Re: [PEDA] Twisted-pair tracks
Cam: How did you test it? What were you testing for? Twisted pairs are good for reducing interference from other sources. They are less effective at reducing interference emitted by your device. The principle is that EM fields fluxing through the loop area formed by your signal conductors generates currents (interference) in your signal conductors. By reducing the loop area you can reduce the interference. But the loop area cannot be reduced to zero, so there will always be some interference. To reduce interference even further, twist the conductors around each other. What this does is make the EM flux currents cancel, since the flux current going through one twist of the loop cancels the flux current going through the adjacent loop. This is why twisted pair is used by 10/100baseT ethernet. You can run 300 ft. ethernet drops of CAT5 and still get good signal integrity, even with all the trash in the EM spectrum (and getting trashier every day). On a PCB, you would have to use vias for your loop crossovers. You would also need to make sure the lengths of each track segment are identical, since symmetry is vital to make the concept work. And put the traces as close together as possible and still be manufacturable. If you are also trying to achieve a tightly controlled impedance, that would be tricky... Best regards, Ivan Baggett Bagotronix Inc. website: www.bagotronix.com - Original Message - From: Cam Andruik [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Monday, September 16, 2002 12:47 PM Subject: [PEDA] Twisted-pair tracks Has anyone ever tried to fake creating a twisted-pair on a PCB? I tried doing so on a recent board and our testing indicates that it did nothing. I think it is a waste of time to even attempt it but some people here think it is helpful. Thanks for any input, Cam Andruik Harding Instruments * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * To post a message: mailto:[EMAIL PROTECTED] * * To leave this list visit: * http://www.techservinc.com/protelusers/leave.html * * Contact the list manager: * mailto:[EMAIL PROTECTED] * * Forum Guidelines Rules: * http://www.techservinc.com/protelusers/forumrules.html * * Browse or Search previous postings: * http://www.mail-archive.com/proteledaforum@techservinc.com * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Re: [PEDA] Twisted-pair tracks
Cam Andruik wrote: Has anyone ever tried to fake creating a twisted-pair on a PCB? I tried doing so on a recent board and our testing indicates that it did nothing. I think it is a waste of time to even attempt it but some people here think it is helpful. I did this on one board that had analog differential signals right next to 10 A, 80 V pulses with 40-100 nS Tr. It probably works, as there was no measurable crosstalk on the analog signals. I didn't however, make the board both ways and compare. I wouldn't even bother with two comparable low-level signals. Jon * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * To post a message: mailto:[EMAIL PROTECTED] * * To leave this list visit: * http://www.techservinc.com/protelusers/leave.html * * Contact the list manager: * mailto:[EMAIL PROTECTED] * * Forum Guidelines Rules: * http://www.techservinc.com/protelusers/forumrules.html * * Browse or Search previous postings: * http://www.mail-archive.com/proteledaforum@techservinc.com * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Re: [PEDA] Twisted-pair tracks
I also looked at this for long LVDS on printed flex cabling and the effect was marginal. This was not a full twist (utilizing vias) but rather superimposed snaking traces on a thin base material. Instead we opted for controlled impedance traces grouped to minimize cross talk or isolated with guard traces. The only place I've used this with appreciable benefit is in minimizing magnetic signatures in sensitive instruments, and there the primary benefit was to reduce the number of areas to calculate and sum. John Jon Elson wrote: Cam Andruik wrote: Has anyone ever tried to fake creating a twisted-pair on a PCB? I tried doing so on a recent board and our testing indicates that it did nothing. I think it is a waste of time to even attempt it but some people here think it is helpful. I did this on one board that had analog differential signals right next to 10 A, 80 V pulses with 40-100 nS Tr. It probably works, as there was no measurable crosstalk on the analog signals. I didn't however, make the board both ways and compare. I wouldn't even bother with two comparable low-level signals. Jon -- John M. Cardone Electro-Mechanical Dsgn. Engr. Grp. M/S 278-100 Mechanical Engineering Section, 352 4800 Oak Grove Dr.NASA / Jet Propulsion Laboratory Pasadena, Ca 91109 Tel: 818.354.5407 MailTo:[EMAIL PROTECTED] Fax: 818.393.6400 Cell: 818.653.7818 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * To post a message: mailto:[EMAIL PROTECTED] * * To leave this list visit: * http://www.techservinc.com/protelusers/leave.html * * Contact the list manager: * mailto:[EMAIL PROTECTED] * * Forum Guidelines Rules: * http://www.techservinc.com/protelusers/forumrules.html * * Browse or Search previous postings: * http://www.mail-archive.com/proteledaforum@techservinc.com * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
