This question is as much for the good people at Aisler as for those using their services. I am wondering what stackup people have been using for the 6L boards, when designing high dencity/high speed boards. With “stackup”, I am referring to how the layers are used, and not the physical attributes of the layers.
My preferred stackup would look something like this:
L1 ----- Signal / Power L2 ----- Ground L3 ----- Signal L4 ----- Ground L5 ----- Ground L6 ----- Signal / Power
Note that all signal planes are at least 2W separated and that they all have a close reference plane. This makes routing very easy, without cross talk between the layers, and good EMI handling. As a side note, I like having 0.5oz internal layers to prevent bow and tie effects during reflow. Also, I am not a fan of power planes. Instead, I tend to route power rails/fills like any other signal (in terms of EMI etc). Only extra consideration would be to make the traces/fills wide enough.
Alas, my dream stackup is not supported here. The inner layers are 1oz and the stackup looks more like this:
L1 ----- Signal / Power L2 ----- Ground L3 ----- Signal L4 ----- Signal L5 ----- Ground L6 ----- Signal / Power
This is not necessarily bad, just different from what I prefer. Note that layer 3 & 4 have a 0.5mm distance, which would be far enough to allow them to not cause cross talk. However, layer 1 and 3 share the same reference plane. Same with 4 and 6. If routing with this in mind, you end up with four signal layers and two reference planes. At least the inner two layers could be used for high speed signals. Another benefit would be that it is hard to route e.g. 50 ohm traces on the outer layers, because they would need to be so wide that you are not able to create high density boards. On layer 3 and 4, the 50 ohm traces would be thinner, allowing for denser designs. It also makes it easier to manage skew in signal groups, compared to having the signals on the outer layers. However, you would need more vias and there would be stubs on every via.
I still prefer 3+3 signal/reference layers, like I describe above. It’s a safer solution with less cross talk and EMI issues. Nevertheless, I am wondering if anyone could recommend a better stackup than what I have described above, for high density boards. Also, have anyone tried something similar to what I describe above, and how did it turn out? I have not tried this stackup myself, which is why I am asking.
Many thanks in advance for any feedback!