Why do you say signals must stay on layer 1? Are you familiar with layer change caps?
Yes. I dont think they are not good enough for return current for multiple reasons including via inductance, capacitance resonance (a 10 ns rise time can be found on a cheapo MCU. That's 100 MHz, way past your average 0.1 uF mlcc's 2-10 MHz resonance).
The only good way for return current to go from pwr to gnd layers is if they are REALLY close together. In a 10 layer board that has a [s g s s g p s s g s] stackup, you can go from 4 to 7 and vise versa without a stitching via just as you can do so with layers 1 and 3. You can go from layer 1 to 4 with a stitching via, and then move the trace to layer 7 later on, but you cannot go from layer 1 to 7.
Meanwhile a [s g p s] 4 layer stackup is one of the worst stackups you can use, unless on that stackup signals never leave their layers. A [s g p s] 4 layer stackup would work well enough for transitioning layers, but no one in their right mind would do that.
The stackup I suggested and said to never move a signal from layer 1 was sig/pwr/gnd/sig/gnd/sig, which with their "best" stackup for this would look like [s p g s g s]. They also offer it like this [s p g s g s].
Imo the best 6 layer stackup is [s g s p g s], which they dont offer.
Home
Help
Search
Login
Register
