Re: [SI-LIST] : bypass cap question (long, simple)

Jeff Seeger ([email protected])
Tue, 03 Feb 1998 13:11:29 -0500

Christopher Donham wrote:
> > You could also reduce the end cost of the board by reducing the
> > number of layers to 7 instead of 8 by using a split power plane. Thus
> > your stackup could look like this:
> >
> > 1 Horizontal
> > 2 V33 Plane/V5 Plane
> > 3 GND Plane
> > 4 Vertical
> > 5 Horizontal
> > 6 GND Plane
> > 7 Vertical
> Hmmm -- I thought that boards had to be created from layer pairs
> I didn't know that a 7 layer could be easily constructed, and hence
> cost less than an 8 layer board. (I am a novice at these things --
> please correct me if I am wrong. I thought a board was constructed
> by etching metal on either side of a "core" and then pressing it
> together with "prepreg" in between. This would suggest an even
> number of layers).
> I also thought that to keep a board from warping, you wanted a
> stackup that was symmetrical. So having planes 2 and 3 towards
> the top and only plane 6 towards the bottom was also a cause for
> concern. How big does a board have to be for this to be a concern?
Chris, you're absolutely correct for mainstream board fab.
There are cases, such as high-layer-count backplanes, where
this is not true - but this is somewhat unusual.

Taking this further, I agree with Philip Gantt's premise
that splitting the voltage plane can work if you use
caps across the plane split. But I suggest that this is
the way to get to a six layer board:

1 Horizontal
2 V33 Plane/V5 Plane
3 Vertical
4 Horizontal
5 GND Plane
6 Vertical

Just to add a question to this fine discussion, what role
does the capacitor formed by the laminate between reference
planes play? A great many designs are organized in strip-
line pairs, with signals contained in such a cavity.


      Jeff Seeger                             Applied CAD Knowledge Inc
      Chief Technical Officer                      Tyngsboro, MA  01879
      jseeger "at" appliedcad "dot" com                    978 649 9800