|
||||||||
| Application Note AP158 | ||||||||
| Good Reasons to Locate Critical Tracks on Inner Layers of a PCB | ||||||||
Here are several good reasons why you shouldn’t put critical tracks on the surface layer of a PCB. Of course, we all have to compromise sometimes, and critical tracks may end up on an outer layer, but do be aware of the compromise that you make when you place them there. Susceptibility to Electrical Interference from Other External Devices Conversely, inner layer conductors are well shielded from the same radiated signals by the solid copper layers on each side of the conductors. Radiation can sneak in at the edge of the board, between the solid layers, but is attenuated by absorption into the copper surfaces before it can penetrate deep into the heart of the board layer. For that reason critical tracks should be routed away from the edge of the board, and a distance of 10x the layer thickness is often recommended as the ideal minimum. (By the way, tracks also transmit radiation as well as receive it, so the exact opposite can happen. And, as above, the best way to limit the radiation is to position tracks between solid copper planes, which also suppress radiation. As above, radiation can escape through the board edges, and the same 10x layer thickness guideline applies.) Control of the Factors Which Determine Impedance Precise Control of Track Cross-Sectional Shape Field solvers assume either a rectangular or trapezoidal cross section for tracks, as that is the normal result of a high quality etching process. Tracks on inner layers are well represented in this assumption, but plated tracks in outer layers lack the sharp rectangular or trapezoidal profile and, as a result, deviate from the design impedance value. Precise Control of Track Cross-Sectional Dimensions Track Coating Thickness and Profile Design assumptions are made concerning the coating uniformity; for example, that the coating has the same thickness above the track as it does above the dielectric laminate material; the coating has the same thickness above the dielectric laminate between differential tracks as it has above the laminate on the outsides of the track pair; the coating (measured horizontally) on the track sides has the same thickness as the coating on the top surface of the track. In all probability these assumptions are incorrect. Coating profiles and resulting thickness variations depend on a wide variety of parameters such as coating material (and liquid properties), how it is applied (brush or spray), the direction of application, track and spacing dimensions, and more. Generally speaking, although a coating may be, in some sense, “uniformly” applied, it will redistribute itself before and during the curing process to adopt a “normal” thickness over most of the dielectric laminate surface, an increasing thickness between narrowly spaced tracks (especially between differential pairs, which are normally more closely spaced) and a decreasing thickness on track top surfaces. One can easily imagine liquid coating running off the track tops into the valleys, and capillary action drawing it up the vertical track sides, the more so in narrower valleys. The finished effect has a smooth, rounded surface profile whose thickness varies from design value over the entire board surface, resulting, in turn, in significant impedance deviation from design value. The profile could, in principle, be taken into account in a field solver if only the designer and builder could predict and define it. That is simply not practicable, however, and so critical traces are best confined to inner layers where the profile and thickness of the surrounding dielectric is most accurately characterized.
|
||||||||