You are seing a spectrum of MLT3 signal, which is 100 Mbit coding. Increased broadband emission is caused because some of the differential mode (DM) signal is either converted into common mode (CM) due to imbalance in the differential signalling or because you have some power/"ground" "noise", most likely power/"ground" bounce caused by the intentional signal over some common- impedance (e.g. bonding "gnd" wire in a chip).
The parts of the story that have something to do with the first case are balance of the output driver (impedance and signal timing and slope), balance of the components/layout on the way of the signal to the Ethernet connetor, including the (im)balance in the pinout (1-2 vs. 3-6) if you are using regular and not integrated connectors (this happens even at the relatively low frequencies that you have problem with). Lack of good balance in the center-tap of the transformer may also significantly increase DM-CM conversion. If you have "noisy" environment, where a part of the intentional signal is converted into (CM) noise on the power/"ground" structure, then it can increase the CM "driving" voltage (relative to the chassis i.e.enclosure) on the pairs. Center taps on both sides of the transfromer may be paths for CM noise to override the CM chokes and get on the UTP. Be careful with how and where you connect the center taps. You may get some idea of what is happening if you can "force link", in which you activate the Tx without actually establishing link with the AUX equipment. Your SW guys should be able to help you with that. That will keep the Tx active regardless of what you connect. I can not help you with the details here, but I hope you can use this and think of some experiments you may perform to figure out what is causing the problems. Good luck, Neven > Hi Andy, > > Most of the ports are linking to offsite support equipment or looped back to > get > link. The response seems consistent regardless of how it is linked. There is > also some tuning; broadband humps up a little higher in one spot or another > but > not a real big factor. I have tried 50 ft and shorter 3 to 10 ft on all > ports, > and tried ferrites on the offsite portions of the cables. > > Using STP on the offsite portion of the cable will be an interesting test. > > Thanks for the input. > > Rick Linford > > > > > -----Original Message----- > From: Andy White (EWU) [mailto:andy.wh...@ewu.ericsson.se] > Sent: Wednesday, October 09, 2002 3:39 PM > To: Rick Linford > Subject: RE: Ethernet Radiated Emissions > > Hi Rick, > Is the problem only when you link more than 2 cables together? What is the > length of link cabling when adding cables? I susepect that the BB noise is > directly associated with the increased cabling and the lengths becoming > resonant. > If cable lengths are the issue for the BB, the RE problem is not that easy to > resolve. Changing to another UTP will not really improve it, a section of STP > (on a link or some of the links) would probably help. I would give it a try. > Perhaps a cable ferrite may work on a link cable as well to reduce the BB > lump > amplitude. 100bT uses a 25MHz source so it may be possible to add some > filtering > at the pcb but this will probably affect the ENET operation/functionality. > Hope you find a simple solution. > I liked your 'thanks to the marketing dept for the opportunity to learn' > comment > - classic :) > all the best > Andy > ____________________________________ > Andy White, > Staff EMC Engineer > Ericsson Wireless Communications Inc. > San Diego, CA 92121 > Tel 858 332 6214 / 877 877 7799 ext 26214 > Fax 858 332 7311 > e-mail andy.wh...@ewu.ericsson.se > ____________________________________ > -----Original Message----- > From: rlinf...@sonicwall.com [ mailto:rlinf...@sonicwall.com] > Sent: Wednesday, October 09, 2002 1:55 PM > To: emc-p...@majordomo.ieee.org > Subject: Ethernet Radiated Emissions > Having worked on Ethernet (100 MB UTP) EMC for a few years, I now find myself > lacking understanding of Ethernet communications and how this may be > affecting > radiated emissions. > What is seen on the OATs is broadband noise between 50 and 150 MHz. With two > cables connected and linked, broadband is not even out of the noise floor. > With > the third and subsequent cables linked the broadband increases With all > cables > attached and not linked there is no broadband. Reducing the resolution shows > peaks every 60 kHz. Probing on a linked signal shows peaks every 30 kHz with > every other peak (60 kHz) being 20 dB higher. Probed several other types of > Ethernet equipment, NICs and switches and this seem consistent across all > products and manufactures. I feel the filters and layout is quite good, in > that, > with cables attached but no link there is no emission. In probing on the > board > it self the 30 kHz signature is present only on the transmit and > receive all > other signals and voltages appear clean. > Is this the nature of Ethernet 100 MB over Unshielded Twisted Pair? > Could this be the fault of the UTP cables? If so any suggestions on the type > of > UTP cable to purchase? > Thanks in advance for your time and responses. > Rick Linford > And thanks to a marketing department for this opportunity to learn, by > believing > 6 ports of 100 MB is need in the home. > > ------------------------------------------- This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ To cancel your subscription, send mail to: majord...@ieee.org with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Ron Pickard: emc-p...@hypercom.com Dave Heald: davehe...@attbi.com For policy questions, send mail to: Richard Nute: ri...@ieee.org Jim Bacher: j.bac...@ieee.org All emc-pstc postings are archived and searchable on the web at: http://ieeepstc.mindcruiser.com/ Click on "browse" and then "emc-pstc mailing list"
