Chris, I suggest you do the following experiment (I've done it so many times to toy with EMC engineers). Take a system board with your favorite sub nf caps, do an EMI scan, replace every single sub nf caps with .1uf or .01uf on the same PCB and redo your scan. Let's see if they make any difference. I never lost my bets but you can be the first one. -----Original Message----- From: Chris Padilla [mailto:cpad@xxxxxxxxx] Sent: Tuesday, November 27, 2001 3:41 PM To: chris.cheng@xxxxxxxxxxxx Cc: si-list@xxxxxxxxxxxxx Subject: Re: [SI-LIST] Re: Sanmina patent on dielectric thicknesses under 4 mi ls I use plenty of sub nf caps to help curb emission problems in a lot of my systems and they do work just fine and are necessary. However, you point specifically to core noise (presumabley on some large ASIC or processor or something) and while that is not an area I specifically deal with, it certainly comes into play. EMI is a system level phenomenon and the size of your PCB most certainly can resonate at frequencies well above 100 MHz. I fear I am missing your point here.... ----->Chris EMC Engineer Cisco Systems >Ron, >May I ask what is the intend use for the sub nf caps ? > >By no means I claim there is absolutely no need for sub-nf caps, >however, I always question the need for sub-nf caps for EMI or >noise decoupling in a digital system. I believe the place for >such decoupling is either on die or package not on the PCB. >In fear of sounding like a broken record again, I claimed : >a) High frequency core noise cannot propagation through the >package back to the PCB. Same reason you have sso limitation >on a package, its just as bad to go out of than go into a >package. Core noise above 100MHz can not be decoupled >outside the package. >b) I/O switching noise is related to the image current >return path which exist between the power plane sandwiching >the signal layers and dictated by the impedance control >of the stackup. > >I've used Zycon before thinking they can lower EMI or >power noise but the reality is they don't and the reasons >are stated above. > >If you have other reasons to use distributed sub nf >decoupling caps, I will keep by mouth shut. And I practice >what I preach. I never use decoupling below 1nf in any of >my systems and they pass class B just fine. > > > ------------------------------------------------------------------ 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 archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu