good evening, > > I wonder if anyone has any experience, insight, > opinion on this subject. > The engineer would like his longer (> 6") > differential pairs "twisted". > Basically run the pair in a dual strip line (pairs > running on top of > each other) when the pair reaches 4 inches swap > layers on the pair. > > So for example - say we have a 6 layer card. L1 is > top, L2 is GND L3 > signal, L4 signal, L5 VCC and L6 bottom. > Differential pair (lets call it > Diff > pair) > running > on Signal 3 and Signal 4. Diff Pair+ on Signal 3, > Diff Pair- on Signal > 4. When the pair reaches 4 inches swap layers. Now > Diff pair- on Signal > 3 and Diff > pair+ > on Signal > 4. Repeat layer swap every 4 inches until end of > net. > > The only caveat he has on this is that this is not > to be done on Gigabit > lines. Other than that he would like to see this > model on his > differential routing. > > I have not heard of this nor can I find any info on > this. I'm curious if > anyone is doing this. Is this accepted practice? Is > it beneficial? > > Thanks for sharing your experience on this matter > > Dennis > > ------------------------------------------------------------------------ > --------- > DesignerCouncil Mail List provided as a free service > by IPC using > LISTSERV 1.8d To unsubscribe, send a message to > [log in to unmask] with > following text in the BODY (NOT the subject field): > SIGNOFF > DesignerCouncil. To set a vacation stop for delivery > of DesignerCouncil > send: SET DesignerCouncil NOMAIL Search previous > postings at: > www.ipc.org > On-Line Resources & Databases > > E-mail Archives > Please visit IPC web site > http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for > additional > information, or contact Keach Sasamori at > [log in to unmask] or 847-509-9700 > ext.5315 > ------------------------------------------------------------------------ > --------- > > ------------------------------------------------------------------------ > --------- > DesignerCouncil Mail List provided as a free service > by IPC using > LISTSERV 1.8d To unsubscribe, send a message to > [log in to unmask] with > following text in the BODY (NOT the subject field): > SIGNOFF > DesignerCouncil. To set a vacation stop for delivery > of DesignerCouncil > send: SET DesignerCouncil NOMAIL Search previous > postings at: > www.ipc.org > On-Line Resources & Databases > E-mail > Archives Please > visit IPC web site > http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for > additional information, or contact Keach Sasamori at > [log in to unmask] or > 847-509-9700 ext.5315 > ------------------------------------------------------------------------ > --------- > > ---------------------------------------------------------------------------- > ----- > DesignerCouncil Mail List provided as a free service > by IPC using LISTSERV > 1.8d > To unsubscribe, send a message to [log in to unmask] > with following text in > the BODY (NOT the subject field): SIGNOFF > DesignerCouncil. > To set a vacation stop for delivery of > DesignerCouncil send: SET > DesignerCouncil NOMAIL > Search previous postings at: www.ipc.org > On-Line > Resources & Databases > > E-mail Archives > Please visit IPC web site > http://www.ipc.org/contentpage.asp?Pageid=4.3.16 > for additional information, or contact Keach > Sasamori at [log in to unmask] or > 847-509-9700 ext.5315 > ---------------------------------------------------------------------------- > ----- > > ---------------------------------------------------------------------------- > ----- > DesignerCouncil Mail List provided as a free service > by IPC using LISTSERV > 1.8d > To unsubscribe, send a message to [log in to unmask] > with following text in > the BODY (NOT the subject field): SIGNOFF > DesignerCouncil. > To set a vacation stop for delivery of > DesignerCouncil send: SET > DesignerCouncil NOMAIL > Search previous postings at: www.ipc.org > On-Line > Resources & Databases > > E-mail Archives > Please visit IPC web site > http://www.ipc.org/contentpage.asp?Pageid=4.3.16 > for additional information, or contact Keach > Sasamori at [log in to unmask] or > 847-509-9700 ext.5315 > ---------------------------------------------------------------------------- > ----- > > --------------------------------------------------------------------------------- > DesignerCouncil Mail List provided as a free service > by IPC using LISTSERV 1.8d > To unsubscribe, send a message to [log in to unmask] > with following text in > the BODY (NOT the subject field): SIGNOFF > DesignerCouncil. > To set a vacation stop for delivery of > DesignerCouncil send: SET DesignerCouncil NOMAIL > Search previous postings at: www.ipc.org > On-Line > Resources & Databases > E-mail Archives > Please visit IPC web site > http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for > additional information, or contact Keach Sasamori at > [log in to unmask] or 847-509-9700 ext.5315 > --------------------------------------------------------------------------------- __________________________________ Do you Yahoo!? Protect your identity with Yahoo! Mail AddressGuard http://antispam.yahoo.com/whatsnewfree --------------------------------------------------------------------------------- DesignerCouncil Mail List provided as a free service by IPC using LISTSERV 1.8d To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF DesignerCouncil. To set a vacation stop for delivery of DesignerCouncil send: SET DesignerCouncil NOMAIL Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315 ---------------------------------------------------------------------------------