Hello all, I have had quite alot of grief with using small valued ceramic capacitors (in the 100pF-1000pF range) in parallel with large valued capacitors due to the parallel resonant tank circuits that are caused by the interaction of the capacitor parasitics. This would cause our products to fail radiated emissions, and tweaking the capacitor values would only cause the resonant point to shift -- improving one harmonic but making another even worse. I now use 0.1uF ceramic capacitors exclusively as my smallest amount of decoupling capacitor, and have never had resonant issues again. Since the self resonant frequency of a 0.1uF ceramic capacitor is around 30MHz, you are operating above the series resonance of these caps, so minimizing the inductance is paramount. I have done some measurements of this tank circuit phenomenon a few years back, and have written an article about it. The link to this is http://www.spectrumcircuits.com/articles/articles.html. Howard Johnson, the author of the SI book "High-Speed Digital Design, a Handbook of Black Magic" also discusses why it is only the impedance of the capacitors that is important for decoupling, not the self resonant point. His article is at http://www.signalintegrity.com/news/2_3.htm. Wayne Cooke Signal Integrity Engineer, Innovance Networks 19 Fairmont Ave., Ottawa, Ont. K1Y 1X4 email: wcooke@xxxxxxxxxxxxx -----Original Message----- From: npischl [mailto:npischl@xxxxxxxxx] Sent: Friday, May 17, 2002 4:43 PM To: si-list@xxxxxxxxxxxxx Subject: [SI-LIST] Re: Decoupling capacitors Chris, what I believe is hidden between your lines is that you didn't use the caps all over the board that are 10x or 100x from each other's value. I am sure you needed to !appropriately! place and mix the values. Sometimes you need low-value caps to tune at a particular "high frequency" in order to provide efficient bypass and lower emissions. However, the boards that are designed with whole series of caps that are 10x from each other, e.g.0.1, 0.01, 0.001 etc. are sure to have wronlgy selected cap values (they may still work though :). The series resonances of these caps just can't coincide with the harmonics of the signals. And in such cases, the "benefit" of having some "high-frequency" bypass is counteracted by paralel resonances that such an arrangement will cause. It is of paramount importance to get the L down first, by proper layout. Once you have done it, then worry about the proper selection and mix of the C-values, which will depend on the exact spectral content of the signals on the board. All from above was in the first place from te point of view of reducing board emissions, and "high-frequency" in this sanse may be much higher than the useful signal spectrum of 1/2Tr or similarly defined. Desinging for the useful spectrum, which the SI community is mostly concerned with and some excellent contributirs on this list have been preaching, requires more than that. Neven --- In si-list@xxxx, Chris Padilla <cpad@xxxx> wrote: > > This is my experience as well. In several cost reduced boards in which > they immediately strip out "superfluous" layers, i.e. ground planes, I've > had to make it up with many, appropriately placed/routed (sometimes > appropriately mixed) sub-100 pF capacitors. Thank goodness my 10/100 and > 10/100/1000 48-port ethernet boards with miles of Cat5 UTP loading them do > not need to pass Class B...Class A can be challenging enough! > > Chris Padilla > EMC Engineer > Cisco Systems > > At 10:20 AM 5/17/2002 -0700, Lieby David wrote: > > >I have some experience with problems in RFI. To get some systems to pass > >FCC class B with enough margin it is necessary to add low value caps. The > >best is setting them at two orders of magnitude in sequence from 10pf to > >whatever. I like 10pf, 1nf, 0.1uf, 10uf placed in groups and a couple of > >extras at the corners of the board or plane. > > > >They really do make a difference in radiated emissions, whether they help > >SI or not > > > >dav0 > >dlieby@xxxx > >dav0@xxxx > > > >"mittalr@xxxx" wrote: > > > > > > > > > On a slightly tangential topic. Have you ever wondered the need for > > 100pf caps. Everyone just seems to sprinkle them as a safety measure but > > I see no need for them because > > > - the IC by itself will have about 50-100pF. I checked with many IC > > guys and the ICs where you really need decoupling are generally big > > enough to have that much decoupling on-chip. > > > - On top of this you have inter-plane caps. > > > > > > I have seen many of my designs get away with no 100pF caps. I would use > > 1nF and 0.1uF, rather than 100pf, 10nF and 0.1uF. Anyone have any other > > experience... > > > > > > ------------------------------------------------ > > > Join Excite! - http://www.excite.com > > > The most personalized portal on the Web! > > > > > > ---------------------------------------------------------------- -- > > > To unsubscribe from si-list: > > > si-list-request@xxxx 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@xxxx 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 > > > > >------------------------------------------------------------------ > >To unsubscribe from si-list: > >si-list-request@xxxx 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@xxxx 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 > > > > ------------------------------------------------------------------ > To unsubscribe from si-list: > si-list-request@xxxx 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@xxxx 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 ------------------------------------------------------------------ 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 ------------------------------------------------------------------ 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