Steve, Well put. I'd route the two signals over the plane named "ground" for the memory to reduce chances of coupling supply noise (ripple) into the clock as well. Lee W. Ritchey Speeding Edge P. O. Box 2194 Glen Ellen, CA 95442 Phone- 707-568-3983 FAX- 707-568-3504 I just used the energy it took to be angry to write some blues. Count Basie > [Original Message] > From: steve weir <weirsp@xxxxxxxxxx> > To: <james.f.peterson@xxxxxxxxxxxxx>; <si-list@xxxxxxxxxxxxx> > Date: 8/17/2004 4:41:53 AM > Subject: [SI-LIST] Re: routing of DDR clock signals > > Jim, > > This is a favorite hot button of Lee's. They are a pair of single-ended > difference signals as you describe. With the very rare exception of a > tandem net forming an isoscoles triangle with the pair, the pair is > susceptible to aggressor noise just like any other signal. The only > difference is that the gain sensitivity is one-half ( see the formula below ). > > To maintain reliable timing the rising and falling edges need to cross at > the receiver. In order to match that timing, and not disturb the signals, > we want to: > > 1) Match the physical circumstances of the routing path, which means > routing both signals reasonably close together, but not so close that there > is significant odd mode coupling that will complicate the termination ( > unless we specifically design for and account for that, see b. ). We need > to terminate them properly, and have two choices: > > a. Couple loosely, we ignore the odd mode termination as insignificant > energy flows in that path. > b. Couple tightly, in which case we would want to terminate both the odd > and even modes, which is more engineering effort. This might be justified > if we are tight on space. Otherwise, it is easier to just use the loosely > coupled solution. > > 2) Keep both traces far enough from aggressor nets that the timing does not > get disturbed. > > Remember that fundamentally what we are dealing with really is an analog > differential amplifier at the receiver. The amplifier will have its own > limitations in terms of common-mode rejection, and operating point > dependent delay. So, to keep timing jitter free, we want to limit the > common mode excursions, which depend on SSO at both ends and the path > matching, and differential noise that can be translated to a time offset, > approximately as: > > Tnoise = Vnoise * { Tfall, Trise } / ( 2 * Vswing ) > > We are pretty much at the mercy of whatever silicon we have in-hand on the > SSO front. But the crosstalk, and route matching are fully within our control. > > Steve. > > At 04:17 AM 8/17/2004 -0700, Peterson, James F (FL51) wrote: > >Steve, > > > >I'm a little confused by defining the DDR signals CK and CK# as > >differential. > >Does the return current from CK want to flow back into CK#? (I don't think > >it does.) > >I believe these are both single-ended buffers that try to be exactly 180 > >degrees out of phase from each other (which is the reason we need to match > >the routed length so closely). > >I don't believe we want any coupling between these two critical nets and > >there is no need for a diff-Z. > >( all : Since I'm currently implementing a DDR interface, please let me know > >if this incorrect! thanks.) > > > >Jim Peterson > >Honeywell > > > >-----Original Message----- > >From: steve weir [mailto:weirsp@xxxxxxxxxx] > >Sent: Monday, August 16, 2004 8:11 AM > >To: hencox@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx > >Subject: [SI-LIST] Re: routing of DDR clock signals > > > > > >Henrik, if you are new at this, you will probably get the best results by=20 > >treating the pair as a loosely coupled differential pair. That means=20 > >providing greater than normal horizontal spacing between the CK and CK#=20 > >than normal signals. Then the differential impedance will be close to 2X=20 > >the normal mode impedance of each line. Additionally, timing aberrations=20 > >caused by crosstalk can be reduced by affording 5H spacing between either=20 > >CK or CK# and the nearest outside neighbor signal. Remember that you are=20 > >matching electrical length. The easiest way to do that is to keep the=20 > >physical routes reasonably close together and then match physical length. > > > >Steve > > > >At 01:38 PM 8/16/2004 +0200, Henrik Gild wrote: > > >Hi! > > > > > >This is my first post to this list. Looking at some previous posts I'm sure > > >you'll be able to answer this one for me: > > > > > >When routing DDR SDRAM signals (ordinary DDR1, 166MHz, DDR333), should the > > >CK and CK# lines be treated as a differential pair or just like any other > > >signals (like DQ or DQS) only that they should be of equal length? I had a > > >look in the datasheet on the MT46V16M16 from Micron and it didn't say > > >anything about differential clock lines, just that the signal edges has to > > >coincide time-wise. But the clock signals were described as differential in > > >their IBIS model... > > > > > >Thanks > > >/Henrik Gild=E5, B.S.CE > > > > > >_________________________________________________________________ > > >Chat: Ha en fest p=E5 Habbo Hotel > > >http://habbohotel.msn.se/habbo/sv/channelizer Checka in h=E4r! > > > > > >------------------------------------------------------------------ > > >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 FAQ wiki page is located at: > > > http://si-list.org/wiki/wiki.pl?Si-List_FAQ > > > > > >List technical documents are available at: > > > http://www.si-list.org > > > > > >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 FAQ wiki page is located at: > > http://si-list.org/wiki/wiki.pl?Si-List_FAQ > > > >List technical documents are available at: > > http://www.si-list.org > > > >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 FAQ wiki page is located at: > > http://si-list.org/wiki/wiki.pl?Si-List_FAQ > > > >List technical documents are available at: > > http://www.si-list.org > > > >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 FAQ wiki page is located at: > http://si-list.org/wiki/wiki.pl?Si-List_FAQ > > List technical documents are available at: > http://www.si-list.org > > 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 FAQ wiki page is located at: http://si-list.org/wiki/wiki.pl?Si-List_FAQ List technical documents are available at: http://www.si-list.org 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