Hi Jon, I used to design PCBs for automotive and off-highway applications. Most of the time, these were 4-6 layers (6 layers considered as luxury). Often, I had to split the power plane into 3 or more separate voltage planes. There are also a number of PC motherboards that have been routed on a 4-layer PCB. From SI perspective, it depends on how much of a "bump" you can tolerate on your signal/return path, and the sensitivity of the signaling scheme to this amount of bump/discontinuity. Generally, I would remain aware of signal and power return paths, and power and signal (and isolation from noise) integrity should follow. For example, I would reference the processor and the memory and other peripherals interfacing with it to the same power/GND plane. E.g., a 2.5-volt plane should be used as the reference for the I/O that is powered by the 2.5V. Also, isolation of different power supply rails, be it digital-digital or digital-analog, has proven to be achievable while maintaining adequate routing requirements w.r.t. return path impedance. If package can provide the coupling that Steve has mentioned (I.e., the power/return accommodate the low-impedance needed for power integrity by means of adequate on-board decoupling), then it maybe possible to use traces for power signals and forgo all power planes in favor of GND planes as the reference plane--Except when isolation is required... Regards, Oscar On 4/6/06, steve weir <weirsi@xxxxxxxxxx> wrote: > > Jon, your question covers a broad range of possibilities. Here are > two general cases I think are worth considering: > > 1. Suppose that we would like maximum route flexibility, that we > would like to route signals against any plane layer and switch > reference layers without concern. Then, we would need to tightly > couple the various planes together across our signal spectrum. In > the real world we can't get perfect coupling from DC to daylight. So > first we have to determine whether this approach is feasible, and > second we need to figure out what we need do to make it work so that > it better suits our goals than the alternatives. There isn't a > global answer. It boils down to numbers for specific cases. > > 2. Next suppose that we want to keep digital noise out of an analog > domain. To do that well we need to understand what needs to be > coupled, how we are going to couple it, and then how to really > isolate the rest. > > Steve > > At 08:42 PM 4/5/2006, Jon Anderson wrote: > >Greetings. > > > >I am curious to get some insight/opinions on power distribution > >approaches and the benefits of various configurations from an SI > >standpoint. Specifically, I am wondering how a split power plane > >within a single board layer affects SI, and the best methods for > >mitigating integrity issues (for example the use of multiple ground > >planes, or don't do split power planes since it's an SI no-no). > > > >It may be best to present an example. It is not unusual to see > >separate power (or ground) planes for analog and digital sections of > >a board. This is obviously done to provide isolation and limit the > >effects one section has upon the other. I am considering the > >tradeoffs of using multiple supply voltages/planes within a single > >board layer (which may be necessary in the case of an ) versus having > >to do unique layers for each voltage. Such a configuration might be > >necessary in the case of an ASIC that has multiple supply voltages > >but a common ground reference. Also, is it best to use matched > >power/ground layers for each voltage, even if the grounds are > >commonly referenced? If you do a split plane, are you as concerned > >with separate ground planes for each voltage? > > > >I understand that there are non-SI benefits to limiting the number of > >board layers (cost is the obvious one), but I'm strictly looking at > >this from a signal/power integrity standpoint. I also understand > >that the answers are likely application dependent, so I'm just > >looking for generalized reasoning why a specific configuration may or > >may not work well. I fully realize this is a series of very open- > >ended questions. > > > >For sake of arguement (and to simplify the situation), let's say that > >you can route all signals on the top and bottom layer, so any > >internal board layers would only be power or ground planes. I > >realize that the question becomes more complicated if you are looking > >to have internal ground planes specifically for isolation of signal > >planes. > > > >Thanks in advance. > > > > > > > >------------------------------------------------------------------ > >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